Integrating Preoperative NLR and PLR with Perioperative Clinical Factors: A Nomogram for Predicting Postoperative CRP Elevation After Laparoscopic Hysterectomy

Introduction

Laparoscopic hysterectomy has become a cornerstone in the surgical management of benign gynecological conditions, offering significant advantages over laparotomy, including reduced intraoperative blood loss, less postoperative pain, shorter hospital stays, and faster recovery times.¹ Despite its minimally invasive nature, the procedure still induces a significant surgical stress response, manifesting as a systemic inflammatory reaction.

Among the various biomarkers of this response, C-reactive protein (CRP), an acute-phase reactant synthesized by the hepatocytes in response to interleukin-6 (IL-6), has emerged as a highly sensitive and objective indicator of tissue trauma and surgical stress.² Elevated postoperative CRP levels have been consistently associated with an increased risk of complications, such as surgical site infections, prolonged ileus, and overall poorer recovery outcomes, not only in major abdominal surgery but also in pelvic procedures.^3,4 Accordingly, identifying patients at risk of an exaggerated postoperative CRP response may facilitate tailored perioperative care and early intervention.

The magnitude of the postoperative inflammatory response is influenced by a complex interplay of patient-specific and procedure-related factors. Patient physiology, such as elevated body mass index (BMI), is a well-established risk factor, as adipose tissue secretes pro-inflammatory cytokines (eg, IL-6 and TNF-α), creating a state of chronic low-grade inflammation that can amplify the acute-phase response to surgery.⁵ Preoperative vital signs, particularly heart rate (HR), may reflect autonomic tone and cardiopulmonary reserve, potentially modulating the neuroendocrine response to surgical stress.⁶

Intraoperative factors, including surgical duration, fluid administration, and blood loss, are also determinants of surgical trauma. Prolonged operation time and significant bleeding contribute directly to inflammation through increased tissue damage and the release of damage-associated molecular patterns (DAMPs).^7,8

In recent years, hematologic indices derived from routine complete blood count (CBC), such as the neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR), have gained prominence as accessible and cost-effective biomarkers of systemic inflammation. The NLR reflects the balance between innate immune activation (neutrophilia) and adaptive immunity (lymphopenia), with a high ratio indicating a pro-inflammatory state and physiological stress.⁹ The PLR integrates information from the coagulation and immune systems, as platelets actively participate in inflammatory signaling.¹⁰ Although these indices have demonstrated prognostic value in oncologic and cardiovascular settings, their utility in forecasting postoperative inflammatory trajectories after elective benign gynecologic surgery remains less explored.

Currently, there is a lack of a comprehensive prediction tool that synthesizes these inflammatory indices (NLR and PLR) to forecast the risk of significant postoperative CRP elevation. Most existing perioperative models focus on postoperative complications (eg, infection or delayed recovery) rather than the early intensity of the inflammatory response itself, which may serve as an actionable intermediate signal before complications become clinically apparent.^11–13 Therefore, CRP-centered prediction may complement complication-based models by enabling earlier postoperative risk stratification and more timely supportive measures.

Given the multifactorial nature of postoperative inflammation, predictive modeling using least absolute shrinkage and selection operator (LASSO) regression and nomograms offers a robust approach to risk stratification. LASSO regression performs variable selection and regularization to identify relevant predictors while reducing model complexity.¹⁴ Nomograms provide a visual and quantitative tool for clinicians to estimate individualized risk based on multiple predictors.¹⁵ Previous investigations have implemented similar methodologies in surgical contexts, illustrating their value in predicting outcomes and informing perioperative management.^16,17

Therefore, this study aimed to develop and internally validate a user-friendly perioperative nomogram that integrates preoperative inflammatory indices and key perioperative clinical factors to predict postoperative CRP elevation following laparoscopic hysterectomy. By providing an individualized risk estimate shortly after surgery, this tool may enhance clinical decision-making, support patient counseling, and facilitate personalized perioperative management strategies to improve recovery.

Materials and Methods

Patient Data

After approval by the Ethics Committee of Fujian Maternity and Child Health Hospital (approval number: No. 2023KY148), we retrospectively collected clinical data from patients who underwent laparoscopic hysterectomy between January 2018 and December 2022 at Fujian Maternity and Child Health Hospital (Fuzhou, Fujian, China). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants. The study flowchart is shown in Figure 1.

