Abstract
Type 2 myocardial infarction (type 2 MI) caused by demand ischemia and characterized by an imbalance between myocardial oxygen supply and demand without acute athero-thrombosis, is common in critically ill patients, including those with septic shock. This study aimed to evaluate the relationship between type 2 MI and inpatient mortality in patients hospitalized with septic shock using the National Inpatient Sample (NIS) from 2016 to 2020. A propensity-matched cohort was constructed, consisting of 31,905 hospitalizations with septic shock and type 2 MI (study group) and an equal number without type 2 MI (control group).
Overall, type 2 MI was not associated with a statistically significant difference in mortality (27% vs 26.1%, OR 1.05, 95% CI 0.97–1.14, p = 0.23). Subgroup analysis revealed an increased mortality risk in younger patients (aged 18–39) with type 2 MI (OR 1.93, 95% CI 1.09–3.43, p = 0.024). Conversely, patients with coronary artery disease (CAD) and type 2 MI demonstrated a lower mortality risk (OR 0.82, 95% CI 0.70–0.96, p = 0.012). Other comorbidities, including chronic heart failure, tachyarrhythmias, chronic kidney disease, and iron deficiency anemia, did not significantly alter mortality outcomes. These findings suggest that type 2 MI may be correlated with inpatient mortality only in specific subgroups of patients with septic shock, particularly younger patients and those with CAD. Our findings underscore the importance of demographic characteristics and comorbidities as key determinants of outcomes. Further research is needed to elucidate underlying mechanisms and improve clinical management in these populations.
Introduction
Demand ischemia is defined as an imbalance between myocardial oxygen supply and demand, unrelated to acute athero-thrombosis. 1 Despite its clinical significance, it is often overlooked once diagnosed. 2 Limited high-quality studies and the lack of diagnostic or therapeutic guidelines contribute to this oversight.3,4
Emerging evidence indicates that demand ischemia and its sequela, type 2 myocardial infarction (type 2 MI), despite occurring without a thrombotic event, are associated with poorer outcomes, particularly in critically ill patients.4,5 These patients frequently exhibit myocardial injury, as indicated by elevated cardiac troponin I levels, which is often unrecognized and associated with increased morbidity and mortality. 4 Demand ischemia and type 2 MI among critically ill patients are also independently associated with both short- and long-term mortality, even after adjusting for disease severity. 5 The relationship between type 2 MI and mortality in septic shock has not been evaluated, despite its association with poorer outcomes in both critically and noncritically ill patients. We set out to determine whether type 2 MI is correlated with inpatient mortality in patients with septic shock, as it could hold potential for prognostication if so.
Methodology
We utilized a retrospective cohort study design. All septic-shock-related hospitalizations identified from the Healthcare Cost and Utilization Project's (HCUP) National Inpatient Sample (NIS) from 2016 to 2020 were included in this study. There is no precise ICD-10-CM code for “troponin elevation” alone due to the need for further etiology specifications under administrative coding requirements, so consequently, hospitalizations with a principal diagnosis of septic shock and a secondary diagnosis of type 2 MI (ICD-10-CM I21.A1) were classified under the study or “exposed” group. The propensity scoring method was used to select hospitalizations with a principal diagnosis of septic shock and no secondary diagnosis of type 2 MI as the control or “unexposed” group. The scoring was based on a multivariate logistic regression model accounting for age, gender, race, primary insurance payer, hospital type, hospital bed size, hospital region, and hospital teaching status. Using 8-to-1-digit match, we paired each hospitalization in the study group with one in the control group. Before employing the propensity scoring method, the original sample sizes of the study and control groups were 31,925 and 964,890 respectively. After matching, both groups consisted of 31,905 hospitalizations each. We used this matching method to mitigate selection bias and control for patient and institutional imbalances.
Due to the heterogeneity of conditions known to be associated with type 2 MI, we performed univariate analysis of mortality risk, then stratified by subgroups of interest to probe patient characteristics and comorbidities likely to influence a potential relationship between type 2 MI and inpatient mortality. The selection of these characteristics and comorbidities was based on literature about the most common principal diagnoses in type-2-MI-related hospitalizations with in-hospital mortality. 6 They are listed in Table 1.
Baseline Patient Characteristics Across the Study and Control Groups After Propensity Score Matching.
