Development and Validation of a Nomogram for Predicting Epidural-Related Maternal Fever Following Labor Analgesia: A Retrospective Cohort Study

Development and Validation of a Nomogram for Predicting Epidural-Related Maternal Fever Following Labor Analgesia: A Retrospective Cohort Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Development and Validation of a Nomogram for Predicting Epidural-Related Maternal Fever Following Labor Analgesia: A Retrospective Cohort Study Ting Chen, Qian Zhou, Xiaohong Zheng, Yuerong Lin, Linshen Huang, and 4 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6854966/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Objectives : Epidural-related maternal fever (ERMF) is a frequent complication following labor analgesia, posing diagnostic challenges due to its overlap with infectious causes. Reliable tools for predicting individualized risk are currently lacking. This study aimed to develop and validate a clinical nomogram to predict the risk of ERMF in parturients undergoing epidural labor analgesia. Methods : A retrospective cohort study was conducted involving parturients receiving epidural analgesia from January 2022 to December 2023 at a single university-affiliated hospital. Eligible participants (n=373) were randomly allocated using a 7:3 ratio into training and validation datasets. Potential predictor variables were screened using the LASSO regression, and significant predictors were identified via multivariable logistic regression analysis within the training set. The nomogram's performance was rigorously evaluated using the Area Under the Receiver Operating Characteristic curve (AUC) for discrimination, calibration plots with associated statistical tests for agreement, and Decision Curve Analysis (DCA) for clinical utility. Results : ERMF was observed in 321 (86.1%) parturients included in the study. Seven independent predictors were incorporated into the final nomogram: body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell count, epidural fentanyl dosage, and baseline temperature. The nomogram demonstrated good discriminative performance with an AUC of 0.847 (95% CI: 0.778–0.917) in the training set and 0.772 (95% CI: 0.649–0.896) in the validation set. Calibration analyses indicated excellent agreement between predicted probabilities and observed ERMF frequencies in both datasets (P > 0.05). DCA confirmed positive net clinical benefit across wide and clinically relevant ranges of threshold probabilities (1.5%-71% in training; 4.5%-54.5% in validation). Discussion : This study successfully developed and internally validated the first nomogram designed to predict individualized risk for ERMF. By integrating readily available clinical factors, the nomogram demonstrates robust predictive accuracy and calibration, alongside potential clinical utility. It represents a valuable tool to support enhanced risk stratification and inform clinical management decisions for parturients receiving epidural labor analgesia, pending further external validation. Epidural-Related Maternal Fever Nomogram Prediction Model Risk Stratification Labor Epidural Analgesia Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Effective management of labor pain is a cornerstone of modern obstetric care, significantly enhancing the maternal experience during childbirth. Among the available analgesic modalities, neuraxial techniques, particularly epidural analgesia, are widely regarded as the gold standard, providing superior pain relief compared to systemic opioids or other methods and facilitating maternal coping during labor. 1 However, the administration of epidural analgesia is associated with several potential side effects and complications, among which intrapartum maternal fever is frequently observed. 2 , 3 A substantial proportion of such febrile episodes occurring after epidural initiation is now recognized as epidural-related maternal fever (ERMF), a phenomenon generally considered distinct from infectious etiologies and attributed primarily to non-infectious inflammatory mechanisms triggered by the epidural procedure or its physiological consequences. 4 , 5 Given the high utilization rates of labor epidural analgesia globally, 6 ERMF represents a common clinical event demanding appropriate understanding and management within obstetric practice. 3 The occurrence of intrapartum fever following epidural analgesia presents a significant clinical challenge due to the inherent difficulty in reliably distinguishing non-infectious ERMF from potentially serious underlying infections, such as chorioamnionitis, based solely on maternal temperature elevation. 2 This diagnostic ambiguity frequently prompts a cascade of clinical responses aimed at excluding infection, often including comprehensive maternal evaluations involving blood cultures and inflammatory marker assessments, as well as extensive neonatal sepsis workups, potentially encompassing blood tests, lumbar puncture, and the administration of empirical antibiotics. 2 , 7 , 8 Such interventions, while undertaken with cautionary intent, contribute substantially to increased healthcare expenditures, generate considerable maternal anxiety, can interfere with early mother-infant interaction, and may expose both mother and neonate to the risks associated with unnecessary investigations and treatments when the underlying cause is non-infectious ERMF. 2 , 8 The current lack of dependable methods to prospectively identify parturients specifically at high risk for developing non-infectious ERMF significantly hinders the ability to mitigate this cascade of interventions. 4 Addressing this clinical need requires improved strategies for risk stratification, yet a notable gap exists in the availability of validated tools designed for the personalized prediction of ERMF in clinical settings 2 , 4 . While numerous observational studies have identified various potential risk factors associated with intrapartum fever during epidural analgesia – encompassing maternal factors, labor characteristics, and specifics of the epidural technique 2 – these factors are often evaluated in isolation, and their collective predictive power has not been effectively synthesized into a readily usable format for individualized risk assessment at the bedside. Predictive modeling offers a robust approach to integrate multiple relevant variables simultaneously, thereby providing a more nuanced and accurate estimation of individual risk. 9 Among various modeling techniques, the nomogram stands out as a particularly valuable tool in clinical practice due to its intuitive graphical format, which facilitates the straightforward calculation of individualized probabilities and has demonstrated utility across diverse medical fields for risk prediction and decision support. 10 Therefore, the primary objective of the present study was to address the aforementioned gap by developing and rigorously validating a novel clinical nomogram specifically designed to predict the individualized risk of epidural-related maternal fever. Utilizing readily ascertainable data encompassing maternal baseline characteristics, specific obstetric history elements, and key intrapartum clinical and treatment variables obtained from a retrospective cohort of parturients receiving epidural labor analgesia, this investigation aimed to construct a statistically robust and clinically pragmatic predictive tool. The successful development and validation of such a nomogram is intended to provide clinicians with an evidence-based instrument to aid in the identification of parturients at elevated risk for ERMF, potentially enabling more targeted surveillance, informed counseling, and rationalized decision-making regarding subsequent investigations or interventions following epidural placement. 2. Materials and Methods 2.1 Study Design and Population This investigation was structured as a retrospective cohort study conducted at The First Affiliated Hospital of Fujian Medical University. Ethical oversight and approval for the study protocol were obtained from the Institutional Ethics Committee (Approval No. MRCTA, ECFAH of FMU [2025] 508), and this study was registered on ClinicalTrials.gov on May 25, 2025 (Identifier: ChiCTR2500102964), ensuring adherence to established ethical principles governing human subject research. The study methodology complied with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to promote rigorous and transparent reporting of this observational research. 