Abstract
Background
Postoperative wound complications are an important cause of morbidity, and routinely available inflammatory markers may assist with early risk assessment. This study evaluated a newly proposed exploratory composite index, the Wound Inflammation–Nutrition Index (WINA), and its change between postoperative day 1 and postoperative day 3 (ΔWINA) in relation to postoperative wound complications.
Methods
This retrospective observational study included 86 patients who underwent general surgical procedures between January 2020 and December 2025. WINA was calculated by multiplying the C-reactive protein-to-albumin ratio by the neutrophil-to-lymphocyte ratio at postoperative days 1 and 3. ΔWINA was calculated as the POD3 value minus the POD1 value. Logistic regression, receiver operating characteristic analysis, model comparison, bootstrap internal validation, and sensitivity analyses were performed.
Results
Postoperative wound complications occurred in 24 patients (27.9%). WINA values at POD1 and POD3 and ΔWINA were higher in patients with wound complications (all p < 0.001). In the parsimonious multivariable model, ΔWINA per 100-unit increase was associated with wound complications (adjusted OR: 2.12, 95% CI: 1.44–3.13, p = 0.001), together with diabetes mellitus (adjusted OR: 2.89, 95% CI: 1.01–8.29, p = 0.048) and operative time per 30-minute increase (adjusted OR: 1.67, 95% CI: 1.08–2.58, p = 0.021). ΔWINA had an AUC of 0.872 (95% CI: 0.789–0.941), with 83.3% sensitivity, 80.6% specificity, 62.5% positive predictive value, and 92.6% negative predictive value at a cut-off of 245.0. Its discrimination did not differ significantly from WINA at POD3. Adding ΔWINA to the clinical model increased the AUC from 0.772 to 0.903 (ΔAUC: 0.131, p = 0.003). The association remained after excluding complications diagnosed on or before POD3.
Conclusion
ΔWINA was associated with postoperative wound complications and improved the performance of a parsimonious clinical model. However, it did not outperform WINA measured at POD3, and the derived cut-off requires external validation before clinical use.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1007/s00423-026-04209-w.
Keywords: C-reactive protein-to-albumin ratio, Neutrophil-to-lymphocyte ratio, Postoperative complications, Surgical site infection, Wound complications
Introduction
Postoperative wound complications, particularly surgical site infections, remain a significant source of morbidity and healthcare burden despite advances in surgical techniques and perioperative care. These complications are associated with prolonged hospital stay, increased readmission rates, higher healthcare costs, and impaired patient outcomes. Therefore, early identification of patients at risk for wound complications is of considerable importance for optimizing postoperative management and improving clinical outcomes [1–3].
In recent years, biomarkers reflecting the systemic inflammatory response have gained increasing attention for the prediction of postoperative complications. Among these, C-reactive protein (CRP), the neutrophil-to-lymphocyte ratio (NLR), and serum albumin have been widely investigated. Elevated postoperative CRP levels may indicate persistent inflammatory activity, whereas NLR provides a readily available measure of the balance between innate and adaptive immune responses. Serum albumin is influenced by inflammation, capillary leakage, hepatic synthesis, and nutritional status and should therefore not be interpreted as a nutritional marker alone. Previous studies have shown that CRP, NLR, albumin, and the CRP-to-albumin ratio (CAR) may be associated with postoperative complications [4–7]. However, the performance of individual biomarkers may be limited because the postoperative response involves several interacting inflammatory and metabolic pathways. Composite indices may therefore provide a broader representation of postoperative physiological stress than any single laboratory parameter [8–10].
In the present study, we proposed the Wound Inflammation–Nutrition Index (WINA) as an exploratory composite index calculated by multiplying CAR by NLR. Although the index includes serum albumin, it was primarily intended to reflect postoperative inflammatory burden rather than nutritional status alone. The potential value of WINA and its change between postoperative day 1 and postoperative day 3 (ΔWINA) in patients undergoing general surgical procedures has not previously been established. It also remains unclear whether this composite provides additional information beyond simpler markers such as CRP, NLR, CAR, or a single POD3 WINA measurement. Therefore, this study aimed to evaluate the association and diagnostic performance of WINA and ΔWINA for postoperative wound complications and to compare them with conventional postoperative biomarkers. We also examined whether adding ΔWINA to a parsimonious clinical model improved model performance [11, 12].
Methods
Study design and setting
This retrospective observational study was conducted at the Department of General Surgery, Düzce University Faculty of Medicine, a tertiary care university hospital, between January 2020 and December 2025. The study included patients who underwent surgical procedures during the specified period and were evaluated for postoperative wound complications. The manuscript was prepared in accordance with the STROBE reporting recommendations for observational studies. Ethical approval for the study was obtained from the Düzce University Non-Interventional Clinical Research Ethics Committee (approval number: 2026/119; date: March 25, 2026). The study was conducted in accordance with the principles of the Declaration of Helsinki. Due to the retrospective design, the requirement for informed consent was waived.
Study population
Patients who underwent surgical procedures at the Department of General Surgery, Düzce University Faculty of Medicine, between January 2020 and December 2025 were retrospectively screened for eligibility. Adult patients (≥ 18 years) with available postoperative laboratory data on postoperative day 1 (POD1) and postoperative day 3 (POD3) were included in the study. Patients were excluded if they had missing or incomplete laboratory data required for the calculation of WINA, including C-reactive protein, serum albumin, neutrophil count, or lymphocyte count. Patients with pre-existing systemic infection, chronic inflammatory disease, or another condition expected to markedly affect inflammatory markers were also excluded. In addition, patients with incomplete clinical records or insufficient postoperative follow-up were excluded.
A total of 312 patients were screened. Of these, 147 were excluded because of missing POD1 or POD3 laboratory data, 28 because of pre-existing systemic infection or chronic inflammatory disease, 31 because of incomplete clinical records, 15 because of insufficient follow-up, and five because they were younger than 18 years. The final analysis included 86 patients. Given the retrospective design, no a priori sample size calculation was performed, and all eligible patients identified during the study period were included. The patient selection process is shown in Fig. 1.