Figure 1 Flowchart depicting the process of patient selection, cohort division, model development, and validation for a predictive model in patients undergoing laparoscopic hysterectomy. Initially, 1238 patients who underwent laparoscopic hysterectomy were identified; 39 were excluded based on inclusion/exclusion criteria, resulting in a final dataset of 1199 patients. This dataset was split into a training cohort and a validation cohort. For the training cohort, univariate regression, LASSO regression, and multivariate regression were performed to develop a nomogram prediction model. The validation cohort and the developed model were evaluated using receiver operating characteristic (ROC) curves, calibration analysis, and decision curve analysis (DCA), culminating in study conclusions.

Inclusion and Exclusion Criteria

The inclusion criteria were: (1) patients aged ≥18 years; (2) availability of complete preoperative clinical data, including BMI, vital signs (heart rate), and complete blood count (CBC) within one week before surgery for the calculation of NLR, PLR, and SII; (3) availability of intraoperative data, including surgical duration, infusion quantity, and estimated blood loss; and (4) availability of postoperative CRP measurements (typically taken 24–48 hours after surgery) for outcome assessment.

Patients were excluded based on the following criteria: (1) conversion to laparotomy; (2) presence of preoperative active infection, inflammatory diseases, or autoimmune disorders; (3) diagnosis of malignant disease; (4) use of anti-inflammatory or immunosuppressive medications within one week prior to surgery; or (5) missing data for any of the predictive variables or the outcome measure.

Data Collection and Variables

Demographic, clinical, laboratory, and intraoperative data were extracted from the electronic medical records and anesthesia information management system. Preoperative variables included age, height, weight, body mass index (BMI), and preoperative heart rate (HR, bpm). Intraoperative variables included surgical duration (hours), total intraoperative infusion quantity (crystalloid + colloid, mL), and estimated blood loss (mL), which were recorded at the end of surgery and can therefore be used for immediate perioperative risk estimation after the procedure. Preoperative inflammatory indices were calculated from the complete blood count (CBC) obtained within one week before surgery: neutrophil-to-lymphocyte ratio (NLR) = absolute neutrophil count / absolute lymphocyte count; platelet-to-lymphocyte ratio (PLR) = platelet count / absolute lymphocyte count; systemic immune-inflammation index (SII) = (platelet count × neutrophil count) / lymphocyte count. Outcome: Postoperative CRP elevation was defined as a CRP value >10 mg/L, based on institutional practice and published literature.

Statistical Analysis

The total dataset was randomly split into a training cohort (70%) for model development and an internal validation cohort (30%) for model evaluation. The normality of continuous variables was assessed using the Shapiro–Wilk test. Normally distributed data were presented as mean ± standard deviation and compared using the Student’s t-test. Non-normally distributed data were presented as median (interquartile range, IQR) and compared using the Mann–Whitney U-test. Categorical variables were presented as numbers (percentages) and compared using the Chi-square or Fisher’s exact test. Patients with missing data for any candidate predictor or the outcome were excluded (complete-case analysis); therefore, no imputation was performed.

In the training cohort, the Least Absolute Shrinkage and Selection Operator (LASSO) regression with 10-fold cross-validation was used to select the most predictive variables and reduce overfitting. Prior to LASSO, continuous predictors were standardized (mean=0, SD=1) to place them on a comparable scale (glmnet standardizes predictors by default). Variables with non-zero coefficients were then entered into a multivariable logistic regression model using original measurement units for interpretability. Results are reported as odds ratios (ORs) with 95% confidence intervals (CIs).

A nomogram was constructed based on the final multivariable logistic regression model to visualize the prediction tool. The model’s performance was evaluated in both cohorts. Discrimination was assessed by the area under the receiver operating characteristic curve (AUC). Calibration was evaluated using calibration curves with bootstrap resampling (1000 repetitions) to compare predicted probabilities against observed outcomes. Clinical utility was assessed using decision curve analysis (DCA), which quantifies the net benefit across different threshold probabilities. A two-tailed P-value < 0.05 was considered statistically significant. All statistical analyses were performed using R software (version 4.2.2; R Foundation for Statistical Computing) with packages including glmnet, rms, pROC, and rmda.

