A Need for Prospective Long-Term Data to Characterize Cardiovascular Risk Factors in Sepsis Survivors

As in-hospital mortality from sepsis decreases, there is an imminent need to understand the complex medical burdens of sepsis survivors (1). One of the major medical burdens faced by sepsis survivors are complications of cardiovascular disease. Epidemiological studies have shown higher risk of stroke, myocardial infarction, atrial fibrillation, and other cardiovascular complications after a hospitalization for sepsis (2–4).

In this issue of Critical Care Medicine, the study by Angriman et al (5) evaluates the risk factors associated with the observed cardiovascular burden in sepsis survivors. The authors used several administrative databases in Ontario, Canada, between 2008 and 2017 to identify patients without baseline cardiovascular disease who were hospitalized for and survived an episode of sepsis. The patients were evaluated for the subsequent postdischarge occurrence of major cardiovascular events, defined as stroke, myocardial infarction, or cardiac death, over the observation period between 1 and 5 years. The authors evaluated the association between baseline clinical characteristics and characteristics of the sepsis episode with major cardiovascular events. In total, 268,259 sepsis survivors without baseline cardiovascular disease were included, of whom 10.4% experienced a major cardiovascular event. The authors found that there were both baseline characteristics and characteristics of the sepsis episode that were associated with cardiovascular events. The baseline characteristics were not surprising and included increased age, male sex, and comorbidities such as diabetes mellitus and hypertension. The characteristics of the sepsis episode associated with cardiovascular events included pneumonia compared with other sites of infection, presence of septic shock, need for ICU admission, and need for renal replacement therapy. Additionally, higher creatinine and troponin levels were associated with cardiovascular events. Finally, discharge prescription for a statin was associated with a decreased hazard ratio for a major cardiovascular event. The strengths of the study include: 1) the unification of data across multiple provincial administrative databases to evaluate diverse risk factors from comorbidities to laboratories to social determinants of health, 2) the use of a cause-specific Cox regression model framework, and 3) multiple sensitivity analyses to account for varying definitions of baseline cardiovascular disease and sepsis.

Angriman et al (5) present a well-designed study that adds to the important and growing body of literature revealing increased cardiovascular disease burden in sepsis survivors (2–4). However, the use of retrospective data to evaluate long-term cardiovascular outcomes results in several limitations. First, the study design relies heavily on International Classification of Diseases (ICD) codes for cohort selection, comorbidity characterization, and outcome definition. Even with sensitivity analyses, there is an inherent limitation to ICD codes. Second, the datasets lack well-known risk factors for cardiovascular disease such as obesity, family history, and tobacco use. Third, there are important hospitalization risk factors that were not available such as vasopressor and inotrope use. The ICD-based definitions and the unmeasured risk factors limit the ability to infer causal pathways from the data. For instance, the need for renal replacement therapy was found to be a cardiovascular risk factor. However, it is possible that patients who needed renal replacement therapy during a sepsis hospitalization already had underlying cardiovascular disease not captured by ICD coding and prescription medicine use. Pneumonia was noted as a cardiovascular risk factor, but a pulmonary infection may be a surrogate for other risk factors such as tobacco use that are associated with cardiovascular events.

Despite the limitations, the study by Angriman et al (5) presents interesting findings and highlights the need for: 1) prospective collection of long-term granular data in sepsis survivors and 2) long-term follow-up of patients with sepsis in randomized controlled trials (RCTs). There is a need for prospective collection of long-term data to clarify the risk factors for cardiovascular disease in sepsis survivors. Prospective data are essential for accurate identification of patients without baseline cardiovascular disease, characterization of the baseline and hospitalization risk factors, and evaluation of the outcomes of interest. The creation of collaborative multi-institutional datasets could provide the basis for investigating cardiovascular risk profiles of sepsis survivors. The findings from this type of robust data could provide the basis for predictive models to create an atherosclerotic cardiovascular disease equivalent for sepsis survivors (6). The study by Angriman et al (5) also notes a decreased risk for cardiovascular events with discharge prescription of a statin. Although, this is only hypothesis-generating, this finding more broadly suggests that therapies given to patients with sepsis likely have long-term effects that we are currently not studying. For instance, most studies evaluating interventions in critical illness including statins focus on short-term treatment and near-term outcomes (7, 8). It is possible that long-term follow-up of sepsis survivors would reveal unexpected beneficial or harmful effects from these interventions. RCTs with a focus on long-term outcomes in sepsis patients will be a critical step in understanding how interventions modify cardiovascular disease risk in sepsis survivors.

The study by Angriman et al (5) is part of an expanding field of research into cardiovascular disease in sepsis survivors and highlights important future directions. There is now mounting evidence that sepsis survivors have an increased cardiovascular disease burden. What remains is gathering robust data that will allow us to identify modifiable risk factors associated with post-sepsis cardiovascular disease.