U.S. patent application number 17/164628 was filed with the patent office on 2021-12-30 for biomarkers for early determination of a critical or life threatening response to illness and/or treatment response.
The applicant listed for this patent is Fio Corporation. Invention is credited to Andrea Conroy, Laura Erdman, Kevin Kain, W. Conrad Liles.
Application Number | 20210405059 17/164628 |
Document ID | / |
Family ID | 1000005828104 |
Filed Date | 2021-12-30 |
United States Patent
Application |
20210405059 |
Kind Code |
A1 |
Kain; Kevin ; et
al. |
December 30, 2021 |
BIOMARKERS FOR EARLY DETERMINATION OF A CRITICAL OR LIFE
THREATENING RESPONSE TO ILLNESS AND/OR TREATMENT RESPONSE
Abstract
The invention relates to the use of novel biomarkers and
biomarker combinations having utility in the early determination of
an individual's critical and/or life threatening response to
illness and/or in predicting outcome of said illness. The
measurement of expression levels of the products of the biomarkers
and combinations of biomarkers of the invention have utility in
making the determination of an individual's critical and/or life
threatening response to illness. In some embodiments, the biomarker
and biomarker combinations are agnostic and are independent of the
pre-identification and/or determination of the cause or nature of
the illness. In some embodiments, the biomarkers and biomarker
combinations can be utilized to select treatment and/or monitor the
effectiveness of treatment interventions for an individual who has
a critical illness.
Inventors: |
Kain; Kevin; (Toronto,
CA) ; Liles; W. Conrad; (Seattle, WA) ;
Erdman; Laura; (Toronto, CA) ; Conroy; Andrea;
(Toronto, CA) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Fio Corporation |
Toronto |
|
CA |
|
|
Family ID: |
1000005828104 |
Appl. No.: |
17/164628 |
Filed: |
February 1, 2021 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14916758 |
Mar 4, 2016 |
10921328 |
|
|
PCT/CA2014/050841 |
Sep 5, 2014 |
|
|
|
17164628 |
|
|
|
|
14019447 |
Sep 5, 2013 |
|
|
|
14916758 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
G01N 2800/50 20130101;
G01N 2800/52 20130101; Y02A 50/30 20180101; G01N 33/6842
20130101 |
International
Class: |
G01N 33/68 20060101
G01N033/68 |
Claims
1. A method of determining a likelihood of a test individual having
a critical and/or life threatening response to a suspected illness,
said method comprising: (i) detecting and quantifying a level of
each of two protein biomarkers in a sample from the test
individual, wherein the test individual has not been diagnosed or
differentially diagnosed as having the suspected illness, wherein
said protein biomarkers are: angiopoietin-2 (Ang-2) and soluble
triggering receptor expressed on myeloid cells-1 (sTREM-1)); (ii)
comparing said quantified levels of said protein biomarkers to
control levels of said protein biomarkers from a control
population; and (iii) determining the differential levels for each
biomarker in the comparison of step (ii) so as to make a
determination as to whether said test individual is at an increased
risk of having the critical and/or life threatening response.
2-3. (canceled)
4. The method of claim 1, wherein the determination of step (iii)
is indicative of said individual requiring the application of a
treatment protocol as a result of the increased risk
identified.
5. The method of claim 4, wherein said individual is subjected to
the treatment protocol.
6. The method of claim 1, wherein said control population is a
population of individuals having a critical and/or life threatening
illness.
7. The method of claim 1, wherein said control population is a
population of individuals having a critical and/or life threatening
illness, wherein the individuals have not developing a critical
and/or life threatening response to the illness.
8. The method of claim 1, wherein said control population is a
population of individuals having a critical and/or life threatening
illness, wherein the individuals have developing a critical and/or
life threatening response to the illness.
9. The method of claim 1, wherein said control population is a
population of individuals who are normal.
10. The method of claim 1, wherein said control population is a
population of individuals that do not have an illness which is
critical and/or life threatening.
11. The method of claim 1, wherein said control population is a
population of individuals wherein the members of the control
population do not have an illness which is critical and/or life
threatening.
12. The method of claim 1, wherein the illness is a pneumonia, an
upper respiratory tract infection, a lower respiratory tract
infection, an influenza, an E. coli infection, a bacteremia, a
rickettsial infection, salmonellosis, a streptococcal infection, a
staphylococcus infection, malaria, sepsis, Dengue fever, west nile
virus, toxic shock syndrome, leptospirosis, or a viral hemorrhagic
fever.
13. A method of determining a likelihood of a test individual
having a critical and/or life threatening response to a suspected
illness, said method comprising: (i) detecting and quantifying a
level of each of two protein biomarkers in a sample from the test
individual, wherein the test individual has not been diagnosed or
differentially diagnosed as having the suspected illness, wherein
said protein biomarkers are: angiopoietin-2 (Ang-2 and soluble
triggering receptor expressed on myeloid cells-1 (sTREM-1); (ii)
utilizing the quantified levels of each said protein biomarkers
from said sample in a classifier derived from testing said protein
biomarkers in one or more control populations; and (iii) making a
determination as to whether said individual is an increased risk of
a life threatening response as a result of application of said
classifier.
14-15. (canceled)
16. The method of claim 13, wherein the determination of step (iii)
is indicative of said individual requiring the application of a
treatment protocol as a result of the increased risk
identified.
17. The method of claim 16, wherein said individual is subjected to
the treatment protocol.
18. The method of claim 13, wherein said classifier is derived
using at least two control populations, a population of individuals
having the suspected illness, and not having a critical and/or life
threatening response to said suspected illness and a population of
individuals having the suspected illness and having a critical
and/or life threatening response to said illness.
19. The method of claim 13, wherein said classifier is derived
using at least two control populations, a population of individuals
having an illness which can be critical and/or life threatening,
and not having a critical and/or life threatening response to said
illness and a population of individuals having an illness which can
be critical and/or life threatening, and having a critical and/or
life threatening response to said illness.
20. The method of claim 13, wherein said classifier is derived
using at least two control populations, a population of individuals
considered normal and a population of individuals having an illness
which can be critical and/or life threatening, and having a
critical and/or life threatening response to said illness.
21. The method of claim 13, wherein the suspected illness is a
pneumonia, an upper respiratory tract infection, a lower
respiratory tract infection, an influenza, an E. coli infection, a
bacteremia, a rickettsial infection, salmonellosis, a streptococcal
infection, a staphylococcus infection, malaria, sepsis, Dengue
fever, west nile virus, toxic shock syndrome, leptospirosis, or a
viral hemorrhagic fever.
22-42. (canceled)
43. A composition comprising a collection of two antibodies and a
suitable buffer, said composition capable of selectively binding to
two protein biomarkers from a sample isolated from a test
individual suspected of having an illness, wherein the protein
biomarkers are: angiopoietin-2 (Ang-2) and soluble triggering
receptor expressed on myeloid cells-1 (sTREM-1) and wherein the
composition is used to quantify the level of said protein
biomarkers in said sample and determine whether said test
individual is at an increased risk of having a critical and/or life
threatening response to the illness.
44-59. (canceled)
60. A method of determining whether the administration of a
treatment protocol is likely to be useful in a test individual
having presenting with one or more symptoms of an illness, said
method comprising: (i) detecting and quantifying a level of each of
two protein biomarkers in a sample from the test individual,
wherein the illness of the test individual has not been diagnosed
or differentially diagnosed, wherein said protein biomarkers are:
angiopoietin-2 (Ang-2) and soluble triggering receptor expressed on
myeloid cells-1 (sTREM-1) (ii) utilizing the quantified levels of
each said protein biomarkers from said sample in a classifier
derived from testing said protein biomarkers in one or more control
populations (iii) making a determination as to whether said test
individual is likely to benefit from the treatment protocol as a
result of application of said classifier.
61-69. (canceled)
70. The method of claim 21, wherein the suspected illness is
sepsis.
Description
1. FIELD OF THE INVENTION
[0001] Encompassed within the scope of the invention is the use of
novel biomarkers and biomarker combinations having utility in the
early determination of an individual's critical and/or life
threatening response to illness and/or in predicting outcome of
said illness. In some embodiments, the biomarker and biomarker
combinations are agnostic and are independent of the
pre-identification and/or determination of the cause or nature of
the illness. In some embodiments, the biomarkers and biomarker
combinations can be utilized to monitor the effectiveness of
treatment interventions for an individual who has a critical
illness or to select a treatment intervention which is likely to be
effective for the individual, independently of the
pre-identification and/or determination of the cause or nature of
the illness.
2. BACKGROUND OF THE INVENTION
Diagnosis and Treatment
[0002] Diagnosis, in the medical context, is the act or process of
identifying or determining the nature and/or cause of an illness by
identifying the condition(s) (including the diseases and/or
injuries) responsible through evaluation of one or more factors
which can include patient history, physical examination, review of
symptoms and review of data from one or more laboratory tests.
While it is not always possible to identify the exact nature or
cause of the illness, differential diagnosis may also be utilized
in an attempt to eliminate one or more possible causes in order to
select the most likely cause.
[0003] Once a diagnosis or differential diagnosis has been made,
treatment options are considered, and a treatment strategy chosen.
In some cases, treatment may begin before diagnosis has been
completed (for example, treatment pending receipt of lab results).
In other cases, the cause of the illness may remain elusive, but
nevertheless treatment is selected on the basis of the symptoms
which the individual presents. When the diagnosis, differential
diagnosis, or symptoms are indicative of a condition which has the
potential to be critical and/or life threatening, the management
strategy may include additional considerations to ensure the best
possible clinical outcome including rapid triage, referral,
admission to hospital, enhanced monitoring, admission to an
intensive care unit, and the like.
The Agnostic Approach to Diagnosis and Treatment
[0004] The traditional model of selecting a treatment strategy
based solely on the pre-determined origin or cause of the illness
has some significant drawbacks. While identifying the cause helps
to ensure that the selected course of treatment is disease, injury,
or at least symptom specific, it often fails to recognize the
importance that the individual's unique response to their condition
plays in defining the course and severity of the illness. It also
places an emphasis on diagnostic predetermination of disease or
illness which may be incompatible with the availability and/or
financial burden associated with appropriate diagnostic
methods.
[0005] The "agnostic" approach to treatment challenges the
traditional paradigm of selecting a treatment strategy based on the
origin or cause of illness. The agnostic approach is chosen not
necessarily because the cause or origin is unknowable (as in the
religious context), or because diagnosis cannot be of assistance,
but because knowing as early as possible and/or without the
benefits of diagnosis whether an individual will respond critically
and/or in a life threatening manner to illness can provide a more
effective and rapid method to triage and select appropriate
treatment tailored to the individual.
Individual's Response to Illness
[0006] It is well recognized that not all individuals respond to an
illness in the same manner. Many develop only mild and self-limited
disease, while a small proportion may progress to a critical and/or
life threatening stage. At presentation to medical care, it can be
difficult to determine who will do well without intervention, or
with only minimal intervention, and who needs admission and
specialized management in order to improve clinical outcome. For
example, in the case the H1N1 influenza pandemic, it was estimated
that approximately 61 million individuals in the United States were
infected with H1N1 (during the period from April 2009 to April
2010), but only a small percentage of those cases resulted in
death. Of the 61 million individuals infected, approximately
274,000 individuals were admitted for hospitalization (0.449%), and
12,470 thousand deaths occurred (0.012%) (Emerging Infection
Programs Data released May 14, 2010 from the Centre for Disease
Control; Deaths rounded to the nearest ten. Hospitalizations have
been rounded to the nearest thousand and cases have been rounded to
the nearest million). Clearly some individuals were more able to
fight the H1N1 infection than others.
[0007] Despite this diversity of response, it has been difficult,
even with retrospective analysis, to determine what specific
factors and characteristics contributed to the differential outcome
in these individuals. For example, a retrospective study was
performed on worldwide data available prior to Jul. 16, 2009 on the
684 deaths reported as of that date (Vaillant, L. et al,
Eurosurveillance, Vol. 14, Issue 33, p. 1-6 (2009)) and the age of
the patients were reviewed, by country. In that study it was found
that while overall most deaths (51%) occurred in the age group of
20-49, the impact of age, and the age group most impacted varied in
different countries, making it difficult to draw predictive
conclusions.
[0008] Another example of an illness which has life threatening
potential is sepsis (septicemia). Sepsis is a systemic inflammatory
response to a presumed infection, and may result from numerous
diverse diseases or etiologies. In some cases severe sepsis may
develop wherein the syndrome is also associated with organ
dysfunction, hypoperfusion, or hypotension.
[0009] Because only a small fraction of individuals with an illness
proceed to have a critical and/or life threatening response, an
ability to differentiate those individuals who require urgent
triage and intensive treatment from those individuals who do not,
would be of significant advantage.
[0010] Current attempts to selectively treat individuals who are
most vulnerable for a life threatening response to an illness
occurs by first diagnosis said illness, and then either
pre-classifying individuals based on known risk factors (e.g. age,
existing co-morbidities and the like) and/or by monitoring
individuals for early indications that suggest the illness is
proceeding in a life threatening manner. For example, a prospective
cohort study conducted in 2 phases at 2 general hospitals in Brazil
found that by increased monitoring of in-hospital patents using
currently existing measurable indicators for detection specific to
sepsis, and providing treatment accordingly, the mortality rate for
patients was reduced from 61.7% to 38.2% (Wesphal, G. A., et al.
"Reduced mortality after the implementation of a protocol for the
early detection of severe sepsis" Journal of Critical Care (2011)
26 p. 76-81).
[0011] Nevertheless, reliance on risk factors remains vastly
inadequate as a means of selecting individuals who are likely to
have a life threatening response (see Vaillant, L et al. supra),
and existing measurable indicators that an individual is having a
life threatening response often requires extensive and costly
monitoring of patients and can take too long to be of clinical use
in managing the patient. Furthermore, relying on diagnosis prior to
monitoring or providing treatment can increase costs and cause
unnecessary delay. This is problematic, particularly in cases where
resources are limited, such as in developing countries, but applies
equally to developed countries given the costs associated with
critical care.
[0012] For example, in the case of H1N1 treatment, Durben et. al
modeled the costs from a societal perspective for the treatment of
the Ontario population (assuming no preventative vaccination) and
determined a total cost of $1.10 billion dollars with approximately
87 million dollars being allocated to various aspects of hospital
care (Durben et al. (2011) "A cost effectiveness analysis of the
H1N1 vaccine strategy for Ontario, Canada" Journal of Infectious
Diseases and Immunity Vol. 3(3) p. 40-49). The early and accurate
identification and stratification of those individuals more likely
to have a poor response to the infection could have focused
resources on those most likely to benefit from them and away from
the majority of infected individuals who recovered well without
specific medical intervention. This strategy would presumably have
decreased these projected costs quite significantly.
[0013] Thus, what is needed in the art is one or more biomarkers
which provide greater certainty than current models of an
individual's increased risk of progressing to a critical and/or
life threatening response to illness, and/or to identify an
individual as needing treatment intervention, so as to select
and/or modify an appropriate treatment protocol for said
individual. Preferably these biomarkers would recognize the
increased risk as early as possible so as to allow the greatest
potential for treatment intervention. It would also be particularly
helpful if the biomarkers were agnostic and had utility
irrespective of the illness, so it would be unnecessary to first
diagnose the illness. Also, the ability to use one or more
biomarkers to monitor the impact of the treatment protocol on the
progress of a life threatening response would permit modification
of the treatment protocol as necessary would also be of significant
benefit.
3. SUMMARY
[0014] In one aspect, what is disclosed are biomarkers and
biomarker combinations which provide an indication of an
individual's response to illness, the severity of that response,
and whether they already have, or are progressing to, a critical
and/or life threatening form of illness. In another aspect the
biomarker and biomarker combinations are capable of providing an
early indication of the severity of an individual's response to
illness which is not predicated upon first determining the cause or
source of the illness. In yet another aspect, what is disclosed are
biomarkers and biomarker combinations which can be used prior to,
or in place of, diagnosis or differential diagnosis in order to
select an appropriate treatment protocol. In yet another aspect,
what is disclosed are biomarkers and biomarker combinations which
provide an early indication of the impact of the treatment protocol
on the individual's risk or progress of their life threatening
response.
[0015] In another aspect is a composition comprising a collection
of two or more antibodies and a suitable buffer, the composition
capable of selectively binding to at least two protein biomarkers
from a sample isolated from a test individual, where the protein
biomarkers are those in Table 1. In another aspect the composition
is a composition comprising three or more antibodies and the
composition is capable of selectively binding to at least three
protein biomarkers from a sample isolated from the test individual,
where the protein biomarkers are those in Table 1. In another
aspect the composition comprises a collection of two or more
antibodies and a suitable buffer, the composition is capable of
selectively binding to at least two protein biomarkers from a
sample isolated from a test individual, and the protein biomarkers
are C5a, VEGF, sFlt-1, CHI3L1, CRP, Ang-like3, FactorD, or IL8bpa).