11 The analysis focused on parturients who received epidural labor analgesia between January 2022 and December 2023. Inclusion criteria mandated that participants were classified as American Society of Anesthesiologists (ASA) physical status I or II, 12 aged between 22 and 45 years, carrying a singleton pregnancy at 37 to 42 weeks of gestation, and deemed suitable for both a trial of vaginal delivery by an obstetrician and epidural labor analgesia by an anesthesiologist. Furthermore, only parturients with complete clinical documentation available in the hospital records were included, signifying voluntary participation subsequent to confirmation of eligibility. Exclusion criteria were applied to parturients with pregnancy-induced hypertension or gestational diabetes mellitus necessitating pharmacological management, those with a documented history of chronic pain disorders or long-term analgesic medication use, individuals presenting contraindications to neuraxial anesthesia techniques, those diagnosed with pre-existing neurological or significant psychiatric conditions, concurrent enrollment in other clinical trials, or any other circumstances judged by the research team to render the participant unsuitable for inclusion in the study cohort. The primary outcome measure, intrapartum fever, was operationally defined as a maternal oral temperature exceeding 37.0°C during labor following the initiation of epidural analgesia. All epidural procedures were uniformly administered by experienced anesthesiologists adhering to standardized institutional protocols to minimize procedural variability. 2.2 Data Collection and Variable Selection Clinical data for eligible parturients were meticulously extracted retrospectively from multiple integrated electronic health record systems within the institution, specifically the Zhiye Hospital Information System, the Maternal Healthcare Management System, and the MFM-CNS Central Monitoring System. This comprehensive data retrieval process encompassed all parturients receiving epidural labor analgesia during the study period from January 2022 to December 2023. Cases satisfying the predefined selection criteria were systematically compiled into a dedicated epidural analgesia registry database for subsequent analysis. The selection of potential predictive variables was informed by a thorough review of pertinent medical literature and established clinical evaluations pertinent to epidural analgesia and intrapartum fever 2 , 4 . The variables collected encompassed a wide range of demographic, obstetric, labor-related, and biochemical parameters, including: maternal age, body weight upon admission for labor, comprehensive obstetric history (parity, gravidity), baseline maternal temperature prior to epidural initiation, administration of oxytocin for labor augmentation or induction, presence of meconium staining indicating amniotic fluid contamination, duration of the active phase of labor, assessed uterine contraction intensity, the total number of documented vaginal examinations performed during labor, the time interval from spontaneous or artificial rupture of membranes to delivery, the latency period from the initiation of epidural analgesia to the onset of fever (if applicable), serum concentrations of Interleukin-6 (IL-6), Procalcitonin (PCT), and C-reactive protein (CRP), peripheral white blood cell (WBC) count, specific details of the epidural analgesia protocol employed (including maintenance infusion rates and bolus strategies), and the initial preload doses of ropivacaine and fentanyl administered via the epidural catheter. To ensure the integrity and reliability of the dataset, patient records exhibiting a missing data rate of 20% or higher for the selected variables were excluded from the final analysis. Statistical techniques were employed to manage data quality issues: identified outliers were addressed using the trimming method, whereby extreme values falling outside a predefined range were removed, and missing values for included variables were handled using multiple imputation techniques to generate plausible substitute values based on observed data patterns, thereby preserving sample size and reducing potential bias. 13 A flowchart detailing the study participant selection process and overall study design is presented elsewhere (Fig. 1 ). 2.3 Statistical Analysis and Model Development The entire cohort of 373 eligible parturients was partitioned utilizing a computer-generated random allocation sequence into a training dataset and a validation dataset, maintaining a 7:3 ratio, a standard practice in prediction model development to ensure sufficient data for both model derivation and independent testing. 14 This resulted in distinct cohorts designated for initial model construction and subsequent performance evaluation. Prior to model development, baseline demographic and clinical characteristics were formally compared between the training and validation sets using appropriate statistical tests (detailed further in section 2.6 ) to ascertain comparability; no statistically significant differences were detected, confirming the validity of the random allocation process. The identification of salient predictive factors for epidural-related maternal fever was executed exclusively utilizing the training dataset through a sequential analytical approach. Initially, the Least Absolute Shrinkage and Selection Operator (LASSO) regression method was applied. 15 This technique was chosen for its proficiency in managing potential multicollinearity among predictors and its capacity to perform automated feature selection by shrinking the regression coefficients of less contributive variables towards zero, thereby effectively reducing the dimensionality of the predictor space, which is particularly advantageous when dealing with numerous potential covariates. Subsequently, variables retained following the LASSO procedure, identified by their non-zero coefficients, were entered into a multivariable logistic regression analysis. This step aimed to further refine the model by quantifying the independent association strength of each selected predictor with the binary outcome (presence or absence of epidural-related maternal fever) and to identify the final set of statistically significant predictors based on a predetermined significance threshold (P < 0.05). The final predictive model, incorporating the significant predictors identified through the multivariable logistic regression, was subsequently operationalized and visually represented as a nomogram. This graphical tool was constructed based on the regression coefficients derived from the training cohort data, providing a user-friendly interface for estimating the individualized probability of developing intrapartum fever following epidural analgesia. All statistical computations for model development, including LASSO and logistic regression analyses, as well as nomogram generation, were performed using R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria), 16 specifically employing functions within the 'rms' package. 17 A two-tailed P-value less than 0.05 was consistently considered indicative of statistical significance throughout all analytical procedures. 2.4 Model Validation and Performance Assessment The predictive accuracy and clinical applicability of the developed nomogram were rigorously evaluated through a multifaceted validation process conducted independently on both the training dataset (to assess initial fit) and, more critically, the reserved validation dataset (to gauge generalizability and performance on unseen data). 14 The model's discriminative ability, defined as its capacity to correctly differentiate between parturients who experienced epidural-related maternal fever and those who did not, was quantitatively assessed using Receiver Operating Characteristic (ROC) curve analysis. 18 The Area Under the ROC Curve (AUC) was calculated as the primary metric for discrimination, with values ranging from 0.5 (no discrimination better than chance) to 1.0 (perfect discrimination); 95% confidence intervals (CIs) for the AUC were also computed to reflect the precision of the estimate. Concurrently, the calibration of the nomogram was meticulously examined to determine the extent of agreement between the predicted probabilities generated by the model and the actual observed frequencies of intrapartum fever across different risk strata. 19 Calibration was visually assessed using calibration plots, which graph predicted probabilities against observed proportions, with perfect calibration indicated by points falling along a 45-degree diagonal line. Quantitative assessment of calibration involved statistical testing, such as the Spiegelhalter Z test (as referenced in the original protocol documentation) 20 or comparable goodness-of-fit statistics, to formally evaluate the concordance between predictions and observations. Furthermore, the potential clinical utility and practical value of implementing the nomogram in guiding clinical decisions were evaluated using Decision Curve Analysis (DCA). 21 DCA assesses the net benefit derived from using the prediction model across a range of clinically relevant threshold probabilities for intervention (e.g., initiating prophylactic measures or heightened surveillance), comparing the model's utility against default strategies of treating all patients or treating none. To enhance the robustness of the performance estimates and mitigate the risk of overfitting, particularly when evaluating performance on the training set, all validation procedures, including the calculation of AUC, calibration metrics, and DCA net benefit ranges, incorporated bootstrap resampling techniques with 1000 iterations. 14 This internal validation approach provides more stable and less biased estimates of the nomogram's true performance characteristics. 