Results

Patient Characteristics

According to the inclusion and exclusion criteria, a total of 1199 patients who underwent laparoscopic hysterectomy were included in this study. The entire cohort was randomly divided into a training cohort (n = 839, 70%) for model development and an internal validation cohort (n = 360, 30%) for model evaluation.

The baseline demographic and clinical characteristics of both cohorts are summarized in Table 1. The mean age of the patients was 43 ± 9 years in both groups (P = 0.569). Anthropometric and vital sign measures, including height, BMI, body temperature, heart rate, respiratory rate, and blood pressure, were well-balanced between the two cohorts with no statistically significant differences (all P > 0.05). A statistically significant but clinically modest difference was noted in surgical duration, which was slightly longer in the validation cohort (2.42 ± 0.92 hours) compared to the training cohort (2.28 ± 0.98 hours; P = 0.022). Intraoperative management parameters, including the use of vasoactive drugs, volumes of crystalline and colloidal fluid administration, total infusion quantity, urine output, and amount of bleeding, were consistent across both cohorts (all P > 0.05). Hematologic parameters (HCT, HGB, RBC) showed minor but statistically significant differences, while the inflammatory indices of interest—NLR, PLR, and SII—did not differ significantly between the training and validation cohorts (P = 0.527, 0.708, and 0.735, respectively). The overall balance of baseline variables supports the validity of the subsequent model development and internal validation. Detailed baseline characteristics stratified by postoperative CRP elevation status within each cohort are provided in Supplementary Table 1.

Table 1 Baseline Characteristics of Patients in the Training and Validation Cohorts

Variable Selection and Model Development

Univariate analyses were first performed in the training cohort to explore the association between each candidate predictor and postoperative CRP elevation. All candidate variables, including BMI, HR, surgical duration, infusion quantity, estimated blood loss, NLR, PLR, and SII, differed significantly between patients with and without postoperative CRP elevation (all P < 0.05; Supplementary Table 1).

To avoid overfitting and identify the most informative features, least absolute shrinkage and selection operator (LASSO) regression with 10-fold cross-validation was applied (Figure 2). The optimal lambda (λ) value was selected using the minimum binomial deviance criterion. Eight variables with non-zero coefficients were retained: BMI, HR, surgical duration, infusion quantity, estimated blood loss, NLR, PLR, and SII (Figure 3A and Supplementary Table 2). Receiver operating characteristic (ROC) curves for each single predictor are presented in Figure 3B, and the corresponding AUC values with 95% confidence intervals are summarized in Supplementary Table 3.

Figure 2 Ten-fold cross-validation for selecting the optimal tuning parameter (λ) in the LASSO regression. (A) Binomial deviance plot for LASSO regression. Numbers at the top denote the count of non-zero coefficients at respective λ values. (B) LASSO coefficient profile plot. Each line represents the coefficient of a predictor variable as a function of log(λ).

Figure 3 Lasso-Selected Predictors and Corresponding Coefficients and ROC curve for prediction with a single independent variable. (A) Coefficient plot showing the magnitude of coefficients for candidate predictive variables, including neutrophil-to-lymphocyte ratio (NLR), surgical duration (h), body mass index (BMI), platelet-to-lymphocyte ratio (PLR), estimated blood loss, systemic immune-inflammation index (SII), infusion quantity, and heart rate (HR). (B) ROC curves depicting the diagnostic performance of different indicators (BMI, HR, surgical duration, infusion quantity, estimated blood loss, NLR, PLR, SII) for predicting the target outcome. The area under the curve (AUC) and corresponding 95% confidence intervals (95% CIs) are indicated for each variable.

These eight variables were subsequently entered into a multivariable logistic regression model. The final analysis identified six independent predictors of postoperative CRP elevation: higher BMI (OR 1.08, 95% CI 1.02–1.14, P = 0.008), lower preoperative heart rate (OR 0.97, 95% CI 0.96–0.99, P < 0.001), greater infusion quantity (OR 1.00, 95% CI 1.00–1.00, P = 0.009), greater amount of bleeding (OR 1.00, 95% CI 1.00–1.01, P = 0.044), higher NLR (OR 2.37, 95% CI 1.53–3.67, P < 0.001), and higher PLR (OR 1.01, 95% CI 1.00–1.01, P = 0.003). Surgical duration and SII were not statistically significant independent predictors in the final multivariable model (P =0.225 and P =0.549, respectively) (Table 2).

