|
TROPOPOP: Population models of high sensitivity cardiac troponins
Introduction to the tropopop project
This is a project to develop a population model of the chronic serum
levels cardiac troponin (cTn) in the general population. By `chronic
serum levels' we refer to the baseline, unchanging level that is
relatively constant, over weeks or months, in apparently healthy
subjects, or patients in a stable chronic condition. In this phase,
the project does not aim to model acute increases in the troponin
level, that occur, for example, during acute coronary syndromes, but
to provide a personalised expected range of cTn to which apparent
increases may be compared.
The primary anticipated use of this model is to support the
differential diagnosis of cTn increases, in particular small
increases, less then three times the upper reference limit (URL). Knowing
the expected range of the cTn in a particular normal or pathological
population with greater accuracy would allow the better identification
of unexpectedly high values and lead to the exploration of the cause
of the increase.
Methods
In short, we perform a meta-regression of studies and datasets that
characterise the general population, from which marginal probability
distributions (MPD) of the cTn for some variables, usually specific to each
study, in a section of the population that is also specific, are
extracted. The meta-regression will result in a joint probability
distribution (JPD) of the cTn, defined on a set of variables
(gender, age, morphologic parameters of the heart, other investigations,
comorbidities, treatment).
The project involves five types of activities:
- extraction of MPDs from models, aggregate datasets from the literature
and also by developing models of microdata where available;
- implementation of distinct MPDs as batch and online applications;
- development of the JPD based on some of the MPDs and implementation of the main application;
- extraction of new MPDs from the JPD for validation with other datasets;
- clinical studies of the predictive power of un/expected cTn results
in patients.
Results
So far (october 2024) we implemented two MPDs (marginal probability
distributions of the cTn), described below.
Tropopop2 model
This is the second version of an online
implementation based on the `H' regression model described in McKie et al, 2013. It has an interface that will
be used also in the final JPD model: the user provides the variables
that are known in a patient and the application computes the
probability distribution of the cTn in the subpopulation that has the
same values, assuming that the unknown variables have the same
distribution as in the general population, conditional on the known
ones. The 99th centile of this patient-specific distribution can be
used as a personalised URL for diagnostic purposes.
This MPD is limited to a few variables (age, gender, blood
pressure, left ventricular mass index and total serum cholesterol)
and, moreover, it is based on a population sample of adults aged
40-70. Extrapolations outside this range are possibly unreliable.
General population limits by age and gender
In the table below we propose preliminary upper reference limits, by age and gender,
for cardiac troponins T and I. These limits reflect the expected 99th
centile in each age and gender group. The line labeled ``Limits:''
represents the 99th centile limits designated by the manufacturer of
the testing kit for the whole male or female population. All values in
the table are in ng/L.
These limits are computed from an MPD constructed on the dataset
reported in McEvoy et al, 2023.
| MALES | FEMALES |
cTnT | cTnI |
cTnT | cTnI |
| Roche | Abbott | Siemens | Ortho |
Roche | Abbott | Siemens | Ortho |
Limits: | 22 | 35 | 58 | 12 | 14 | 17 | 39 | 9 |
Age | | |
1 | 11 | 10 | 25 | 17 | 10 | 10 | 24 | 17 |
2 | 10 | 8 | 23 | 12 | 9 | 8 | 23 | 12 |
3 | 10 | 8 | 21 | 9 | 9 | 7 | 21 | 9 |
4 | 9 | 7 | 20 | 7 | 9 | 7 | 20 | 7 |
5 | 9 | 6 | 19 | 6 | 9 | 6 | 19 | 5 |
6 | 9 | 6 | 19 | 5 | 9 | 5 | 18 | 4 |
7 | 9 | 6 | 19 | 4 | 9 | 5 | 18 | 4 |
8 | 9 | 5 | 19 | 4 | 9 | 5 | 17 | 3 |