2.5 Data Presentation and Descriptive Statistics Prior to statistical analysis, all continuous variables collected for this study underwent assessment for normality of distribution, typically employing methods such as the Shapiro-Wilk test 22 or visual inspection of histograms and Q-Q plots, to determine the appropriate methods for summarization and subsequent comparative analysis. Based on these assessments, continuous data conforming to a normal distribution were presented as the mean ± standard deviation (SD), providing a measure of central tendency and dispersion suitable for symmetrically distributed data. Conversely, continuous variables exhibiting skewed or non-normal distributions were summarized using the median [M] accompanied by the interquartile range (IQR), specifically detailing the 25th percentile (Q25) and the 75th percentile (Q75), which offers a more robust representation of central tendency and spread for such data. Categorical variables, including demographic factors, clinical characteristics, and binary outcomes like the presence or absence of fever, were presented as absolute frequencies (n) and relative percentages (%), allowing for a clear depiction of the distribution of these characteristics within the study population and its subgroups (training and validation sets). These descriptive statistical approaches were applied following the rigorous data quality control measures outlined previously (Section 2.2 ), including the handling of missing data through multiple imputation and the management of outliers via the trimming method, ensuring that the presented summaries accurately reflect the characteristics of the final analytical cohort. 2.6 Inferential Statistical Analysis Comparative analyses between relevant groups (such as comparing baseline characteristics between the training and validation sets, or comparing predictor variables between febrile and afebrile parturients in univariate analysis, typically presented in the Results section) were conducted using appropriate inferential statistical tests selected based on the data type and distribution. For comparisons of continuous variables between two independent groups, the Independent Student's t-test was employed when the data were normally distributed and exhibited equal variances (homoscedasticity, typically assessed using Levene's test or similar 23 ). In instances where normally distributed data displayed unequal variances, Welch's t-test, an adaptation robust to heteroscedasticity, was utilized 24 . For continuous variables that did not follow a normal distribution, the non-parametric Mann-Whitney U test (also known as the Wilcoxon rank-sum test) was applied to compare the central tendencies of the two groups. 25 Comparisons involving categorical variables were performed using the Pearson's chi-squared (χ²) test. In situations where the assumptions for the χ² test were not met, specifically when expected cell frequencies were low (commonly defined as less than 5 in more than 20% of cells), 26 Fisher's exact test was employed as a more appropriate alternative, particularly for 2x2 contingency tables. Across all inferential statistical procedures undertaken in this study, a two-tailed probability value (P-value) of less than 0.05 was consistently adopted as the threshold for determining statistical significance. All statistical analyses, encompassing descriptive summaries and inferential comparisons, as well as the previously detailed model development and validation steps, were executed using the R statistical software environment, version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria). 16 3. Results 3.1 Study Cohort Characteristics The final study population comprised 373 eligible parturients who received epidural labor analgesia during the specified study period. Following eligibility screening, these participants were randomly allocated, using a 7:3 ratio, into a training cohort (n = 261) designated for model development and an independent validation cohort (n = 112) reserved for testing the model's generalizability. Within the total analyzed cohort, a notably high incidence of the primary outcome was observed, with 321 parturients (approximately 86.1%) developing epidural-related maternal fever (defined as maternal oral temperature > 37.0°C) during labor after epidural initiation. Crucially, a comparative statistical analysis of baseline demographic and pertinent clinical characteristics between the randomly formed training and validation cohorts revealed no statistically significant differences across all measured variables (all P > 0.05). This finding confirms the successful randomization process and ensures that the two cohorts were comparable at the outset, supporting the validity of using the validation cohort for unbiased assessment of the prediction model derived from the training cohort. Comprehensive details regarding the baseline characteristics of the study participants, potentially stratified by the occurrence of fever and explicitly comparing the training versus validation sets, are systematically presented (Table 1, Table 2). 3.2 Identification of Predictors for Epidural-Related Maternal Fever The process of identifying significant factors associated with the development of epidural-related maternal fever was conducted using the data from the training cohort (n=261) through a structured two-phase analytical strategy designed to handle the multiplicity of clinical variables effectively. In the initial phase, Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis was employed as a robust feature selection technique. This method facilitated preliminary variable screening by identifying a subset of potentially influential predictors from a larger pool, while simultaneously mitigating multicollinearity and controlling for potential overfitting, thereby enhancing model stability. The LASSO procedure, including the determination of the optimal regularization parameter (lambda) via fivefold cross-validation, is illustrated graphically (Figure 2A-B). This analysis yielded seven candidate variables demonstrating non-zero coefficients, indicating their potential predictive value: maternal body weight upon admission, prior delivery history (parity), the total number of antenatal vaginal examinations performed, assessed uterine contraction intensity, peripheral white blood cell (WBC) count, the cumulative fentanyl dosage administered via the epidural, and baseline maternal temperature recorded before epidural placement. In the second phase, these seven LASSO-selected variables were incorporated into a multivariable logistic regression model to ascertain their independent predictive contributions while adjusting for each other. The analysis yielded the following adjusted odds ratios (ORs) with corresponding 95% confidence intervals (CIs) and P-values for each variable: body weight (OR=1.034, 95% CI: 0.991–1.079, P=0.124), delivery history (OR=7.541, 95% CI: 1.408–141.4, P=0.058), number of antenatal vaginal examinations (OR=2.860, 95% CI: 1.807–4.768, P<0.001), uterine contraction intensity (OR=2.153, 95% CI: 0.896–5.533, P=0.096), WBC Count (OR=1.196, 95% CI: 1.059–1.356, P=0.004), fentanyl dosage (OR=0.731, 95% CI: 0.513–1.031, P=0.076), and baseline temperature (OR=0.283, 95% CI: 0.023–3.329, P=0.316). Based on the conventional statistical significance threshold (P<0.05), the number of antenatal vaginal examinations and the WBC count emerged as statistically significant independent predictors of epidural-related maternal fever within this multivariate model. The remaining variables, despite being selected by LASSO and included in the final model construction, did not demonstrate statistically significant independent associations at the P<0.05 level in this analysis. Complete, detailed results from the multivariable logistic regression analysis performed on the training cohort are systematically documented (Table 3). 3.3 Development of the Predictive Nomogram Subsequent to the identification of key predictive factors through statistical analysis of the training cohort data, a practical predictive tool was developed in the form of a nomogram. This nomogram was specifically constructed to facilitate individualized risk assessment for epidural-related maternal fever in parturients undergoing labor with epidural analgesia. The development process integrated the seven predictor variables that emerged from the combined LASSO screening and multivariable logistic regression analyses: maternal body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell (WBC) count, cumulative epidural fentanyl dosage, and baseline maternal temperature. The resulting nomogram provides a graphical representation of the multivariable logistic regression model, visually quantifying the weighted contribution of each of these distinct risk factors. By summing the points assigned to each predictor variable based on an individual parturient's characteristics, clinicians can derive an estimated probability of developing intrapartum fever following epidural placement. The complete visual construct of this predictive nomogram, illustrating the points scale for each included variable and the corresponding total points to probability conversion axis, is presented for reference (Figure 3). This tool is intended to serve as a clinical decision-support aid for personalized risk stratification. 