National estimates were calculated after accounting for the sample design elements (clusters, strata, and trend weights) provided by the NIS. Categorical variables are reported as weighted N and percentages. Rao-Scott modified chi-square test was used to test the difference of distribution for categorical variables (Table 1). All statistical analyses were performed using SAS® 9.4 (SAS Institute Inc., Cary, NC, USA) on Windows 10. Statistical significance was defined by the two-sided test with significance being determined by p-value < 0.05.
As the study was based entirely on secondary data analysis, there was no direct involvement of patients or the public in the design, conduct, or reporting of the research. Institutional review board approval was not required.
Results
Baseline characteristics across the study and control groups after propensity score matching are reported in Table 1. When comparing the study group (patients in septic shock with type 2 MI) to the control group (patients in septic shock without type 2 MI), there was no statistically significant difference in age, gender, chronic respiratory failure (CRF) or iron deficiency anemia (IDA). Patients in septic shock with type 2 MI were more likely than those without type 2 MI to have coronary artery disease (CAD), chronic heart failure (CHF), tachyarrhythmias and chronic kidney disease (CKD). These differences were statistically significant.
The primary outcome (inpatient mortality) is reported in Table 2. It shows that among patients hospitalized for septic shock, type 2 MI was not associated with a statistically significant difference in inpatient mortality.
Comparison of Primary Outcome (Inpatient Mortality).
Subgroup analysis is reported in Table 3. It shows that in our sample, type 2 MI was not associated with inpatient mortality except in two subgroups. The first subgroup was patients with septic shock aged 18 to 39 years. They were more likely to die if diagnosed with type 2 MI, a statistically significant difference (OR 1.93, 95% CI, 1.09 to 3.43, p-value 0.024). This was the only comorbidity subgroup in our study for whom type 2 MI was associated with a significant increase in inpatient mortality. When compared to patients without CAD, patients with CAD had a lower risk of mortality if diagnosed with type 2 MI (OR in patients with CAD 0.82, 95% CI 0.70 to 0.96, p-value 0. 012, OR in patients without CAD 1.16, 95% CI 1.06 to 1.27, p-value 0.001)
Comparison of the Primary Outcome (Inpatient Mortality) Within the Specified Subgroups of Patients in Septic Shock.
Discussion
We investigated whether type 2 MI, a condition defined by an imbalance between myocardial oxygen supply and demand, was correlated with inpatient mortality in patients with septic shock. Our findings indicate that among patients hospitalized with a principal diagnosis of septic shock, type 2 MI was not associated with inpatient mortality in the general population. This is somewhat surprising when considering prior evidence of unexpectedly high incidence of myocardial injury, and associated increases in morbidity and mortality among critically ill patients.4,5 Our findings are, however, in line with previous evidence that unrecognized myocardial injury in critically ill patients is more common in young patients. 4
Patients without CAD demonstrated a higher mortality risk when diagnosed with type 2 MI (OR 1.16, p-value 0.001). The increase in mortality risk in patients without CAD is similar to the lower mortality risk seen in patients with CAD (OR 0.82 p-value 0. 012). This observation aligns with the hypothesis that patients with established CAD may receive more aggressive monitoring and management when troponin elevations are detected, potentially leading to improved outcomes as similar patterns have been observed in other studies evaluating myocardial injury in critically ill populations.7,8
Prior studies have demonstrated that elevated cardiac troponin is associated with increased morbidity and mortality in critically ill patients.9–11 In contrast, our findings suggest that while type 2 MI in septic shock is not associated with higher mortality overall, certain subgroups such as younger patients may be particularly vulnerable. Younger patients (aged 18 to 39 years) demonstrated a significantly increased inpatient mortality risk when diagnosed with type 2 MI (OR 1.93, p-value 0.024). This may be attributable to differences in underlying physiology, comorbidities, or healthcare access in this population. Similar to our observation about CAD, clinicians may also be less likely to ascribe significance to elevated levels of troponin in younger patients, leading to poorer outcomes from unrecognized myocardial injury. The particularly strong association between type 2 MI and inpatient mortality observed among younger patients in our cohort warrants additional consideration. One possible explanation is that in the absence of significant underlying coronary artery disease, younger patients may require a more profound physiologic insult—such as severe hemodynamic compromise, heightened inflammatory burden, or direct myocardial injury from endotoxins or cytokines—to generate a measurable rise in cardiac troponins. In this context, the presence of a positive cardiac biomarker may serve as a more specific indicator of systemic stress or myocardial strain, thereby enhancing its prognostic value in this subgroup. However, unlike studies utilizing prospective clinical data, we lacked access to key physiologic markers such as lactate levels or APACHE scores, limiting our ability to assess whether troponin elevation in these patients simply reflects other objective indicators of critical illness. It is thus plausible that elevated troponin functions as a surrogate marker for disease severity rather than as an independent driver of mortality risk. Regardless, our findings reinforce the importance of routine early troponin assessment in patients with sepsis or septic shock, particularly given its potential role in identifying those at higher risk of adverse outcomes.