In another aspect the composition comprises a collection of three
or more antibodies and a suitable buffer, the composition is
capable of selectively binding to at least three protein biomarkers
from a sample isolated from a test individual, and the protein
biomarkers are C5a, VEGF, sFlt-1, CH13L1, CRP, Ang-like3, Factor D,
or IL18bpa). In yet another aspect, the sample is a whole blood
sample, a serum sample or a plasma sample. In another aspect the
composition comprises a collection of two or more antibodies and a
suitable buffer, the composition is capable of selectively binding
to at least two protein biomarkers from a sample isolated from a
test individual, and the protein biomarkers are CRP, PCT, CHI3L1,
P-Selectin, vWF, Ang3L1, Tie-2, Endoglin, and IL18bpa. In another
aspect the composition comprises a collection of three or more
antibodies and a suitable buffer, the composition is capable of
selectively binding to at least three protein biomarkers from a
sample isolated from a test individual, and the protein biomarkers
CRP, PCT, CHI3L1, P-Selectin, vWF, Ang3L1, Tie-2, Endoglin, and
IL18bpa. In yet another aspect, the sample is a whole blood sample,
a serum sample or a plasma sample
[0016] In some embodiments, the compositions are used to (i) detect
and quantify a level of the two or more protein biomarkers in the
sample, (ii) compare the quantified level to control levels of the
protein biomarkers in a control population, (iii) determine the
presence of differential levels for the two or more biomarkers so
as to make a determination that the individual is at a
significantly increased risk of having a critical and/or life
threatening response to illness as compared with the control
population. In some embodiments, the detecting and quantifying
utilizes one or more devices to transform the sample into data
indicative of the levels of each of the two or more protein
biomarkers. In some embodiments, the device is an enzyme linked
immunoassay which is utilized to transform the sample into data. In
some embodiments, the test individual is subjected to a treatment
protocol on the basis of the determination in step (iii).
[0017] In some embodiments, the control population is an population
of individuals having the same illness as the test individual. In
some embodiments, the control population is a population of
individuals having the same illness as the test individual, and not
developing a critical and/or life threatening response to the
illness. In some embodiments, the control population is a
population of individuals who are normal. In some embodiments, the
control population is a population of individuals wherein the
majority of members of the control population do not have the same
illness as the test individual. In some embodiments, the
populations noted above are unbiased populations.
[0018] In some embodiments, there is a method of determining the
likelihood that a test individual has or will develop a critical
and/or life threatening response to illness, where the method
includes (i) detecting and quantifying a level of each of two or
more protein biomarkers in a sample, where the protein biomarkers
are those in Table 1 (ii) comparing the quantified levels of said
protein biomarkers to control levels of the protein biomarkers from
a control population (iii) determine the presence of differential
levels for the two or more biomarkers based on the comparison in
step (ii) so as to make a determination that the individual is an
increased risk of having a critical and/or life threatening
response to illness when compared with the control population.
[0019] In some embodiments, the determination is made that the
individual is at a significantly increased risk. In some
embodiments, the detecting and quantifying of step (i) utilizes one
or more devices to transform the sample into data indicative of the
levels of each of the two or more protein biomarkers. In some
embodiments, the one or more devices is an enzyme linked
immunoassay. In some embodiments, the individual is subjected to a
treatment protocol on the basis of the determination made. In some
embodiments, the control population is an unbiased population of
individuals having the same illness as the test individual. In some
embodiments, the control population is a population of individuals
having the same illness as the test individual, and not developing
a critical and/or life threatening response to the illness. In some
embodiments, the control population is a population of individuals
who are normal. In some embodiments, the control population is a
population of individuals wherein the majority of members of the
control population do not have the same illness as the test
individual.
[0020] In some embodiments, there is a method of determining the
likelihood that a test individual will develop a critical and/or
life threatening response to illness, where the method includes (i)
detecting and quantifying a level of each of two or more protein
biomarkers in a sample, where the protein biomarkers arm those in
Table 1 (ii) using the quantified levels of each of the protein
biomarkers from the sample in a classifier where the classifier was
generated using two populations, a first population who developed a
critical and/or life threatening response to illness and a second
control population, (iii) making a determination as to whether the
quantified levels are indicative of the individual being more
similar to the first population or the second control population so
as to determine whether the individual is at an increased risk of
developing a critical and/or life threatening response to
illness.
[0021] In some embodiments, the determination is made that the
individual is at a significantly increased risk. In some
embodiments, the detecting and quantifying of step (i) utilizes one
or more devices to transform the sample into data indicative of the
levels of each of the two or more protein biomarkers. In some
embodiments, the one or more devices is an enzyme linked
immunoassay. In some embodiments, the individual is subjected to a
treatment protocol on the basis of the determination made. In some
embodiments, the second control population is an unbiased
population of individuals having the same illness as the test
individual. In some embodiments, the second control population is a
population of individuals having the same illness as the test
individual, and not developing a critical and/or life threatening
response to the illness. In some embodiments, the second control
population is a population of individuals who are normal. In some
embodiments, the second control population is a population of
individuals wherein the majority of members of the control
population do not have the same illness as the test individual.
[0022] In some embodiments, the test individual has not been
diagnosed or differentially diagnosed with an illness which has the
potential to become critical and/or life threatening prior to use
of compositions or methods as disclosed.
[0023] In some embodiments, the compositions are used to (i) detect
and quantify a level of the two or more protein biomarkers in the
sample, (ii) compare the quantified level to control levels of the
protein biomarkers in a control population, (iii) determine the
presence of differential levels for the two or more biomarkers so
as to make a determination that a treatment protocol should be
administered to the individual. In some embodiments, the detecting
and quantifying utilizes one or more devices to transform the
sample into data indicative of the levels of each of the two or
more protein biomarkers. In some embodiments, the device is an
enzyme linked immunoassay which is utilized to transform the sample
into data.
[0024] In some embodiments, the control population is a population
of individuals having an illness for which it is appropriate to
administer the treatment protocol. In some embodiments the control
population is a population of individuals for which the
administration of the treatment protocol is unnecessary. In some
embodiments, the control population is a population of individuals
wherein the majority of members of the control population are those
to whom it is appropriate to administer the treatment protocol. In
some embodiments, the populations noted above are unbiased
populations. In some embodiments, the control population is a
population of individuals having a bacterial infection which can be
treated with antibiotic. In some embodiments, the control
population is a population of individuals having a viral infection
for which antibiotics would not be effective.
4. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] The objects and features of the invention can be better
understood with reference to the following detailed description and
drawings.
[0026] FIGS. 1A and 1B in one embodiment, compares protein
biomarker levels isolated from plasma in children who have been
diagnosed as having malaria (including individuals who can be
subclassified as having either cerebral malaria (CM) or severe
malarial anemia (SMA)) and who survived the malaria, as compared
with the protein biomarker levels isolated from plasma in children
who died from the malaria and demonstrate a statistically
significant difference as between the two phenotypic groups. FIG.
1A shows the results from biomarker Ang-2, sICAM-1, sFlt-1, CHI3L1,
IP-10, sTie-2, and PCT. FIG. 1B shows the results from biomarker
sTREM-1. * indicates a statistical difference in the protein levels
with a p value of <0.05. ** indicates p values of <0.01.
[0027] FIG. 2A, in one embodiment, demonstrates the receiver
operating characteristic (ROC) curves generated using the selected
biomarkers sICAM-1, sFlt-1, Ang-2, PCT, IP-10, sTREM-1, and CHI3L1
to differentiate between fatal and non-fatal malaria. Dashed
reference lines represent the ROC curve for a test with no
discriminatory ability. Area under the ROC curve is noted in each
graph with the 95% confidence interval shown below in parentheses P
values are indicated *p<0.05, **p<0.01.
[0028] FIG. 2B, in one embodiment, demonstrates the receiver
operating characteristic (ROC) curve for parasetimia diagnosis
alone. Dashed reference lines represent the ROC curve for a test
with no discriminatory ability. Area under the ROC curve is noted
in each graph with the 95% confidence interval shown below in
parentheses. P values are indicated *p<0.05.
[0029] FIG. 3, in one embodiment, demonstrates a classification
tree analysis used to predict outcome of severe malaria infection
with host biomarkers where six biomarkers were entered into the
CRT, and the resulting CRT using IP-10, Ang-2, and sICAM-1 resulted
as shown with cut-off points as determined Prior probabilities of
survival and death were specified (94.3% and 5.7% respectively).
The cut-points selected by the analysis are indicated between
parent and child nodes. Below each terminal node (ie no further
branching), the predicted categorization of all patients in that
node is indicated. The model yields 100% sensitivity and 92.5%
specificity for predicting mortality (cross validated
misclassification rate 15.4% with standard error 4.9%).
[0030] FIG. 4A, in one embodiment, demonstrates the absolute and
median concentrations of angiopoietin-1 (Ang-1) and angiopoietin-2
(Ang-2), as well as the ratio between the two (Ang-2:Ang-1
expressed as log base 10) in acute and convalescent plasma from
patients with or without STSS. *P<0.05; **P<0.01.
[0031] FIG. 4B, in one embodiment, demonstrates the receiver
operating characteristic curves for each of Ang-1, Ang-2 and the
ratio between the two, comparing patients with STSS in the acute
phase of illness to those without STSS, also in the acute phase of
illness.
[0032] FIG. 5, in one embodiment, shows Angiopoietin-1 and -2
(Ang-1 and Ang-2) concentrations, and the ratio between the two
(Ang-2:Ang-1), in matched acute and convalescent plasma samples
from patients with invasive Group A streptococcal infection and
STSS.
[0033] FIG. 6A, in one embodiment, is a histogram showing the
relationship between mortality (%) and measured Ang-1 levels on
admission.
[0034] FIG. 6B in one embodiment, shows a receiver operating
characteristic (ROC) curve illustrating added sensitivity and
specificity in predicting 28-day mortality when comparing plasma
Ang-1 levels, MOD score or age with the combination of the three
variables.
[0035] FIG. 7A, in one embodiment, shows the comparison of Ang-2
levels with MOD score as predictors of mortality in patients with
severe sepsis.
[0036] FIG. 7B, in one embodiment, shows the comparison of Ang-2
levels taken one day prior to assessing the MOD score in patients
with severe sepsis.
[0037] FIG. 8A, in one embodiment, shows the levels of
Angiopoietin-1 (Ang-1), Angiopoietin-2 (Ang-2) and the Ang-2:Ang-1
ratio in children with uncomplicated E. coli O157:H7 infection
(infected), children prior to the diagnosis of HUS (pre-HUS), and
children demonstrating HUS at the time of diagnosis (HUS).
*p<0.05, **p<0.01 unfilled circles indicate outliers
(1.5.times. interquartile range [IQR], filled circles indicate
extreme outliers (3.times.IQR).
[0038] FIG. 8B, in one embodiment, shows Receiver Operating
Characteristic (ROC) curves for Ang-1, Ang-2 and Ang-1:Ang-2 ratio
as comparing children with uncomplicated infection and those with
the pre-HUS phase of illness, with the null hypothesis being that
the area under the curve is 0.5 p=0.01 for Ang-1.
[0039] FIG. 9, in one embodiment, shows the CRT analysis of Model 1
of Example 15, from Table 10, wherein the ability of biomarkers
CRP, Endoglin and P-selectin 1 to differentiate between children
having pneumonia (as confirmed by chest x-ray) and children
characterized as having "clinical" pneumonia pursuant to WHO
standards, but not having pneumonia in accordance with chest x-ray
criteria is shown, as is the incremental benefits of each biomarker
when layered onto the decision tree of the previous biomarker.
5. DETAILED DESCRIPTION
5.1 Definitions
[0040] The following definitions are provided for specific terms
which are used in the following written description.
[0041] As used herein, the "amino terminal region of a polypeptide"
refers to the polypeptide sequence of a protein biomarker. As used
herein, the "amino terminal region" refers to a consecutive, or
nearly consecutive stretch of amino acids located near the amino
terminus of a polypeptide and is not shorter than 3 amino acids in
length and not longer than 350 amino acids in length. Other
possible lengths of the "amino terminal" region of a polypeptide
include but are not limited to 5, 10, 20, 25, 50, 100 and 200 amino
acids.
[0042] The term "antibody" encompasses monoclonal and polyclonal
antibodies and also encompasses antigen-binding fragments of an
antibody. The term "antigen-binding fragment" of an antibody (or
simply "antibody portion," or "antibody fragment"), as used herein,
refers to one or more fragments of a full-length antibody that
retain the ability to specifically bind to a polypeptide encoded by
one of the genes of a biomarker of the invention. Examples of
binding fragments encompassed within the term "antigen-binding
fragment" of an antibody include (i) a Fab fragment, a monovalent
fragment consisting of the VL, VH, CL and CH1 domains; (ii) a
F(ab').sub.2 fragment, a bivalent fragment comprising two Fab
fragments linked by a disulfide bridge at the hinge region; (iii) a
Fd fragment consisting of the VH and CH1 domains; (iv) a Fv
fragment consisting of the VL and VH domains of a single arm of an
antibody, (v) a dAb fragment (Ward et al., (1989) Nature
341:544-546), which consists of a VH domain; and (vi) an isolated
complementarity determining region (CDR). Furthermore, although the
two domains of the Fv fragment, VL and VH, are coded for by
separate genes, they can be joined, using recombinant methods, by a
synthetic linker that enables them to be made as a single protein
chain in which the VL and VH regions pair to form monovalent
molecules (known as single chain Fv (scFv); see e.g., Bird et al.
(1988) Science 242:423-426; and Huston et al. (1988) I1 Proc. Natl.
Acad. Sci. USA 85.5879-5883). Such single chain antibodies are also
intended to be encompassed within the term "antigen-binding
fragment" of an antibody. These antibody fragments are obtained
using conventional techniques known to those with skill in the art,
and the fragments are screened for utility in the same manner as
are intact antibodies. The antibody can be monospecific, e.g., a
monoclonal antibody, or antigen-binding fragment thereof. The term
"monospecific antibody" refers to an antibody that displays a
single binding specificity and affinity for a particular target,
e.g., epitope. This term includes a "monoclonal antibody" or
"monoclonal antibody composition," which as used herein refer to a
preparation of antibodies or fragments thereof of single molecular
composition.
[0043] As used herein an "array" contemplates a set of protein
biomarkers, or antibodies complementary to protein biomarkers, or
combinations thereof immobilized to a support. An array can also
include fragments of protein biomarkers or fragments of antibodies
immobilized to a support wherein the fragment still allows the
selective binding of the protein or antibody fragment to its
complementary binding partner.
[0044] As used herein, the "carboxy terminal region of a
polypeptide" refers to the polypeptide sequences of a protein
biomarker. As used herein, the "carboxy terminal region" refers to
a consecutive, or nearly consecutive stretch of amino acids located
near the carboxy terminus of a polypeptide and is not shorter than
3 amino acids in length and not longer than 350 amino acids in
length. Other possible lengths of the "amino terminal" region of a
polypeptide include but are not limited to 5, 10, 20, 25, 50, 100
and 200 amino acids. The "carboxy terminal" region does not
normally include the polyA tail, if one is present in the protein
biomarker.
[0045] As used herein, the term "classifier" includes a
mathematical model generated on its ability to differentiate
between at least two different traits with respect to an
individual's response to illness. Classifiers can include logistic
regression, classification and/or regression tree analysis, or
other known mathematical models, and are generated using at least
two populations wherein the phenotype of the populations is known.
In some embodiments, a first population has been confirmed as
demonstrating a critical and/or life threatening response to
illness, and the second population is a control population as
defined herein. The classifier, so generated, can be used with data
from a test individual to generate a numerical output which is
indicative of whether the individual is at risk of developing a
critical and/or life threatening response to illness, (or is
already developing a critical and/or life threatening response to
illness), or not.
[0046] As used herein the term "complementary binding partner"
includes a compound which selectively binds to a protein biomarker
and includes nucleic acid aptamers, peptide aptamers, a peptibody,
a mimetic, an inhibitor, and any compound that binds to the protein
biomarker in vivo, an antibody including a monoclonal and/or
polyclonal antibody.
[0047] As used herein the term "control population" is considered
in reference to the test individual since the levels of the
biomarker and biomarker combinations in the test individual must be
compared to levels in the control population to determine the
likelihood of the test individual having a critical and/or life
threatening response, and/or to predict the outcome of the
response. Control populations can either be negative control
populations or positive control populations. In some embodiments,
the control population is a negative control population, the test
individual has been diagnosed with an illness, and the control
population is a population of individuals who have had the illness
of the test individual and have not developed a critical or life
threatening response. In some embodiments, the test individual has
been diagnosed with an illness and the control population is a
population of normal individuals. In some embodiments, the test
individual has been diagnosed with an illness and the control
population is an unbiased population of individuals with said
illness. In some embodiments, the control population is a positive
control population, the test individual has been diagnosed with an
illness, and the control population is a population of individuals
who have had the illness and have developed a critical or life
threatening response. In any of the above embodiments, the control
population may be an unbiased population.
