Help Wanted for Sepsis: Biomarkers in Low- and Middle-Income Countries Please Apply*

Sepsis, characterized as the dysregulated host response to infection, is an important contributor to pediatric morbidity and mortality worldwide. Published data report an estimated 25 million cases of pediatric sepsis and nearly 3.5 million deaths in children less than 19 globally in 2017 (1). The burden of pediatric sepsis, however, is unequally distributed, with low- and middle-income countries (LMICs) bearing a far higher incidence of pediatric sepsis and pediatric sepsis-related mortality when compared with high-income countries (1). While the reasons for these discrepancies are multifactorial, it remains that the epidemiological distribution of pediatric sepsis globally is unequal. Pediatric sepsis is itself a heterogeneous condition with various causes and phenotypes, which has led to difficulty identifying consistently effective therapies. The use of biomarkers to better profile the host biochemical response to infection has enhanced the identification of sepsis phenotypes and stratification of risk of mortality (2), which in turn enriches investigational studies into the management of pediatric sepsis. However, given the resources and infrastructure required to conduct specialized biochemical assays for many biomarkers, the vast majority of biomarker-based information is derived from samples obtained in resource-rich areas of the world (3). A disconnect therefore exists between our evolving understanding of pediatric sepsis and the children most affected by this illness. Undeniably, the scientific field will benefit from prioritizing new knowledge in LMICs to best understand which aspects of sepsis are generalizable and areas where epigenetic and other differences may exist. This disconnect raises multiple areas of potential investigation. It is unknown whether existing biomarker-based models are generalizable globally when considering the epidemiological differences in pediatric sepsis worldwide. Further, it is unknown if the biochemical profiles of children with sepsis in LMICs, areas with a higher burden of pediatric sepsis and pediatric sepsis-related mortality, are similar to those observed in higher-income areas of the world. Finally, it is unclear how feasible it is to conduct resource-intensive investigations, such as those involving biomarkers, in LMICs. All sepsis is not the same, all patients are not the same, and until we evaluate the disease mechanistically, within specific causal pathways, and across epigenetic factors, we will manage the syndrome nonspecifically rather than using a more targeted and precise approach. With these considerations in mind, Ishaque et al (4), in this issue of Pediatric Critical Care Medicine, investigated the utility of the Pediatric Sepsis Biomarker Risk Model (PERSEVERE-II), a biomarker-based pediatric sepsis risk stratification model validated in the United States, for sepsis risk stratification among children presenting to a single pediatric hospital in Pakistan. Patient blood samples were prospectively collected from children meeting pediatric modifications of adult Sepsis-3 criteria within the first 24 hours of PICU admission, after which biomarker level measurements were performed on these samples at sponsoring institutions in the United States (5). In this cohort of pediatric sepsis subjects from Pakistan, PERSEVERE-II performed similarly in discriminating mortality when compared with its performance in validation cohorts in the United States, with an area under the receiver operating characteristic (AUROC) curve of 0.83 (0.72–0.94). Other commonly used models to estimate illness severity, Pediatric Logistic Organ Dysfunction 2, and pediatric Sequential Organ Failure Assessment score (5,6). Since the performance of PERSEVERE-II was unknown a priori in this cohort, additional biomarkers for endotheliopathy, coagulopathy, and lung injury were preliminarily evaluated with intent to enrich PERSEVERE-II with these markers if better risk stratification was achieved by including these markers or if the performance of PERSEVERE-II was unreliable in this cohort. Incorporating these additional biomarkers resulted in a higher, although nonsignificant, AUROC when compared with that of PERSEVERE-II, although this current study is still enrolling. Taken together, the authors conclude that PERSEVERE-II may be generalizable to LMICs for risk of mortality stratification in pediatric sepsis. While the scope of the study by Ishaque et al (4) is limited to a single institution in Pakistan, and therefore the results may not be generalizable to other regions of the world, this work is also a step toward increasing LMIC representation within medical research. The lack of significance in the study by Ishaque et al (4) does not demonstrate a failure for the use of biomarkers but rather a success to develop future studies within this space. However, to provide even further equity, future studies will hopefully be able to be performed within LMICs rather than shipped back to the United States. This current study is the first step in creating such equitable research infrastructure. Likewise, a lack of difference seen by the authors may in fact reflect a myriad of factors including the group’s use of a single, static biomarker plan rather than serial sampling. Serial samples will allow for understanding of dynamic causal pathways that may influence immune, hemostatic, and endothelial function as sepsis in patients endures an evolution to clinical trajectory that are likely missed with single time point biomarkers. The work by Ishaque et al (4) utilized samples from areas of the world outside high-income countries may reveal important patterns that could be missed in cohorts exclusively from high-income countries. Additionally, incorporating data from individuals most affected by a particular illness may increase the ability of investigators to find meaningful results with potential for benefit in LMICs. Further, by validating models in LMICs that have the potential to enrich clinical research, new tools or models may be more available to enhance ongoing research in institutions based in LMICs. As logistical barriers to international research efforts become smaller, collaborative efforts between individuals in high-income and LMICs may begin to reveal globally conserved patterns, which may point at core pathophysiologic processes in critically ill children. Indeed, the notion that a biomarker-based model developed in a resource-rich area of the world can be validated in another area with significantly different sepsis epidemiology and outcomes suggests that these biomarkers may reflect common pathophysiologic pathways. By validating risk models such as PERSEVERE-II globally, generalizable findings may emerge, which may help further focus research efforts to identify important therapeutic targets. And modulating these targets may reduce the abhorrent mortality seen in the study by Ishaque et al (4) in children of more than 15%, whereas that in resource-rich areas is less than 2%. While the study by Ishaque et al (4) focuses on the validation of one biomarker-based risk stratification model for pediatric sepsis in one hospital in a LMIC, it highlights the feasibility of an enormous potential in collaborative research efforts between institutions from differently resourced areas of the world. Other institutions may have interest in pursuing similar investigations, possibly including more advanced methodologies such as clinical trials. There may also be opportunities for enrichment of the diagnostic and therapeutic capabilities at collaborating institutions in LMICs, potentially directly leading to a positive impact on patient care in these settings. Ultimately, these mutually beneficial relationships may lead to new understanding and the improved care of critically ill children globally. Development of clinical and laboratory-based biomarkers may further refine a high mortality syndrome and lead toward further leveling the playing field for high-level research as the field moves to precision medicine. Pediatric sepsis phenotyping is currently hiring, and more biomarkers in diverse locations and populations, including LMICs, must apply to be employed in future practice.