Understanding Heterogeneity in Acute Care Trials: Resource Availability Impacts Outcomes

Sepsis is common, global, and deadly. Although most of its victims live in low- and middle-income countries (LMICs) (1), clinical trials of treatment approaches are disproportionately conducted in high-income countries (HICs) (2). This disparity in access to research is an important and modifiable contributor to the inequitable distribution of the global burden of sepsis.

Multiple factors—host, pathogen, and healthcare system factors—can impact the effectiveness of specific sepsis treatments; these differ around the world and are only rarely incorporated into trial design (Figure 1). Malnutrition and the human immunodeficiency virus may be more prevalent in LMICs. In HICs, sepsis is primarily driven by bacterial infection, whereas in LMICs parasitic and mycobacterial infections are more common (3). Antimicrobial resistance causes excess mortality worldwide, but its impact is most notable in Sub-Saharan Africa (4). Healthcare systems in LMICs often lack essential investigations, monitors, medications, invasive forms of source control, and advanced organ support (5). Even among survivors of sepsis, those in LMICs may have less access to post–acute care recovery resources, such as rehabilitation.

Figure 1.

Multiple factors associated with variability in patient risk, disease characteristics, and the clinical and socioeconomic context of care can impact the response of a given patient to a specific intervention. Understanding these modifying factors is essential to applying the results of population-based studies and clinical trials to decision making in particular patients.

Fluid resuscitation is a basic element of hemodynamic support in patients with sepsis (6); however, approaches to its use are variable. In this issue of the Journal, Gendreau and colleagues (pp. 517–528) hypothesized that higher-volume fluid resuscitation might be harmful in settings with limited healthcare resources, because of a decreased system capacity to recognize and respond to such negative consequences of fluid resuscitation as volume overload (7). They conducted a systematic review of randomized controlled trials of people with sepsis, comparing mortality in those receiving higher-volume and lower-volume fluid resuscitation. The analysis included a prespecified meta-regression by income group of the country where the trial was conducted. The main finding was that higher-volume compared with lower-volume fluid resuscitation was associated with increased mortality in LMICs (odds ratio, 1.47; 95% confidence interval, 1.14–1.90) but not in HICs (odds ratio, 1.00; 95% confidence interval, 0.87–1.16). Higher-volume fluid resuscitation was also associated with increased use of ventilation in LMICs but not HICs, and income group explained 60% of the heterogeneity in the meta-analysis. Sensitivity analyses showed increased mortality with higher-volume fluid resuscitation in settings with limited access to ventilation. A limitation of the work is that the interventions studied differed, making it difficult to attribute the observed effects of each trial solely to the volume of fluid resuscitation received. Despite this limitation, their findings have several important implications.

First, the data show no benefit for higher-volume fluid resuscitation regardless of geography or economic status, only harm in resource-limited settings, suggesting to the authors that the harms of higher-volume fluid resuscitation are masked by compensating clinical processes—particularly mechanical ventilation—that are available only in HIC settings. Additional potential explanations include the possibility of delayed presentation preventing the early administration of intravenous fluids (8) or other differences in the microbiologic features of the disease, in regional host genetic profiles, or in local approaches to clinical management. Regardless of explanation, the message is clear: fluid administration in sepsis must be judicious, and future work must address how this is best achieved.

Second, the analysis reflects a common emerging theme—that minimizing the harms of intervention is as important as showing superiority in critical care trials. Many aspects of critical care become common practice before randomized evaluation, and impressive improvements in patient outcomes often relate to identifying how to use these interventions safely (9, 10), an issue particularly important to patient management in LMICs. Future research focused on understanding why higher-volume fluid resuscitation is associated with harm in LMICs could help improve care in countries at all income levels.

Third, randomized controlled trials and syntheses of those trials show the mean effect of a therapy in a population, rather than the effect that might be experienced by any individual in that population. These findings do not imply that higher-volume fluid resuscitation should never be used but that on average, especially in LMIC settings, lower-volume fluid resuscitation should be favored. The observation that there is variability or heterogeneity in response to treatment is common in trials studying acutely ill patients and exemplified by reported variability in the response of patients with coronavirus disease (COVID-19) to corticosteroids (11), heparin (12, 13), or anakinra (14). An important new frontier in acute care research is the development of strategies and tools to detect and evaluate these reliably, to refine treatment guidelines emerging from randomized trials and ensure that patients benefit maximally from the findings of these trials.

Finally, the findings underline the importance of conducting acute care research at a global scale. If there is a differential effect of fluid resuscitation according to country income group, how many other fundamentals of acute care medicine have similar heterogeneity? If care based on experiments conducted in HICs is harmful in LMICs, uncovering this through research is an opportunity to improve outcomes for millions of people.

But there are challenges. Barriers to conducting research in LMICs include a shortage of financial and human capacity, difficulties navigating ethical and regulatory systems, limited preexisting research infrastructure, and competing demands on the part of healthcare systems (15). To overcome these barriers, potentially helpful actions include increasing the diversity of editorial boards (16); facilitating international collaboration within and across trials using harmonized protocols or platform, basket, and umbrella trial structures (17); ensuring equitable collaboration between HIC and LMIC researchers (18); integrating clinical trials with robust data collection and reporting systems (3); and empowering clinical research networks in resource-limited settings in the global south.

In conclusion, the work of Gendreau and colleagues demonstrated that higher-volume fluid resuscitation was not better than lower-volume fluid resuscitation and that instead it was associated with harm in LMICs. These findings have important implications for fluid resuscitation in sepsis worldwide and highlight the need to explore sources of variability in clinical trials and the urgency of global collaboration in acute care research.