Table 2 Results of Multivariable Logistic Regression Analysis for Predicting Postoperative CRP Elevation (Training Cohort)

Construction and Performance of the Nomogram

Based on the results of the multivariable analysis, a perioperative nomogram was constructed to provide a visual tool for individualized prediction of postoperative CRP elevation (Figure 4). The nomogram incorporates the six independent predictors: BMI, HR, infusion quantity, estimated blood loss, NLR, and PLR.

Figure 4 Perioperative nomogram for predicting postoperative CRP elevation.Note: To use the nomogram, locate the patient’s value for each variable, draw a line upward to the “Points” axis to determine the score for each variable, sum all the points, and then locate the total points on the “Total Points” axis. A line drawn downward to the “Risk” axis will indicate the individual’s predicted probability of postoperative CRP elevation.

The discriminatory performance of the nomogram was excellent. In the training cohort, the area under the ROC curve (AUC) was 0.772 (95% CI: 0.733–0.811). The model maintained good discrimination in the internal validation cohort, with an AUC of 0.741 (95% CI: 0.677–0.804) (Figure 5).

Figure 5 Receiver operating characteristic (ROC) curves of the nomogram in the training and internal validation cohorts. (A) Training cohort, with an area under the curve (AUC) of 0.772 (95% confidence interval [CI]: 0.733–0.811). (B) Validation cohort, with an area under the curve (AUC) of 0.741 (95% confidence interval [CI]: 0.677–0.804).

Calibration curves for the probability of CRP elevation showed good agreement between the nomogram predictions and the actual observed outcomes in both the training and validation cohorts (Figure 6A and B). The decision curve analysis (DCA) demonstrated that the nomogram provides a positive net benefit across a wide range of threshold probabilities, confirming its potential clinical utility for decision-making (Figure 7A and B).

Figure 6 Calibration curves of the nomogram. (A) Training cohort. (B) Validation cohort. The dotted diagonal line represents the ideal prediction, and the solid line represents the performance of the nomogram (closer to the diagonal indicates better calibration).

Figure 7 Decision curve analysis (DCA) for the nomogram. (A) Training cohort. (B) Validation cohort. The y-axis represents the net benefit. The solid black line represents the assumption that no patients have the event, and the solid grey line represents the assumption that all patients have the event. The red line shows the net benefit of using the nomogram across different threshold probabilities.

Discussion

In this study, we developed and internally validated a perioperative nomogram that integrates readily obtainable preoperative inflammatory indices (NLR and PLR) and key clinical factors (BMI, HR, intraoperative infusion volume, and estimated blood loss) to predict postoperative CRP elevation in patients undergoing laparoscopic hysterectomy. The model demonstrated robust performance, with good discrimination (AUC > 0.74 in both cohorts) and calibration, and was translated into a practical, visual nomogram to aid clinical decision-making.

The most salient finding of our study is the powerful, independent predictive value of the neutrophil-to-lymphocyte ratio (NLR). With the highest odds ratio (OR: 2.37) in our model, a preoperatively elevated NLR was the strongest driver of a significant postoperative inflammatory response. This finding is highly plausible from a pathophysiological perspective. NLR is a composite marker that reflects a dual state of innate immune system activation (neutrophilia) and relative immunosuppression or stress (lymphopenia).^9,18 Surgery induces a rapid release of neutrophils from bone marrow reserves and promotes the secretion of catecholamines and cortisol, which directly cause a redistribution of lymphocytes and a decrease in their circulating count.^19,20 Therefore, a patient presenting with a high NLR even before surgery is likely to have a pre-existing subclinical pro-inflammatory state or diminished immunological resilience, priming them for an exaggerated response to the physiological stress of surgery. Our results align with a growing body of evidence across surgical disciplines, including colorectal and cardiothoracic surgery, where preoperative NLR has been consistently linked to higher rates of postoperative complications and prolonged hospitalization, often mediated by a more pronounced inflammatory cascade.^21–23

Similarly, the platelet-to-lymphocyte ratio (PLR) remained a significant, albeit weaker, independent predictor. Platelets are increasingly recognized as active mediators of inflammation, not just coagulation, through the release of cytokines, chemokines, and other inflammatory mediators.²⁴ A high PLR may thus represent a hypercoagulable and pro-inflammatory milieu. The fact that the more complex systemic immune-inflammation index (SII) was not retained in the final model (P = 0.549) suggests that NLR and PLR capture the most clinically relevant inflammatory signal for this specific outcome, and that adding platelet count again (through SII) may not provide substantial additional information beyond what is offered by the simpler PLR.

