# The contribution of minimally invasive tissue sampling compared to antemortem-derived cause of death determination among inpatient child deaths: the minimally invasive tissue sampling in Malawi study

## TL;DR

This study shows that minimally invasive tissue sampling significantly improves the accuracy of determining causes of death in hospitalized children in Malawi.

## Contribution

The study demonstrates that MITS provides new and refined diagnoses in nearly all cases, even after thorough clinical evaluation.

## Key findings

- MITS added new or adjusted diagnoses in 97% of cases, mostly related to undernutrition or infections.
- Pathogens were identified in 89% of infectious conditions, with Klebsiella pneumoniae being the most common.
- Only 3% of causes of death remained unchanged after including MITS data.

## Abstract

Improved causes of death (CoD) understanding in low- and middle-income countries is needed to reduce child mortality. Compared to full autopsy, minimally invasive tissue sampling (MITS), using transcutaneous needle sampling, is a feasible, socially acceptable, and validated method. We aimed to quantify the additional contribution of MITS to CoD attribution based on clinical records and inpatient research data with intensive patient characterisation.

We enrolled children aged seven days to 59 months who died while on admission for acute illness and/or severe malnutrition to Queen Elizabeth Central Hospital in Blantyre, Malawi. Standard MITS procedures included histologic, immunohistochemical, and microbiologic testing. Phase 1 CoD determination was based on medical records alone, Phase 2 also included research data, and Phase 3 included all data, including from MITS.

We enrolled 29 children. Based on clinical notes alone (Phase 1), we identified 60 causal and 39 contributing conditions. Of the 45 (45%) infectious conditions, pathogens were identified in 15 (33%). Only one patient’s (3%) CoD was unchanged compared to including all data (Phase 3). Further, we identified 69 new (n = 43) or adjusted (n = 26) diagnoses among 28 cases (97%); the majority were undernutrition-related (n = 22, 32%) or infectious (n = 41, 59%) conditions. Overall, the majority of final Phase 3 conditions were also undernutrition-related (n = 46, 32%) or infectious (n = 61, 43%) and a pathogen was identified in 54 (89%) of the infectious conditions. Klebsiella pneumoniae was the most prevalent aetiology in both pneumonia and sepsis.

The addition of MITS to clinical and inpatient research data led to almost all (97%) of cases receiving new and/or refined diagnoses, including microbe identification in infectious conditions. Pathogens not specifically addressed by current clinical guidelines, such as Klebisiella pneumoniae, were commonly identified. Our findings support the utility of MITS to understand CoD even after thorough clinical characterisation of children during hospitalisation.