9 | 9 | 5 | 20 | 3 | 9 | 4 | 17 | 3 |
10 | 10 | 5 | 21 | 3 | 9 | 4 | 16 | 3 |
11 | 10 | 5 | 22 | 3 | 9 | 4 | 16 | 2 |
12 | 11 | 6 | 24 | 4 | 9 | 4 | 16 | 2 |
13 | 11 | 6 | 26 | 4 | 9 | 4 | 16 | 2 |
14 | 12 | 7 | 30 | 5 | 9 | 4 | 16 | 3 |
15 | 13 | 7 | 34 | 6 | 9 | 4 | 17 | 3 |
16 | 14 | 8 | 34 | 5 | 9 | 5 | 15 | 3 |
17 | 14 | 8 | 34 | 5 | 9 | 5 | 15 | 3 |
18 | 14 | 8 | 34 | 5 | 9 | 5 | 15 | 3 |
19 | 14 | 9 | 34 | 5 | 9 | 5 | 15 | 3 |
20 | 14 | 9 | 34 | 5 | 9 | 5 | 15 | 3 |
21 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
22 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
23 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
24 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
25 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
26 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
27 | 14 | 9 | 34 | 5 | 9 | 5 | 16 | 3 |
28 | 14 | 9 | 34 | 5 | 9 | 5 | 17 | 3 |
29 | 14 | 9 | 34 | 6 | 9 | 6 | 17 | 3 |
30 | 14 | 10 | 34 | 6 | 9 | 6 | 17 | 3 |
31 | 15 | 10 | 34 | 6 | 9 | 6 | 17 | 3 |
32 | 15 | 10 | 34 | 6 | 9 | 6 | 17 | 3 |
33 | 15 | 10 | 35 | 6 | 10 | 6 | 17 | 3 |
34 | 15 | 10 | 35 | 6 | 10 | 6 | 18 | 3 |
35 | 15 | 10 | 35 | 6 | 10 | 6 | 18 | 3 |
36 | 15 | 10 | 35 | 6 | 10 | 6 | 18 | 3 |
37 | 16 | 10 | 35 | 6 | 10 | 6 | 18 | 3 |
38 | 16 | 11 | 36 | 7 | 10 | 7 | 19 | 3 |
39 | 16 | 11 | 36 | 7 | 10 | 7 | 19 | 4 |
40 | 16 | 11 | 36 | 7 | 11 | 7 | 19 | 4 |
41 | 17 | 11 | 37 | 7 | 11 | 7 | 20 | 4 |
42 | 17 | 11 | 37 | 7 | 11 | 7 | 20 | 4 |
43 | 17 | 12 | 37 | 7 | 11 | 7 | 20 | 4 |
44 | 18 | 12 | 38 | 8 | 12 | 8 | 21 | 4 |
45 | 18 | 12 | 38 | 8 | 12 | 8 | 21 | 4 |
46 | 18 | 12 | 39 | 8 | 12 | 8 | 22 | 4 |
47 | 19 | 13 | 39 | 8 | 12 | 8 | 22 | 5 |
48 | 19 | 13 | 40 | 9 | 13 | 8 | 22 | 5 |
49 | 20 | 13 | 40 | 9 | 13 | 9 | 23 | 5 |
50 | 20 | 13 | 41 | 9 | 13 | 9 | 23 | 5 |
51 | 21 | 14 | 41 | 9 | 14 | 9 | 24 | 5 |
52 | 21 | 14 | 42 | 10 | 14 | 9 | 25 | 5 |
53 | 22 | 14 | 42 | 10 | 15 | 10 | 25 | 6 |
54 | 22 | 15 | 43 | 10 | 15 | 10 | 26 | 6 |
55 | 23 | 15 | 44 | 11 | 16 | 10 | 26 | 6 |
56 | 24 | 15 | 44 | 11 | 16 | 10 | 27 | 6 |
57 | 25 | 16 | 45 | 12 | 17 | 11 | 28 | 7 |
58 | 25 | 16 | 46 | 12 | 17 | 11 | 28 | 7 |
59 | 26 | 16 | 47 | 13 | 18 | 11 | 29 | 7 |
60 | 27 | 17 | 48 | 13 | 18 | 12 | 30 | 8 |
61 | 28 | 17 | 49 | 14 | 19 | 12 | 31 | 8 |
62 | 29 | 18 | 50 | 14 | 20 | 12 | 32 | 8 |
63 | 30 | 18 | 51 | 15 | 21 | 13 | 33 | 9 |
64 | 31 | 19 | 52 | 16 | 21 | 13 | 34 | 9 |
65 | 32 | 19 | 53 | 16 | 22 | 14 | 35 | 10 |
66 | 34 | 20 | 54 | 17 | 23 | 14 | 36 | 10 |
67 | 35 | 20 | 55 | 18 | 24 | 15 | 37 | 11 |
68 | 36 | 21 | 56 | 19 | 25 | 15 | 38 | 11 |
69 | 38 | 21 | 58 | 20 | 26 | 16 | 39 | 12 |
70 | 40 | 22 | 59 | 21 | 28 | 16 | 41 | 12 |
71 | 41 | 23 | 60 | 22 | 29 | 17 | 42 | 13 |
72 | 43 | 23 | 62 | 23 | 30 | 17 | 43 | 14 |
73 | 45 | 24 | 63 | 24 | 32 | 18 | 45 | 15 |
74 | 47 | 25 | 65 | 26 | 33 | 19 | 46 | 16 |
75 | 49 | 26 | 67 | 27 | 35 | 19 | 48 | 16 |
76 | 52 | 26 | 69 | 29 | 37 | 20 | 50 | 17 |
77 | 54 | 27 | 70 | 30 | 38 | 21 | 52 | 19 |
78 | 57 | 28 | 72 | 32 | 40 | 22 | 53 | 20 |
79 | 60 | 29 | 74 | 34 | 43 | 22 | 56 | 21 |
80 | 63 | 30 | 77 | 36 | 45 | 23 | 58 | 22 |
81 | 66 | 31 | 79 | 38 | 47 | 24 | 60 | 24 |
82 | 70 | 32 | 81 | 40 | 50 | 25 | 62 | 25 |
83 | 73 | 34 | 84 | 43 | 53 | 26 | 65 | 27 |
84 | 77 | 35 | 86 | 46 | 56 | 27 | 67 | 29 |
85 | 82 | 36 | 89 | 49 | 59 | 28 | 70 | 31 |
References
[McKie et al, 2013]
Clin Chem. 2013 Jul;59(7):1099-107. doi:
10.1373/clinchem.2012.198614. Epub 2013 Apr 16. Defining
high-sensitivity cardiac troponin concentrations in the community Paul
M McKie, Denise M Heublein, Christopher G Scott, Mary Lou Gantzer,
Ramila A Mehta, Richard J Rodeheffer, Margaret M Redfield, John C
Burnett Jr, Allan S Jaffe PMID: 23592511 DOI:
10.1373/clinchem.2012.198614
[McEvoy et al, 2023] Eur Heart J. 2023 Jul 21; 44(28): 2595–2605.
Jun 2. doi: 10.1093/eurheartj/ehad328 PMCID: PMC10361011 PMID:
37264651 High-sensitivity troponins and mortality in the general
population John W McEvoy, Natalie Daya, Olive
Tang, Michael Fang, Chiadi E Ndumele, Josef Coresh, Robert H
Christenson, and Elizabeth Selvin
|
|