Interestingly, our analysis found that gender, CHF, tachyarrhythmias, CRF, CKD, and IDA did not appear to influence the relationship between type 2 MI and inpatient mortality. This suggests that the prognostic implications of type 2 MI in septic shock are relatively independent of these specific comorbidities, highlighting that patient demographics and comorbidities should be considered when assessing the prognostic implications of troponin elevation in septic shock. Future research should explore the mechanisms underlying these associations and evaluate potential interventions aimed at mitigating the adverse outcomes associated with type 2 MI in these vulnerable populations.
Our findings align with those of Choi et al (2025), a large prospective cohort study using the Korean Sepsis Alliance registry, which found that troponin elevation at sepsis recognition was not significantly associated with increased hospital mortality or worse clinical outcomes in patients with sepsis after propensity score matching. 12 In contrast, Sheyin et al (2015) conducted an earlier meta-analysis of 17 studies which concluded that troponin elevation in patients with sepsis was associated with significantly higher mortality (RR 1.91, 95% CI: 1.65–2.22). 13 However, their analysis did not document usage of propensity matching and included studies with heterogeneous definitions of sepsis and troponin assays, although higher mortality did persist in the subgroup analysis that looked at studies that adjusted for known cardiovascular risk factors. Building on this, Gajardo et al (2025) published an updated meta-analysis that included only studies defining sepsis by sepsis-3 criteria, severe sepsis by sepsis-1 or sepsis-2, or septic shock by any of the above criteria. They also excluded disease-specific cohorts such as cancer, and required use of high-sensitivity troponin assays within the first 24 h. 14 These more stringent inclusion criteria led to an adjusted odds ratio of 1.06 (95% CI: 0.99–1.13), suggesting that once accounting for confounders and consistent methodology, troponin elevation may not independently predict mortality in sepsis. Remarkably, this mirrors our unadjusted but propensity-matched odds ratio of 1.05 (95% CI: 0.97–1.14), lending support to the idea that the prognostic utility of troponin in sepsis is highly context-dependent and is likely significantly influenced by overall patient characteristics, sepsis definitions, and assay sensitivity. These findings collectively suggest that troponin elevation, though common in sepsis, may not uniformly reflect poor prognosis once baseline risk is adequately accounted for—highlighting the need for more granular tools to risk-stratify septic patients.
Limitations, Mitigating Considerations and Strengths
Our study is limited by its retrospective design and reliance on administrative data from the National Inpatient Sample (NIS), which may introduce misclassification bias. Specifically, while we used the ICD-10-CM code I21.A1 to identify patients with type 2 myocardial infarction (MI), the NIS lacks the clinical granularity to confirm whether diagnostic criteria consistent with the universal definition of MI—such as a rise and/or fall in cardiac biomarkers with ischemic symptoms or echocardiographic wall motion abnormalities—were present. As such, we cannot confirm whether these events represent true NSTEMIs, and the potential for coding errors or upcoding for financial or administrative reasons remains a notable limitation. This limitation is particularly relevant given well known concern among cardiac quality assurance teams about the inflation of NSTEMI prevalence in critically ill populations, where troponin elevations can occur in up to one-third of ICU patients even in the absence of overt ischemia. 15
Furthermore, because of the absence of laboratory and clinical data in the NIS, we are unable to ascertain the etiology of troponin elevation—whether it reflects true type 2 MI from supply-demand mismatch, direct myocardial toxicity from sepsis-related inflammation or endotoxins, or even false positives from heterophilic antibodies. While we intentionally use the term “type 2 MI” to remain consistent with the ICD-10-CM code utilized, we acknowledge that this categorization may imperfectly represent the spectrum of myocardial injury in sepsis. Similarly, although we focused specifically on patients with septic shock, the lack of laboratory and clinical parameters limits our ability to confirm whether diagnostic criteria for sepsis or septic shock met current consensus definitions. This again reflects a broader limitation of large administrative databases, which rely on the diagnostic accuracy and coding fidelity of the treating physicians and hospitals.