3.4 Assessment of Nomogram Discrimination The capacity of the developed nomogram to accurately differentiate between parturients who subsequently developed epidural-related maternal fever and those who remained afebrile was rigorously evaluated using Receiver Operating Characteristic (ROC) curve analysis in both the training and validation datasets. In the training cohort (n=261), the nomogram demonstrated excellent discriminative performance, yielding an Area Under the ROC Curve (AUC) of 0.847 (95% Confidence Interval [CI]: 0.778–0.917). This high AUC value within the development dataset suggests a strong ability of the model to distinguish between the outcome groups based on the included predictors. When applied to the independent validation cohort (n=112), the model maintained good discriminative ability, achieving an AUC of 0.772 (95% CI: 0.649–0.896). While indicating a slight decrease in performance compared to the training set, as is typically expected, this result confirms the model's robust capacity for discriminating risk in a separate sample of parturients not used for its initial development. These findings consistently underscore the nomogram's substantial ability to differentiate parturients at higher versus lower risk for developing epidural-related maternal fever. The corresponding ROC curves visually depicting these discrimination results for both the training and validation sets are provided (Figure 4). 3.5 Assessment of Nomogram Calibration The calibration of the predictive nomogram, which evaluates the concordance between the predicted probabilities of epidural-related maternal fever and the actual observed frequencies of the event, was assessed in both the training and validation cohorts. Visual inspection of the calibration plots revealed excellent agreement between the nomogram's predictions and the observed outcomes across the spectrum of risk in the training set. This visual finding was corroborated by quantitative statistical testing; the Spiegelhalter Z test yielded a P-value of 0.974, indicating no statistically significant difference between the predicted probabilities and the observed frequencies, thus confirming good calibration within the data used for model development. Similarly, assessment in the independent validation cohort demonstrated strong calibration performance. The calibration plot for the validation set also showed close alignment between predicted and observed probabilities, further supported by a non-significant result from the Spiegelhalter Z test (P=0.109). Collectively, these calibration analyses performed on both datasets suggest that the nomogram provides accurate and reliable risk probability estimates for epidural-related maternal fever, demonstrating that the predicted risks align closely with the actual observed risks across different patient strata. The graphical calibration plots illustrating these findings for both cohorts are provided (Figure 5). 3.6 Clinical Utility Assessment via Decision Curve Analysis To evaluate the potential clinical applicability and net benefit derived from using the developed nomogram to guide clinical decisions, Decision Curve Analysis (DCA) was conducted for both the training and validation datasets. DCA assesses the clinical consequences of adopting the prediction model across a range of threshold probabilities, comparing the net benefit of using the nomogram against default strategies, such as classifying all parturients as high-risk (treat all) or all as low-risk (treat none). In the training cohort, the DCA revealed that employing the nomogram to predict epidural-related maternal fever yielded a positive net benefit across a wide range of threshold probabilities, specifically from 1.5% up to 71%. This indicates that within this substantial risk threshold range, utilizing the nomogram for decision-making would be more advantageous than the default strategies. Importantly, the analysis performed on the independent validation cohort corroborated the clinical utility of the nomogram, demonstrating a positive net benefit over threshold probabilities ranging from 4.5% to 54.5%. These findings from the DCA confirm that the nomogram holds potential clinical value, offering improvements in net benefit across clinically relevant risk thresholds in both the development and validation settings. This suggests the model's potential utility in optimizing clinical management strategies related to the risk of epidural-related maternal fever. The graphical representations of the DCA curves for both cohorts are displayed (Figure 6). 4. Discussion This study successfully developed and internally validated the clinical nomogram specifically designed for the prediction of epidural-related maternal fever (ERMF) in parturients undergoing labor analgesia. By integrating seven readily accessible clinical variables – maternal body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell (WBC) count, cumulative epidural fentanyl dosage, and baseline maternal temperature – identified through a rigorous variable selection process involving LASSO regression and multivariable logistic analysis, the nomogram demonstrated commendable predictive performance. The model exhibited good discriminative capacity, effectively distinguishing between parturients who developed ERMF and those who did not, alongside reliable calibration, indicating accurate probability estimation across different risk strata. These findings, observed consistently in both the training and independent validation cohorts, represent a significant advancement towards providing a quantitative, individualized risk assessment tool to address a common and challenging clinical scenario in contemporary obstetric practice 2 , 4 . The inclusion of specific predictors within the final nomogram warrants interpretation regarding their potential mechanistic links to the pathophysiology of ERMF, which is predominantly considered a non-infectious inflammatory response. 4 , 5 The statistically significant association identified between an increased number of antenatal vaginal examinations and ERMF risk aligns with hypotheses suggesting that repeated examinations may increase bacterial translocation or induce localized inflammatory responses, even in the absence of overt clinical infection. 27 Similarly, the elevated WBC count, another significant predictor, likely reflects a heightened systemic inflammatory state, potentially pre-existing or developing in response to the stresses of labor and potentially amplified by epidural analgesia itself. 5 While maternal body weight and delivery history (likely reflecting nulliparity given the odds ratio trend) did not achieve conventional statistical significance in the final multivariable model, their selection by LASSO suggests predictive contribution; obesity is recognized as a chronic pro-inflammatory condition, 28 and nulliparity is often associated with longer labor durations and potentially different inflammatory signaling 2 . The roles of uterine contraction intensity, fentanyl dosage, and baseline temperature appeared less pronounced or more complex in the final adjusted model, suggesting intricate interactions or the need for further investigation into their specific contributions to non-infectious inflammatory pathways during labor with epidural analgesia. The findings of this investigation resonate with, yet also extend, the existing body of literature concerning risk factors for intrapartum fever. Several factors incorporated into the nomogram, such as nulliparity, elevated maternal weight, increased number of vaginal examinations, and higher WBC counts, have been inconsistently reported as potential risk factors in previous studies focusing broadly on intrapartum fever or specifically on fever following epidural initiation 2 . The current study provides further support for the relevance of these factors within the specific context of ERMF prediction. However, the primary contribution and novelty lie not merely in identifying individual risk factors, but in demonstrating their collective predictive utility through the development and validation of an integrated multivariable prediction model. To date, predictive tools combining multiple factors into a composite risk score for ERMF have been lacking 2 , 4 . The construction of this nomogram, therefore, represents a crucial step forward, moving beyond associative analyses of single variables towards a synthesized, quantitative approach for personalized risk estimation, providing a framework potentially more robust than reliance on isolated clinical indicators. The practical implications of this validated nomogram warrant careful consideration for potential clinical application. If externally validated, this tool could serve as a valuable adjunct to clinical judgment at the point of care, potentially calculated using readily available data shortly after epidural placement or during the course of labor. Its primary utility lies in providing an objective, individualized estimation of ERMF risk, thereby enabling clinicians to prospectively identify parturients situated at the higher end of the risk spectrum. Such identification could facilitate tailored patient counseling regarding the likelihood of developing a non-infectious fever and inform decisions about heightened surveillance, such