[0048] In some embodiments, the utility of the biomarkers and
biomarker combinations is independent of the cause or source of the
illness of the test individual. Control populations can still
either be negative control populations or positive control
populations. In some embodiments the test individual has not been
diagnosed and/or differentially diagnosed with an illness prior to
testing the biomarker and/or biomarker combinations. In some
embodiments, the individual has not been diagnosed and/or
differentially diagnosed with an illness that can be critical
and/or life threatening prior to testing. In some embodiments, the
control population is a negative control population of individuals
who have had an illness and have not developed a critical and/or
life threatening response. In these embodiments, the illness does
not have to be the same as the illness of the test individual (if
the illness had been diagnosed and/or differentially diagnosed). In
some embodiments, none of the members of the control population
have had the same illness as the test individual. In yet other
embodiments, the majority of the members of the control population
have not had the same illness as the test individual. In yet other
embodiments 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more of the
control population does not have the same illness as the test
individual. In yet other embodiments 20%, 30%, 40%, 50%, 60%, 70%,
80%, 90% or more of the control population has the same illness as
the test individual. In some embodiments, the test individual has
not been diagnosed with an illness prior to testing the biomarker
and/or biomarker combinations and the control population is a
population of normal individuals. In some embodiments, the test
individual has not been diagnosed with an illness prior to testing,
and the control population is a positive control population of
individuals who have had an illness and have developed a critical
or life threatening response to said illness. In some embodiments,
none of the members of the control population have had the same
illness as the test individual. In yet other embodiments, the
majority of the members of the control population have not had the
same illness as the test individual. In yet other embodiments 20%,
30%, 40%, 50%, 60%, 70%, 80%, 90% or more of the control population
does not have the same illness as the test individual. In yet other
embodiments 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more of the
control population has the same illness as the test individual. In
yet other embodiments, none of the members of the control
population have been diagnosed and/or differentially diagnosed with
an illness which is critical and/or life threatening. In yet other
embodiments, a majority of the members of the control population
have not been diagnosed and/or differentially diagnosed with an
illness which is critical and/or life threatening. In yet other
embodiments 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90% or more of the
control population has not been diagnosed or differentially
diagnosed with an illness which is critical and/or life
threatening. In yet other embodiments, the control population is a
population of individuals who have not been diagnosed and/or
differentially diagnosed with an illness which can be critical
and/or life threatening. In yet other embodiments, the control
population is a population of individuals who have not been
diagnosed and/or differentially diagnosed with any illness which is
likely to be critical and/or life threatening. In some embodiments,
the control population is selected from a region or geographic area
comparable with the test subjects and the status of the control
population with respect to the critical and/or life threatening
illness is determined on the basis of the illnesses that are
indigenous to that region or geographic area. In any of the above
embodiments, the control population may be an unbiased
population.
[0049] As used herein "diagnosis" refers to the act or process of
identifying or determining the nature and/or cause of an illness by
identifying the condition(s) (including the diseases and/or
injuries) responsible through evaluation of one or more factors
which can include patient history, physical examination, review of
symptoms and review of data from one or more laboratory tests.
[0050] As used herein "diagnosed with an illness" refers to having
confirmed the nature and/or cause of the illness by identifying the
agent, disease, or injury responsible for one or more of the
symptoms exhibited by said individual, and/or having utilized the
diagnostic test(s) and/or benchmarks that are considered the most
appropriate tests to be applied to diagnose said illness available
under optimum conditions, as defined by conditions that exist in a
typical North American hospital, and that have been adopted by as
the "gold standard" test for such hospital in determining such
illness.
[0051] As used herein "differentially diagnosed with an illness"
refers to having narrowed down the nature and/or cause of the
illness sufficiently to ensure that the patient will receive the
same treatment that the patient would have received if the nature
and/or cause of the illness was known with certainty, or had been
diagnosed utilizing the diagnostic test(s) and/or benchmarks that
are considered the most appropriate tests to be applied to diagnose
said illness available under optimum conditions, as defined by
conditions that exist in a typical North American hospital, and
that have been adopted by as the "gold standard" test for such
hospital in determining such illness.
[0052] As used herein, "illness" refers to a condition which has as
one possible outcome a critical and/or life threatening outcome,
including death. In some embodiments, illness encompasses disorders
of endothelial cell function. In some embodiments, illness is one
which results from an infection such as a parasitic infection, a
viral infection, a bacterial infection, and/or results from
bioactive molecules including microbial toxins. In some embodiments
illness includes conditions wherein one of the causes of the
condition is a significant burn or physical trauma. In other
embodiments illness includes exposure to a biothreat agent such as
anthrax. In other embodiments illness includes exposure to agents
which can cause acute lung injury, such as smoke. In other
embodiments an illness can include disease caused by weaponized
microbes and/or biothreat agents, in some embodiments which cannot
be diagnosed using traditional diagnosis techniques. For example,
the virulence factor or toxin of the microbe and/or biothreat agent
has been modified and inserted into a harmless carrier bacteria,
virus or other carrier agent (Trojan horse effect). Examples of
illnesses include but are not restricted to pneumonias and lower
respiratory tract infections, influenza, E. coli infections and its
complications such as hemolytic uremic syndrome, bacteremias,
rickettsial infections, salmonellosis, streptococcal infections,
staphylococcus infections, malaria, sepsis, Dengue fever, west nile
virus, toxic shock syndrome, leptospirosis, agents causing viral
hemorrhagic fever (e.g, Ebola, Marburg), and microbes or biothreat
agents, including those that have been altered to obscure
traditional diagnosis.
[0053] "Differential levels" refers to protein biomarker levels
which demonstrate a statistically significant difference in the
level when compared with the levels of the protein biomarker in a
control population, wherein the difference is at least 10% or more,
for example, 20%, 30%, 40%, or 50%, 60%, 70%, 80%, 90% or more or
1.5 fold, 2 fold, 2.5 fold, 3.0 fold, 3.5 fold, or more in protein
levels relative to the levels in a control population.
[0054] Differentially increased levels" refers to protein biomarker
levels which demonstrate a statistically significant increased
level when compared with the levels of the protein biomarker in a
control population, wherein the increase in levels is at least 10%
or more, for example, 20%, 30%, 40%, or 50%, 60%, 70%, 80%, 90% or
more or 1.5 fold, 2 fold, 2.5 fold, 3.0 fold, 3.5 fold, or more
increase in protein levels relative to the levels in a control
population.
[0055] "Differentially decreased levels" refers to protein
biomarker levels which demonstrate a statistically significant
decreased level when compared with the levels of the protein
biomarker in a control population, wherein the decrease in levels
is at least 10% or more, for example, 20%, 30%, 40%, or 50%, 60%,
70%, 80%, 90% or more or 1.5 fold, 2 fold, 2.5 fold, 3.0 fold, 3.5
fold, or more decrease in protein levels relative to the levels in
a control population.
[0056] As used herein "an individual's response to illness"
indicates an individual's ability to garner resources to control
and/or battle the illness and determines the course of the illness
within the individual. The individual's response to illness can be
influenced by their innate and acquired immune response, genetic
background, medical history, health status, age, sex, and
pre-existing or co-existing illnesses and/or treatments. In
addition, the course of the illness is also affected by the
treatment protocol applied for the illness itself. Irrespective of
the specific factors which influence the individual's response to
illness, the response impacts the course of the illness in that
individual.
[0057] As used herein "a critical and/or life threatening response
to illness" is indicative of an individual's response to the
illness such that the individual is at an increased risk of death
as compared with the risk of death in an unbiased population of
individuals who suffer the illness. In some embodiments the
increased risk of death is a "significantly increased risk" which
means that the increase in risk as compared to an unbiased
population of individuals having the illness is greater than 50%,
60%, 70%, 80%, 85%, 90%, 95% or more.
[0058] As used herein, the "internal region of a polypeptide"
refers to the polypeptide sequences of a protein biomarker. As used
herein, the "internal region" refers to a consecutive, or nearly
consecutive stretch of amino acids located within the internal
region of a polypeptide and is not shorter than 3 amino acids in
length and not longer than 350 amino acids in length. Other
possible lengths of the "internal" region of a polypeptide include
but are not limited to 5, 10, 20, 25, 50, 100 and 200 amino
acids.
[0059] As used herein, "normal" refers to an individual, a group of
individuals, or a population of individuals who have not shown any
symptoms of illness as defined herein and/or do not have an
illness.
[0060] As used herein, "patient" or "individual" refers to a
human.
[0061] As used herein, "protein biomarker" refers to the form of
the protein, including fragments, which are expressed and
potentially processed and exist in sufficient quantity and for
sufficient time so as to be capable of being measured in humans
using a compound which selectively binds to the protein. Biomarkers
may be capable of being used individually, or in combination with
other biomarkers, additively or synergistically to provide
information as to an individual's response to illness. As used
herein "protein biomarker fragments" may include the "amino
terminal region of a polypeptide", the "carboxy terminal" region of
a polypeptide" or the "internal polypeptide region of a
polypeptide"
[0062] As used herein, the terms "purified" in the context of a
protein biomarker and/or an complementary binding partner (e.g., a
peptide, polypeptide, protein or antibody) refers to a compound
which is substantially free of cellular material and in some
embodiments, substantially free of heterologous agents (i.e.,
contaminating proteins) from the cells or tissue source from which
it is derived, or substantially free of chemical precursors or
other chemicals when chemically synthesized. The language
"substantially free of cellular material" includes preparations of
a proteins in which the proteins are separated from cellular
components of the cells from which it is isolated or recombinantly
produced. Thus, a compound that is substantially free of cellular
material includes preparations of a compounds having less than
about 30%, 20%, 10%, or 5% (by dry weight) of heterologous proteins
(e.g., protein, polypeptide, peptide, or antibody; also referred to
as a "contaminating protein") When the compound is recombinantly
produced, it is also preferably substantially free of culture
medium, i.e., culture medium represents less than about 20%, 10%,
or 5% of the volume of the protein preparation. When the compound
is produced by chemical synthesis, it is preferably substantially
free of chemical precursors or other chemicals, i.e., it is
separated from chemical precursors or other chemicals which are
involved in the synthesis of the compound. Accordingly, such
preparations of a compound have less than about 30%, 20%, 10%, 5%
(by dry weight) of chemical precursors or compounds other than the
compound of interest.
[0063] As used herein, the term "selectively binds" refers to the
specific interaction between a protein biomarker and complementary
binding partner which is able to interact with the protein
biomarker in specific manner, and preferentially to other proteins.
Selective binding of a protein biomarker and a complementary
binding partner and includes the specific interaction of an
antibody with a protein biomarker, including the binding of a
monoclonal antibody and/or a polyclonal antibody to a protein
biomarker preferentially in comparison to non-specific binding.
Selective binding can also include binding between the protein
biomarker and a nucleic acid or peptide aptamer, a peptibody, or
the like. For example, a region, portion or structure of a first
protein molecule recognizes and binds to a region, portion or
structure on a second protein molecule preferentially to the
binding of a non-specific third protein. "Selective binding".
"Selective binding", as the term is used herein, means that a
molecule binds its specific binding partner with at least 2-fold
greater affinity, and preferably at least 10-fold, 20-fold,
50-fold, 100-fold or higher affinity than it binds a non-specific
molecule.
[0064] As used herein, the term "suspected illness" means an
illness which has not been diagnosed and/or differentially
diagnosed.
[0065] As used herein, the term "a therapeutic protocol" or
"treatment protocol", refers to a treatment and/or monitoring
strategy which an individual is subjected to, and can be as a
result of traditional diagnosis, differential diagnosis,
identification of symptoms and/or as a result of use of the protein
biomarkers of the invention and can include the application of one
or more drug therapies or strategies, medical monitoring which can
include increased nursing care, admission to hospital or clinic,
admission to an intensive care unit, and or combinations
thereof.
[0066] By "an unbiased population" as used herein is meant a
population of individuals who have a specific illness, but have not
been pre-selected on the basis of one or more known risk factors
for response to the specific illness (for example, age, sex,
existing co-morbidities and the like).
[0067] As used herein. "a plurality of" or "a set of" refers to
more than two, for example, 3 or more, 4 or more, 5 or more, 6 or
more, 7 or more, 8 or more, 9 or more 10 or more etc.
[0068] As used herein, the terms "treat". "treatment" and
"treating" refer to the reduction or amelioration of the
progression, severity and/or duration of episodes and/or symptoms
of illness.
5.2 Detailed Summary
[0069] We have reviewed various illnesses, each of distinctly
different etiologies, which nevertheless have in common the
potential to progress to a stage which is critical and/or life
threatening. Another commonality amongst these illnesses is the
fact that not all individuals, despite being properly diagnosed,
progress to the critical and/or life threatening form of the
illness. Although it has been known that the individual response to
illness plays a significant role in disease progression, it has
been difficult to accurately predict which individuals will
demonstrate a critical and/or life threatening response, even once
the illness has been diagnosed. We have surprisingly identified
certain proteins biomarkers, many of which are involved in
endothelial activation and/or inflammation, that are found
circulating in the blood of individuals who progress to the
critical and/or life threatening stage of illness at different
levels than the biomarkers are found in individuals who will not
demonstrate a critical and/or life threatening response to illness.
The biomarkers are often found at different levels even in the very
early stages of illness, and often before other known indicators of
disease severity can be measured. More surprisingly, we have found
that these biomarkers have utility across a diverse group of
illnesses suggesting that these biomarkers have utility even if the
individual has not yet been diagnosed or differentially diagnosed
with a specific illness, making the application of these biomarker
particularly useful in situations where: diagnosis is not possible
(such as in cases of weaponized microbes or biothreat agents which
have been designed to prevent identification), diagnosis may be too
costly (such as in developing worlds), diagnosis can delay
appropriate treatment, or diagnosis results in overabundance of
treatment. As such, we have identified proteins that represent
early indicators that an individual is unable to respond
effectively to illness and will progress to a critical and/or life
threatening stage of illness. Because these proteins are
differentially found across such diverse diseases, they have the
ability to be used apriori to diagnosis allowing more timely and
cost effective interventions than would otherwise be available.
[0070] The practice of the present invention employs, in-part
conventional techniques of protein chemistry and molecular biology
which are within the skill of the art. Such techniques are
explained fully in the literature. See, e.g., Sambrook, Fritsch
& Maniatis, 1989, Molecular Cloning: A Laboratory Manual,
Second Edition; Oligonucleotide Synthesis (M J Gait, ed., 1984);
Nucleic Acid Hybridization (B. D. Harnes & S. J. Higgins, eds.,
1984); A Practical Guide to Molecular Cloning (B. Perbal, 1984);
and a series, Methods in Enzymology (Academic Press, Inc.); Short
Protocols In Molecular Biology, (Ausubel et al., ed., 1995). All
patents, patent applications, and publications mentioned herein,
both supra and infra, are hereby incorporated by reference in their
entireties.
5.3 Control and Test Samples
[0071] In some embodiments, all that is required is a drop of
blood. This drop of blood can be obtained, for example, from a
simple pinprick. In some embodiments, any amount of blood is
collected that is sufficient to detect the expression of one, two,
three, four, five, six, seven or more of the genes in Table 1. In
some embodiments, the amount of blood that is collected is 1 ul or
less, 0.5 ul or less, 0.1 ul or less, or 0.01 ul or less. In some
embodiments more blood is available and in some embodiments, more
blood can be used to effect the methods of the present invention.
As such, in various specific embodiments, 0.001 ml, 0.005 ml, 0.01
ml, 0.05 ml, 0.1 ml, 0.15 ml, 0.2 ml, 0.25 ml, 0.5 ml, 0.75 ml, 1
ml, 1.5 ml, 2 ml, 3 ml, 4 ml, 5 ml, 10 ml, 15 ml or more of blood
is collected from a subject. In another embodiment, 0.001 ml to 15
ml, 0.01 ml to 10 ml, 0.1 ml to 10 ml, 0.1 ml to 5 ml, 1 to 5 ml of
blood is collected from a subject.
[0072] In some embodiments, whole blood is utilized. In some
embodiments of the present invention, whole blood collected from a
subject is fractionated (i.e., separated into components) and only
a particular fraction is utilized. In some embodiments only blood
serum is used, wherein the serum is separated from the remaining
blood sample by isolating the liquid fraction of blood which has
been allowed to clot. In some embodiments plasma samples are used,
wherein the blood has been pre-treated with an anticoagulant, such
as EDTA, sodium citrate (including buffered or non-buffered),
heparin, or the like and the supernatant collected and utilized. In
some embodiments, the blood is subjected to Ficoll-Hypaque
(Pharmacia) gradient centrifugation and the peripheral blood
mononuclear cells (PBMC's) are used. Other fractions and/or
fractionating techniques known in the art may also be used, for
example, blood cells can be sorted using a using a fluorescence
activated cell sorter (FACS) e.g. Kamarch, 1987, Methods Enzymol
151:150-165).
5.4 Biomarker and Biomarker Combinations
[0073] Table 1 provides a list of proteins which are useful as
biomarkers either individually or in combination.
