Redefining Cardiac Biomarkers in Predicting Mortality of Inpatients With COVID-19

Juan‐Juan Qin, Xu Cheng, Feng Zhou, Fang Lei, Gauri Akolkar, Jingjing Cai, Xiao-Jing Zhang, Alice Blet, Jing Xie, Peng Zhang, Ye-Mao Liu, Zizhen Huang, Ling-Ping Zhao, Lijin Lin, Xia Meng, Ming-Ming Chen, Xiaohui Song, Liangjie Bai, Ze Chen, Xingyuan Zhang, Xiang Da, Jing Chen, Qingbo Xu, Xinliang Ma, Rhian M. Touyz, Chen Gao, Haitao Wang, Li Liu, Weiming Mao, Pengcheng Luo, Youqin Yan, Ping Ye, Manhua Chen, Guohua Chen, Lihua Zhu, Zhi‐Gang She, Xiaodong Huang, Yufeng Yuan, Bing-Hong Zhang, Yibin Wang, Peter P. Liu, Hongliang LiHypertension 2020 · DOI

TL;DR

This study reveals that standard laboratory cutoffs for cardiac biomarkers significantly underestimate the risk of mortality in COVID-19 patients. By establishing new, lower thresholds, clinicians can better identify high-risk individuals and potentially improve patient outcomes.

What Problem Does This Paper Address?

The early stages of the COVID-19 pandemic saw a surge in research trying to understand its impact on various organ systems. While it was clear that the virus could cause significant damage, the specific role and predictive power of cardiac biomarkers in COVID-19 patients, particularly concerning mortality, remained unclear. Existing cardiac biomarker thresholds, established for conditions like heart attacks, might not be sensitive enough to detect the subtle cardiac injury associated with COVID-19, leading to a potential underestimation of risk and delayed intervention. This paper aims to clarify the prognostic value of common cardiac biomarkers in COVID-19 patients and to redefine their cutoff values for predicting mortality.

Methods

This was a multicenter retrospective cohort study involving 3219 patients diagnosed with COVID-19 admitted to nine hospitals in Hubei Province, China, between December 31, 2019, and March 4, 2020. The primary endpoint was 28-day all-cause mortality.

Key methodological aspects include:

  • Mixed-effects Cox Model: Used to assess the association between cardiac biomarker levels and the risk of 28-day mortality, while accounting for patient-level data and hospital site as a random effect. This model was adjusted for age, sex, and common comorbidities (hypertension, diabetes mellitus, coronary heart disease, and cerebrovascular disease).
  • Receiver Operating Characteristic (ROC) Curve Analysis: Employed to determine the optimal prognostic cutoff values for each cardiac biomarker. This analysis helped quantify the performance of each biomarker in discriminating between patients at high and low risk of mortality.
  • Normalization of Biomarker Levels: To account for variations in laboratory techniques and instruments across different hospitals, biomarker levels were normalized to their respective upper limits of normal (ULN).
  • Trajectory Analysis: The temporal relationship between cardiac biomarker elevation and inflammatory markers (CRP, neutrophil count, IL-6) was analyzed to understand the mechanistic progression of cardiac injury in relation to the overall disease course.

Datasets

The study utilized data from 3219 COVID-19 patients admitted to nine hospitals in Hubei Province, China. This data was collected from electronic medical records and included demographic information, comorbidities, clinical characteristics, chest CT scan findings, laboratory data (including cardiac biomarkers and inflammatory markers), and clinical outcomes. No new benchmarks were introduced; the study focused on re-evaluating existing biomarkers within the context of COVID-19.

Key Insights and Ideas

  • Standard Cutoffs are Insufficient: The most striking finding is that the "normal" laboratory thresholds for cardiac biomarkers are too high to effectively predict mortality in COVID-19 patients. This suggests that even mild elevations, which would be considered normal in other contexts, are highly significant in COVID-19.
  • Lowered Thresholds for Prognosis: The study proposes new, significantly lower cutoff values for hs-cTnI, CK-MB, (NT-pro)BNP, CK, and MYO that are much more predictive of 28-day mortality in COVID-19. These new thresholds are approximately 19%-50% of the currently recommended levels.
  • Early Warning Signs: Cardiac injury markers appear to be early indicators of severe disease progression. Their elevation often coincides with rising inflammatory markers like CRP and neutrophils, and precedes the rise of IL-6, suggesting a cascade where cardiac involvement is an early, critical event.
  • Cardiac Injury is a Major Driver of Mortality: The strong associations between elevated cardiac biomarkers and increased mortality highlight that cardiac involvement is not just a comorbidity but a significant factor contributing to poor outcomes in COVID-19.