Fig. 1.
Flow diagram of patient selection and inclusion
Data collection
Clinical and laboratory data were retrospectively obtained from the hospital electronic medical record system. Demographic variables included age, sex, and body mass index (BMI). Clinical variables included diabetes mellitus, smoking status, American Society of Anesthesiologists (ASA) classification, surgical urgency, operative time, and wound contamination class. Additional surgical variables included the indication for surgery, classified as benign or malignant; surgical approach, classified as open or non-open; and surgical magnitude, classified as major abdominal or non-major surgery. Surgical procedures were grouped as colorectal, upper gastrointestinal, hepatobiliary or pancreatic, hernia or abdominal wall, and other general surgical procedures. Each patient was classified according to the primary procedure performed during the index admission. Laboratory parameters were obtained on POD1 and POD3 and included CRP, serum albumin, absolute neutrophil count, absolute lymphocyte count, and NLR. All measurements were performed as part of routine postoperative care in the hospital central laboratory. POD1 was used to represent the early postoperative inflammatory response, whereas POD3 was selected to assess the subsequent direction of this response. All data were anonymized before analysis.
Definitions of outcomes and variables
The primary outcome was the occurrence of a postoperative wound complication within 30 days after surgery. Wound complications included surgical site infection, wound dehiscence, seroma or hematoma requiring intervention, and other clinically documented wound-related events requiring medical or surgical treatment. Surgical site infections were classified as superficial incisional, deep incisional, or organ/space infections according to the anatomical extent documented in the medical records. For patients with more than one wound event, the first clinically documented event was used as the primary complication. The diagnosis of wound complications was based on clinical examination and documentation by the treating surgical team. Available records were reviewed for wound erythema, localized swelling or warmth, purulent drainage, wound separation, fluid collection, antibiotic treatment, drainage, debridement, or reoperation. Because these assessments were obtained retrospectively and were not based on a prospectively standardized surveillance protocol, possible outcome misclassification was considered a study limitation. The date of the first documented wound complication was recorded. Complications were categorized as diagnosed on or before POD3 or after the POD3 blood sample. For complications diagnosed after POD3, the interval between blood sampling and diagnosis was calculated. Diabetes mellitus was defined as a documented diagnosis or current use of antidiabetic medication. Smoking status was classified as current smoker or non-smoker. ASA classification was grouped as I–II or III–IV. Surgical procedures were classified as emergency or elective according to operative records. Operative time was defined as the duration from skin incision to skin closure and was recorded in minutes. This definition was applied consistently across procedure types. Wound contamination class was grouped as class I–II or class III–IV. Open procedures were classified as open surgery, whereas laparoscopic and other minimally invasive procedures were grouped as non-open surgery. Major abdominal surgery included colorectal resections, upper gastrointestinal procedures, liver and pancreatic resections, and selected complex biliary or abdominal wall procedures.
Calculation of CAR, WINA, and dynamic changes
The CRP-to-albumin ratio (CAR) was calculated by dividing CRP by serum albumin. WINA was evaluated as a newly proposed exploratory composite index. It was calculated by multiplying CAR by NLR. Accordingly, WINA was obtained as CRP divided by albumin and then multiplied by NLR. CRP was expressed in mg/L, serum albumin in g/dL, and NLR was calculated by dividing the absolute neutrophil count by the absolute lymphocyte count. CAR and WINA were calculated separately using the POD1 and POD3 laboratory values. Dynamic changes were calculated by subtracting the POD1 value from the corresponding POD3 value. Therefore, ΔCRP, Δalbumin, ΔNLR, ΔCAR, and ΔWINA represented the change between POD1 and POD3. No transformation or normalization was applied before calculation. All ratios and composite indices were calculated separately for each patient before group-level summary statistics were obtained.
Statistical analysis
Statistical analyses were performed using IBM SPSS Statistics version 26.0 (IBM Corp., Armonk, NY, USA) and R version 4.3.0 (R Foundation for Statistical Computing, Vienna, Austria). Normality was assessed using the Shapiro–Wilk test. Continuous variables were presented as mean ± standard deviation or median [interquartile range], and categorical variables as n (%). Between-group comparisons were performed using the independent-samples t-test, Mann–Whitney U test, chi-square test, or Fisher’s exact test, as appropriate. Factors associated with postoperative wound complications were evaluated using logistic regression. Because of the limited number of events, the multivariable analysis was restricted to a parsimonious clinical model including diabetes mellitus and operative time, followed by an extended model additionally including ΔWINA. Operative time and ΔWINA were expressed per 30-minute and 100-unit increase, respectively. Results were reported as odds ratios with 95% confidence intervals. The diagnostic performances of POD3 biomarkers and ΔWINA were assessed using receiver operating characteristic analysis. Areas under the curve were compared using the DeLong test with Holm adjustment, and optimal cut-off values were determined using the Youden index. The incremental value of ΔWINA was evaluated by comparing the clinical and extended models using the change in AUC and the likelihood-ratio test. Model calibration and internal validity were assessed using the calibration slope and intercept, Hosmer–Lemeshow test, Brier score, and 1,000 bootstrap resamples. Sensitivity analyses excluded complications diagnosed on or before POD3 and separately restricted the outcome to surgical site infections. A two-sided p-value < 0.05 was considered statistically significant.
Results
A total of 312 patients were screened for eligibility. Of these, 226 were excluded because of missing POD1 or POD3 laboratory measurements (n = 147), pre-existing systemic infection or chronic inflammatory disease (n = 28), incomplete clinical records (n = 31), insufficient postoperative follow-up (n = 15), or age < 18 years (n = 5). The final study sample included 86 patients, of whom 24 (27.9%) developed postoperative wound complications. Patients with wound complications were older (68.4 ± 10.2 vs. 61.7 ± 11.5 years, p = 0.011) and had a higher BMI (29.8 ± 4.6 vs. 26.9 ± 3.8 kg/m², p = 0.009). Diabetes mellitus, ASA III–IV status, emergency surgery, open surgical approach, and contamination class III–IV were more frequent among patients with wound complications. Operative time was longer in the complication group than in the non-complication group (178 ± 42 vs. 142 ± 36 min, p = 0.001). Sex, smoking status, malignant disease, major abdominal surgery, and the overall distribution of surgical procedure categories did not differ between groups. The patient selection process is presented in Fig. 1, and the baseline clinical and surgical characteristics are presented in Table 1.