[0074] The biomarkers may be used to determine an individual's
status with respect to their developing a critical and/or life
threatening response to illness. In some cases the biomarkers are
individually useful in helping to assess the likelihood of an
individual having a critical and/or life threatening response to
illness. In some cases the biomarkers are useful in helping to
assess whether an individual is at a significantly increased risk
of a critical and/or life threatening response. In yet other
instances the biomarkers are useful in helping to assess whether an
individual is not at a significantly increased risk of having a
critical and/or life threatening response. In yet other instances,
the biomarkers are useful in determining an appropriate treatment
protocol. In yet other instances, the biomarkers are useful in
assessing the impact of a treatment protocol on an individual who
has a significantly increased risk of a critical and/or life
threatening response. In some cases, the biomarkers are useful in
determining the likelihood of an individual demonstrating an
improvement in their critical and/or life threatening response. The
biomarkers are thought to be useful as early indicators of critical
and/or life threatening illness because many play roles in
endothelial activation and vascular leak, angiogenesis, thrombosis,
and inflammation.
TABLE-US-00001 TABLE 1 Symbol/Alternative Protein Name Symbols
Complement fragment C5a C5a Angiopoietin-1 Ang-1 Angiopoietin-2
Ang-2 10 kDa interferon gamma- IP-10 induced protein Soluble
intercellular sICAM-1 adhesion molecule-1 Vascular endothelial VEGF
growth factor A soluble Fms-like tyrosine sFlt-1 kinase receptor-1
(also known as soluble VEGFR1 - Vascular Endothelial Growth Factor
Receptor 1) Chitinase-3-like protein 1 CHI3L1 Soluble triggering
sTREM-1 receptor expressed on myeloid cells-1 C-reactive protein
CRP Procalcitonin PCT Angiopoietin-like Ang-like 3; protein 3 Ang-3
like 1; Ang3L1 Complement factor D Factor D Interleukin 18 Binding
IL18bp; Protein IL18bpa Endoglin End; endoglin p-selectin P-sel;
Pselectin; Endothelial soluble sTie 2; Tie-2 Receptor von
Willebrand Factor vWF
[0075] Combinations of biomarkers of the present invention includes
any combination of 2, 3, 4, 5, 6, 7, 8, 9, 10, I1, 12, 13, 14, 15,
16, 17, or all of the biomarkers listed in Table 1 can be used. For
instance, the number of possible combinations of a subset m of n
proteins in Table 1 above is described in Feller, Intro to
Probability Theory, Third Edition, volume 1, 1968, ed. J. Wiley,
using the general formula:
m!/(n)!(m-n)!
[0076] In one embodiment of the invention, where n is 2 and m is
14, the number of combinations of protein markers selected from
Table 1 is:
14 ! 2 ! .times. .times. ( 14 .times. - .times. 2 ) ! = 14 .times.
x .times. .times. 13 .times. x .times. .times. 12 .times. x .times.
.times. 11 .times. x .times. .times. 10 .times. x .times. .times. 9
.times. x .times. .times. 8 .times. x .times. .times. 7 .times. x
.times. .times. 6 .times. x .times. .times. 5 .times. x .times.
.times. 4 .times. x .times. .times. 3 .times. x .times. .times. 2
.times. x .times. .times. 1 ( 2 .times. x .times. .times. 1 )
.times. .times. ( 12 .times. x .times. .times. 11 .times. x .times.
.times. 10 .times. x .times. .times. 9 .times. x .times. .times. 8
.times. x .times. .times. 7 .times. x .times. .times. 6 .times. x
.times. .times. 5 .times. x .times. .times. 4 .times. x .times.
.times. 3 .times. x .times. .times. 2 .times. x .times. .times. 1 )
= 91 ##EQU00001##
unique two-gene combinations.
[0077] In another embodiment of the invention, where n is 2 and m
is 18, the number of combinations of protein markers selected from
Table 1 is:
18 ! 2 ! .times. .times. ( 18 .times. - .times. 2 ) ! = 18 .times.
x .times. .times. 17 .times. x .times. .times. 16 .times. x .times.
.times. 15 .times. x .times. .times. 14 .times. x .times. .times.
13 .times. x .times. .times. 12 .times. x .times. .times. 11
.times. x .times. .times. 10 .times. x .times. .times. 9 .times. x
.times. .times. 8 .times. x .times. .times. 7 .times. x .times.
.times. 6 .times. x .times. .times. 5 .times. x .times. .times. 4
.times. x .times. .times. 3 .times. x .times. .times. 2 .times. x
.times. .times. 1 ( 2 .times. x .times. .times. 1 ) .times. .times.
( 16 .times. x .times. .times. 15 .times. x .times. .times. 14
.times. x .times. .times. 13 .times. x .times. .times. 12 .times. x
.times. .times. 11 .times. x .times. .times. 10 .times. x .times.
.times. 9 .times. x .times. .times. 8 .times. x .times. .times. 7
.times. x .times. .times. 6 .times. x .times. .times. 5 .times. x
.times. .times. 4 .times. x .times. .times. 3 .times. x .times.
.times. 2 .times. x .times. .times. 1 ) = 153 ##EQU00002##
The measurement of the gene expression of each of these two-gene
combinations, in an additive manner, can be used as described
herein. In another embodiment there are 14!/3!(14-3)! or 364 unique
three-gene combinations and the measurement of each of these
three-gene combinations, in an additive manner, can be used as
described herein.
5.5 Biomarker Quantification
[0078] Protein biomarkers to be quantified are often first isolated
from a sample using techniques which are well known to those of
skill in the art. Protein isolation methods can, for example, be
such as those described in Harlow and Lane (Harlow, E. and Lane,
D., Antibodies: A Laboratory Manual, Cold Spring Harbor Laboratory
Press, Cold Spring Harbor, N.Y. (1988)).
[0079] Detection of quantity or level of the biomarkers in a sample
can occur either directly in said sample, or upon further isolation
or purification of extracted proteins using one or more techniques
known in the art including density gradient centrifugation,
ultra-centrifugation, concentration, dialysis, chromatography,
precipitation, electrophoresis, flow preparation electrophoresis,
selective banding and the like. Commercially available products for
purification of proteins from samples, including blood, are also
well known in the art including Qiagen.RTM.'s AllPrep
DNA/RNA/Protein Mini Kit, and Molecular Research Centre's
(MRC.RTM.) Tri-Reagent.RTM. BD-RNA/DNA Protein Isolation Blood
Derivative.
[0080] Protein biomarkers of a sample can also be differentiated
upon purification or partial purification using such standard
techniques such as a sodium dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE), potentially in combination with western
blotting. Quantities of protein biomarkers can be determined using
techniques known in the art. Useful ways to determine such levels
include, but are not limited to, Western blots, protein
microarrays, and Enzyme-Linked Immunosorbent Assays ("ELISA") and
the like. A number of different types of other useful assays that
measure the presence of a protein biomarker are well known in the
art Immunoassays may be homogeneous, i.e. performed in a single
phase, or heterogeneous, where antigen or antibody is linked to an
insoluble solid support upon which the assay is performed. Sandwich
or competitive assays may be performed. The reaction steps may be
performed simultaneously or sequentially. Threshold assays may be
performed, where a predetermined amount of analyte is removed from
the sample using a capture reagent before the assay is performed,
and only analyte levels of above the specified concentration are
detected. Assay formats include, but are not limited to, for
example, assays performed in test tubes, wells or on
immunochromatographic test strips, as well as dipstick, lateral
flow or migratory format immunoassays. Such examples are not
intended to limit the potential means for determining the level of
a protein biomarker in a sample.
[0081] Agents for detecting a protein biomarker may utilize a
complementary binding partner capable of binding to a protein of
interest. A suitable complementary binding partner can include a
nucleic acid aptamer, a peptide aptamer, a peptibody, a mimetic, a
polyclonal antibody, a monoclonal antibody or any other protein or
nucleic acid, or fragment thereof which is known to have specific
interaction with the protein biomarker either in vivo or in vitro,
or combinations thereof.
[0082] Complementary binding partners, including antibodies, can be
conjugated to non-limiting materials such as magnetic compounds,
paramagnetic compounds, other proteins such as avidin and/or
biotin, nucleic acids, antibody fragments, or combinations thereof
and/or can be disposed on an appropriate surfaces to allow
detection including glass, polystyrene, polypropylene,
polyethylene, dextran, nylon, amylases, natural and modified
celluloses, polyacrylamides, gabbros, and magnetite NPV membrane,
plastic, including a support intended to be used as a dipstick or a
support useful for a microarray.
[0083] One or more complementary binding partners used for
quantification of the protein biomarker can be operably linked
(attached via either covalent or non-covalent methods) to a
detectable label Methods for linking said detectable label to a
complementary binding partner is well known in the art (see, e.g.,
Wong, S. S., Chemistry of Protein Conjugation and Cross-Linking,
CRC Press 1991, Burkhart et al., The Chemistry and Application of
Amino Crosslinking Agents or Aminoplasts, John Wiley & Sons
Inc., New York City, N Y., 1999).
[0084] Useful labels can include, without limitation, fluorophores
(e.g., fluorescein (FITC), phycoerythrin, rhodamine), chemical
dyes, fluorescent dies or compounds that are radioactive,
chemiluminescent, magnetic, paramagnetic, promagnetic, or enzymes
that yield a product that may be colored, chemiluminescent, or
magnetic. The signal is detectable by any suitable means, including
spectroscopic, photochemical, biochemical, immunochemical,
electrical, optical or chemical means. In certain cases, the signal
is detectable by two or more means.
[0085] All protein biomarkers are easily purified from blood, and
can be readily used to generate monoclonal and/or polyclonal
antibodies using traditional techniques for antibody generation
well known in the art. Monoclonal antibodies can be prepared, e.g.,
using hybridoma methods, such as those described by Kohler and
Milstein, Nature, 256:495 (1975) or can be made by recombinant DNA
methods (U.S. Pat. No. 4,816,567). See also Goding, Monoclonal
Antibodies Principles and Practise, (New York: Academic Press,
1986), pp. 59-103. Kozbor, J. Immunol., 133:3001 (1984); Brodeur et
al., Monoclonal Antibody Production Techniques and Applications
(Marcel Dekker, Inc.: New York, 1987) pp. 51-63.
[0086] Monoclonal and/or polyclonal antibodies that have been used
or are known to be available as potentially useful complementary
binding partners for detecting the protein biomarkers are disclosed
in Table 2 herein.
TABLE-US-00002 TABLE 2 Protein Commercially Available Protein Name
Symbol Antibody Reference Complement C5a Abcam .RTM. fragment C5a
ab11878 Angiopoietin-1 Ang-1 Abcam .RTM. ab8451 Angiopoietin-2
Ang-2 Abcam .RTM. ab8452 10 kDa interferon IP-10 Abcam .RTM.
gamma-induced ab8098 protein Soluble sICAM-1 R&D Systems .RTM.
intercellular Mab720 adhesion molecule-1 Vascular VEGF Abcam .RTM.
endothelial growth Ab46154 factor A Soluble vascular sFlt-1 R&D
Systems .RTM. endothelial growth Mab321 factor receptor 1
Chitinase-3-like CHI3L1 Abcam .RTM. protein 1 Ab93034 Soluble
triggering sTREM-1 Abcam .RTM. receptor expressed Ab93717 on
myeloid cells-1 C-reactive protein CRP Abcam .RTM. Ab76434
Procalcitonin PCT Abcam .RTM. Ab53897 Angiopoietin-like Ang-like 3
R&D Systems .RTM. protein 3 MAb38291 Complement factor Factor D
R&D Systems .RTM. D Mab1824 Interleukin 18 IL18bpa Abcam .RTM.
Binding Protein Ab52914 Endoglin END R&D Systems .RTM. Mab13201
P-Selectin Psel Santa Cruz Biotechnology Inc. sc-8419 Endothelial
soluble sTie 2 Abcam .RTM. Tie-2 Receptor Ab10349 von Willebrand
vWF Santa Cruz Factor Biotechnology In. sc-365712
5.6 Use of Biomarkers and Biomarker Combinations
[0087] As taught herein, one or more biomarkers or biomarker
combinations can be used to determine the likelihood of a test
individual having, or not having a critical and/or life threatening
response to illness. In one aspect, the test individual has been
diagnosed or differentially diagnosed, prior to use of the
biomarkers or biomarker combinations. In another aspect, the test
individual has not been diagnosed or differentially diagnosed prior
to the use of the biomarkers or biomarker combinations. In other
aspects, the test individual has been diagnosed with one or more
symptoms indicative of having an illness, but the source or cause
of the illness, and/or the appropriate treatment, remains unknown
prior to the use of the biomarker or biomarker combinations.
[0088] In some embodiments, the biomarker and biomarker
combinations determine that the test individual has an increased
risk of having a critical and/or life threatening response. In some
embodiments, the biomarker and biomarker combinations determine
that the test individual has a decreased risk of having a critical
and/or life threatening response. In some embodiments, the
biomarker and biomarker combinations determine that the test
individual has is at a significantly increased risk of having a
critical and/or life threatening response. In some embodiments, the
biomarker and biomarker combinations determine that the test
individual has a significantly decreased risk of having a critical
and/or life threatening response. The increased risk or decreased
risk is in comparison to a control population. In some embodiments,
the control population is a negative control population of
individuals not having an increased risk of a critical and/or life
threatening response to illness. In some embodiments, the control
population is a positive control population of individuals having
an increased risk of a critical and/or life threatening response to
illness. In some embodiments, the control population is a
population of individuals who have had the illness of the test
individual and have not developed a critical or life threatening
response. In some embodiments the control population is population
of normal individuals. In some embodiments, the control population
is a population of individuals with the same illness as the test
individual. In some embodiments, the control population is a
population of individuals who have had the illness and have
developed a critical or life threatening response. In some
embodiments, the control population is a population of individuals
who have not been diagnosed or differentially diagnosed as having
any illness which may be critical or life threatening. In some
embodiments the population is unbiased with respect to any of the
above.
[0089] In some embodiments, the biomarker and biomarker
combinations can be used to determine that the test individual
would benefit from a specific treatment protocol. In some
embodiments, the test individual is not diagnosed or differentially
diagnosed as having an illness for which a treatment protocol is
warranted, but nevertheless the biomarker and/or biomarker
combinations can be used to determine that there is an increased
likelihood that the test individual would benefit from the
application of the treatment protocol. In some embodiments, the
test individual is not diagnosed or differentially diagnosed as
having an illness for which a treatment protocol is warranted, but
nevertheless the biomarker and/or biomarker combinations can be
used to determine that there is an increased likelihood that the
test individual would not benefit from the application of the
treatment protocol. In some embodiments, the control population is
a negative control population of individuals who have an illness
that would not benefit from the treatment protocol. In some
embodiments, the control population is a positive control
population of individuals having an illness that would benefit from
the treatment protocol. In some embodiments, the control population
is a positive control population of individuals with the same
illness as the test individual. In some embodiments, the control
population is a positive control population of individuals with a
different illness as the test individual, but nevertheless having
an illness which would benefit from the treatment protocol. In some
embodiments, the control population is a negative control
population of individuals who have an illness that would not
benefit from the treatment protocol.
[0090] In order to determine the likelihood of an individual having
a critical and/or life threatening response to an illness, the
levels of one or more of the protein biomarkers of Table 1 in a
sample are detecting and quantified and compared with the
quantified control levels of said one or more protein biomarkers in
a control population. In order to determine the likelihood of an
individual benefitting from the application of a treatment protocol
effective for a critical and/or life threatening illness, the
levels of one or more of the protein biomarkers of Table 1 in a
sample are detecting and quantified and compared with the
quantified control levels of said one or more protein biomarkers in
a control population.
[0091] For each individual protein biomarker, where the level of
the protein biomarker in the test individual is significantly
different (where by significantly different is meant a
statistically significant difference) from the level of the protein
biomarker in the control population, it aids in the determination
that the test individual is likely to have a different response to
a critical and/or life threatening response to illness than the
control individual. In some embodiments, the results from a single
biomarker may be sufficient to determine that the test individual
is at an increased or decreased risk of having a critical and/or
life threatening response to illness. Whether a single biomarker is
sufficient to determine that the test individual is at an increased
or decreased risk of having a critical and/or life threatening
response to illness will depend upon the desired sensitivity and/or
specificity of the test results. In some embodiments, it will be
sufficient that the sensitivity is greater than 51% and the
specificity is greater than 51%. In other embodiments, the
sensitivity of the test results must be greater than 55%, 60%, 65%,
70%, 75%, 80%, 85%, 90%, 95%, 96%, 97%, 98%, 99% or must be 100%.
In some embodiments the specificity of the test results must be
greater than 55%, 60%, 65%, 70%, 75%, 80%, 85%, 90%, 95%, 96%, 97%,
98%, 99% or must be 100%.
[0092] In some embodiments, in order to achieve the desired
sensitivity and/or specificity of the test results, two or more
biomarkers, three or more biomarkers, four or more biomarker, five
or more biomarkers, six or more biomarkers, seven or more
biomarkers, eight or more biomarkers, nine or more biomarkers, ten
or more biomarkers, 11 or more biomarkers, 12 or more biomarkers,
13 or more biomarkers, 14 or more biomarkers, 15 or more
biomarkers, 16 or more biomarkers, 17 or more biomarkers or all
biomarkers must be used in combination.
[0093] In some embodiments, each of said two or more biomarkers,
three or more biomarkers, four or more biomarker, five or more
biomarkers, six or more biomarkers, seven or more biomarkers, eight
or more biomarkers, nine or more biomarkers, ten or more
biomarkers, 11 or more biomarkers, 12 or more biomarkers, 13 or
more biomarkers, 14 or more biomarkers, or more biomarkers, 16 or
more biomarkers, 17 or more biomarkers, or all biomarkers are
weighted equally to make a determination with respect to the status
of a test individual.
