Diagnostic accuracy of high-sensitivity cardiac troponin in the diagnosis of acute myocardial infarction at the Aga Khan University hospital Nairobi, Kenya
Date of Award
Master of Medicine (MMed)
Dr. Geoffrey Omuse
Dr. Daniel Maina
Dr. Mzee Ngunga
Pathology (East Africa)
Methodology: This cross-sectional study was carried out at Aga Khan University Hospital, Nairobi between July 1st 2020 and December 31st 2021. Participants were adults aged 18 years and above presenting to the emergency department with signs and symptoms suggesting an acute myocardial infarction. Patients were identified on the hospital management information system (CARE 2000) using test results for baseline high-sensitivity cardiac troponin I. The final diagnosis as well as secondary clinical data was obtained from both physical and electronic patient medical records. The reference standard for the diagnosis of acute myocardial infarction was taken as the final clinical diagnosis made by the physician. The physician makes this diagnosis using a composite of the patient’s presenting signs and symptoms, risk factors, electrocardiography (ECG) findings and cardiac troponin concentrations as no single clinical or laboratory parameter can be used to accurately diagnose an AMI. Sensitivity, specificity, positive and negative predictive values were calculated using manufacturer recommended cut offs. Diagnostic accuracy was determined using receiver operating characteristic (ROC) curve with area under the curve (AUC). Youden index was used to determine the cut-off value that best predicted the diagnosis of acute myocardial infarction and linear regression analysis was used to determine the association between the clinical-pathological characteristics and troponin levels. A p-value < 0.05 was considered statistically significant.
Results: The high-sensitivity cardiac troponin I assay had a diagnostic accuracy of 0.97 (95% CI 0.955-0.982). The sensitivity and the specificity at the 99th percentile cut-off for both males and females was 90% (95% CI 87,91) and 93% (95% CI 85,98), respectively. The sensitivity and the specificity at the 99th percentile cut-off for males was 94% (95% CI 86,98) and 93% (95% CI 90,95), respectively while females had a sensitivity and specificity at the 99th percentile cut-off of 100% (95% CI 85,100) and 90% (95% CI 87,93), respectively. For both males and females, the negative and positive predictive values were 99%(95% CI 99,100) and 41% (95% CI 34,49), respectively. For the sex-specific cut-offs at the 99th percentile the negative and positive predictive values were 99% (95% CI 99,100), 69% (95% CI 60,77) for males, respectively and 100% (95% CI 99,100) and 34% (95% CI 23,47) for females, respectively. We subsequently obtained an optimal cut-off of 55.3 ng/l and found significant association between troponin levels and SARS-COV2 status and the HEART score.
Conclusion: The diagnostic accuracy of the Siemens Healthineers© high-sensitivity cardiac troponin I had good performance and was comparable to the manufacturer stated performance. This study derived a higher cut-off for the diagnosis of AMI in our setting. Before adopting our derived cut-offs, a multidisciplinary meeting with all relevant stakeholders at Aga Khan University Hospital, Nairobi needs to be held to evaluate the potential impact of such a change. The HEART score was higher in those diagnosed with AMI and therefore patients with a high HEART score should be investigated further to confirm or rule out a diagnosis of AMI. The higher high-sensitivity cardiac troponin I in patients with a positive SARS-COV-2 PCR supports the finding that SARS-COV-2 causes cardiac injury and the need to consider SARS-COV-2 as a differential when evaluating patients suspected to have acute coronary syndrome.
Chege, A. W. (2022). Diagnostic accuracy of high-sensitivity cardiac troponin in the diagnosis of acute myocardial infarction at the Aga Khan University hospital Nairobi, Kenya (Unpublished master's dissertation). Aga Khan University, East Africa.