Injury mechanism patterns, mortality predictors, and resource utilization in pediatric trauma in Karachi, Pakistan: A seven-year registry study.

BACKGROUND

Pakistan faces a particularly severe pediatric trauma burden due to its large youth population, ubiquitous motorcycle use, and limited traffic safety enforcement. Low- and middle-income countries bear the vast majority of the global injury burden, yet data derived from their trauma registries remain sparse.

METHODS

We conducted a retrospective analysis of a prospectively maintained pediatric trauma registry in Karachi, Pakistan (2018-2024; n = 4576; outcomes cohort n = 3406). Injury patterns and resource utilization were compared across five age groups. Severity-adjusted mortality was benchmarked using observed-to-expected (O:E) ratios derived from the Revised Trauma Score. Multivariable logistic regression identified independent predictors of in-hospital mortality.

RESULTS

Of 4576 patients in the full cohort, motorcycle crashes were the most common mechanism (34%), followed by falls (33%) and pedestrian injuries (12%); falls predominated in children under 10 and motorcycle injuries in adolescents aged 11-18. Males comprised 82% overall and 93% of adolescents aged 15-18. Among 3406 patients with known disposition, crude mortality was 15.1% and varied by age group (χ²=44.1, p < 0.001), ranging from 10.6% in children aged 11-14-34.9% in infants. Resource utilization included blood transfusion (33.4%), operative intervention (20.2%), ICU admission (11.9%), and intubation (6.4%); mortality was 22.5% in high-resource-use patients (ICU admission, intubation, or operative intervention) vs. 11.4% in the remainder (p < 0.001). Among 2480 patients with calculable RTS, the O:E ratio was 1.41 (95% CI 1.27-1.57; 337 observed vs. 238 expected deaths), peaking in 2021 (2.13) and lowest in 2023 (0.77). GCS was the strongest mortality predictor (OR 0.71 per point, 95% CI 0.68-0.73); polytrauma was also associated with higher odds (aOR 2.90, 95% CI 1.13-7.43; 9 deaths among 41 cases). One-third of the full cohort arrived via inter-facility transfer (33.3%, n = 1526). After multivariable adjustment, transfer remained independently associated with higher mortality (aOR 1.88, 95% CI 1.42-2.49).

CONCLUSIONS

Pediatric trauma mortality in Karachi greatly exceeds RTS-based predictions without anatomic adjustment, especially among the youngest patients. These findings highlight the need for trauma center regionalization, strengthened prehospital systems, and age-specific resuscitation protocols. Systematic O:E monitoring embedded within Pakistan's trauma system represents an immediately deployable quality improvement tool.