[0094] In some embodiments, in order to achieve the desired
sensitivity and/or specificity, each of said biomarkers in the
combination may be weighted differently as determined by a
classifier using at least two populations, wherein at least one
population has been pre-determined to have a critical and/or life
threatening response to an illness, and at least one population has
been pre-determined to not have a critical and or life threatening
response to an illness.
[0095] In some embodiments the classifier is built using logistic
regression as the mathematical model. In other embodiments, the
classification and/or regression tree analysis is used
5.7 Kits
[0096] The present invention provides kits for measuring the levels
of at least 1, at least 2, at least 3, at least 4, at least 5, at
least 6, at least 7, at least 8, at least 9, at least 10, at least
11, at least 12, at least 13, at least 14, at least 15, at least
16, at least 17, or any or all combinations of the protein
biomarkers of the invention Such kits comprise materials and
reagents required for measuring the levels of such protein
biomarkers. As such, the kits provide one or more complementary
binding proteins to measure the level of said biomarkers of said
combinations. In some embodiments the complementary binding
proteins are monoclonal antibodies, and the kit includes antibodies
which bind specifically to each of biomarkers to be measured. The
kits may additional comprise one or more additional reagents
employed in the various methods, such as (1) one or more labelled
or non-labelled antibodies which can bind the complementary binding
proteins in said kit (e.g. Anti-mouse antibodies (1) labeling
reagents ((2) one or more buffer mediums, e.g., hybridization and
washing buffers; (3) protein purification reagents; (4) signal
generation and detection reagents, e.g., streptavidin-alkaline
phosphatase conjugate, chemifluorescent or chemiluminescent
substrate, and the like. In particular embodiments, the kits
comprise prelabeled quality controlled protein for use as a
control.
[0097] In some embodiments, an antibody based kit can comprise, for
example: (1) at least one first antibody (which may or may not be
attached to a support) which binds to a specific protein biomarker;
(2) a second, different antibody which binds to either the protein
biomarker, or the first antibody and is conjugated to a detectable
label (e.g., a fluorescent label, radioactive isotope or enzyme).
The antibody-based kits may also comprise beads for conducting an
immunoprecipitation. Each component of the antibody-based kits is
generally in its own suitable container. Thus, these kits generally
comprise distinct containers suitable for each antibody. Further,
the antibody-based kits may comprise instructions for performing
the assay and methods for interpreting and analyzing the data
resulting from the performance of the assay. In a specific
embodiment, the kits contain instructions for determining the
likelihood an individual is at an increased risk of a critical
and/or life threatening response to illness.
5.8 Examples
Example 1 Individual Biomarkers Predictive of Outcome in
Pre-Diagnosed Malaria
[0098] A retrospective case-control study was performed at Mulago
Hospital in Kampala studying children with malaria as the illness.
Children were enrolled between the ages of 6 months and 12 years
old who presenting with clinical signs and symptoms of malaria
wherein the diagnosis was confirmed by detecting the presence of P.
falciparum infections by microscopic analysis were utilized.
Children with co-morbidities such as sickle cell trait/disease, HIV
co-infection or severe malnutrition were excluded. Using plasma
banked samples, various protein biomarkers were isolated and
measured in the plasma from the approximately 100 Ugandan children
where the diagnosis was confirmed as either cerebral malaria (CM)
or severe malarial anemia (SMA) (both illness that can progress to
life threatening disease). The levels of each of a selection of
specific protein biomarkers was measured in the banked plasma
samples and compared as between children who were known to have
survived the malaria as compared with the levels of these protein
biomarkers in children who died.
[0099] Plasma samples were isolated from whole blood after
treatment with sodium citrate anticoagulant, and were stored at
-20.degree. C. prior to testing. ELISAs were used to quantify the
levels of various potential biomarkers including Ang-2, CRP,
sTREM-1, IP-10, sFlt-1, sICAM-1, and PCT, in said samples. ELISAs
were performed in accordance with manufacturer's instructions with
the following changes: assays were performed in a volume of 50
.mu.L/well; plasma samples were incubated overnight at 4.degree.
C.; and ELISAs were developed using Extravidin.RTM.-Alkaline
Phosphatase (Sigma, 1:1000 dilution, 45 min incubation) followed by
addition of p-Nitrophenyl phosphate substrate (Sigma) and optical
density readings at 405 nm. Assays were developed with
tetramethylbenzidine, stopped with H.sub.2SO4, and read at 450 nm.
Samples with concentrations below the limit of detection were
designated as twice the background level. Background signal was
determined from blank wells included on each plate (assay buffer
added instead of sample), and background optical density was
subtracted from all samples and standards prior to analysis.
Samples with optical densities below the lowest detectable standard
were assigned the value of that standard.
[0100] GraphPad Prism v4, SPSS v18, and MedCalc software were used
for analysis. For clinical and demographic variables, differences
between groups were assessed using the Chi-square test (categorical
variables) or the Kruskal-Wallis test with Dunn's multiple
comparison post-hoc tests (continuous variables). The Mann-Whitney
U test was used to compare biomarker levels between groups, and p
values were corrected for multiple comparisons using Holm's
correction.
[0101] Levels of protein biomarkers were compared as between
children who survived the malaria as compared with children who
died from the malaria and are presented as dot plots with medians
shown in FIG. 1A. FIG. 1B demonstrates results on the same
population for the biomarker sTREM-1, and the dotplot categorizes
the individuals has having either survived or died. A Mann Whitney
U test was performed for each comparison to determine the
statistical significance of the difference as between the two
populations of levels, and those biomarkers showing a statistically
significant difference between the two populations is shown with a
*(p<0.05) or **(p<0.01) in FIG. 1A, 1B. Within this small
sample size, sTie-2 did not reach statistical significance.
Nevertheless, given the close interaction between sTie-2 (as the
receptor to Ang-2), the fact that Ang-2 did show a statistically
significant response, and given the differential trend seen for
sTie-2 (despite not reaching statistical significance) we
reasonably predict that this biomarker will demonstrate utility
when tested with sample populations in greater numbers.
[0102] Receiver operating characteristic curves were generated
using the non-parametric method of Delong et. al (DeLong E R,
DeLong D M, Clarke-Pearson D L (1988) Comparing the areas under two
or more correlated receiver operating characteristic curves; a
nonparametric approach. Biometrics 44:837-845). Data is shown for
biomarkers sICAM-1, sFlt-1, Ang-2, PCT, IP-10, and sTREM-1 in FIG.
2A. As would be understood the area under the ROC curve is
indicative of the ability of each biomarker to differentiate
between the likelihood of an individual dying and not dying. Shown
in dashed reference lines is an ROC curve for a test which has no
discriminatory ability. The area under the ROC curve is noted and
its statistical significance as either *p<0.05 or **p<0.01
shown. In parenthesis is the 95% confidence intervals for the area
under the curve. FIG. 2B shows the ROC curve for parasitemia, which
is currently relied upon to assess the individual's response to
malaria. Parasitemia predicts the quantitative content of parasites
in the blood and is used as a measurement of parasite load in the
organism and an indication of the degree of an active parasitic
infection. As can be seen, each of the biomarkers noted is better
at predicting death than the currently utilized index of
parasitemia.
[0103] To evaluate the biomarkers further, the Youden index was
used to obtain a cut-point for each biomarker, and clinical
performance measures evaluated for these dichotomized biomarkers
(Table 3). All parameters presented in Table 3 are presented with
95% confidence intervals shown in brackets. All cut points were
determined using the Youden Index
(J-max[sensitivity+specificity-1]). For each biomarker is shown the
PLR, positive likelihood ratio, NLR the negative likelihood ratio,
PPV, the positive predictive value and NPV, the negative predictive
value. PPVs and NPVs were based on estimates that 5.7% of CM and
SMA diagnosed patients at the Mulago hospital died of the malaria
infection, sTREM-1 achieved the highest sensitivity (95.7%) but had
low specificity (43.8%), while IP-10 predicted death with the
highest overall accuracy (82.6% sensitivity, 85% specificity).
TABLE-US-00003 TABLE 3 Clinical Performance of Biomarkers for
Predicting Mortality Among Children with Severe Malaria Sensitivity
Specificity PPV NPV Cut-point (%) (%) PLR NLR (%) (%) Ang-2 >5.6
ng/ml 78.3 78.8 3.7 0.3 18.2 98.4 (56.3-92.5) (68.2-87.1) (2.9-4.7)
(0.1-0.7) (5.8-38.7) (92.4-99.9) sICAM >645.3 ng/ml 87.0 75.0
3.5 0.2 17.4 99.0 (66.4-97.2) (64.1-84.0) (2.8-4.3) (0.06-0.5)
(5.9-35.9) (93.2-100) sFlt-1 >1066.3 pg/ml 82.6 57.5 1.9 0.3
10.5 98.2 (61.2-95.0) (45.9-68.5) (1.5-2.5) (0.1-0.8) (3.4-23.1)
(90.4-100) PCT >43.1 ng/ml 56.5 82.5 3.2 0.5 16.3 96.9
(34.5-76.8) (72.4-90.1) (2.2-4.7) (0.3-1.0) (3.8-39.5) (90.5-99.5)
IP-10 >831.2 pg/ml 82.6 85.0 5.5 0.2 25 98.8 (61.2-95.0)
(75.3-92.0) (4.5-6.8) (0.07-0.6) (8.3-49.8) (93.4-100) sTREM-1
>289.9 pg/ml 95.7 43.8 1.7 0.1 9.3 99.4 (78.1-99.9) (32.7-55.3)
(1.3-2.2) (0.01-0.7) (3.3-19.6) (90.5-100)
Example 2 Biomarker Combinations Predictive of Mortality in
Pre-Diagnosed Malaria
[0104] Data was obtained as described in Example 1. The use of
biomarker combinations improved the ability to predict the
likelihood of an individual's life threatening response in malaria.
In this example, a modest number of deaths in the study precluded
using multivariable logistic regression analysis to create
classifiers with more than 2-3 independent variables (Harrell F E,
Jr., Lee K L, Mark D B (1996) Multivariable prognostic models:
issues in developing models, evaluating assumptions and adequacy,
and measuring and reducing errors. Stat Med 15:361-387). Therefore,
as performed in other conditions. (Morrow D A, Braunwald E (2003)
Future of biomarkers in acute coronary syndromes: moving toward a
multimarker strategy. Circulation 108:250-252; Vinueza C A, Chauhan
S P, Barker L, Hendrix N W, Scardo J A (2000) Predicting the
success of a trial of labor with a simple scoring system. J Reprod
Med 45:332-336), six biomarkers were combined (Ang-2, sICAM-1,
sFlt-1, PCT, IP-10 and TREM-1) into a single score. For each
marker, one point was assigned if the measured value was greater
than the corresponding cut-point, and zero points were assigned if
it was lower. A cumulative "biomarker score" was calculated for
each patient by summing the points for all six markers. No two
dichotomized biomarkers were highly correlated (Spearman's rho
<0.6; data not shown), suggesting that each biomarker would
contribute unique information to the score since biomarkers which
are not correlated indicate that the biomarkers each add new
information as compared with single biomarkers alone.
[0105] Biomarker score was highly positively correlated with risk
of death (data not shown; Spearman's rho=0.96, p=0.003). Scores
were elevated among fatalities compared to survivors (median
(interquartile range): 5 (4-6) and 1 (0-2.5), respectively, data
not shown.
[0106] In a univariate logistic regression model, the biomarker
score was a significant predictor of death with an odds ratio of
7.9 (95% CI 4.6-54.4) (Table 4, Model 1). After adjustment to
exclude parasitemia and age, which have been associated with
malaria mortality as predictive factors, the score remained
significant with an adjusted odds ratio of 7.8 (4.7-134) (Table 4,
Model 2).
[0107] ROC curve analysis and cut-point determination were
performed as above for various biomarker combinations to determine
their utility in as predictive indicators of outcome of illness.
Table 5 shows the data resulting from some of the biomarker
combinations tested. All combinations demonstrated some utility as
predictive indicators of outcome of illness. Additional
combinations are shown in Table 6 and Table 6A and 6B. All
parameters in the tables are presented with 95% CIs in parentheses.
Cut-points were determined using the Youden Index
(J=max[sensitivity+specificity-1]). PLR indicates the positive
likelihood ratio; NLR indicates the negative likelihood ratio; PPV
is the positive predictive value; and NPV is the negative
predictive value.
[0108] Using logistic regression on the six biomarker combination
of Ang-2, sICAM-1, sFlt-1, PCT, IP-10 and TREM-1, the AUC was 0.96
(0.90-0.99) (data not shown), and a score .gtoreq.4 was found to
have a 95.7% sensitive and 88.8% specific for predicting death in
the samples tested (Table 5, row 1) For logistic regression,
linearity of an independent variable with the log odds of the
dependent was assessed by including a Box-Tidwell transformation
into the model and ensuring that this term was not significant.
Bootstrapping (1000 sample draws) was used to generate variance
estimates for the cut point. Model goodness-of-fit was assessed by
the Hosmer-Lemeshow test and calibration slope analysis (Steyerberg
E W, Eijkemans M J, Harrell F E, Jr., Habbema J D (2001) Prognostic
modeling with logistic regression analysis: in search of a sensible
strategy in small data sets. Med Decis Making 21:45-56). Positive
and negative predictive values were calculated using the reported
case fatality rate of 5.7% for microscopy-confirmed CM and SMA
cases. (Hosmer D W, Lemeshow S. Applied Logistic Regression. 2nd
ed. New York: John Wiley & Sons, Inc, 2000). PPVs and NPVs were
based on estimates that 5.7% of CM and SMA patients at the Mulago
hospital where samples were obtained die of the malaria infection.
While the positive predictive value for the six biomarker
combination was low (33.9/6) given a fatality rate of 5.7%, the
negative predictive value (NPV) was 99.7%, indicating that a child
with a score .ltoreq.3 will likely respond well to standard
treatment protocols.
TABLE-US-00004 TABLE 4 Association of biomarker score with outcome
among children with severe malaria: logistic regression..sup.a p OR
Hosmer-Lemeshow test Variable b (95% CI) SE Wald df value (95% CI)
Chi square df p value Model 1.sup.b Biomarker score 2.1 (1.5-4.0)
2.3 18.6 1 0.001 7.9 (4.6-54.4) 3.3 5 0.66 Model 2.sup.c Biomarker
score.sup.d 2.1 (1.6-4.9) 21.5 18.2 1 0.001 7.8 (4.7-134) 1.1 8 1.0
Log parasitemia.sup.e 0.050 ((-1.1)-1.3) 2.8 0.010 1 0.91 1.1
(0.35-3.6) Age 0.053 ((-0.61)-1.2) 8.5 0.052 1 0.89 1.1 (0.55-3.3)
.sup.aThe reference category was "survival." .sup.bPseudo-R.sup.2
(Cox & Snell) 0.473 and calibration slope 0.98.
.sup.cPseudo-R.sup.2 (Cox & Snell) 0.474 and calibration slope
1.0. .sup.dBiomarker score and log parasitemia had a significant
but low correlation (Spearman's rho 0.292, p < 0.01).
.sup.eParasitemia was log-transformed in order to achieve linearity
with the log-odds of the dependent variable. SE, standard error;
OR, odds ratio.
