BackgroundEmerging evidence suggests that minimal acute kidney injury (stage 1 AKI) is associated with increased hospital mortality rates. However, for those who do not meet the AKI diagnostic criteria, whether a small increase in serum creatinine (SCr) levels is associated with an increased mortality rate in elderly patients is not known. Therefore, we aimed to investigate small elevations in SCr of <26.5 µmol/L within 48 h after invasive mechanical ventilation (MV) on the short-term mortality of critically ill patients in the geriatric population. MethodsWe conducted a retrospective, observational, multicenter cohort study enrolling consecutive elderly patients (≥75 years) who received invasive MV from January 2008 to December 2020. Recursive partitioning was used to calculate the ratio of SCr rise from baseline within 48 h after MV and divided into six groups, (1) <10%, (2) 10%–<20%, (3) 20%–<30%, (4) 30%–<40%, (5) 40%–<50%, and (6) ≥50%, where the reference interval was defined as the ratio <10% based on an analysis, which confirmed that the lowest mortality risk was found in this range. Clinical data and laboratory data were noted. Their general conditions and clinical characteristics were compared between the six groups. Prognostic survival factors were identified using Cox regression analysis. Kaplan–Meier survival analysis was employed for the accumulative survival rate. ResultsA total of 1292 patients (1171 men) with a median age of 89 (interquartile range: 85–92) with MV were suitable for further analysis. In all, 376 patients had any stage of early AKI, and 916 patients had no AKI. Among 916 non-AKI patients, 349 patients were in the ratio <10%, 291 in the 10%–<20% group, 169 in the 20%–<30% group, 68 in the 30%–<40% group, 25 in the 40%–<50% group, and 14 in the ≥50% group. The 28-day mortality rates in the six groups from the lowest (<10%) to the highest (≥50%) were 8.0%, 16.8%, 28.4%, 54.4%, 80.0%, and 85.7%, respectively. In the multivariable-adjusted analysis, patients with a ratio of 10%–<20% (hazard ratio [HR]=2.244; 95% confidence interval [CI]: 1.410 to 3.572; P=0.001), 20%–<30% (HR=3.822; 95% CI: 2.433 to 6.194; P <0.001), 30%–<40% (HR=10.472; 95% CI: 6.379 to 17.190; P <0.001), 40%–<50% (HR=13.887; 95% CI: 7.624 to 25.292; P <0.001), and ≥50% (HR=13.618; 95% CI: 6.832 to 27.144; P <0.001) had relatively higher 28-day mortality rates. The 90-day mortality rates in the six strata were 30.1%, 35.1%, 45.0%, 60.3%, 80.0%, and 85.7%, respectively. Significant interactions were also observed between the ratio and 90-day mortality: patients with a ratio of 10%–<20% (HR=1.322; 95% CI: 1.006 to 1.738; P=0.045), 20%–<30% (HR=1.823; 95% CI: 1.356 to 2.452; P <0.001), 30%–<40% (HR=3.751; 95% CI: 2.601 to 5.410; P <0.001), 40%–<50% (HR=5.735; 95% CI: 3.447 to 9.541; P <0.001), and ≥50% (HR=6.305; 95% CI: 3.430 to 11.588; P <0.001) had relatively higher 90-day mortality rates. ConclusionsOur study suggests that a ≥ 10% SCr rise from baseline within 48 h after MV was independently associated with short-term all-cause mortality in mechanically ventilated elderly patients.