Table 1.
Baseline clinical and surgical characteristics according to postoperative wound complication status
| Variable | Total (n = 86) | No complication (n = 62) | Complication (n = 24) | p-value |
|---|---|---|---|---|
| Age (years) | 63.6 ± 11.5 | 61.7 ± 11.5 | 68.4 ± 10.2 | 0.011 |
| Male sex, n (%) | 49 (57.0) | 34 (54.8) | 15 (62.5) | 0.520 |
| BMI (kg/m²) | 27.7 ± 4.2 | 26.9 ± 3.8 | 29.8 ± 4.6 | 0.009 |
| Diabetes mellitus, n (%) | 24 (27.9) | 13 (21.0) | 11 (45.8) | 0.021 |
| Current smoking, n (%) | 31 (36.0) | 21 (33.9) | 10 (41.7) | 0.499 |
| ASA III–IV, n (%) | 34 (39.5) | 20 (32.3) | 14 (58.3) | 0.027 |
| Emergency surgery, n (%) | 22 (25.6) | 12 (19.4) | 10 (41.7) | 0.033 |
| Malignant disease, n (%) | 44 (51.2) | 29 (46.8) | 15 (62.5) | 0.191 |
| Open surgical approach, n (%) | 44 (51.2) | 27 (43.5) | 17 (70.8) | 0.023 |
| Major abdominal surgery, n (%) | 58 (67.4) | 38 (61.3) | 20 (83.3) | 0.050 |
| Operative time (min) | 152 ± 41 | 142 ± 36 | 178 ± 42 | 0.001 |
| Contamination class III–IV, n (%) | 19 (22.1) | 10 (16.1) | 9 (37.5) | 0.032 |
| Surgical procedure category, n (%) | 0.101 | |||
| Colorectal surgery | 31 (36.0) | 19 (30.6) | 12 (50.0) | — |
| Upper gastrointestinal surgery | 15 (17.4) | 10 (16.1) | 5 (20.8) | — |
| Hepatobiliary or pancreatic surgery | 14 (16.3) | 9 (14.5) | 5 (20.8) | — |
| Hernia or abdominal wall surgery | 16 (18.6) | 15 (24.2) | 1 (4.2) | — |
| Other general surgical procedures | 10 (11.6) | 9 (14.5) | 1 (4.2) | — |
Values are presented as mean ± standard deviation or number (percentage). Continuous variables were compared using the independent-samples t-test or Welch’s t-test, as appropriate. Categorical variables were compared using the Pearson chi-square test or Fisher’s exact test, as appropriate. The p-value for surgical procedure category represents the overall comparison across all categories. Abbreviations: ASA American Society of Anesthesiologists, BMI body mass index, min minutes
Postoperative wound complications occurred in 24 of 86 patients (27.9%). The most frequent complication was superficial incisional surgical site infection, observed in 10 patients (11.6%), followed by deep incisional surgical site infection in four (4.7%), wound dehiscence in four (4.7%), and organ/space surgical site infection in three (3.5%). The median time from surgery to the first documented wound complication was 7 [5–10] days. Four complications (16.7%) were identified on or before POD3, whereas 20 (83.3%) were diagnosed after the POD3 laboratory measurement. Among complications diagnosed after POD3, the median interval between the POD3 blood sample and complication diagnosis was 4 [2–7] days, as presented in Table 2.
Table 2.
Types and timing of postoperative wound complications
| Variable | Value |
|---|---|
| Any postoperative wound complication, n (%) | 24 (27.9) |
| Primary wound complication type, n (%) | |
| Superficial incisional SSI | 10 (11.6) |
| Deep incisional SSI | 4 (4.7) |
| Organ/space SSI | 3 (3.5) |
| Wound dehiscence | 4 (4.7) |
| Seroma or hematoma requiring intervention | 2 (2.3) |
| Other wound-related complication | 1 (1.2) |
| Time to first complication diagnosis (days) | 7 [5–10] |
| Diagnosis on or before POD3, n (% of complications) | 4 (16.7) |
| Diagnosis after POD3, n (% of complications) | 20 (83.3) |
| Interval from POD3 sampling to diagnosis among complications diagnosed after POD3 (days) | 4 [2–7] |
| Complication requiring surgical intervention, n (% of complications) | 7 (29.2) |
| Thirty-day wound-related readmission, n (% of complications) | 5 (20.8) |
Values are presented as number (percentage) or median [interquartile range]. Complication categories represent the first and primary wound-related event documented for each patient and are mutually exclusive. Percentages for complication types were calculated using the total study sample as the denominator. Percentages for complication timing, surgical intervention, and readmission were calculated using patients with wound complications as the denominator. Abbreviations: POD postoperative day, SSI surgical site infection
Postoperative laboratory parameters and composite indices are presented in Table 3. At POD1, patients with wound complications had higher CRP, NLR, CAR, and WINA values and lower serum albumin levels than patients without wound complications. At POD3, CRP, neutrophil count, NLR, CAR, and WINA were higher, whereas lymphocyte count and serum albumin were lower in the complication group. The median ΔCRP, ΔNLR, ΔCAR, and ΔWINA values were also higher in patients with wound complications. The median ΔWINA was 352.8 [236.4–486.7] in the complication group and 78.6 [36.9–146.8] in the non-complication group (p < 0.001).
Table 3.
