This single-center retrospective series analyzed 731 surgical episodes from 709 adults who underwent cardiac surgery with cardiopulmonary bypass. Investigators reported SHR as the first ICU glucose divided by the HbA1c-derived estimated average glucose. The abstract did not provide details on intraoperative or postoperative glucose management protocols, so those elements are not available here. The primary outcome was an unweighted composite that pooled clinically significant events: 30-day mortality, stroke, Kidney Disease: Improving Global Outcomes (KDIGO) stage 3 acute kidney injury, reoperation for bleeding, prolonged mechanical ventilation, and deep sternal wound infection. This composite was chosen to capture major postoperative complications relevant to cardiac surgical patients.
The study team used multivariable logistic regression to adjust for clinically relevant covariates; the abstract lists adjusted effect estimates but does not specify every covariate included in the final model in this summary. After adjustment, stress hyperglycemia ratio (SHR) measured at ICU admission was associated with the composite outcome. Investigators reported an odds ratio of 1.103 per 0.1-unit increase in SHR (95% confidence interval 1.008–1.208; P = 0.033). By contrast, absolute ICU admission glucose was not significantly associated with the composite outcome after multivariable adjustment: odds ratio 1.019 per 10 mg/dL increase (95% CI 0.946–1.096; P = 0.626). The authors therefore framed SHR as the metric that retained an adjusted association with major postoperative complications, while simple admission glucose did not.
Ninety surgical episodes met the composite endpoint, representing 12.3% of the cohort. The abstract provides the primary adjusted effect estimates for both SHR and absolute glucose, as noted above. Investigators did not report additional granular subgroup counts or breakdown of which individual composite components drove the overall event rate in the abstract text. That level of detail—component-specific frequencies and the distribution of SHR by event status—is not reported in the abstract and thus cannot be extrapolated here.
Authors noted that confidence intervals for SHR included 1.0 in several sensitivity analyses, indicating that the association was not uniformly robust across all tested models. The abstract does not enumerate the specific sensitivity analyses performed nor the alternate model specifications that produced confidence intervals crossing unity. The study design was retrospective and single-center; the investigators obtained ethical approval from their institutional review board and used an opt-out consent approach given the retrospective nature. Trial registration was not required and none was reported. The absence of prospective measurement and the potential for unmeasured confounding are intrinsic limitations to the design and were acknowledged implicitly by the authors' call for prospective evaluation.
The Ethics Review Board of Hamamatsu University School of Medicine approved the study (approval number cited in the abstract). The authors declared no conflicts of interest. The abstract clarifies that the requirement for written informed consent was waived because of the retrospective design. Important methodological details—such as the full list of covariates in adjusted models, perioperative glycemic control protocols, timing and method of the first ICU glucose sample, and HbA1c assay specifics—were not reported in the abstract. Those omissions limit the ability to assess potential measurement bias and residual confounding from the summary alone.
Investigators concluded that SHR at ICU admission was associated with major postoperative complications after cardiac surgery with cardiopulmonary bypass in their primary analysis and recommended prospective evaluation of SHR as an early postoperative risk marker. Open questions that remain—based on the abstract—include how SHR performs compared with established risk scores, whether SHR adds incremental predictive value beyond routine clinical variables, and how perioperative glucose management might modify the observed associations. The abstract ends with a recommendation for prospective study rather than a statement of clinical adoption or specific glucose-management changes.