The research, conducted by Shiwei Wu of the First Affiliated Hospital of Nanchang University and colleagues, focused on postoperative pancreatic fistula, a complication that occurs when pancreatic juice leaks from the cut surface of the gland and pools in or around the abdomen. Because the pancreas produces enzymes capable of digesting tissue, such leaks can erode blood vessels, trigger abscesses, delay recovery, and in severe cases become life-threatening. The team specifically examined clinically relevant postoperative pancreatic fistula, abbreviated CR-POPF, defined under the 2016 criteria of the International Study Group of Pancreatic Surgery as Grade B or Grade C leaks — those that demand active intervention such as drainage, antibiotics, or reoperation, rather than merely being a biochemical curiosity on a lab report.

To probe whether the state of a patient’s immune and nutritional status before surgery could foreshadow this complication, the investigators assembled a cohort of 300 consecutive patients who underwent distal pancreatectomy between June 2022 and June 2024. Distal pancreatectomy is a common operation for tumors of the body and tail of the pancreas, and it carries one of the highest fistula rates of any abdominal procedure, precisely because the pancreatic stump left behind must seal itself against a torrent of enzymatic secretions. The researchers chose three preoperative laboratory indices that have attracted growing interest across surgical oncology: the C-reactive protein-to-albumin ratio, which captures the balance between systemic inflammation and protein reserves; the inflammatory burden index, or IBI, derived from C-reactive protein and lymphocyte counts; and the CALLY index, which combines C-reactive protein, albumin, and lymphocyte values into a single composite measure of the inflammatory-nutritional-immune milieu.

The headline finding was sobering but not surprising to pancreatic surgeons: clinically relevant pancreatic fistula occurred in 84 of the 300 patients, a rate of 28 percent. The overwhelming majority of these events — 81 of 84, or 96.4 percent — were Grade B leaks requiring intervention, while three patients, 3.6 percent, suffered the more severe Grade C fistulas associated with organ failure or the need for major reoperation. That roughly one in four patients in the cohort experienced a leak demanding treatment underscores why the search for reliable preoperative risk markers remains so urgent in this field.

Using multivariable logistic regression to adjust for other clinical factors, the team found that each of the three indices was independently associated with fistula risk, and in the directions one would intuitively expect. Higher CAR values — signaling more inflammation relative to available albumin — were associated with dramatically higher odds of CR-POPF, with an odds ratio of 230.244 and a 95 percent confidence interval stretching from 45.053 to 1176.652, a result that reached statistical significance at P less than 0.001. Higher inflammatory burden index values likewise carried increased odds, with an odds ratio of 1.015 per unit increase. The CALLY index pointed the opposite way: patients with higher CALLY scores, reflecting a more favorable combination of low inflammation and preserved nutritional and immune reserves, had lower odds of fistula, with an odds ratio of 0.979 per unit.

Restricted cubic spline analysis added an important layer of nuance by mapping the dose-response relationship between each index and fistula risk. Rather than assuming a simple straight-line relationship, this technique allows researchers to visualize how risk changes across the full range of a biomarker, revealing whether thresholds exist or whether risk climbs smoothly. The results confirmed that the associations were not statistical artifacts of arbitrary cutoff points but reflected continuous gradients of risk tied to the underlying biology of inflammation, nutrition, and immune competence.

The most consequential part of the study, however, was the head-to-head comparison of predictive performance. The researchers built a baseline preoperative model using standard clinical variables and then tested how much each inflammation-nutrition index improved it. The baseline model alone achieved an area under the receiver operating characteristic curve — the AUC, a standard measure of discrimination where 0.5 represents coin-flip accuracy and 1.0 represents perfection — of 0.786. Adding CAR lifted that figure to 0.894, a statistically significant gain (P less than 0.001), while adding IBI raised it to 0.821 (P equal to 0.004) and adding CALLY to 0.827 (P equal to 0.016). In other words, a single ratio of two cheap, universally available blood tests nearly pushed the model into the range where it could meaningfully guide individual clinical decisions.

Recognizing that models built and tested on the same dataset tend to look better than they truly are, the team subjected their models to internal validation using 1,000-resample bootstrapping, a statistical technique that repeatedly resamples the data to estimate how much the apparent performance is inflated by overfitting. The bootstrap-corrected AUCs were 0.763 for the baseline model, 0.879 for the model extended with CAR, 0.796 with IBI, and 0.805 with CALLY. The CAR-extended model also achieved the lowest Brier score — a measure combining discrimination and calibration, where lower is better — at 0.120, and decision curve analysis showed it delivered the greatest net benefit across much of the examined range of risk thresholds, meaning that patients and clinicians would, on balance, make better decisions using this model than using alternatives.

What makes this study compelling is not merely the statistical performance but the biological plausibility behind it. C-reactive protein is a sensitive marker of systemic inflammation, while albumin reflects both hepatic synthetic function and nutritional status; lymphocytes are the cellular backbone of adaptive immunity and wound healing. A patient entering surgery with high inflammation and depleted protein reserves faces a hostile healing environment: the fragile pancreatic stump must seal against enzymatic flow, and impaired immune function and poor nutrition are well-known enemies of tissue repair. The CAR ratio, by dividing an inflammation marker by a nutrition marker, distills this entire physiological story into one number that costs pennies to obtain.

The authors are careful to frame their conclusions appropriately. This was a retrospective, single-center study of 300 patients, and the enormous odds ratio for CAR, while striking, comes with wide confidence intervals that reflect the mathematics of rare-event modeling in a modest cohort. The team explicitly notes that external validation in independent, multi-institutional cohorts is needed before the findings can be translated into routine practice. Still, the implications are tantalizing. If confirmed, a preoperative blood panel could help surgeons stratify risk before the operation begins — informing choices about operative approach, drain placement, the use of somatostatin analogs to suppress pancreatic secretion, postoperative surveillance intensity, and candid conversations with patients about what to expect. In an operation where nearly three in ten patients develop a leak requiring treatment, the ability to see that risk coming from a routine blood draw, days before the scalpel touches tissue, would represent a quiet but genuinely transformative advance in pancreatic surgery.

Subject of Research: Preoperative inflammation-nutrition blood indices as predictors of postoperative pancreatic fistula after distal pancreatectomy

Article Title: Correlation analysis between preoperative inflammation-nutrition-related indices and the risk of pancreatic fistula formation following distal pancreatectomy

Article References: Wu, S., Zhang, D., Zhu, Y., Li, N., Wang, D., Zou, J., Hu, X., Jin, J., He, J., Xue, X., & Qin, L. (2026). Correlation analysis between preoperative inflammation-nutrition-related indices and the risk of pancreatic fistula formation following distal pancreatectomy. PLOS One, 21(10), e0360418. https://doi.org/10.1371/journal.pone.0360418

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0360418

Keywords: pancreatic fistula, distal pancreatectomy, C-reactive protein-to-albumin ratio, inflammatory burden index, CALLY index, postoperative complications, pancreatic surgery, risk prediction model, preoperative assessment, PLOS One, retrospective study, surgical oncology