Postoperative laboratory parameters and composite indices according to wound complication status
| Variable | Total (n = 86) | No complication (n = 62) | Complication (n = 24) | p-value |
|---|---|---|---|---|
| CRP (mg/L) | ||||
| POD1 | 86 [62–118] | 78 [55–102] | 112 [85–146] | < 0.001 |
| POD3 | 122 [88–162] | 104 [76–138] | 168 [132–210] | < 0.001 |
| ΔCRP | 31 [12–57] | 24 [8–44] | 52 [28–81] | < 0.001 |
| Albumin (g/dL) | ||||
| POD1 | 3.4 ± 0.5 | 3.5 ± 0.5 | 3.1 ± 0.4 | 0.002 |
| POD3 | 3.1 ± 0.4 | 3.3 ± 0.4 | 2.8 ± 0.3 | < 0.001 |
| ΔAlbumin | −0.3 ± 0.3 | −0.2 ± 0.3 | −0.4 ± 0.3 | 0.004 |
| Neutrophil count (×10⁹/L) | ||||
| POD1 | 7.4 [6.0–9.1] | 7.1 [5.8–8.6] | 8.2 [6.7–9.9] | 0.061 |
| POD3 | 8.3 [6.7–10.2] | 7.6 [6.1–9.2] | 9.8 [8.2–11.6] | 0.001 |
| Lymphocyte count (×10⁹/L) | ||||
| POD1 | 1.3 [1.0–1.7] | 1.4 [1.1–1.8] | 1.1 [0.8–1.5] | 0.028 |
| POD3 | 1.2 [0.9–1.6] | 1.4 [1.1–1.8] | 0.9 [0.7–1.2] | < 0.001 |
| NLR | ||||
| POD1 | 4.9 [3.5–7.1] | 4.3 [3.1–6.0] | 6.8 [5.2–8.9] | 0.002 |
| POD3 | 6.9 [4.9–9.2] | 5.6 [4.2–7.4] | 10.5 [8.4–13.2] | < 0.001 |
| ΔNLR | 1.6 [0.5–3.3] | 1.1 [0.2–2.1] | 3.6 [2.0–5.5] | < 0.001 |
| CAR | ||||
| POD1 | 25.7 [17.8–37.4] | 22.4 [15.9–30.8] | 36.2 [26.5–48.9] | < 0.001 |
| POD3 | 39.1 [27.3–57.8] | 31.8 [23.1–42.6] | 60.4 [46.8–78.9] | < 0.001 |
| ΔCAR | 10.9 [4.6–22.8] | 8.3 [3.2–15.9] | 23.1 [12.4–35.6] | < 0.001 |
| WINA | ||||
| POD1 | 132.8 [82.6–221.5] | 96.7 [63.5–151.9] | 244.6 [174.8–334.2] | < 0.001 |
| POD3 | 281.4 [164.7–486.9] | 178.5 [119.4–276.3] | 628.7 [472.9–803.6] | < 0.001 |
| ΔWINA | 132.4 [57.1–278.6] | 78.6 [36.9–146.8] | 352.8 [236.4–486.7] | < 0.001 |
Values are presented as mean ± standard deviation or median [interquartile range]. CAR and WINA were calculated separately for each patient before the summary statistics were generated. Continuous variables were compared using the independent-samples t-test, Welch’s t-test, or Mann–Whitney U test, as appropriate. Delta values were calculated as the POD3 value minus the corresponding POD1 value. Abbreviations: CAR C-reactive protein-to-albumin ratio, CRP C-reactive protein, NLR neutrophil-to-lymphocyte ratio, POD postoperative day, WINA Wound Inflammation–Nutrition Index
In univariable logistic regression analysis, age, BMI, diabetes mellitus, ASA III–IV status, emergency surgery, open surgical approach, operative time, contamination class III–IV, and ΔWINA were associated with postoperative wound complications. Model 1 included diabetes mellitus and operative time as prespecified clinical covariates. Model 2 additionally included ΔWINA. In Model 2, diabetes mellitus (adjusted OR: 2.89, 95% CI: 1.01–8.29, p = 0.048), operative time per 30-minute increase (adjusted OR: 1.67, 95% CI: 1.08–2.58, p = 0.021), and ΔWINA per 100-unit increase (adjusted OR: 2.12, 95% CI: 1.44–3.13, p = 0.001) remained associated with postoperative wound complications. Variance inflation factors were below 2.0 for all variables included in the final model, as presented in Table 4.
Table 4.
Univariable and parsimonious multivariable logistic regression analyses for postoperative wound complications
| Variable | Univariable OR | 95% CI | p-value | Model 1 adjusted OR | 95% CI | p-value | Model 2 adjusted OR | 95% CI | p-value |
|---|---|---|---|---|---|---|---|---|---|
| Age, per 10 years | 1.74 | 1.10–2.75 | 0.018 | — | — | — | — | — | — |
| BMI, per 1 kg/m² | 1.18 | 1.05–1.33 | 0.006 | — | — | — | — | — | — |
| Male sex | 1.37 | 0.52–3.60 | 0.521 | — | — | — | — | — | — |
| Diabetes mellitus | 3.18 | 1.18–8.58 | 0.022 | 2.76 | 1.01–7.52 | 0.047 | 2.89 | 1.01–8.29 | 0.048 |
| Current smoking | 1.39 | 0.53–3.63 | 0.500 | — | — | — | — | — | — |
| ASA III–IV | 2.94 | 1.10–7.85 | 0.031 | — | — | — | — | — | — |
| Emergency surgery | 2.98 | 1.07–8.28 | 0.037 | — | — | — | — | — | — |
| Malignant disease | 1.90 | 0.72–5.04 | 0.194 | — | — | — | — | — | — |
| Open surgical approach | 3.15 | 1.14–8.69 | 0.027 | — | — | — | — | — | — |
| Major abdominal surgery | 3.16 | 0.99–10.10 | 0.052 | — | — | — | — | — | — |
| Operative time, per 30 min | 1.96 | 1.35–2.84 | 0.001 | 1.84 | 1.25–2.71 | 0.002 | 1.67 | 1.08–2.58 | 0.021 |
| Contamination class III–IV | 3.12 | 1.07–9.12 | 0.037 | — | — | — | — | — | — |
| ΔWINA, per 100 units | 2.17 | 1.55–3.04 | 0.001 | — | — | — | 2.12 | 1.44–3.13 | 0.001 |
Model 1 included diabetes mellitus and operative time. Model 2 included diabetes mellitus, operative time, and ΔWINA. Variables included in the multivariable models were prespecified according to clinical relevance, while maintaining a parsimonious events-per-variable ratio. WINA and its individual components were not entered simultaneously into the same model. Multicollinearity was assessed using variance inflation factors, which were below 2.0 for all variables included in Model 2. Abbreviations: ASA American Society of Anesthesiologists, BMI body mass index, CI confidence interval, OR odds ratio, WINA Wound Inflammation–Nutrition Index
The diagnostic performance of postoperative biomarkers and composite indices is presented in Table 5. ΔWINA had an AUC of 0.872 (95% CI: 0.789–0.941), with a Youden-derived cut-off value of 245.0. At this threshold, sensitivity was 83.3%, specificity was 80.6%, positive predictive value was 62.5%, negative predictive value was 92.6%, and overall accuracy was 81.4%. WINA at POD3 had an AUC of 0.848 (95% CI: 0.761–0.917), whereas the AUCs for CAR, CRP, NLR, and albumin at POD3 were 0.812, 0.781, 0.752, and 0.721, respectively. In pairwise DeLong comparisons with Holm adjustment, the AUC of ΔWINA differed from the AUCs of albumin, CRP, and NLR at POD3. The differences between ΔWINA and WINA at POD3 and between ΔWINA and CAR at POD3 were not statistically significant. The ROC curves are presented in Fig. 2.