A further limitation of our study is the inability to determine the specific generation of cardiac troponin assay used in the diagnosis of type 2 MI. Given the study period and the nature of administrative data, it is likely that most hospitals contributing to the NIS dataset during this time were using fourth-generation assays, rather than the fifth-generation high-sensitivity assays (hs-cTn) that are now more commonly utilized. This distinction is important, as fifth-generation assays are known to detect lower levels of myocardial injury and are associated with increased sensitivity but reduced specificity, particularly in non-cardiac admissions. As such, conclusions drawn from data predominantly reflecting older-generation assays may not fully apply to current clinical contexts where hs-cTn is more widespread. However, it is important to note that our study used the ICD-10-CM code I21.A1, which reflects a physician-assigned diagnosis of type 2 MI rather than a specific troponin threshold, thereby introducing a level of standardization that may mitigate assay variability.
Despite the above limitations, the relevance of our findings regarding the relationship between type 2 MI and mortality is reinforced by its similarity to that seen in the meta-analysis conducted by Gajardo et al (2025), which specifically included only studies that used high-sensitivity cardiac troponin assays within the first 24 h of presentation and drew from intensive care and emergency department data across multiple international settings including the USA, UK, Finland, Sweden, Germany, Italy, Spain, South Korea, China and Israel. The striking similarity between our mortality association (propensity matched OR 1.05) and theirs (adjusted OR 1.06) lends credence to the robustness and contemporary applicability of our findings, despite the lack of assay-specific data in the administrative database.
Our study has multiple strengths. The use of a nationally representative sample enhances generalizability, and the inclusion of over 63,000 admissions with septic shock provides good statistical power. Importantly, our use of propensity score matching helps address confounding and improves the validity of our comparisons. Additionally, the consistency of our findings with those of Choi et al (2025), further reinforces the clinical relevance of our results despite the inherent limitations of administrative data. These consistencies suggest that while coding variability and misclassification are real concerns, they likely have not significantly biased the overall trends observed in our analysis.
Conclusions
Our study demonstrates that type 2 MI is not associated with increased inpatient mortality in the general population of patients with septic shock. However, a notable exception exists among younger patients aged 18 to 39 years, who exhibit nearly double the inpatient mortality risk when type 2 MI is present. Additionally, patients without CAD have a higher inpatient mortality risk, whereas those with CAD demonstrate a lower risk when type 2 MI is identified. Gender, CHF, tachyarrhythmias, CRF, CKD, and IDA do not appear to influence the relationship between type 2 MI and inpatient mortality.
Taken together, these findings underscore the importance of demographic characteristics and comorbidities as key determinants of outcomes in septic patients with elevated cardiac biomarkers. They also draw attention to a potential prognostic significance of biomarker elevation in younger adults—a group in which troponin testing is not routinely performed in the setting of sepsis. Although reliance on administrative data limits our ability to validate biomarker generation and clinical context, our results and its congruence with other literature on this topic supports further investigation into the use of early cardiac biomarker testing as a targeted risk stratification tool, particularly in atypical or lower-risk sepsis subgroups such as younger patients. Future investigations should focus on understanding the underlying mechanisms contributing to increased inpatient mortality in younger patients and those without CAD, while also evaluating the real-world integration of troponin testing into early sepsis triage protocols.
Footnotes
Acknowledgements
None.
Consent for Publication
Not applicable.
Consent to Participate
Not applicable.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Ethical Considerations
Ethical approval was not required.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Statements and Declarations
While this manuscript was drafted by the lead authors, artificial intelligence technology was employed to edit the writing for clarity. All applicable sentences have been assessed for accuracy and the authors take full responsibility for the accuracy of the manuscript.
Supplemental Material
Supplemental material for this article is available online.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
