Nonlinear Relationship between Early Postoperative Pulse Pressure Variation and Acute Kidney Injury Risk in Older...

Acute kidney injury (AKI) frequently complicates the postoperative course of older patients in surgical intensive care units (ICUs), contributing to increased mortality and prolonged hospital stays. Traditional AKI diagnostic methods, relying on changes in serum creatinine and urine output, often result in delayed detection. Pulse pressure variation (PPV), a dynamic hemodynamic parameter derived from arterial waveform analysis, reflects preload responsiveness and can provide early, complementary information on circulatory status and end-organ perfusion. Although both elevated and diminished PPV levels have been associated with adverse outcomes, the specific nature of the association between early postoperative PPV and AKI risk remains uncertain. This study investigated the relationship between PPV during the first 12 and 24 h after ICU admission and the development of AKI, with a focus on potential nonlinear associations and optimal PPV thresholds.

This retrospective cohort analysis utilized data from the MIMIC-IV and MIMIC-III databases. We included patients aged 60 years and older who underwent cardiothoracic surgery and had available PPV measurements within the first 24 h following ICU admission. The primary dataset was derived from MIMIC-IV, while MIMIC-III served as an external validation cohort. The main outcome was postoperative AKI, as identified by KDIGO criteria. Logistic regression models adjusted for relevant clinical covariates were used to assess associations between PPV at 12 and 24 h (PPV₁₂h and PPV₂₄h) and AKI. Nonlinear associations were examined using restricted cubic spline (RCS) modeling. Discriminative ability was assessed using area under the receiver operating characteristic curve (AUC), with DeLong's test used to compare model performance.

Of the 5,445 patients included, 2,574 (47.3%) developed AKI after surgery. Those with AKI were generally older and had more comorbidities and higher illness severity scores. In multivariable analyses incorporating RCS terms, a U-shaped relationship was identified between PPV₁₂h and AKI (p for nonlinearity = 0.0224). The lowest predicted risk was observed at a PPV₁₂h of 0.144, corresponding to a 39.4% relative reduction in AKI risk. This association was independently validated in a separate cohort (n = 897), where a PPV₁₂h of 0.142 was associated with a 59.8% risk reduction. In contrast, PPV₁₂h provided better predictive value than PPV₂₄h, which showed no significant association with AKI. The inclusion of PPV₁₂h significantly enhanced the discrimination of predictive models.

Early postoperative PPV demonstrates a U-shaped correlation with AKI risk in older patients undergoing cardiothoracic surgery, with both low and high values associated with higher incidence. A PPV₁₂h around 0.14 was consistently linked to the lowest risk in both primary and validation cohorts. These findings remained robust after adjusting for multiple confounders. PPV assessed at 24 h, however, showed limited predictive utility. Routine early PPV monitoring may, therefore, serve as a valuable adjunct for perioperative hemodynamic assessment, potentially facilitating timely interventions to mitigate postoperative renal complications.