TABLE-US-00005 TABLE 5 Clinical performance of biomarker
combinations for predicting mortality among children with severe
malaria..sup.a Number of individuals utilized Threshold in
generating the (positives based Sensitivity Specificity Biomarker
combination data (n). on ROC curves) (%) (%) PPV NPV IP-10, sICAM1
104 2/2 77.3 96.6 85 94.4 IP-10, sICAM1 98 (exclude non- 2/2 93.8
96.3 83.3 98.8 CM/SMA fatal) ANG-2, IP10, sICAM1 104 2/3 86.4 87.5
63.3 96.3 ANG-2, IP10, sICAM1 98 (exclude non- 2/3 93.8 86.6 57.7
98.6 CM/SMA fatal) ANG-2, IP10, CHI3L1 77 2/3 93.8 82.0 57.7 98.0
ANG-2, IP10, sTREM1 77 2/3 93.8 85.2 62.5 98.1 ANG-2, sICAM1,
CHI3L1 77 2/3 93.8 93.4 78.9 98.3 ANG-2, sICAM1, sTREM1 77 2/3 93.8
88.5 68.2 98.2 ANG-2, CHI3L1, sTREM1 77 2/3 81.3 85.2 59.1 94.5
IP10, sICAM1, CHI3L1 77 2/3 93.8 88.5 68.2 98.2 IP10, sICAM1,
sTREM1 77 2/3 93.8 86.9 65.2 98.1 sICAM1, CHI3L1, sTREM1 77 2/3
93.8 86.9 65.2 98.1 ANG-2, IP10, sICAM1, CHI3L1 77 2/4 100.0 80.3
57.1 100.0 ANG-2, IP10, sICAM1, sTREM1 77 2/4 100.0 78.7 55.2 100.0
ANG-2, sICAM1, CHI3L1, sTREM1 77 2/4 100.0 82.0 59.3 100.0 IP10,
sICAM1, CHI3L1, sTREM1 77 2/4 100.0 77.0 53.3 100.0 ANG-2, IP10,
CHI3L1, sTREM1 77 2/4 100.0 73.8 50.0 100.0 ANG-2, IP10, sICAM1,
CHI3L1 77 3/4 87.5 93.4 77.8 96.6 ANG-2, IP10, sICAM1, sTREM1 77
3/4 87.5 95.1 82.4 96.7 ANG-2, sICAM1, CHI3L1, sTREM1 77 3/4 81.3
93.4 76.5 95.0 IP10, sICAM1, CHI3L1, sTREM1 77 3/4 87.5 95.1 82.4
96.7 ANG-2, IP10, CHI3L1, sTREM1 77 3/4 81.3 93.4 76.5 95.0 ANG-2,
IP10, sICAM1, CHI3L1, 77 3/5 100 91.8 76.2 100 sTREM1
TABLE-US-00006 TABLE 6 Clinical performance of selected biomarker
combinations for predicting mortality among children with severe
malaria..sup.a Sensitivity Specificity PPV NPV Combination
Cut-point.sup.b (%) (%) PLR.sup.c NLR (%).sup.d (%) (Ang-2,
.gtoreq.4 95.7 88.8 8.5 0.05 33.9 99.7 sICAM-1, sFlt- (78.1-99.9)
(79.7-94.7) (7.6-9.6) (0.007-0.4) (12.8-61.3) (95.2-100) 1, PCT,
IP-10) Ang-2, PCT, .gtoreq.2 91.3 88.8 8.1 0.1 32.9 99.4 sICAM-1
(72.0-98.9) (79.7-94.7) (7.0-9.4) (0.02-0.4) (12.1-60.3) (94.7-100)
Ang-2, 1P-10, .gtoreq.2 91.3 86.3 6.6 0.1 28.6 99.4 PCT (72.0-98.9)
(76.7-92.9) (5.7-7.7) (0.02-0.4) (10.2-54.4) (94.6-100) PCT, IP-10,
.gtoreq.2 91.3 81.3 4.9 0.1 22.7 99.4 sTREM-1 (72.0-98.9)
(71.0-89.1) (4.1-5.7) (0.03-0.4) (8.1-44.8) (94.2-100)
TABLE-US-00007 TABLE 6A Biomarker combination Cut-point Sens Spec
PLR NLR PPV NPV Biomarker score .gtoreq.4 95.7 88.8 8.5 0.049 21.9
99.8 (all 6 markers) (78.1-99.9) (79.7-94.7) (7.6-9.6) (0.007-0.4)
(4.9-51.3) (95.6-100.0) ANG-2, IP-10, .gtoreq.2 100 81.2 5.3 0 15.0
100 CHI3L1 (85.2-100) (71.0-89.1) (4.8-5.9) (3.4-37.2) (95.5-100)
ANG-2, .gtoreq.2 95.7 81.3 5.1 0.054 14.4 99.8 sICAM-1, (78.1-99.9)
(71.0-89.1) (4.4-5.8) (0.007-0.4) (3.1-36.5) (95.2-100) CHI3L1
ANG-2, .gtoreq.2 91.3 88.8 8.1 0.098 21.2 99.7 sICAM-1, PCT
(72.0-98.9) (79.7-94.7) (7.0-9.4) (0.02-0.4) (4.5-50.5) (95.3-100)
ANG-2, IP-10, .gtoreq.2 91.3 86.3 6.6 0.10 18.0 99.7 PCT
(72.0-98.9) (76.7-92.9) (5.7-7.7) (0.02-0.4) (3.7-44.8) (95.2-100)
sICAM-1, .gtoreq.2 91.3 83.8 5.6 0.10 15.7 99.7 IP-10, CHI3L1
(72.0-98.9) (73.8-91.1) (4.8-6.6) (0.03-0.4) (3.3-39.4) (95.0-100)
sICAM-1, PCT, .gtoreq.2 91.3 80.0 4.6 0.11 13.1 99.6 CHI3L1
(72.0-98.9) (69.6-88.1) (3.9-5.4) (0.03-0.4) (2.7-34.3) (94.8-100)
sICAM-1, 2 91.3 85.0 6.1 0.09 16.8 99.7 CHI3L1 (72.0-98.9)
(75.3-92.0) (5.2-7.1) (0.02-0.4) (3.4-42.4) (95.1-100) (alternative
dichotomization)
TABLE-US-00008 TABLE 6B # +ve Sen Spec Biomarkers BMs (%) (%) NPV
sICAM-1, IP-10 2/2 93.8 95.8 94 CHI3L1, sTREM-1, 2/3 93.8 84.8 98
sICAM-1 ANG-2, CHI3L1, 2/4 100 82.5 100 sTREM-1, sICAM-1 CHI3L1,
sTREM-1, 3/5 100 87.9 100 ANG-2, IP-10, sICAM-1 CHI3L1, sTREM-1,
4/6 100 90.9 100 ANG-2, IP-10, sICAM-1, sFLT-1
Example 3--Use of Classification Tre Analysis as an Alternative
Classifier Predictive of Mortality in Pre-Diagnosed Malaria
[0109] To explore other synergistic combinatorial strategies,
wherein weighting of each biomarker may vary, classification tree
analysis was used, which selects and organizes independent
variables into a decision tree that optimally predicts the
dependent measure. Initially, a model based on IP-10 and sTREM-1
was generated with 43.5% sensitivity and 100% specificity for
predicting mortality (FIG. 3). Since in some instances high
sensitivity would be of particular importance, the analysis
assigning the cost of misclassifying a death as a survivor was
weighted as being 10 times greater than the cost of misclassifying
a survivor as a death A model based on IP-10, Ang-2, and sICAM-1
was generated with 100% sensitivity and 92.5% specificity for
predicting outcome (cross-validated misclassification rate 15.4%,
standard error 4.9%). In summary, combining dichotomized biomarkers
using a scoring system or a classification tree predicted severe
malaria mortality in our patient population with high accuracy.
Example 4 Individual Biomarkers and Biomarker Combinations
Predictive of Patients Developing Toxic Shock Syndrome in Patients
with Invasive S. pyogenes Disease
[0110] A prospective, population-based surveillance for invasive
group A streptococcal disease was undertaken in Ontario, Canada via
mandatory laboratory reporting of S. pyogenes isolates from
normally sterile sites and thirty-seven patients, enrolled between
1999 and 2009, were included in the study. Informed consent was
obtained to collect bacterial isolates and plasma samples, as well
as detailed clinical data from interviews with the attending
physicians and patient chart review. Patients were considered to
have S. pyogenes infections which resulted in streptococcal toxic
shock syndrome (STSS) (a critical and/or life threatening form of
an S. pyogenes infection) if they met the current consensus of
indicator symptoms including: hypotension in combination with at
least two of coagulopathy, acute renal failure, elevated serum
aminotransferases, acute respiratory distress syndrome (ARDS),
rash, or necrotizing fasciitis. Of the 37 patients, 16 were
considered to have invasive streptococcal infection and toxic shock
(STSS), while 21 were determined to have invasive streptococcal
infection alone (no STSS). The underlying source of the infection
was similar between the two groups, with the majority of patients
in both groups having skin and soft tissue infections (7 patients
(44%) with STSS and 12 patients (57%) with invasive streptococcal
infection alone). Presenting group A streptococcal infections in
the remaining patients included respiratory tract infections,
bacteremia without an identified source, post-partum infection, and
peritonitis, and did not differ significantly between the groups.
The two groups were significantly different only in the symptomatic
diagnostic criteria for STSS; hypotension was present in 100% of
patients with STSS and 33% of patients without (P<0.0001). Five
patients with invasive infection and STSS died as compared to one
patient with invasive infection alone (31% versus 5%, P=0.06).
[0111] Acute phase plasma samples were collected upon study
enrollment and stored at minus 70.degree. C. until use. Plasma
concentrations of angiopoietins-1 and -2 were measured by ELISA
(R&D Systems. Minneapolis Minn.) according to the
manufacturer's instructions. The upper and lower limits of
detection for the assays were 10,000 pg/mL and 9.77 pg/mL for Ang-1
and 2520 pg/mL and 2.46 pg/mL for Ang-2, respectively. Samples were
diluted in assay diluent (1:20 for Ang-1 and 1:4 for Ang-2) to fall
within the range of the standard curves.
[0112] Angiopoietin dysregulation (a correlated decrease in Ang-1
levels and an increase in Ang-2 levels) was associated with an
increased likelihood of the individual having the invasive group A
streptococcal disease with STSS as compared with individuals having
invasive group A streptococcal disease without STSS (FIG. 4A and
FIG. 4B). The median plasma concentration of Ang-1 was lower during
the acute phase of illness in patients pre-diagnosed with invasive
infection and STSS than in those pre-diagnosed with invasive
streptococcal infection alone (13,915 pg/mL vs. 29,084 pg/mL),
while the median plasma concentration of Ang-2 was higher (5752
pg/mL vs. 1337 pg/mL). As a result, the normally low Ang-2:Ang-1
ratio was significantly higher amongst patients with invasive
infection and STSS as compared to those with invasive streptococcal
infection alone (0.437 versus 0.048, P<0.05).
[0113] Receiver operating characteristic (ROC) curves were
generated for Ang-1, Ang-2, and the Ang-2:Ang-1 ratio, and the area
under the ROC curves indicated that the degree of magnitude of
Ang-1/2 dysregulation accurately differentiated those individuals
with STSS from those without STSS (FIG. 4B). Although the ROC curve
for plasma Ang-1 concentration did not differ significantly from
chance (AUC: 0.683, P=0.07), the ROC curves for plasma Ang-2 (AUC:
0.759, P=0.009) and for the Ang-2:Ang-1 ratio (AUC: 0.791, P=0.003)
revealed that both discriminated between patients with STSS and
those with invasive streptococcal infection alone (no STSS) and it
is anticipated that the ROC curve for plasma Ang-1 would also be
discriminatory upon an increased sample size since the ROC curve
for plasma Ang-1 concentration trended despite not reaching
statistical difference (AUC: 0.683, P=0.07).
Example 5 Individual Biomarkers and Biomarker Combinations
Predictive of Response in Patients Having Group a Streptococcal
Disease
[0114] Using the samples and methods as outlined in Example 4, we
further measured the biomarkers Ang-1, Ang-2 and the ratio of
Ang-1/Ang-2 as the patients convalesced to demonstrate the
potential for the biomarkers to function as indicators of response
to treatment. Ang-1/2 dysregulation was seen to resolve consistent
with convalescence in both groups of patients (FIG. 4A). In the
cohort of patients with STSS, the median plasma concentration of
Ang-1 rose from 13,519 pg/mL to 21,115 pg/mL, the median plasma
concentration of Ang-2 decreased fell from 5752 pg/mL to 378 pg/mL
(P<0.01), and the median Ang-2:Ang-1 ratio fell from 0.437 to
0.019 (P<0.05).
[0115] Furthermore, in individual patients with STSS, the matched
acute and convalescent plasma Ang-2 concentrations and the
Ang-2:Ang-1 ratios also differed significantly (FIG. 5) The same
pattern was observed in the cohort of patients with invasive
streptococcal disease without STSS, the changes in Ang-1/2
concentrations although the changes were more modest. The median
plasma concentration of Ang-1 in this group increased from 29,084
pg/mL to 31,743 pg/mL, while the Ang-2 concentration declined from
1337 pg/mL to 535 pg/mL, and the Ang-2:Ang-1 ratio decreased from
0.048 to 0.027.
Example 6 Individual Biomarkers and Biomarker Combinations
Predictive of Outcome in Pre-Diagnosed Sepsis
[0116] A multicenter retrospective analysis was performed on
prospectively collected biological and clinical data so as to
identify molecular markers demonstrating an increased likelihood of
patients dying from severe sepsis. Samples were collected from
three tertiary hospital intensive care units (ICU) associated with
Hamilton General Hospital in Hamilton, Canada.
[0117] Seventy patients with severe sepsis enrolled within 24-hours
of admission to the ICU and were followed until day 28, discharge
or death. Clinical data and plasma samples were available on
admission for all patients and daily for 1 week, then weekly
thereafter for 43 of the 70 patients.
[0118] Patients were diagnosed as having severe sepsis if they met
the modified American College of Chest Physicians/Society of
Critical Care Medicine criteria for sepsis known in the art
(Bernard G R, Vincent J-L, Laterre P-F, et al. Efficacy and safety
of recombinant human activated protein C for severe sepsis. N Engl
J Med 2001; 344(10):699-709; Bone R C, Sibbald W J, Sprung C L. The
ACCP-SCCM consensus conference on sepsis and organ failure. Chest
1992; 101(6):1481-1483.) Patients were included if they had known
or suspected infection as well as at least three of four modified
SIRS criteria and at least one of five criteria for organ
dysfunction.
[0119] Venous blood (4.5 ml) collected from indwelling catheters
was transferred into 15 ml polypropylene tubes containing 0.5 ml of
0.105 M buffered trisodium citrate (pH 5.4) and 100 .mu.l of 1 M
benzamidine HCl and centrifuged at 1,500 g for 10 min (20.degree.
C.). Plasma for analysis was stored in aliquots at -80.degree. C.
Commercial enzyme-linked immunoassays (ELISAs) were used to measure
levels of biomarkers. Ang-1 and Ang-2 (R&D Systems.
Minneapolis, Minn., USA) were measured on available samples from
days 1 to 7, 14, and 28. ESEL (R&D Systems, Minneapolis, Minn.
USA), sICAM-1 (R&D Systems. Minneapolis, Minn., USA) and vWF
(antibody: Dako. Carpinteria, Calif., USA, standard: American
Diagnostica, Stamford, Conn., USA), levels were measured on days 1
and 3. All standards, controls and test samples were assayed in
duplicate and averaged prior to interpretation. Concentrations were
interpolated from four parameter logistic fit curves generated
using a standard curve of recombinant human proteins.
[0120] It was determined that patients with low Ang-1 plasma levels
(.ltoreq.5.5 ng/mL) at admission were less likely to survive than
those with high Ang-1 levels (.gtoreq.5.6 ng/ml; relative risk 0.49
[95% CI: 0.25-0.98], p=0.046 (FIG. 6A).
[0121] Ang-1 levels .ltoreq.5.5 ng/mL also remained a significant
predictor of mortality at 28 days in a multivariate logistic
regression model (adjusted odds ratio 0.282 [95% confidence
interval (CI): 0.086-0.93], p=0.037) using known clinical
indicators of increased risk of mortality. Age is a known risk
factor leading to increased likelihood of death from sepsis.
Similarly Multiorgan Dysfunction (MOD) score exists as the current
method of measuring and quantifying organ disfunction, either as a
risk factor for death, a measure of severity of illness, or a
measure of increased risk for morbidity over time. The multivariate
logistic regression model used age (p=0.008) and MOD score
(p=0.014) as additional clinical biomarkers, suggesting that Ang-1
provides independent prognostic information above and beyond age
and MOD scores alone.
[0122] This finding is supported by receiver operating
characteristic (ROC) curve analysis (FIG. 6B) illustrating the
apparent added sensitivity and specificity in predicting 28-day
mortality when comparing plasma Ang-1 levels (area under the ROC
curve (AUROC): 0.62 [95% CI: 0.50-76]). MOD score (AUROC: 0.64 [95%
CI: 0.51-0.77]) or age (AUROC: 0.68 [95% CI: 0.55-0.80]) with the
combination of the three variables (AUROC: 0.79 [95% CI:
0.67-0.90]).
Example 7 Individual Biomarkers and Biomarker Combinations as Early
Predictors of Risk of Mortality in Patients with Sepsis
[0123] As noted, the current standard for determining an
individuals increased likelihood of death from sepsis is the
Multiorgan Dysfunction (MOD) score. Using samples and methods as
described in Example 6, the level of Ang-2 was measured and
correlated with the MOD score across the population of individuals
tested. As noted in FIG. 7A, the level of Ang-2 correlated (as
noted on the y axes in ng/ml) when compared with the MOD score (as
noted on the x axis) as a predictor of mortality, with a
statistical significance of p<0.0001 as tested using as a single
biomarker was demonstrated. The ability of the Ang-2 levels to act
as an earlier predictor of mortality was analyzed by similarly
comparing the level of Ang-2 (ng/ml) taken from patients one day
prior to the evaluation of the patient as determined by MOD score.
As can be seen in FIG. 7B, Ang-2 levels measured on day x predicted
the clinical condition on the next hospital day (i.e day x+1).
There was a strong statistical correlation (P<0.0001) between
the Ang-2 levels performed on day x compared to the MOD score on
the next hospital day (day x+1), indicating Ang-2 is an earlier
indicator of disease progression and risk of mortality than the
current standard of the MOD score.