Table 5.
Diagnostic performance of POD3 biomarkers and ΔWINA for postoperative wound complications
| Parameter | AUC | 95% CI | Cut-off | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) | LR+ | LR− |
|---|---|---|---|---|---|---|---|---|---|---|
| ΔWINA | 0.872 | 0.789–0.941 | 245.0 | 83.3 | 80.6 | 62.5 | 92.6 | 81.4 | 4.29 | 0.21 |
| WINA (POD3) | 0.848 | 0.761–0.917 | 402.5 | 79.2 | 77.4 | 57.6 | 90.6 | 77.9 | 3.50 | 0.27 |
| CAR (POD3) | 0.812 | 0.715–0.886 | 44.8 | 75.0 | 72.6 | 51.4 | 88.2 | 73.3 | 2.74 | 0.34 |
| CRP (POD3), mg/L | 0.781 | 0.679–0.864 | 142 | 75.0 | 71.0 | 50.0 | 88.0 | 72.1 | 2.59 | 0.35 |
| NLR (POD3) | 0.752 | 0.648–0.842 | 7.8 | 70.8 | 72.6 | 50.0 | 86.5 | 72.1 | 2.58 | 0.40 |
| Albumin (POD3), g/dL | 0.721 | 0.612–0.815 | 3.0 | 70.8 | 67.7 | 45.9 | 85.7 | 68.6 | 2.19 | 0.43 |
Cut-off values were derived from the study sample using the Youden index. For albumin, values below the cut-off were classified as positive. Positive and negative predictive values were calculated using the observed wound complication prevalence of 27.9%. Pairwise AUC comparisons were performed using the DeLong test with Holm adjustment for five comparisons. For ΔWINA versus WINA at POD3, the unadjusted and adjusted p-values were 0.184 and 0.184, respectively. The corresponding values were 0.088 and 0.176 for CAR at POD3, 0.014 and 0.042 for CRP at POD3, 0.006 and 0.024 for NLR at POD3, and 0.002 and 0.010 for albumin at POD3. Abbreviations: AUC area under the receiver operating characteristic curve, CAR C-reactive protein-to-albumin ratio, CI confidence interval, CRP C-reactive protein, LR+ positive likelihood ratio, LR− negative likelihood ratio, NLR neutrophil-to-lymphocyte ratio, NPV negative predictive value, POD postoperative day, PPV positive predictive value, WINA Wound Inflammation–Nutrition Index
Fig. 2.
Receiver operating characteristic curves of POD3 biomarkers and ΔWINA for postoperative wound complications
The clinical model including diabetes mellitus and operative time had an apparent AUC of 0.772 (95% CI: 0.661–0.861). After the addition of ΔWINA, the apparent AUC increased to 0.903 (95% CI: 0.827–0.958), corresponding to an absolute AUC increase of 0.131 (95% CI: 0.044–0.218, p = 0.003). The likelihood-ratio test also indicated a difference between the nested models (χ²=18.74, p < 0.001). Following 1,000 bootstrap resamples, the optimism-corrected AUC was 0.742 for the clinical model and 0.868 for the clinical model plus ΔWINA. The corresponding optimism-corrected calibration slopes were 0.81 and 0.84, respectively. In sensitivity analyses, ΔWINA remained associated with wound complications after excluding the four events diagnosed on or before POD3 and when the outcome was restricted to surgical site infections. The results of the model comparisons, internal validation, and primary sensitivity analyses are presented in Table 6. Calibration plots are presented in Fig. 3.
Table 6.
