Work

Early Shock-Index Change and a Short Renal-Surveillance Window After Acute Myocardial Infarction: Contemporary-Cohort Findings With Secondary Transportability Analyses

Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background Acute kidney injury (AKI) after acute myocardial infarction (AMI) is usually recognized only after serum creatinine has risen. Clinicians need an earlier, low-burden signal that can focus renal surveillance while the patient is still in the first hours of critical care. We examined whether early change in routinely recorded shock index (SI) marks such a window. Methods We conducted a retrospective 6-hour landmark cohort study. MIMIC-IV and eICU-CRD were the contemporary core cohorts; MIMIC-III and SICdb were secondary transportability analyses. ΔSI was the median SI at 4–6 hours minus the median at 0–2 hours. The outcome was the first recorded fulfillment of creatinine-defined AKI after the landmark, defined by the Kidney Disease: Improving Global Outcomes creatinine criteria as an increase of at least 0.3 mg/dL within 48 hours or at least 1.5 times the lowest creatinine in the preceding 7 days; urine-output criteria were unavailable. Database-specific cause-specific Cox models adjusted for early SI and prespecified clinical covariates. Results The contemporary cohorts included 876 MIMIC-IV and 4,052 eICU patients. Per 0.1 higher ΔSI—equivalently, a higher 4–6-hour SI conditional on the 0–2-hour SI—the adjusted hazard of subsequent recorded AKI was 10% higher in the prespecified MIMIC-IV hospital-boundary analysis (hazard ratio [HR], 1.100; 95% confidence interval [CI], 1.011–1.197) and 23% higher as an average linear summary in eICU (HR, 1.230; 95% CI, 1.137–1.331), where the association was nonlinear (P < 0.001). At 6.75 days, cumulative AKI incidence was 44.8% versus 41.4% in MIMIC-IV and 14.4% versus 11.9% in eICU for increasing versus non-increasing SI. The MIMIC-IV estimate crossed the null under the common ICU-transfer boundary, and secondary transportability analyses were null. Delaying risk-set entry to hour 24 attenuated both contemporary-cohort estimates. Stand-alone discrimination gain in MIMIC-IV was small (ΔC-index = 0.0040). Conclusions Early ΔSI identified a short period in which routinely recorded heart rate and blood pressure carried information about the next recorded creatinine-defined AKI event in two contemporary cohorts. Because the calculation requires no new laboratory test, it may help target when clinicians review perfusion, congestion, urine output, nephrotoxic exposure, and the need for renal-function reassessment. The signal is not a universal threshold or stand-alone alarm; prospective silent-mode studies should test its local lead time, workload, and clinical benefit. Clinical trial registration Not applicable.
Language: English
Subfield: Nephrology
Field: Medicine
Open Access status: green
OpenAlex ID: W7236807890
Cites: 38
Cited by:
Related to: 10