Example 8 Individual Biomarkers and Biomarker Combinations
Predictive of Patients of Having Hemolytic Uremic Syndrome as a
Result of an E. Coli Infection
[0124] A population-based surveillance study for E. coli O157:H7
infection in children less than 10 years of age was undertaken in
Washington, Oregon, Idaho, and Wyoming through mandatory laboratory
reporting of positive stool cultures. Seventy-eight children,
enrolled between 1998 and 2005, from whom a positive stool culture
was obtained within the first 7 days of illness were included for
this analysis. Phlebotomy was conducted at enrollment and as
clinically indicated thereafter. HUS was diagnosed as hemolytic
anemia (a hematocrit <30% with evidence of schistocytes on
peripheral blood film), thrombocytopenia (platelet count <150
000/mm3), and renal insufficiency (serum creatinine above the
age-adjusted upper limit of normal); participants who had not met
these criteria by day 14 of illness were considered to have had
uncomplicated infection.
[0125] 84 serum samples were tested: 26 from patients on the day of
diagnosis of HUS, 8 from patients who would subsequently be
diagnosed with HUS but had not yet met diagnostic criteria
(pre-HUS), and 50 from patients with uncomplicated infection. Six
patients had samples taken both prior to (pre-HUS) and on the day
of HUS diagnosis.
[0126] Serum samples were stored in aliquots at -80.degree. C.
until use. To measure angiopoietin levels in cell culture
supernatant, HMVEC were grown to confluence in complete medium in
6-well plates. Complete medium was replaced with basal medium
lacking serum and growth factors on the day of toxin treatment.
Shiga toxin or vehicle was added 4 hours later, and aliquots of
medium were taken at 24 hours following toxin addition, centrifuged
to remove dead cells, and likewise stored at -80.degree. C. until
use.
[0127] Serum and supernatant concentrations of Ang-1 and Ang-2 were
measured by ELISA (R&D.RTM. Systems, Minneapolis Minn.) as per
the manufacturer's instructions. The technical upper limits of
detection were 10,000 pg/mL for Ang-1 and 2520 pg/mL for Ang-2,
yielding effective upper limits of detection of 200,000 pg/mL and
10.080 pg/mL, respectively, for the dilutions employed in the
assay. Lower limits of detection for the assay were 9.77 pg/mL for
Ang-1 and 2.46 pg/mL for Ang-2.
[0128] Angiopoietin dysregulation (decreased Ang-1 and increased
Ang-2) was found to be associated with illness severity. The median
serum Ang-1 concentration in patients with uncomplicated infection
was significantly higher than in those patients with HUS (77, 357
pg/mL [interquartile range (IQR): 53, 437-114, 889 pg/mL] versus
10, 622 pg/mL [IQR: 3464-43, 523 pg/mL]), P<0.001 (FIG. 8A).
Conversely, the median serum Ang-2 concentration was significantly
lower in those with uncomplicated infection than in those with HUS
(1140 pg/mL [IQR: 845-1492 pg/mL] versus 1959 pg/mL [IQR: 1057-2855
pg/mL]), P<0.05 Finally, the Ang-2:Ang-1 ratio was 0.014 (IQR:
0.011-0.023) in patients with uncomplicated infection, and more
than 10-fold higher, at 0.18 (IQR).
[0129] In addition, the serum Ang-1 concentration at the time of
presentation to hospital effectively discriminated between two
populations of clinically indistinguishable children: 1) those with
uncomplicated hemorrhagic colitis and 2) those with hemorrhagic
colitis who would eventually develop HUS (Area under the Receiver
operating characteristic (ROC) curve [AUC]: 0.785, 95% confidence
interval (CI): 0.641-0.923; P=0.01) (FIG. 8B).
[0130] The serum Ang-1 and Ang-2 concentrations reported here for
children with uncomplicated infection are comparable to those found
in the serum of healthy children and adults, and are in keeping
with the clinical observation that there is little if any
endothelial activation present in these patients. In contrast, the
relative deficit of Ang-1 and excess of Ang-2 found in children
with HUS is in keeping with what is anticipated to be significant
endothelial cell activation in these patients.
Example 9 Individual Biomarker of Outcome in Pre-Diagnosed Malaria
and Use in Conjunction with Other Clinical Indicators of
Outcome
[0131] A retrospective case-control study was performed for
children presenting with fever to the Queen Elizabeth Centre
Hospital in Blantyre, Malawi. Children were between 6 months and 14
years of age and recruited between the years 1997 and 2009. EDTA
Plasma samples were obtained subsequent to obtaining informed
consent. Children were characterized based on their status with
respect to Cerebral malaria (CM) and also based on retinal
indicators such as hemorrhages, retinal whitening, or vessel
abnormalities. EDTA Proteins isolated from Plasma samples were
subject to ELISAs to quantify the levels of various potential
biomarkers including Ang-2. Ang-1, and sTie-2.
[0132] Comparisons of continuous variables were performed using the
Mann-Whitney U test and Spearman rank correlation coefficient.
Comparisons of proportions were performed using the Person
chi-square test, linear by linear association, or Fisher's exact
test. Odds rations (ORs_were calculated using Pearson chi-square or
logistic regression models to adjust for covariates. Bonferroni
adjustments were used to account for multiple comparisons.
[0133] Logistic regression and CRT analysis was used to generate
prognostic models using routine clinical parameters in combination
with the protein biomarkers. A clinically predictive model of
mortality was generated using solely the clinical parameters
readily available (Age, BCS, respiratory distress, severe anemia),
and probabilities from this clinical model were used to generate a
c-index (equivalent to the area under the receiver operating
characteristic curves) of 0.73 (95% confidence interval [CI],
0.65-0.79) (data not shown).
[0134] Using these clinical model as a foundation, biomarker tests,
either individually or in combination, were added to determine
whether the biomarkers would significantly improve the predictive
accuracy of the clinical parameters model alone. When the clinical
model was combined with all three biomarkers Ang-1, Ang-2 and
sTie-2, the resulting model had a c-index of 0.79 (95% confidence
interval [CI], 0.72-0.84) which was significantly better than the
clinical model alone (p=0.03) (data not shown)
Example 10 Diagnosis of a Test Individual Using Biomarker
Combination Predictive of a Critical and/or a Life Threatening
Response
[0135] Classifiers of the invention are generated using the
detected levels of protein biomarkers Ang-1, Ang-2, IP10 and CHI3L1
in a population of individuals who demonstrate a critical and/or
life threatening response to illness as compared with the detected
levels of protein biomarkers Ang-2, IP10 and CHI3L1 in a control
population of individuals who are normal. Logistic regression is
applied to differentiate the two populations and generates an
equation which has a sensitivity of 90% and a specificity of
95%.
[0136] Levels of protein biomarkers Ang-2, IP10 and CH13L1 are
determined using a standard ELISA test on a serum sample from a
test individual who may potentially have been exposed to an E. coli
infection, but has not yet been diagnosed with an E. coli
infection. In accordance with the logistic regression equation
generated from the classifier as described, the test individual is
classified as either having or not having a critical and or life
threatening response to illness.
Example 11 Determining the Likelihood of a Test Individual Having a
Critical and/or Life Threatening Response to Disease Using
Biomarker Combination Predictive of a Critical and/or a Life
Threatening Response Despite the Test Individual not being
Diagnosed or Differentially Diagnosed
[0137] Protein levels of the biomarkers noted in Table 1 are
detected in whole blood samples from a population of individuals,
wherein the individuals have a critical illness selected from the
list of malaria, toxic shock syndrome, Group A streptococcal
disease, sepsis, and an E. Coli infection, but where the
individuals do not develop a critical or life threatening response
to the critical illness Protein levels of the biomarkers noted in
Table 1 are also detected in whole blood samples from a second
population of individuals, where the individuals do develop a
critical response to an illness which is selected from the list of
malaria, toxic shock syndrome. Group A streptococcal disease, and
an E. Coli infection. Classifiers are generated using the data
generated from the two populations, in particular ELISA testing is
done on the whole blood samples for each individual of each
population using the antibodies noted in Table 2, and logistic
regression is applied to differentiate the two populations. For
each equation generated, wherein the area under the curve indicates
a sensitivity of greater than 90% and a sensitivity greater than
90%, the classifier is utilized to determine the likelihood that a
test individual suspected of having malaria is likely to have a
critical or life threatening response and should be treated as if
the individual has severe malaria. Those individuals identified are
treated intravenously with drugs and fluids in accordance with the
gold standard treatment for severe malaria as dictated by North
American hospitals.
Example 12 Determining the Likelihood of a Test Individual Having a
Critical and/or Life Threatening Response to Disease Using
Predictive Biomarker Combinations with a Test Individual Suspected
of Having Malaria
[0138] A serum sample is taken from a test individual suspected of
having been exposed to malaria, and displaying flu like symptoms.
ELISA testing is done on the serum sample using each of the
antibodies noted in Table 2. The results of the ELISA testing are
used in conjunction with the biomarker combinations noted in Table
5 and Table 6, and for each biomarker combination, a biomarker
score was determined as done in Example 2 using a one point for
each biomarker of the biomarker combination, wherein the point was
assigned if the measured value was greater than the corresponding
cut-point as determined in Example 2. The results of each biomarker
combination being indicative (with varying degrees of sensitivity
and specificity) whether the test individual has an increased
likelihood of having severe malaria and should be treated
accordingly.
Example 13 Determining the Likelihood of a Test Individual Having a
Critical and/or Life Threatening Response to Disease Using a Test
Individual Suspected of Having Pneumonia
[0139] A serum sample is taken from a test individual suspected of
having pneumonia. ELISA testing is done on the serum sample using
each of the antibodies noted in Table 2 and determining a level of
protein selectively hybridizing to the antibody in the serum
sample. The resulting data is used in conjunction with the
biomarker combinations noted in Example 4, and the levels of
protein in the test sample compared to the levels of protein for
each biomarker of the biomarker combinations in a population of
individuals who have been determined to have pneumonia and have not
developed a life threatening response, and a population of
individuals who have been determined to have pneumonia and have
developed a life threatening response. The biomarker level of said
test individual is compared with said biomarker level in the two
control populations for each biomarker of the combination, and the
combined result is analyzed to determine whether the test
individual is more akin to the control population having pneumonia
and not developing a life threatening response and the control
population having been diagnosed as having pneumonia and developing
a life threatening response, wherein the results being more akin to
the control population having pneumonia and developing a life
threatening response is indicative of the test individual having an
increased likelihood of having or developing a life threatening
response to pneumonia.
Example 14 Determining the Likelihood of a Test Individual Having a
Critical and/or Life Threatening Response to Disease Using a Test
Individual Suspected of Having an E. coli Infection
[0140] A whole blood sample is taken from a test individual
suspected of having an E. coli infection as a result of exposure to
a tainted water supply. As a result of inadequate testing
facilities, the test individual is not diagnosed for Hemolytic
Uremic Syndrome, and is not tested to confirm an E. coli infection.
ELISA testing is done on the serum sample using the antibodies
noted in Table 2 and determining a level of each protein in the
sample corresponding to the biomarkers noted in Table 1. Protein
levels of the biomarkers noted in Table 1 are utilized with
classifiers generated from comparing the levels of said biomarkers
as determined from two separate populations, a population of
individuals who have E. coli infections, but do not develop
Hemolytic Uremic Syndrome, and a population of individuals who have
E. coli infections and have Hemolytic Uremic Syndrome. Classifiers
are chosen which have a sensitivity of greater than 90% and a
sensitivity greater than 90%. The test individual is subsequently
treated for Hemolytic Uremic Syndrome if results of the classifiers
indicate the sample is sufficiently akin to the population of
individuals developing Hemolytic Uremic Syndrome.
Example 15 Determining the Likelihood of a Test Individual Having
Pneumonia Using Agnostic Biomarkers, Individually, and in
Combination, as Shown by the Ability to Differentiate Children
Presenting with Cough and Fever Who have Pneumonia (CXR+) as
Compared with Children Having Clinical Pneumonia Using WHO
Standards
[0141] A prospective study was done with Children presenting to a
community health facility in Africa with fever and upper
respiratory tract symptoms. ELISA testing was done on the serum
samples from these children using antibodies against the following
panel of nine biomarkers selected from Table 1: CRP, PCT, sTie-2,
Endoglin, P-selectin, vWF, CHI3L1, IL18bpa, and Angiopoietin-like
protein 3. The nine biomarkers individually, and in combinations,
were tested for their ability to differentiate between children
later diagnosed as having pneumonia using the north American gold
standard of a chest x-ray (CXR+Pneumonia) (n=30) or children later
diagnosed as having pneumonia by applying WHO Standards of clinical
pneumonia, but did not show pneumonia by chest x-ray (CXR-Pneumonia
n=90). WHO Standards for determining pneumonia rely on a
determination of Tachypnea as determined by measuring respiratory
rates taking into account the age of the child as follows: a
respiratory rate >60 breaths/minute in children <2 months of
age, >50 breaths/minute in children 2 to 12 months of age, and
>40 breaths/minute in children .gtoreq.1 year of age. Children
who were neither CXR+ Pneumonia or CXR- Pneumonia were classified
as having an upper respiratory infection which was not pneumonia
(URTI) (n=90).
[0142] Demographic and clinical characteristics of all children who
presented with fever and upper respiratory tract symptoms are shown
in Table 7.
TABLE-US-00009 TABLE 7 Demographic/clinical characteristics of CXR+
pneumonia, CXR- pneumonia Clinical pneumonia CXR+ pneumonia (CXR-)
n = 30 n = 90 Age (months) 19.4 (3.6, 100.0) 14.6 (2.3, 112.8)
Gender (% male) 36.7% 61.1% Study site (% Dar es 46.7% 36.7%
Salaam) Temperature (.degree. C.) 38.7 (38.0, 40.5) 38.4 (38.0,
40.4) Days of fever prior to 2.5 (1-5) 3 (1-6) presentation
Respiratory rate (/min) 53 (32-90) 50 (40-70) Heart rate (/min)
129.5 (84-169) 124 (91-180) Severe (%) .sup.+ 30.0% 23.3%
Hemoglobin (g/dL) 9.4 (5.5, 17.9) 9.7 (3.8, 13.6) Leukocyte count
(.times.10.sup.9/mL) 23.6 (7.4, 38.7) 11.7 (3.5, 49.9) ***
Neutrophil count (units) 63.4 (35.0, 87.9) 49.8 (8.7, 83.2) **
Continuous variables are represented as: Median (range) .sup.+
indicates symptoms considered "severe" in accordance with WHO
Integrated management of childhood Illness (IMCI) standards.
[0143] Each biomarker was individually tested for its ability to
discriminate between CXR+ pneumonia (pneumoma confirmed by chest
x-ray), and CXR- clinical pneumonia (classified as pneumonia
according to WHO standards, but negative for pneumonia as
determined by chest x-ray). The results of the bivariate analysis
are shown in Table 8.
TABLE-US-00010 CXR+ vs. Clinical pneumonia Biomarker.sup.& n
Odds Ratio (Cl) p-value* Ang-L-3 (ng/mL, log) 120 1.01 (0.39, 2.60)
0.984 CHI3L1 (ng/mL, log) 120 3.30 (1.87, 5.83) <0.001* CRP
(.mu.g/mL, log) 120 3.20 (2.01, 5.11) <0.001* sEndoglin (ng/mL)
120 0.91 (0.79, 1.05) 0.186 IL-18 BP (ng/mL, log) 120 1.23 (0.62,
2.44) 0.546 PCT (ng/mL, log) 120 1.80 (1.27, 2.55) 0.001* pSelectin
(ng/mL, log) 120 1.39 (0.94, 2.04) 0.006 sTie-2 (ng/mL) 120 0.87
(0.32, 2.35) 0.784 vWF (ug/mL) 120 1.38 (0.82, 2.32) 0.231 Age (log
month) 120 1.86 (1.06, 3.24) 0.038 Temperature 120 1.05 (0.99,
1.12) 0.125 Heart rate 120 1.00 (0.99, 1.02) 0.602 Respiratory rate
120 1.01 (0.96, 1.06) 0.742 Site (Dar Es Salaam vs. Ifakara) 120
1.33 (0.65, 3.48) 0.333 WBC 119 7.35 (2.74, 19.73) <0.001* Male
120 0.37 (0.16, 0.87) 0.022* .sup.&Treated as continuous and
tested using logistic regression. For all biomarkers except
sEndoglin, log transformed variables were used. .sup.% P-values in
bold represent statistically significant markers (p < 0.05).
After accounting for multiple comparisons of hypothesized
biomarkers, p-value .ltoreq. 0.0056 (0.05/9) marked with *. .sup.#
Analyzed age, temperature, fever duration, heart rate, respiratory
rate, hemoglobin, WBC, ALT, sex, site, convulsions, dehydration,
jaundice, palm pallor, chest indrawing, nose flapping, grunting,
chest auscultation, wheezing, date, HIV status.
[0144] Individually, biomarkers CRP. PCT. CHI3L1 and P-selectin
were found both by univariate analysis, and by Mann Whitney (data
not shown) to differentiate between the two groups of children.
Table 9 shows the diagnostic cut-off points of each of CRP, PCT,
CHI3L1 and P-selectin as determined by the Receiver Operator Curves
(ROC Curves), and the sensitivity and specificity of the individual
biomarkers. Sensitivity (Sens) and Specificity (Spec) of the
combination, along with the positive likelihood ratio (PLR),
negative likelihood ration (NLR) positive predictive value (PPV)
and negative predictive value (NPV) are shown.