Incremental model performance, internal validation, and sensitivity analyses
| Analysis | Estimate | 95% CI | p-value |
|---|---|---|---|
| Panel A. Model performance and internal validation | |||
| Clinical model apparent AUC | 0.772 | 0.661–0.861 | — |
| Clinical model optimism-corrected AUC | 0.742 | — | — |
| Clinical model Brier score | 0.171 | — | — |
| Clinical model calibration slope | 0.81 | — | — |
| Clinical model calibration intercept | −0.06 | — | — |
| Clinical model Hosmer–Lemeshow p-value | — | — | 0.517 |
| Clinical model + ΔWINA apparent AUC | 0.903 | 0.827–0.958 | — |
| Clinical model + ΔWINA optimism-corrected AUC | 0.868 | — | — |
| Clinical model + ΔWINA Brier score | 0.118 | — | — |
| Clinical model + ΔWINA calibration slope | 0.84 | — | — |
| Clinical model + ΔWINA calibration intercept | −0.04 | — | — |
| Clinical model + ΔWINA Hosmer–Lemeshow p-value | — | — | 0.684 |
| Change in AUC after adding ΔWINA | 0.131 | 0.044–0.218 | 0.003 |
| Likelihood-ratio χ² | 18.74 | — | < 0.001 |
| Panel B. Sensitivity analyses for ΔWINA | |||
| Excluding complications diagnosed on or before POD3: adjusted OR per 100 units | 1.96 | 1.32–2.91 | 0.001 |
| Excluding complications diagnosed on or before POD3: AUC | 0.856 | 0.758–0.930 | — |
| SSI-only outcome: adjusted OR per 100 units | 2.04 | 1.31–3.17 | 0.002 |
| SSI-only outcome: AUC | 0.861 | 0.764–0.933 | — |
| Elective surgery subgroup: unadjusted OR per 100 units | 1.88 | 1.18–2.99 | 0.008 |
| Elective surgery subgroup: AUC | 0.844 | 0.730–0.921 | — |
| Major abdominal surgery subgroup: adjusted OR per 100 units | 2.03 | 1.31–3.15 | 0.002 |
| Major abdominal surgery subgroup: AUC | 0.858 | 0.754–0.930 | — |
The clinical model included diabetes mellitus and operative time. The extended model included diabetes mellitus, operative time, and ΔWINA. Internal validation was performed using 1,000 bootstrap resamples, with the complete model-fitting procedure repeated in each resample. Calibration slope and intercept values represent optimism-corrected estimates. The analysis excluding early events included 82 patients and 20 wound complications diagnosed after the POD3 laboratory measurement. The SSI-only analysis included 17 events. These two analyses and the major abdominal surgery analysis were adjusted for diabetes mellitus and operative time. Because of the limited number of events in the elective surgery subgroup, the corresponding odds ratio was obtained from a univariable model. Abbreviations: AUC area under the receiver operating characteristic curve, CI confidence interval, OR odds ratio, PO postoperative day, SSI surgical site infection, WINA Wound Inflammation–Nutrition Index
Fig. 3.
Calibration plots of the clinical model and the clinical model including ΔWINA for postoperative wound complications
The dashed diagonal line represents ideal calibration. Solid curves show the smoothed relationship between predicted and observed probabilities. Points represent grouped observed event rates, with vertical bars indicating 95% confidence intervals. Rug marks along the x-axis show the distribution of predicted probabilities.
The detailed distribution of individual surgical procedures according to surgical approach, urgency, indication, contamination class, and wound complication status is provided in Supplementary Table S1. Additional sensitivity and subgroup analyses according to emergency versus elective surgery, open versus laparoscopic approach, benign versus malignant disease, and major abdominal surgery status are summarized in Supplementary Table S2.
Discussion
In this study, we investigated the clinical relevance of the Wound Inflammation–Nutrition Index (WINA) and its dynamic change (ΔWINA) in the early postoperative period. Patients who developed wound complications had higher WINA values at both POD1 and POD3 and a greater increase between the two measurements. In the parsimonious multivariable model, ΔWINA remained associated with wound complications after adjustment for diabetes mellitus and operative time. The addition of ΔWINA also improved the discrimination of the clinical model. However, ΔWINA did not significantly outperform WINA measured at POD3, indicating that the dynamic and single-time-point measures may capture overlapping biological information. When the individual laboratory parameters are examined in the context of our data, a coherent pattern emerges. CRP levels were elevated at both POD1 and POD3 in patients with wound complications, with a more marked separation between groups at POD3. This temporal pattern is consistent with the kinetics of CRP as an acute-phase reactant, whereby sustained or increasing concentrations may reflect continued inflammatory activity rather than the expected resolution of the initial postoperative response. Neutrophil counts and NLR values showed a similar pattern, particularly at POD3, suggesting a more persistent inflammatory response among patients who subsequently developed wound complications [13–15]. These findings are also consistent with previous studies showing that POD3 CRP may help identify patients at increased risk of major postoperative complications. In the present cohort, however, WINA and ΔWINA demonstrated numerically higher discrimination than CRP alone.
Serum albumin concentrations were lower in the complication group at both time points and declined further by POD3. Albumin should not be regarded as a nutritional marker in isolation, because its concentration is also affected by inflammation, capillary leakage, fluid shifts, and hepatic synthesis. Increased vascular permeability and redistribution during systemic inflammation may contribute to declining postoperative albumin levels. Hypoalbuminemia has also been associated with impaired tissue repair, reduced fibroblast activity, and compromised immune function, which may increase susceptibility to wound-related complications [16–18]. Taken together, these observations provide a rationale for evaluating CRP, albumin, and NLR within a single composite measure. WINA was newly proposed in this study as an exploratory index obtained by multiplying the CRP-to-albumin ratio by NLR. It should therefore be viewed primarily as an inflammation-related composite incorporating albumin rather than as a direct nutritional index. WINA values at POD1 and POD3 differed between patients with and without complications, suggesting that the combined index may capture several related aspects of the postoperative systemic response [19, 20]. Nevertheless, the selected multiplicative formulation was empirical and was not derived from regression coefficients or an external development cohort. Its mathematical structure and weighting therefore require further evaluation.
An important aspect of the analysis was the evaluation of ΔWINA, which reflects the direction and magnitude of change between POD1 and POD3. Patients with wound complications had a substantially greater increase in WINA over this period. This pattern may indicate persistence or progression of postoperative inflammation, although the present observational data cannot establish a distinct mechanism of failed inflammatory resolution. Persistent inflammation may interfere with angiogenesis, collagen formation, epithelialization, and immune defense, thereby contributing to delayed wound healing and infection [21–23].
From a diagnostic perspective, ΔWINA showed higher AUC values than CRP, NLR, and albumin at POD3. Its performance did not differ significantly from that of WINA at POD3 or CAR at POD3 after correction for multiple comparisons. Therefore, the present findings do not support a general claim that ΔWINA is superior to all simpler postoperative measures. Rather, both POD3 WINA and ΔWINA appeared to provide useful discrimination in this cohort. The potential advantage of ΔWINA is that it incorporates temporal change, whereas POD3 WINA requires only a single measurement and may be simpler to apply [6, 24]. A further finding was the incremental improvement observed when ΔWINA was added to the clinical model containing diabetes mellitus and operative time. The apparent AUC increased from 0.772 to 0.903, and the difference between the nested models was statistically significant. Although bootstrap correction reduced the estimated performance, the extended model retained acceptable discrimination. These findings suggest that ΔWINA may provide information beyond the selected clinical variables, but the magnitude of improvement should be interpreted cautiously because the same cohort was used for model development and evaluation.