as more frequent temperature monitoring. Furthermore, in parturients predicted to be at low risk who subsequently develop a borderline fever, the nomogram score might contribute, alongside the overall clinical picture, to a more confident deferral of extensive infectious workups, thereby potentially reducing unnecessary interventions, associated costs, and maternal anxiety 2 , 8 . It must be emphasized, however, that the nomogram serves as a risk stratification instrument and decision-support aid; it does not supplant clinical acumen or obviate the need for appropriate investigation of maternal fever, particularly when clinical signs suggestive of infection are present. This investigation possesses several methodological strengths that bolster the credibility of its findings. Foremost is the novelty of developing and validating the nomogram specifically tailored for the prediction of ERMF, addressing a recognized gap in clinical prediction tools for this common scenario. The statistical approach employed was robust, utilizing LASSO regression for objective feature selection to handle potential multicollinearity and reduce model complexity, followed by standard multivariable logistic regression for model finalization. The study benefited from the inclusion of a range of pertinent and generally accessible clinical predictors. Crucially, the internal validation process employed a randomly allocated, independent cohort, providing a rigorous assessment of the model's generalizability beyond the training data. 14 Moreover, the evaluation of the nomogram's performance was comprehensive, extending beyond standard measures of discrimination (AUC) to incorporate critical assessments of calibration (agreement between predicted and observed risk via calibration plots and statistical testing) and potential clinical utility (quantified through Decision Curve Analysis), offering a more holistic perspective on the model's potential value in practice. 19 Despite these strengths, several limitations inherent to the study design and execution must be acknowledged. The retrospective nature of the data collection introduces susceptibility to potential information bias, documentation inconsistencies, and the influence of unmeasured confounding variables that could affect the observed associations. 29 Conducted at a single tertiary care center in Fujian, China, the findings' generalizability may be limited, as patient demographics, baseline risks, specific clinical practices, and epidural analgesia protocols might differ significantly in other institutions or geographical regions. The definition of ERMF relied on a specific oral temperature threshold, which may not capture all relevant inflammatory responses or perfectly align with definitions used elsewhere. Furthermore, some predictor variables, such as uterine contraction intensity, may be subject to subjective assessment or measurement variability. Consequently, the critical next step requires external validation of this nomogram in larger, diverse, multi-center prospective cohorts to confirm its predictive accuracy and transportability across different clinical settings and populations. 9 , 14 Future research should also focus on prospectively evaluating the nomogram's real-world impact on clinical decision-making, resource utilization, and patient outcomes, alongside further investigations into the complex pathophysiology of ERMF to potentially refine predictive models or identify targets for intervention. This study successfully developed and internally validated a novel predictive nomogram incorporating seven readily accessible clinical and obstetric factors for estimating the individualized risk of epidural-related maternal fever in parturients. The nomogram demonstrated good discriminative ability and reliable calibration in both training and validation cohorts, suggesting its potential utility as a practical, quantitative decision-support tool. By enabling enhanced risk stratification, this validated model offers a promising approach to aid clinicians in identifying individuals at higher risk for ERMF, potentially facilitating more tailored patient monitoring, informed counseling, and rationalized clinical management strategies within contemporary obstetric anesthesia practice, pending further external validation. Declarations Ethical Approval This retrospective study was approved by the Institutional Ethics Committee of The First Affiliated Hospital of Fujian Medical University (Approval No. MRCTA, ECFAH of FMU [2025] 508) and adhered to the principles of the Declaration of Helsinki. Patient consent handling for this analysis complied with the approved protocol. All the patients have been informed and signed informed consent before the experiments. Consent for publication Consent for publication was obtained from the participants. Competing interests The authors declare that they have no competing interests. Data and Code Availability Individual participant datasets are not publicly available due to patient privacy and ethical restrictions. De-identified summary data may be accessible from the corresponding author upon reasonable request, contingent upon institutional approval and data use agreements. Analyses employed standard functions in R software (v4.4.2), including the 'rms' package; specific code is available upon reasonable request. Funding This work was supported by the Fujian Medical University Research and Development Fund Project (2023QH1114) and the Youth Scientific Research Project of Fujian Provincial Health Commission (2024QNA039). Author Contributions Conceptualization：X.L., L.L., T.C. Methodology: X.L., L.L., T.C. Software: T.C. Formal Analysis: T.C., Q.Z. Data Curation: X.L., L.D., X.Z., Y.L., L.H. Writing – Original Draft: T.C., X.L. Writing – Review & Editing: X.L., L.L., T.C. Supervision, Project Administration: X.L., T.C., Q.Z. All authors have read and approved the final manuscript. Acknowledgements Gratitude is expressed to the participating parturients and the staff of the Departments of Anesthesiology and Obstetrics at The First Affiliated Hospital of Fujian Medical University for their support in data acquisition and institutional assistance. References American College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 209: Obstetric Analgesia and Anesthesia. Obstet Gynecol. 2019 Mar;133(3):e208-e225. DOI: 10.1097/AOG.0000000000003132. Sharpe EE, Arendt KW. Epidural Labor Analgesia and Maternal Fever. Clin Obstet Gynecol. 2017 Sep;60(3):365-374. DOI: 10.1097/GRF.0000000000000270. Selina Patel, Sarah Ciechanowicz, Yair J Blumenfeld, Pervez Sultan. Epidural-related maternal fever: incidence, pathophysiology, outcomes, and management. Am J Obstet Gynecol. 2023 May;228(5S):S1283-S1304.e1. DOI: 10.1016/j.ajog.2022.06.026. Goetzl L. Epidural analgesia and maternal fever: a clinical and research update. Curr Opin Anaesthesiol. 2012 Jun;25(3):292-9. DOI: 10.1097/ACO.0b013e3283530d7c. Katherine W Arendt, B Scott Segal. The association between epidural labor analgesia and maternal fever. Clin Perinatol. 2013 Sep;40(3):385-98. DOI: 10.1016/j.clp.2013.06.002. Millicent Anim-Somuah, Rebecca Md Smyth, Allan M Cyna, Anna Cuthbert. Epidural versus non-epidural or no analgesia for pain management in labour. Cochrane Database Syst Rev. 2018 May 21;5(5):CD000331. DOI: 10.1002/14651858.CD000331.pub4. Lieberman E, Lang JM, Frigoletto F Jr, Richardson DK, Ringer SA, Cohen A. Epidural analgesia, intrapartum fever, and neonatal sepsis evaluation. Pediatrics. 1997 Mar;99(3):415-9. DOI: 10.1542/peds.99.3.415. Sarah Morton, Justin Kua, Christopher J Mullington. Epidural analgesia, intrapartum hyperthermia, and neonatal brain injury: a systematic review and meta-analysis. Br J Anaesth. 2021 Feb;126(2):500-515. DOI: 10.1016/j.bja.2020.09.046. Moons KGM, Royston P, Vergouwe Y, Grobbee DE, Altman DG. Prognosis and prognostic research: what, why, and how? BMJ. 2009 Feb 23;338:b375. DOI: 10.1136/bmj.b375.. Vinod P Balachandran, Mithat Gonen, J Joshua Smith, Ronald P DeMatteo. Nomograms in oncology: more than meets the eye. Lancet Oncol. 2015 May;16(5):e172-80. DOI: 10.1016/S1470-2045(14)71116-7. Erik von Elm, Douglas G Altman, Matthias Egger, Stuart J Pocock, Peter C Gøtzsche, Jan P Vandenbroucke; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Bmj-british Medical Journal. 2007 Oct 20;335(7624):806-8. DOI: 10.1136/bmj.39335.541782.AD. Doyle DJ, Goyal A, Garmon EH. American Society of Anesthesiologists Classification. [Updated 2023 Jul 3]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK441940/. Sterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009 Jun 29;338:b2393. DOI: 10.1136/bmj.b2393. Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): The TRIPOD Statement. Bmj-british Medical Journal. 2015 Jan 7:350:g7594. DOI: 10.1136/bmj.g7594. Alireza Daneshvar, Golalizadeh Mousa. Regression shrinkage and selection via least quantile shrinkage and selection operator. PLoS One, 2023 Feb 16;18(2):e0266267. DOI: 10.1371/journal.pone.0266267 R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2024. URL https://www.R-project.org/. Sandro Sperandei. Understanding logistic regression analysis. Biochem Med. 2014 Feb 15;24(1):12-8. DOI: 10.11613/BM.2014.003. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982 Apr;143(1):29-36. DOI: 10.1148/radiology.143.1.7063747. Steyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010 Jan;21(1):128-38. DOI: 10.1097/EDE.0b013e3181c30fb2. Spiegelhalter DJ. Probabilistic prediction in patient management and clinical trials. Stat Med. 1986 Jul-Sep;5(5):421-33. DOI: 10.1002/sim.4780050506. Vickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006 Nov-Dec;26(6):565-74. DOI: 10.1177/0272989X06295361. María J Blanca, Jaume Arnau, F J García-Castro, Rafael Alarcón, Roser Bono. Non-normal Data in Repeated Measures ANOVA: Impact on Type I Error and Power. Psicothema. 2023 Feb;35(1):21-29. DOI: 10.7334/psicothema2022.292. Levene H. Robust tests for equality of variances. In: Olkin I, Ghurye SG, Hoeffding W, Madow WG, Mann HB, editors. Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling. Stanford (CA): Stanford University Press; 1960. p. 278-92. B L WELCH. The generalisation of student's problems when several different population variances are involved Biometrika. 1947;34(1/2):28-35. DOI: 10.1093/biomet/34.1-2.28. Joshua N Pritikin, Timothy R Brick, Michael C Neale. Multivariate normal maximum likelihood with both ordinal and continuous variables, and data missing at random. Behav Res Methods. 2018 Apr;50(2):490-500. DOI: 10.3758/s13428-017-1011-6. Shuhua Chang, Deli Li, Yongcheng Qi. Pearson's goodness-of-fit tests for sparse distributions. J Appl Stat. 2021 Dec 30;50(5):1078-1093. DOI: 10.1080/02664763.2021.2017413. Gluck O, Mizrachi Y, Ganer Herman H, Bar J, Kovo M, Weiner E. The correlation between the number of vaginal examinations during active labor and febrile morbidity, a retrospective cohort study. BMC Pregnancy Childbirth. 2020 Aug 28;20(1):495. DOI: 10.1186/s12884-020-02925-9. Ellulu MS, Patimah I, Khaza’ai H, Rahmat A, Abed Y. Obesity and inflammation: the linking mechanism and the complications. Arch Med Sci.2017 Jun;13(4):851-863. DOI: 10.5114/aoms.2016.58928. Jan P Vandenbroucke 1, Erik von Elm, Douglas G Altman, Peter C Gøtzsche, Cynthia D Mulrow, Stuart J Pocock, Charles Poole, James J Schlesselman, Matthias Egger; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007 Oct 16;4(10):e297. DOI: 10.1371/journal.pmed.0040297. Tables Tables 1 to 3 are available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files Table123.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviewers agreed at journal 11 May, 2026 Reviews received at journal 28 Jul, 2025 Reviewers agreed at journal 08 Jul, 2025 Reviewers invited by journal 06 Jul, 2025 Editor invited by journal 12 Jun, 2025 Editor assigned by journal 11 Jun, 2025 Submission checks completed at journal 11 Jun, 2025 First submitted to journal 09 Jun, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6854966","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":482386786,"identity":"aabc2956-8ea1-47cd-b95e-725d27630d6f","order_by":0,"name":"Ting Chen","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ting","middleName":"","lastName":"Chen","suffix":""},{"id":482386787,"identity":"60df1e7d-c4ea-4e3b-8b52-a572ad81e2e8","order_by":1,"name":"Qian Zhou","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Qian","middleName":"","lastName":"Zhou","suffix":""},{"id":482386788,"identity":"553fc723-b381-40bc-9702-cdb6bb011cdb","order_by":2,"name":"Xiaohong Zheng","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaohong","middleName":"","lastName":"Zheng","suffix":""},{"id":482386789,"identity":"a3e666a8-4f10-4a22-8a5c-02da77a89600","order_by":3,"name":"Yuerong Lin","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yuerong","middleName":"","lastName":"Lin","suffix":""},{"id":482386790,"identity":"7aaf714c-cce8-422d-b4f5-db6775130c9f","order_by":4,"name":"Linshen Huang","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Linshen","middleName":"","lastName":"Huang","suffix":""},{"id":482386791,"identity":"97039ced-9946-493a-a6b8-442b68ee473c","order_by":5,"name":"Xin Ling","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xin","middleName":"","lastName":"Ling","suffix":""},{"id":482386792,"identity":"31887256-91f8-4684-8062-d4025effcc1b","order_by":6,"name":"Liting Ding","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Liting","middleName":"","lastName":"Ding","suffix":""},{"id":482386793,"identity":"bdcf873b-9f85-4fc7-8e58-37ffa5f911d3","order_by":7,"name":"Xianzhong Lin","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2ElEQVRIie3NMQrCMBSA4RcimapZXxD0Cu0iBcWzCIW6ODh1lIjQyVkKPYU3qBQ6lboKdfAIdXOoYKqLk6mbYP7hJYF8PACT6QfrNQNhwolsbqwFYS/ii/V3BCAV8v35mfB8hG59RLrLbaiCFHgsdVsyH0VYIokXNomKFPCcaAjZZChkuWoI7YYp2DjTEEpCxLp4bqH3VoTRDJElL0JaEYt5rgg9RfzlYVvMLTxpCOe5U2I9RSf29pdbMB7wSENUnT6q6UiARJ2W9r+KXis1h22+mkwm03/2AFWdOlljtB8YAAAAAElFTkSuQmCC","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":true,"prefix":"","firstName":"Xianzhong","middleName":"","lastName":"Lin","suffix":""},{"id":482386794,"identity":"38a0eff3-1e12-4c1a-8964-dcf455ebb2f1","order_by":8,"name":"Lanying Lin","email":"","orcid":"","institution":"The First Affiliated Hospital of Fujian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Lanying","middleName":"","lastName":"Lin","suffix":""}],"badges":[],"createdAt":"2025-06-09 13:38:14","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6854966/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6854966/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":86632309,"identity":"baa06558-22df-4a8b-9ab9-ee6251819043","added_by":"auto","created_at":"2025-07-14 06:37:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":198278,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of study participant selection and allocation.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/245bff926dc414d8f7c0b2cf.png"},{"id":86630873,"identity":"50cae81e-a310-41fc-9dd4-4482fa2abd65","added_by":"auto","created_at":"2025-07-14 06:21:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":83481,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003e[A]\u003c/strong\u003e The optimal parameter [λ] selection in the LASSO model employed fivefold cross-validation using a minimum criteria approach. The optimal values of λ are represented by dotted vertical lines. Among these values, λ = 0.012, corresponding to a logarithm of λ equal to − 4.422, was selected as the optimal choice. \u003cstrong\u003e[B]\u003c/strong\u003e LASSO coefficient profiles of 11 clinical features. The plot was created using a logarithmic scale for the lambda values. A vertical line was added to indicate the lambda value selected through fivefold cross-validation. This optimal lambda value led to the identification of seven features with nonzero coefficients.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/31c112afb6269d09846c515b.png"},{"id":86633444,"identity":"ddbc4751-4e98-41a6-a11e-5c0460b4d9a1","added_by":"auto","created_at":"2025-07-14 06:53:35","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":107721,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram with body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, WBC count, fentanyl dosage and baseline temperature predicts the probability of epidural-related maternal fever.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/df4fcc2e5b26b01088811279.png"},{"id":86630879,"identity":"521aa2a4-6c50-49d7-9f7b-f8f8d66d1429","added_by":"auto","created_at":"2025-07-14 06:21:35","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":135123,"visible":true,"origin":"","legend":"\u003cp\u003eThe area under the receiver operating characteristic curve [AUC] for the discrimination of the model. \u003cstrong\u003e[A]\u003c/strong\u003eThe training set, 0.847 (95%CI: 0.778–0.917). \u003cstrong\u003e[B]\u003c/strong\u003e The validation set, 0.772 (95%CI: 0.649–0.896).\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/2620bddda9f0b4738517d084.png"},{"id":86631846,"identity":"74e576e6-967e-4a0a-8f43-883541cbcf8c","added_by":"auto","created_at":"2025-07-14 06:29:35","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":147352,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration curves for the predicting probability of epidural-related maternal fever in the training set \u003cstrong\u003e[A]\u003c/strong\u003eand \u003cstrong\u003e[B]\u003c/strong\u003e in the validation set, all P-value \u0026gt; 0.05 in the Spiegelhalter Z test suggested an agreement between the predicted probabilities and observed outcomes.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/0ad2498c4cc2c8c31b280aa8.png"},{"id":86630889,"identity":"d5eb477c-489b-4636-9123-a052c414e126","added_by":"auto","created_at":"2025-07-14 06:21:35","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":104647,"visible":true,"origin":"","legend":"\u003cp\u003eDecision curve analysis [DCA] for the epidural-related maternal fever nomogram. The green line represents the assumption of no parturient having fever, while the red line assumes that all parturients experienced fever. The blue line corresponds to the risk nomogram. The analysis was conducted on both the training set\u003cstrong\u003e [A] \u003c/strong\u003eand the validation set\u003cstrong\u003e [B]\u003c/strong\u003e.