TABLE-US-00011 TABLE 9 Cut- AUC* point** Sens Spec PLR NLR PPV NPV
CHI3L1 0.80 >57.0 93.3 64.4 2.6 0.10 39.6 97.5 ng/mL CRP 0.86
>45.9 80.0 81.1 4.2 0.25 51.4 94.2 ug/mL PCT 0.71 >0.51 70.0
70.0 2.3 0.43 36.8 90.3 ng/mL Pselectin 0.62 >59.0 70.0 62.2 1.9
0.48 31.7 89.2 ng/mL *AUC = area under the ROC curve **Cut-points
based on Youden index: J = max{sens + spec - 1}
[0145] Additionally classification and regression tree analysis
(CRT) was performed to demonstrate the utility of the biomarker
combinations of the tested nine biomarkers. While it is anticipated
that numerous combinations and variations have utility, various
criteria were set including the number of biomarkers from which to
select the optimum combination, whether to select a specific
cut-off point for any given biomarker (e.g. dichotomize) or to
allow a continuous range, the minimum number of nodes and maximum
levels per tree to avoid overfitting, and the ability to set
misclassification costs to preferentially avoid e.g. false so as to
select preferential biomarker combinations. Table and Table 11 show
various examples of combination models chosen on the basis of
varied input, and show the Sensitivity (Sens) and Specificity
(Spec) of the combination, along with the positive likelihood ratio
(PLR), negative likelihood ration (NLR) positive predictive value
(PPV) and negative predictive value (NPV). Similarly, FIG. 9
demonstrates the CRT analysis of the combination of Model 1 in
Table 10 in a tree format wherein the combinatorial power added by
each biomarker in differentiating between the two populations (CXR+
pneumonia as compared to Clin pneumonia (or CXR- pneumonia) is
shown.
TABLE-US-00012 TABLE 10 Biomarker entered Other Markers Sens Spec
PLR NLR PPV NPV 1 All 9, Nodes: 10 CRP, 93.3 76.7 4.0 0.1 50.0 97.9
continuous parent, 5 child Endoglin, variables 3x misclassification
Pselectin cost** 2 All 9 Nodes: 10 CRP, 86.7 81.1 4.6 0.2 53.4 96.1
continuous parent, 5 child Endoglin variables 3x misclassif. Cost
Tree limited to 2 levels*** 3 All 9, Nodes: 10 CHI3L1, 93.3 74.4
3.6 0.1 47.7 97.8 continuous parent, 5 child| CRP variables 5x
misclassif. Cost| Tree pruned 4 All 9, Nodes: 10 CRP, 90.0 81.1 4.8
0.1 54.3 97.0 continuous parent, 5 child| CHI3L1, except 2x
misclassif. Endoglin dichotomized Cost CRP, PCT* Tree limited to 2
levels 5 All 9; Nodes: 10 CRP, 80.0 85.6 5.6 0.2 58.1 94.5
continuous parent, 5 child CHI3L1 except 2x misclassif.
dichotomized Cost CRP, PCT* Tree pruned *Dichotomized CRP (40
ug/mL) and PCT (0.5 ng/mL) because POC tests already exist at these
cut-offs **Misclassification costs always in favour of increased
sensitivity for CXR+ pneumonia ***Truncated Model 1 after 2
splits
TABLE-US-00013 TABLE 11 Biomarkers Selected Model entered Model
Parameters Markers Sens Spec PPV NPV 1 Chi3L1, CRP Nodes: 10
parent, 5 child Chi3L1, CRP 70.1 91.1 72.4 90.1 1x
misclassification cost.sup.1 2 Chi3L1, PCT Nodes: 10 parent, 5
child Chi3L1, PCT 53.3 93.3 72.7 85.7 1x misclass. cost 3 Chi3L1,
Tie-2 Nodes: 10 parent, 5 child Chi3L1, Tie-2 40.0 97.8 85.7 83.0
1x misclass. cost 4 Chi3L1, vWF Nodes: 10 parent, 5 child Chi3L1,
vWF 73.3 84.4 61.1 90.5 1x misclass. cost 5-7 a. Chi3L1, Nodes: 10
parent, 5 child Chi3L1.sup.2 53.3 98.9 87.5 79.5 Pselectin 1x
misclass. cost b. Chi3L1, Endoglin c. Chi3L1, IL18bpa 8 Chi3L1,
CRP, Nodes: 10 parent, 5 child CRP 70.0 91.1 72.4 90.1 PCT 1x
misclass. cost 9 Chi3L1, CRP, Nodes: 10 parent, 5 child Chi3L1, CRP
56.7 96.7 85.0 87.0 Pselectin 1x misclass. cost 10 Chi3L1, PCT,
Nodes: 10 parent, 5 child Chi3L1, PCT 53.3 93.3 72.7 85.7
Pselectin, 1x misclass. cost Endoglin 11 All markers.sup.3 Nodes:
10 parent, 5 child CRP, Tie2, 70.0 91.1 72.4 90.1 1x misclass. Cost
Ang3L1 Tree pruned 12 All markers.sup.3 Nodes: 10 parent, 5 child
CRP, Endoglin, 93.3 76.7 50.0 97.9 3x misclass. cost Pselectin 13
All markers.sup.3 Nodes: 10 parent, 5 child CRP, Endoglin 86.7 81.1
53.4 96.1 3x misclass. Cost Tree limited to 2 levels.sup.4 14 All
markers.sup.3 Nodes: 10 parent, 5 child CHI3L1, CRP 93.3 74.4 47.7
97.8 5x misclass. cost Tree pruned 15 All 9, continuous Nodes: 10
parent, 5 child CRP, CHI3L1, 90.0 81.1 54.3 97.0 except 2x
misclassif. cost Endoglin dichotomized Tree limited to 2 levels
CRP, PCT.sup.5 16 All 9; continuous Nodes: 10 parent, 5 child CRP,
CHI3L1 80.0 85.6 58.1 94.5 except 2x misclassif. cost dichotomized
Tree pruned CRP, PCT.sup.5 .sup.1Misclassification costs always in
favour of increased sensitivity for CXR+ pneumonia .sup.2Other
biomarkers not selected in model. .sup.3All markers: Chi3Ll, CRP,
PCT, Endoglin, P-selectin, vWVF, Ang3Ll, Tie-2, IL18bpa
.sup.4Truncated Model 1 after 2 splits .sup.5Dichotomized CRP (40
ug/mL) and PCT (0.5 ng/mL) because POC tests already exist at these
cut-offs
Example 16 Determining the Likelihood of a Test Individual Having
Pneumonia Using Agnostic Biomarkers, Individually, and in
Combination, as Shown by the Ability to Differentiate Children
Presenting with Cough and Fever Who have Pneumonia (CXR+) as
Compared with Children Having Other Upper Respiratory Infections
(URI)
[0146] As described in Example 15, a prospective study was done
with Children presenting to a community health facility in Africa
with fever and upper respiratory tract symptoms. ELISA testing was
done on the serum samples from these children using antibodies
against the following panel of nine biomarkers selected from Table
1: CRP, PCT, sTie-2, Endoglin, P-selectin, vWF, CHI3L1, IL18bpa,
and Ang3L1. The nine biomarkers individually, and in combinations,
were tested for their ability to differentiate between children
later diagnosed as having pneumonia using the north American gold
standard of a chest x-ray (CXR+ Pneumonia) (n=30) or age, sex,
clinical site and date matched children having upper respiratory
infections not confirmed as pneumonia (n-90). Single parameter
biomarkers were evaluated and compared for ability to differentiate
between (a) CXR+pneumonia vs. CXR-clinical pneumonia and (b)
CXR+pneumonia vs. other upper respiratory tract infections (URTI)
with the exclusion of bronchiolitis (ARIs). Results are shown in
Table 12.
TABLE-US-00014 TABLE 12 CXR+ vs. Clinical pneumonia CXR+ vs. Other
ARIs.sup.$ Odds Ratio Odds Ratio Biomarker.sup.& n (CI)
p-value* n (CI) p-value* Ang-Like-3 120 1.01 0.984 120 1.27 0.63
(ng/mL) (0.39, 2.60) (0.48, 3.4) CHI3L1 120 3.30 <0.001* 120
4.39 <0.001* (ng/mL) (1.87, 5.83) (2.10, 9.18) CRP 120 3.20
<0.001* 120 3.36 <0.001* (.mu.g/mL) (2.01, 5.11) (1.88, 6.00)
sEndoglin 120 0.91 0.186 120 0.99 0.803 (ng/mL) (0.79, 1.05) (0.89,
1.10) IL-18 bp 120 1.23 0.546 120 1.54 0.221 (ng/mL) (0.62, 2.44)
(0.77, 3.05) PCT 120 1.80 0.001* 120 1.93 0.002* (ng/mL) (1.27,
2.55) (1.28, 2.91) p-Selectin 120 1.39 0.006 120 1.51 0.067 (ng/mL)
(0.94, 2.04) (0.97, 2.35) sTie-2 120 0.87 0.784 120 3.01 0.049
(ng/mL) (0.32, 2.35) (1.00, 9.01) vWF 120 1.38 0.231 120 1.43 0.196
(ug/mL) (0.82, 2.32) (0.83, 2.45) Age_log.sup.# 120 1.86 0.030 120
1.54 0.436 (1.06, 3.24) (0.52, 4.52) (matching variable)
Temperature 120 1.05 0.125 n/a (0.99, 1.12) Heart rate 120 1.00
0.602 120 1.35 <0.001* (0.99, 1.02) (1.16, 1.57) Respiratory 120
1.01 0.742 120 1.05 <0.001* rate (0.96, 1.06) (1.03, 1.08) WBC
119 7.35 <0.001* 119 8.45 <0.001* (2.74, 19.73) (2.91, 24.55)
Male 120 0.37 0.022* 120 n/a (0.16, 0.87) (matched on sex)
.sup.&Treated as continuous. For all markers except sEndoglin,
log transformed variables were used. Patients in CXR+ vs CXR-
analysis were an unmatched and logistic regression was used. For
CXR+ vs. URTI analysis, conditional logistic regression was used
because patients were matched on age, sex, site and date. .sup.$
Excludes bronchiolitis .sup.%P-values in bold represent
statistically significant markers (p < 0.05). After accounting
for multiple comparisons, p-value .ltoreq. 0.0056 (0.05/9) marked
as *. .sup.#Analyzed age, temperature, fever duration, heart rate,
respiratory rate, hemoglobin, WBC, ALT, sex, site convulsions,
dehydration, jaundice, palm pallor, chest indrawing, nose
flapping.
[0147] Similar individual biomarkers were able to differentiate
CXR+ pneumonia as compared with other upper respiratory tract
infections including biomarkers CRP, PCT, and CHI3L1. P-selectin
was also identified as a statistically significant individual
biomarker when analyzed by Mann Whitney (p=0.044) (data not shown),
but not when analyzed utilizing Kruksill Wallis analysis with
Dunn's post tests (data not shown). In all cases URTI and CXR-
clinical pneumonia were indistinguishable using individual
biomarkers (data not shown). Table 13 shows the diagnostic cut-off
points of each of CRP, PCT, CHI3L1 and P-selectin as determined by
the Receiver Operator Curves (ROC Curves), and the sensitivity and
specificity of the individual biomarkers when comparing CXR+
pneumonia vs. URTI. Sensitivity (Sens) and Specificity (Spec) of
the combination, along with the positive likelihood ratio (PLR),
negative likelihood ration (NLR) positive predictive value (PPV)
and negative predictive value (NPV) are shown.
TABLE-US-00015 TABLE 13 Cut- AUC* point** Sens Spec PLR NLR PPV NPV
CHI3L1 0.80 >57.0 93.3 66.1 2.8 0.1 15.2 99.3 ng/mL CRP 0.87
>31.4 86.7 73.9 3.3 0.2 17.7 98.8 ug/mL PCT 0.70 >0.51 70.0
65.6 2.0 0.5 11.7 97.1 ng/mL Pselectin 0.62 >59.2 70.0 61.7 1.8
0.5 10.6 96.9 ng/mL
[0148] As in Example 15, additional classification and regression
tree analysis (CRT) was performed to demonstrate a selection of
biomarker combinations of the tested nine biomarkers. Table 14 show
selected combination models chosen on the basis of varied input,
and show the Sensitivity (Sens) and Specificity (Spec) of the
combination, along with the positive likelihood ratio (PLR),
negative likelihood ration (NLR) positive predictive value (PPV)
and negative predictive value (NPV).
TABLE-US-00016 TABLE 14 Biomarkers Model entered Other Markers Sens
Spec PLR NLR PPV NPV 1 All 9, Nodes: 20 CRP, 86.7 86.1 6.2 0.2 28.8
99.0 continuous parent, 10 child CHI3L1 variables 5x
misclassification cost** 2 All 9, Nodes: 20 CRP, 80.0 88.3 6.8 0.2
30.8 98.5 continuous parent, 10 child CHI3L1 except 5x
misclassification dichotomized cost** CRF PCT* 3 All 9 Nodes: 20
CRP, 93.3 81.1 4.9 0.1 24.3 99.5 continuous parent, 10 child CHI3L1
except 10x misclassification dichotomized cost CRP, PCT* 4 All 9,
Nodes: 10 CRP, 80.0 91.1 9.0 0.2 36.9 98.6 continuous parent, 5
child CHI3L1, except 5x misclassitication IL18bpa dichotomized cost
CRP, PCT* *Dichotomized CRP (40 ug/mL) and PCT (0.5 ng/mL) because
POC tests already exist at these cut-offs **Misclassification costs
always in favour of increased sensitivity for CXR+ pneumonia
Example 17 Determining Whether Agnostic Biomarkers, Individually
and in Combination are Able to Differentiate Between Individuals
Having a Bacterial Infection which is Treatable with Antibiotics,
and Individuals Having a Viral Infection for which Antibiotics are
not Likely to be Effective
[0149] A prospective study was done with a group of children (n=15)
presenting to a community health facility in Africa with fever and
upper respiratory tract symptoms. ELISA testing was done on the
serum samples from these children using antibodies against the
following panel of nine biomarkers selected from Table 1: CRP, PCT,
sTie-2, Endoglin, P-selectin, vWF, CHI3L1. IL18bpa, and Ang3L1. The
children were later identified as either (i) having bacteremia from
one of E. coli, S. aureus, S. flexneri, Salmonella, Streptococcus,
H. Influenzae, or Acinetobacter (n=16) or (ii) having a viral
infection from one of Epstein Barr virus (EBV) Cytomegalovirus
(CMV). Human herpes virus 6 (HHV6), parvovirus or mumps. Similar
results were seen when the bacterial infections included K.
pneumonia (data not shown).
[0150] Each biomarker was individually tested for its ability to
discriminate between children having a bacterial infection which is
treatable with antibiotics, and children having a viral infection
that would not respond to antibiotics. The results of the bivariate
analysis are shown in Table 15
TABLE-US-00017 TABLE 15 Cut-point Sensitivity Specificity PPV NPV
Biomarker (Youden) (%) (%) (%) (%) Endoglin <12.5 ng/mL 73.3 75
65 81.6 CHI3L1 >29.8 ng/mL 80 62.5 57.5 83.1 CRP >8.6 ug/mL
93.3 68.7 65.4 99.4 TREM1 >71.1 pg/mL 93.3 43.7 51.3 91.2 PCT
>0.4 ng/mL 60 100 100 79.8 P-selectin >48.4 ng/mL 86.7 75
68.7 89.9 ANGL3 >294.6 ng/mL 80 75 67 86 IP10 <477.8 ng/mL
66.7 81.2 69.3 79.4 IL18bpaa <25.5 ng/mL 93.3 62.5 61.2 93.7
[0151] Combinations of the nine biomarker were tested for their
ability to differentiate between children having bacterial
infections (treatable with antibiotics) and children having viral
infections (not benefitting from antibiotics) by applying
classification and regression tree analysis (CRT). For each
biomarker, a point was assigned if the value of the biomarker was
above the set cut-point (as determined using Youden Index). The sum
of all the points was calculated to determine the "biomarker
score". Table 16 show selected combination models chosen on the
basis of the optimal score cut-point. The Sensitivity (Sens) and
Specificity (Spec) of the combination, along with positive
predictive value (PPV) and negative predictive value (NPV) are
shown.
TABLE-US-00018 TABLE 16 Sensi- Speci- Cut-point tivity ficity PPV
NPV Combination (Youden) (%) (%) (%) (%) TREM1 + IL18bpa Score 2
86.7 87.5 81.5 91.2 PCT + END Score .gtoreq.1 93.3 75 70.3 94.7 (or
PCT + Psel) PCT + ANGL3 Score .gtoreq.1 100 75 71.7 100 PCT + IP10
Score .gtoreq.1 100 81.2 77.2 100 End + PCT + IL18bpa Score
.gtoreq.2 93.3 93.7 90.4 95.7 PCT + ANGL3 + Score .gtoreq.2 100
87.5 83.5 100 IL18bpa PCT + ANGL3 + IP10 Score .gtoreq.2 93.3 100
100 95.9
* * * * *