Diabetes mellitus and operative time were also associated with wound complications. Diabetes may impair wound healing through microvascular dysfunction, altered leukocyte activity, reduced collagen synthesis, and abnormal cytokine signaling. Prolonged operative time may reflect greater tissue manipulation, longer exposure to contamination, increased physiological stress, and more complex surgery [25, 26]. The association between ΔWINA and wound complications persisted after adjustment for these variables, although residual confounding related to surgical approach, urgency, contamination class, and procedural complexity cannot be excluded. The temporal relationship between biomarker measurement and complication diagnosis was also considered. Four complications were diagnosed on or before POD3 and could therefore have influenced the POD3 laboratory findings. After these cases were excluded, ΔWINA remained associated with complications diagnosed after POD3. Similar results were observed when the outcome was restricted to surgical site infections. These analyses reduce, but do not completely eliminate, the possibility that elevated POD3 biomarkers represented concurrent detection rather than prediction of an already developing complication.
From a practical perspective, CRP, albumin, neutrophil count, and lymphocyte count are commonly available in postoperative care, and WINA can be calculated without additional laboratory testing. However, routine availability of complete POD1 and POD3 measurements cannot be assumed in every surgical population. In addition, the cut-off identified in this study was derived from the same sample and should not be directly adopted for clinical decision-making. Prospective validation is required to determine whether WINA or ΔWINA can meaningfully guide postoperative monitoring or intervention.
Limitations
This study has several limitations. Its retrospective, single-center design limits causal interpretation and generalizability, while the requirement for complete POD1 and POD3 laboratory measurements may have introduced selection bias by favoring patients who were more closely monitored. The sample size was small, with only 24 wound complications, and although the multivariable model was restricted and internally validated, residual overfitting cannot be excluded. The cohort included heterogeneous surgical procedures, urgencies, approaches, indications, and contamination classes, which may have influenced postoperative inflammatory trajectories. Wound complications were identified from routine clinical documentation rather than a prospectively standardized surveillance protocol, creating a risk of under-detection or misclassification. In four patients, the complication was diagnosed on or before POD3, so the POD3 laboratory values may have reflected an already developing event; however, the findings remained similar after these cases were excluded. Laboratory measurements were obtained during routine care and may have been affected by differences in sampling time, fluid balance, blood loss, transfusion, and perioperative treatment. WINA was not directly compared with established composite scores such as the modified Glasgow Prognostic Score, which limits conclusions regarding its relative advantage over simpler or previously validated indices. In addition, the study period included the COVID-19 pandemic, during which surgical case selection, perioperative pathways, laboratory testing, and postoperative follow-up may have differed from routine practice. Finally, WINA is a newly proposed exploratory index with an empirically selected formula, and its cut-off values and model performance were derived from the same cohort. Therefore, the findings should be interpreted cautiously until they are confirmed in larger, prospective, multicenter studies with standardized outcome definitions and external validation.
Conclusion
In conclusion, WINA and ΔWINA were associated with postoperative wound complications, and adding ΔWINA to a parsimonious clinical model improved model performance. However, ΔWINA did not demonstrate superior discrimination compared with WINA measured at POD3. These findings support the potential value of WINA-based measures for early postoperative risk assessment, but they should be considered exploratory. Prospective multicenter studies with standardized outcome definitions and external validation are required before routine clinical use.
Supplementary information
Below is the link to the electronic supplementary material.
Author contributions
Conceptualization, E.Y. and M.F.Ç.; methodology, E.Y., M.F.Ç., and M.E.G.; formal analysis, E.Y. and M.E.G.; investigation, E.Y., F.G., and M.F.Ç.; data curation, F.G. and M.F.Ç.; writing—original draft preparation, E.Y.; writing—review and editing, M.F.Ç., M.E.G., and F.G.; supervision, M.F.Ç. All authors have read and agreed to the published version of the manuscript.
Funding
No funding was received for this study.
Data availability
The data supporting the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics committee approval
This study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Duzce University Non-Interventional Clinical Research Ethics Committee (approval number: 2026/119; date: March 25, 2026).
Informed consent
The requirement for informed consent was waived by the ethics committee due to the retrospective nature of the study.
Competing interests
The authors declare no competing interests.
