Importance of Dynamic Platelet Changes in Surgical Sepsis: Association with ICU Mortality and Correlation with White Blood Cell Count

Abstract

Background: Platelet consumption and leukocyte activation are integral to the host response in sepsis, yet the prognostic relevance of serial platelet and white blood cell (WBC) trends in surgical septic patients admitted to the intensive care unit (ICU) remains clinically important. This study evaluated whether platelet indices and WBC indices were associated with ICU mortality in adult surgical septic patients.

Objective: To evaluate the association between platelet count and ICU mortality and to correlate platelet and WBC measures among surgical septic patients admitted to intensive care unit.

Methods: This is a single-center hospital-based observational analytical study that included 57 patients with surgical sepsis who were admitted to the intensive care unit (ICU) at No. (1) Military Hospital (700 bedded), Pyin Oo Lwin, from 01/01/2025 to 31/12/2025. Admission values were defined as ICU Day 1 (D1). Platelet nadir was the lowest platelet count recorded from ICU Day 1 to ICU Day 14, and peak WBC was the highest WBC recorded over the same period. The primary endpoint was intra-ICU mortality. Additional derived variables were absolute platelet fall, absolute WBC rise, timing of platelet nadir, and timing of WBC peak. Continuous variables were summarized as median and interquartile range (IQR), compared using the Mann-Whitney U test, and correlations were assessed using Spearman rho. Categorical variables were compared using Fisher exact test or chi-square test where appropriate. Receiver operating characteristic (ROC) analysis with the Youden index was used to identify optimal cut-offs for ICU mortality prediction. A predefined combined platelet-WBC risk group was additionally tested using platelet nadir <=90 x10^9/L and peak WBC >=22.3 x10^9/L. Statistical analysis was performed using IBM SPSS Statistics version 28.0.

Results: Of the 57 patients, 22 died in ICU, yielding an ICU mortality of 38.6%. Non-survivors were older than survivors (median age 58.0 (35.0-68.0) vs 47.0 (40.0-52.0), p = 0.039) and had longer ICU stays (10.0 (6.0-17.0) vs 6.0 (3.0-8.0), p = 0.002). The admission platelet count was lower in non-survivors, but the difference was not statistically significant (median 140.0 (104.0-222.0) vs 160.0 (133.0-380.0), p = 0.222). Platelet nadir was markedly lower in non-survivors (37.5 (20.0-81.5) vs 93.0 (79.0-165.5), p < 0.001), and the absolute platelet fall was greater (p = 0.027). The timing of platelet nadir occurred substantially later in non-survivors (median 6.0 (4.0-8.0) vs 2.0 (2.0-4.5), p < 0.001). Mortality increased across platelet nadir categories (chi-square p = 0.002). Admission WBC did not differ significantly (p = 0.113), whereas peak WBC was significantly higher in non-survivors (32.1 (26.4-42.2) vs 20.5 (15.7-30.9), p = 0.002). In the overall, platelet nadir showed a weak inverse but non-significant correlation with peak WBC (rho -0.098, p = 0.469), whereas platelet nadir correlated positively with admission WBC (rho 0.325, p = 0.014). On ROC analysis, platelet nadir demonstrated the best platelet-based discrimination (AUC 0.768) with an optimal cut-off of 90.0 ×10^9/L, while peak WBC showed the best leukocyte-based discrimination (AUC 0.743) with a cut-off of 22.3 ×10^9/L. When these thresholds were combined, ICU mortality was 0.0% in the low-risk group, 29.2% in the intermediate-risk group, and 71.4% in the high-risk group (chi-square p < 0.001).

Conclusion: In surgical sepsis, dynamic hematologic markers outperformed admission values. A low platelet nadir and high peak WBC were each associated with ICU mortality, and their combination created clinically meaningful mortality risk stratification. Platelet trajectory, particularly the nadir and its delayed occurrence, deserves routine integration into the ICU sepsis assessment rather than reliance on a single admission platelet value.