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/8762cff78b0aa4e7a959c7a8.png"},{"id":86633668,"identity":"2157082d-ab8e-4285-bc13-a182c339d31b","added_by":"auto","created_at":"2025-07-14 06:54:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1475322,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/79f5b01a-fd5a-4070-a17d-7587461122d1.pdf"},{"id":86631843,"identity":"75e801d4-7ec4-47cf-b7e9-b74229b1b6b1","added_by":"auto","created_at":"2025-07-14 06:29:35","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":751304,"visible":true,"origin":"","legend":"","description":"","filename":"Table123.docx","url":"https://assets-eu.researchsquare.com/files/rs-6854966/v1/01185786496fe116940433a9.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Nomogram for Predicting Epidural-Related Maternal Fever Following Labor Analgesia: A Retrospective Cohort Study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eEffective management of labor pain is a cornerstone of modern obstetric care, significantly enhancing the maternal experience during childbirth. Among the available analgesic modalities, neuraxial techniques, particularly epidural analgesia, are widely regarded as the gold standard, providing superior pain relief compared to systemic opioids or other methods and facilitating maternal coping during labor.\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e However, the administration of epidural analgesia is associated with several potential side effects and complications, among which intrapartum maternal fever is frequently observed.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e A substantial proportion of such febrile episodes occurring after epidural initiation is now recognized as epidural-related maternal fever (ERMF), a phenomenon generally considered distinct from infectious etiologies and attributed primarily to non-infectious inflammatory mechanisms triggered by the epidural procedure or its physiological consequences.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e Given the high utilization rates of labor epidural analgesia globally,\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e ERMF represents a common clinical event demanding appropriate understanding and management within obstetric practice.\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eThe occurrence of intrapartum fever following epidural analgesia presents a significant clinical challenge due to the inherent difficulty in reliably distinguishing non-infectious ERMF from potentially serious underlying infections, such as chorioamnionitis, based solely on maternal temperature elevation.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e This diagnostic ambiguity frequently prompts a cascade of clinical responses aimed at excluding infection, often including comprehensive maternal evaluations involving blood cultures and inflammatory marker assessments, as well as extensive neonatal sepsis workups, potentially encompassing blood tests, lumbar puncture, and the administration of empirical antibiotics.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e Such interventions, while undertaken with cautionary intent, contribute substantially to increased healthcare expenditures, generate considerable maternal anxiety, can interfere with early mother-infant interaction, and may expose both mother and neonate to the risks associated with unnecessary investigations and treatments when the underlying cause is non-infectious ERMF.\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e The current lack of dependable methods to prospectively identify parturients specifically at high risk for developing non-infectious ERMF significantly hinders the ability to mitigate this cascade of interventions.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eAddressing this clinical need requires improved strategies for risk stratification, yet a notable gap exists in the availability of validated tools designed for the personalized prediction of ERMF in clinical settings\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. While numerous observational studies have identified various potential risk factors associated with intrapartum fever during epidural analgesia \u0026ndash; encompassing maternal factors, labor characteristics, and specifics of the epidural technique\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e \u0026ndash; these factors are often evaluated in isolation, and their collective predictive power has not been effectively synthesized into a readily usable format for individualized risk assessment at the bedside. Predictive modeling offers a robust approach to integrate multiple relevant variables simultaneously, thereby providing a more nuanced and accurate estimation of individual risk.\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e Among various modeling techniques, the nomogram stands out as a particularly valuable tool in clinical practice due to its intuitive graphical format, which facilitates the straightforward calculation of individualized probabilities and has demonstrated utility across diverse medical fields for risk prediction and decision support.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eTherefore, the primary objective of the present study was to address the aforementioned gap by developing and rigorously validating a novel clinical nomogram specifically designed to predict the individualized risk of epidural-related maternal fever. Utilizing readily ascertainable data encompassing maternal baseline characteristics, specific obstetric history elements, and key intrapartum clinical and treatment variables obtained from a retrospective cohort of parturients receiving epidural labor analgesia, this investigation aimed to construct a statistically robust and clinically pragmatic predictive tool. The successful development and validation of such a nomogram is intended to provide clinicians with an evidence-based instrument to aid in the identification of parturients at elevated risk for ERMF, potentially enabling more targeted surveillance, informed counseling, and rationalized decision-making regarding subsequent investigations or interventions following epidural placement.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Study Design and Population\u003c/h2\u003e\u003cp\u003eThis investigation was structured as a retrospective cohort study conducted at The First Affiliated Hospital of Fujian Medical University. Ethical oversight and approval for the study protocol were obtained from the Institutional Ethics Committee (Approval No. MRCTA, ECFAH of FMU [2025] 508), and this study was registered on ClinicalTrials.gov on May 25, 2025 (Identifier: ChiCTR2500102964), ensuring adherence to established ethical principles governing human subject research. The study methodology complied with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to promote rigorous and transparent reporting of this observational research.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e The analysis focused on parturients who received epidural labor analgesia between January 2022 and December 2023. Inclusion criteria mandated that participants were classified as American Society of Anesthesiologists (ASA) physical status I or II,\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e aged between 22 and 45 years, carrying a singleton pregnancy at 37 to 42 weeks of gestation, and deemed suitable for both a trial of vaginal delivery by an obstetrician and epidural labor analgesia by an anesthesiologist. Furthermore, only parturients with complete clinical documentation available in the hospital records were included, signifying voluntary participation subsequent to confirmation of eligibility. Exclusion criteria were applied to parturients with pregnancy-induced hypertension or gestational diabetes mellitus necessitating pharmacological management, those with a documented history of chronic pain disorders or long-term analgesic medication use, individuals presenting contraindications to neuraxial anesthesia techniques, those diagnosed with pre-existing neurological or significant psychiatric conditions, concurrent enrollment in other clinical trials, or any other circumstances judged by the research team to render the participant unsuitable for inclusion in the study cohort. The primary outcome measure, intrapartum fever, was operationally defined as a maternal oral temperature exceeding 37.0\u0026deg;C during labor following the initiation of epidural analgesia. All epidural procedures were uniformly administered by experienced anesthesiologists adhering to standardized institutional protocols to minimize procedural variability.