Key Results

  • Hazard Ratios for Mortality: After adjusting for confounders, elevated cardiac biomarkers were strongly associated with increased 28-day mortality. The adjusted hazard ratios (HRs) were:
  • hs-cTnI: 7.12 (95% CI, 4.60–11.03)
    • (NT-pro)BNP: 5.11 (95% CI, 3.50–7.47)
    • CK-MB: 4.86 (95% CI, 3.33–7.09)
    • MYO: 4.50 (95% CI, 3.18–6.36)
    • CK: 3.56 (95% CI, 2.53–5.02)
  • New Cutoff Values: The study identified new prognostic cutoff values that were significantly lower than standard laboratory ULNs. For example, the new cutoff for (NT-pro)BNP was approximately 18.9% of its ULN.
  • Prognostic Performance: ROC analysis showed good performance for these biomarkers in predicting mortality. MYO had the highest AUC (0.83), followed by (NT-pro)BNP (0.81), hs-cTnI (0.78), and CK-MB (0.71).
  • Survival Curves: Kaplan-Meier curves clearly demonstrated significantly reduced survival rates for patients with elevated cardiac biomarkers above the newly established cutoffs compared to those with levels below these cutoffs.

Limitations and Future Work

  • Retrospective Design: The authors acknowledge that the retrospective nature of the study prevents establishing a causal link between cardiac injury and COVID-19 mortality.
  • Generalizability of Cutoffs: The established cutoffs were based on a Chinese population and require external validation in diverse cohorts.
  • Incomplete Data: Due to the urgent circumstances of the pandemic, not all cardiac injury-related examinations were consistently performed across all sites, potentially introducing confounding effects.
  • Sample Size for Subgroups: The study's findings might not be fully generalizable to patients with specific complications or comorbidities due to sample size limitations in subgroups.
  • Inpatient Focus: The study only included hospitalized patients, so findings may not apply to outpatients or individuals in isolation.
  • Selection Bias: There's a potential for selection bias, as patients with suspected heart injury might have been more likely to have cardiac biomarkers tested.

Future Work: The authors strongly advocate for additional prospective studies and randomized controlled trials to validate these new risk thresholds and to further investigate the precise impact of myocardial injury on COVID-19 prognosis. They emphasize the clinical utility of monitoring these biomarkers with the redefined cutoffs to identify high-risk patients and guide management.

Summary updated

Abstract

The prognostic power of circulating cardiac biomarkers, their utility, and pattern of release in coronavirus disease 2019 (COVID-19) patients have not been clearly defined. In this multicentered retrospective study, we enrolled 3219 patients with diagnosed COVID-19 admitted to 9 hospitals from December 31, 2019 to March 4, 2020, to estimate the associations and prognostic power of circulating cardiac injury markers with the poor outcomes of COVID-19. In the mixed-effects Cox model, after adjusting for age, sex, and comorbidities, the adjusted hazard ratio of 28-day mortality for hs-cTnI (high-sensitivity cardiac troponin I) was 7.12 ([95% CI, 4.60-11.03] <i>P</i><0.001), (NT-pro)BNP (N-terminal pro-B-type natriuretic peptide or brain natriuretic peptide) was 5.11 ([95% CI, 3.50-7.47] <i>P</i><0.001), CK (creatine phosphokinase)-MB was 4.86 ([95% CI, 3.33-7.09] <i>P</i><0.001), MYO (myoglobin) was 4.50 ([95% CI, 3.18-6.36] <i>P</i><0.001), and CK was 3.56 ([95% CI, 2.53-5.02] <i>P</i><0.001). The cutoffs of those cardiac biomarkers for effective prognosis of 28-day mortality of COVID-19 were found to be much lower than for regular heart disease at about 19%-50% of the currently recommended thresholds. Patients with elevated cardiac injury markers above the newly established cutoffs were associated with significantly increased risk of COVID-19 death. In conclusion, cardiac biomarker elevations are significantly associated with 28-day death in patients with COVID-19. The prognostic cutoff values of these biomarkers might be much lower than the current reference standards. These findings can assist in better management of COVID-19 patients to improve outcomes. Importantly, the newly established cutoff levels of COVID-19-associated cardiac biomarkers may serve as useful criteria for the future prospective studies and clinical trials.
BibTeX
@inproceedings{qin2020redefining,
  author = {Juan‐Juan Qin and Xu Cheng and Feng Zhou and Fang Lei and Gauri Akolkar and Jingjing Cai and Xiao-Jing Zhang and Alice Blet and Jing Xie and Peng Zhang and Ye-Mao Liu and Zizhen Huang and Ling-Ping Zhao and Lijin Lin and Xia Meng and Ming-Ming Chen and Xiaohui Song and Liangjie Bai and Ze Chen and Xingyuan Zhang and Xiang Da and Jing Chen and Qingbo Xu and Xinliang Ma and Rhian M. Touyz and Chen Gao and Haitao Wang and Li Liu and Weiming Mao and Pengcheng Luo and Youqin Yan and Ping Ye and Manhua Chen and Guohua Chen and Lihua Zhu and Zhi‐Gang She and Xiaodong Huang and Yufeng Yuan and Bing-Hong Zhang and Yibin Wang and Peter P. Liu and Hongliang Li},
  title = {{Redefining Cardiac Biomarkers in Predicting Mortality of Inpatients With COVID-19}},
  booktitle = {Hypertension},
  year = {2020},
  doi = {10.1161/hypertensionaha.120.15528},
}
Figure