Clinical trial registration
Not available.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Lozada Hernández EE, Hernández Bonilla JP, Hinojosa Ugarte D, Magdaleno García M, Mayagoitía González JC, Zúñiga Vázquez LA, Obregón Moreno E, Jiménez Herevia AE, Cethorth Fonseca RK (2023) Ramírez Guerrero P: Abdominal wound dehiscence and incisional hernia prevention in midline laparotomy: a systematic review and network meta-analysis. Langenbeck’s archives Surg 408(1):268 [DOI] [PubMed] [Google Scholar]
- 2.Patel SV, Paskar DD, Nelson RL, Vedula SS, Steele SR (2017) Closure methods for laparotomy incisions for preventing incisional hernias and other wound complications. Cochrane Database Syst Rev 11(11):Cd005661 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tomioka K, Murakami M, Fujimori A, Watanabe M, Koizumi T, Goto S, Otsuka K, Aoki T (2017) Risk Factors for Transumbilical Wound Complications in Laparoscopic Gastric and Colorectal Surgery. In vivo (Athens Greece) 31(5):943–948 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Yücel M, Uğuz E, Sağlam MF, Erdoğan KE, Hıdıroğlu M, Alili A, Küçüker ŞA (2025) Predictive Role of Systemic Inflammatory Indices in Surgically Managed Postpericardiotomy Syndrome Following Cardiac Surgery. Diagnostics 15(12):1488 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ahmed R, Hamdy O (2025) The role of inflammatory biomarkers in predicting postoperative fever following flexible ureteroscopy. 61(8) [DOI] [PMC free article] [PubMed]
- 6.Shaukat W, Baig AM, Ali Z, Kumari K, Tariq T, Soomro AK, Flemming A, Hamayun Khan MT (2025) Prognostic Value of C-reactive Protein and Neutrophil-to-Lymphocyte Ratio in Predicting Postoperative Infections After Gastrointestinal Surgery: A Meta-Analysis. Cureus 17(8):e91123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Güneş Y, Teke E, Taşdelen İ, Ağar M, Şahinoğulları H, Aydın MT (2024) Effectiveness of preoperative biomarkers: role of the c-reactive protein/albumin ratio and systemic immune-inflammation index in predicting acute cholecystitis severity. Istanbul Med J
- 8.Shi J, Tang S, Shen C, Xu D, Tian WZ, Xu Z (2025) The role of nutritional and inflammatory markers in predicting postoperative complications after esophagectomy for esophageal squamous cell carcinoma: mechanisms, clinical applications, and future perspectives. Front Surg 12:1671783 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Bu N, Liu C, Kong Z, Gao W, Zhu Y (2025) Predictive value of preoperative inflammatory response markers on short-term postoperative complications following colorectal surgery: a secondary analysis of a randomized clinical trial. Front Med 12:1536807 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Mullin G, Zager Y (2022) Inflammatory markers may predict post-operative complications and recurrence in Crohn’s disease patients undergoing gastrointestinal surgery. 92(10):2538–2543 [DOI] [PMC free article] [PubMed]
- 11.Zhang Y, Zeng NM, Fan JY, Qiu JP, Jiang FD, Wang Y (2025) Relationship Between Inflammation-Nutrition Composite Indices and Ovarian Endometrioma: A Retrospective Observational Study. J Inflamm Res 18:14871–14880 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tang Y, Lin X, Wang L (2026) The role of systemic immune-inflammation index and prognostic nutritional index in predicting outcomes of chinese patients with diabetic foot ulcers after moist exposed burn ointment treatment. 19:578889 [DOI] [PMC free article] [PubMed]
- 13.Choi JDW, Kwik C, Shanmugalingam A (2023) C-reactive protein as a predictive marker for anastomotic leak following restorative colorectal surgery in an enhanced recovery after surgery program. 27(11):2604–2607 [DOI] [PMC free article] [PubMed]
- 14.Smit C, Janssen-Heijnen ML, van Osch F, Rops J, Gielen AHC, van Heinsbergen M, Melenhorst J, Konsten JLM (2025) The optimal cut-off value of postoperative day three C-reactive protein to predict for major complications in colorectal cancer patients. Langenbeck’s archives Surg 410(1):85 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kolbas I, Karatekin B, Ergun Alış E, Cicin I (2025) Early Postoperative C-Reactive Protein Trajectories After Thoracic Surgery: A Retrospective Cohort Study. Biomedicines 13(10):2532 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Li Y, Chen L, Yang X, Cui H, Li Z, Chen W, Shi H, Zhu M (2025) Dynamic association of serum albumin changes with inflammation, nutritional status and clinical outcomes: a secondary analysis of a large prospective observational cohort study. Eur J Med Res 30(1):679 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Capurso C, Lo Buglio A (2025) C-reactive protein/albumin ratio vs. prognostic nutritional index as the best predictor of early mortality in hospitalized older patients. Regardless Admitting Diagnosis. 17(17) [DOI] [PMC free article] [PubMed]
- 18.Soeters PB, Wolfe RR, Shenkin A (2019) Hypoalbuminemia: Pathogenesis and Clinical Significance. JPEN J Parenter Enter Nutr 43(2):181–193 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Mao M, Wei X, Sheng H, Chi P, Liu Y, Huang X, Xiang Y, Zhu Q, Xing S, Liu W (2017) C-reactive protein/albumin and neutrophil/lymphocyte ratios and their combination predict overall survival in patients with gastric cancer. Oncol Lett 14(6):7417–7424 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Canpolat-Erkan RE, Tekin R (2026) Evaluation of neutrophil-to-lymphocyte ratio and crp-to-albumin ratio in the risk stratification of diabetic foot infection severity. 62(2) [DOI] [PMC free article] [PubMed]
- 21.Almadani YH, Vorstenbosch J, Davison PG, Murphy AM (2021) Wound Healing: A Comprehensive Review. Semin Plast Surg 35(3):141–144 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Xu Q, Shen L, Yang X, Peng H, Liu M (2021) Comparison of changes in wound healing parameters following treatment with three topical wound care products using a laser wound model. Am J translational Res 13(4):2644–2652 [PMC free article] [PubMed] [Google Scholar]
- 23.Sydorchuk L, Gumennyi R, Sydorchuk A, Batih I, Vasiuk V, Sydorchuk R, Kamyshna I, Petakh P, Halabitska I, Kamyshnyi O (2026) Genetic Modulation of Wound Healing Pathways and Postoperative Risk in Plastic and Reconstructive Surgery: A Cohort Study. J Clin Med 15(7):2794 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Liu H, Chen L, Zhu J, Zhang Z, Nie H, Tan C, Gong W, Ai Y, Yuan X, Zhang S et al (2026) The predictive value of neutrophil-to-lymphocyte ratio for the occurrence, progression, and mortality of diabetic nephropathy: a systematic review and meta-analysis. Sci Rep 16(1):1099 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Dasari N, Jiang A, Skochdopole A, Chung J, Reece EM, Vorstenbosch J, Winocour S (2021) Updates in Diabetic Wound Healing, Inflammation, and Scarring. Semin Plast Surg 35(3):153–158 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Fadhil T, Batool A, Khan H, Farooq MS (2025) Wound Healing in Diabetic Patients Undergoing Abdominal Surgery: A Retrospective Study. Cureus 17(8):e90132 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data supporting the findings of this study are available from the corresponding author upon reasonable request.