Our findings regarding patient physiology are equally noteworthy. The positive correlation between higher BMI and increased risk of CRP elevation is consistent with well-established literature on obesity as a state of chronic, low-grade inflammation.^25,26 Adipose tissue, particularly visceral fat, functions as an active endocrine organ secreting pro-inflammatory adipokines (eg, IL-6, TNF-α), which can amplify the acute-phase response to surgical trauma.²⁷ Interestingly, surgical duration itself was not an independent predictor in the final model, suggesting that for laparoscopic procedures, the inherent inflammatory phenotype associated with obesity may be a more critical determinant of the postoperative response than the purely temporal length of the operation, once blood loss and fluid administration are accounted for.²⁸

The inverse relationship between preoperative heart rate and the risk of inflammation is a particularly intriguing and somewhat counterintuitive result. A lower resting heart rate is generally a marker of better cardiovascular fitness and higher vagal tone.²⁹ The autonomic nervous system plays a crucial role in modulating the immune response; vagal nerve activity can suppress inflammation through the cholinergic anti-inflammatory pathway.³⁰ Therefore, patients with a lower preoperative heart rate may possess a greater physiological reserve and a more robust regulatory mechanism to dampen the surgical stress response and subsequent inflammatory cascade, leading to a lower peak in CRP. This novel finding highlights the potential role of preoperative autonomic tone as a valuable and easily measurable predictor of surgical outcomes, meriting further investigation.

The clinical implications of our study are direct and meaningful. The proposed nomogram offers a practical, evidence-based tool for early perioperative risk stratification. By estimating an individualized risk of a significant inflammatory response shortly after surgery, clinicians can tailor perioperative management plans. For patients identified as high-risk, strategies could include: 1) Enhanced monitoring: closer postoperative observation for early signs of complications; 2) Optimized analgesia: a prioritized multimodal analgesic regimen, minimizing opioids and maximizing anti-inflammatory agents (eg, NSAIDs), which has been shown to attenuate the surgical stress response;³¹ and 3) Prehabilitation: where feasible, targeted prehabilitation programs focusing on nutritional optimization and supervised exercise to modulate baseline inflammation and improve physiological reserve.³²

Our study has several limitations that must be acknowledged. First, its retrospective and single-center design inherently carries risks of selection bias and may limit generalizability; center-specific perioperative protocols and operator-dependent factors (eg, surgeon technique and anesthetic management) may also influence CRP dynamics. Accordingly, external validation in diverse, multi-institutional prospective cohorts is warranted. Second, while we included a comprehensive set of predictors, other potentially influential variables, such as detailed nutritional status (eg, albumin levels), specific comorbidities (eg, subclinical cardiac disease), or surgical complexity scores, were not accounted for and could enhance model accuracy. Third, the outcome was defined as a binary elevation of CRP. Analyzing peak CRP as a continuous outcome might provide more nuanced information. Finally, while the model’s discrimination is good, it is not perfect (AUC > 0.75), indicating residual unpredictability in the inflammatory response, likely due to unmeasured genetic or biological factors.

Future research should focus on the external validation of this nomogram in diverse patient populations. Furthermore, exploring the integration of novel biomarkers, such as interleukin-6 (IL-6) levels or cell-free DNA, could further refine predictive accuracy. Most importantly, the ultimate test of this model’s value will be its implementation in a clinical trial setting to investigate whether targeting high-risk patients with personalized perioperative interventions (eg, intensified anti-inflammatory protocols) actually leads to improved clinical outcomes, such as reduced complication rates or enhanced recovery.

Conclusion

In summary, we have developed a user-friendly perioperative tool that combines routine clinical and laboratory parameters to predict postoperative inflammatory intensity. By identifying patients at high risk for postoperative CRP elevation early in the perioperative period, this nomogram may facilitate a shift toward more personalized, proactive care, with the potential to improve recovery and optimize resource allocation.