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Data Collection and Variable Selection\u003c/h2\u003e\u003cp\u003eClinical data for eligible parturients were meticulously extracted retrospectively from multiple integrated electronic health record systems within the institution, specifically the Zhiye Hospital Information System, the Maternal Healthcare Management System, and the MFM-CNS Central Monitoring System. This comprehensive data retrieval process encompassed all parturients receiving epidural labor analgesia during the study period from January 2022 to December 2023. Cases satisfying the predefined selection criteria were systematically compiled into a dedicated epidural analgesia registry database for subsequent analysis. The selection of potential predictive variables was informed by a thorough review of pertinent medical literature and established clinical evaluations pertinent to epidural analgesia and intrapartum fever\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. The variables collected encompassed a wide range of demographic, obstetric, labor-related, and biochemical parameters, including: maternal age, body weight upon admission for labor, comprehensive obstetric history (parity, gravidity), baseline maternal temperature prior to epidural initiation, administration of oxytocin for labor augmentation or induction, presence of meconium staining indicating amniotic fluid contamination, duration of the active phase of labor, assessed uterine contraction intensity, the total number of documented vaginal examinations performed during labor, the time interval from spontaneous or artificial rupture of membranes to delivery, the latency period from the initiation of epidural analgesia to the onset of fever (if applicable), serum concentrations of Interleukin-6 (IL-6), Procalcitonin (PCT), and C-reactive protein (CRP), peripheral white blood cell (WBC) count, specific details of the epidural analgesia protocol employed (including maintenance infusion rates and bolus strategies), and the initial preload doses of ropivacaine and fentanyl administered via the epidural catheter. To ensure the integrity and reliability of the dataset, patient records exhibiting a missing data rate of 20% or higher for the selected variables were excluded from the final analysis. Statistical techniques were employed to manage data quality issues: identified outliers were addressed using the trimming method, whereby extreme values falling outside a predefined range were removed, and missing values for included variables were handled using multiple imputation techniques to generate plausible substitute values based on observed data patterns, thereby preserving sample size and reducing potential bias.\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e A flowchart detailing the study participant selection process and overall study design is presented elsewhere (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Statistical Analysis and Model Development\u003c/h2\u003e\u003cp\u003eThe entire cohort of 373 eligible parturients was partitioned utilizing a computer-generated random allocation sequence into a training dataset and a validation dataset, maintaining a 7:3 ratio, a standard practice in prediction model development to ensure sufficient data for both model derivation and independent testing.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e This resulted in distinct cohorts designated for initial model construction and subsequent performance evaluation. Prior to model development, baseline demographic and clinical characteristics were formally compared between the training and validation sets using appropriate statistical tests (detailed further in section \u003cspan refid=\"Sec8\" class=\"InternalRef\"\u003e2.6\u003c/span\u003e) to ascertain comparability; no statistically significant differences were detected, confirming the validity of the random allocation process. The identification of salient predictive factors for epidural-related maternal fever was executed exclusively utilizing the training dataset through a sequential analytical approach. Initially, the Least Absolute Shrinkage and Selection Operator (LASSO) regression method was applied.\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e This technique was chosen for its proficiency in managing potential multicollinearity among predictors and its capacity to perform automated feature selection by shrinking the regression coefficients of less contributive variables towards zero, thereby effectively reducing the dimensionality of the predictor space, which is particularly advantageous when dealing with numerous potential covariates. Subsequently, variables retained following the LASSO procedure, identified by their non-zero coefficients, were entered into a multivariable logistic regression analysis. This step aimed to further refine the model by quantifying the independent association strength of each selected predictor with the binary outcome (presence or absence of epidural-related maternal fever) and to identify the final set of statistically significant predictors based on a predetermined significance threshold (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The final predictive model, incorporating the significant predictors identified through the multivariable logistic regression, was subsequently operationalized and visually represented as a nomogram. This graphical tool was constructed based on the regression coefficients derived from the training cohort data, providing a user-friendly interface for estimating the individualized probability of developing intrapartum fever following epidural analgesia. All statistical computations for model development, including LASSO and logistic regression analyses, as well as nomogram generation, were performed using R software (version 4.4.2; R Foundation for Statistical Computing, Vienna, Austria),\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e specifically employing functions within the 'rms' package.\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e A two-tailed P-value less than 0.05 was consistently considered indicative of statistical significance throughout all analytical procedures.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Model Validation and Performance Assessment\u003c/h2\u003e\u003cp\u003eThe predictive accuracy and clinical applicability of the developed nomogram were rigorously evaluated through a multifaceted validation process conducted independently on both the training dataset (to assess initial fit) and, more critically, the reserved validation dataset (to gauge generalizability and performance on unseen data).\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e The model's discriminative ability, defined as its capacity to correctly differentiate between parturients who experienced epidural-related maternal fever and those who did not, was quantitatively assessed using Receiver Operating Characteristic (ROC) curve analysis.\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e The Area Under the ROC Curve (AUC) was calculated as the primary metric for discrimination, with values ranging from 0.5 (no discrimination better than chance) to 1.0 (perfect discrimination); 95% confidence intervals (CIs) for the AUC were also computed to reflect the precision of the estimate. Concurrently, the calibration of the nomogram was meticulously examined to determine the extent of agreement between the predicted probabilities generated by the model and the actual observed frequencies of intrapartum fever across different risk strata.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e Calibration was visually assessed using calibration plots, which graph predicted probabilities against observed proportions, with perfect calibration indicated by points falling along a 45-degree diagonal line. Quantitative assessment of calibration involved statistical testing, such as the Spiegelhalter Z test (as referenced in the original protocol documentation)\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e or comparable goodness-of-fit statistics, to formally evaluate the concordance between predictions and observations. Furthermore, the potential clinical utility and practical value of implementing the nomogram in guiding clinical decisions were evaluated using Decision Curve Analysis (DCA).\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e DCA assesses the net benefit derived from using the prediction model across a range of clinically relevant threshold probabilities for intervention (e.g., initiating prophylactic measures or heightened surveillance), comparing the model's utility against default strategies of treating all patients or treating none. To enhance the robustness of the performance estimates and mitigate the risk of overfitting, particularly when evaluating performance on the training set, all validation procedures, including the calculation of AUC, calibration metrics, and DCA net benefit ranges, incorporated bootstrap resampling techniques with 1000 iterations.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e This internal validation approach provides more stable and less biased estimates of the nomogram's true performance characteristics.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Data Presentation and Descriptive Statistics\u003c/h2\u003e\u003cp\u003ePrior to statistical analysis, all continuous variables collected for this study underwent assessment for normality of distribution, typically employing methods such as the Shapiro-Wilk test\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e or visual inspection of histograms and Q-Q plots, to determine the appropriate methods for summarization and subsequent comparative analysis. Based on these assessments, continuous data conforming to a normal distribution were presented as the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD), providing a measure of central tendency and dispersion suitable for symmetrically distributed data. Conversely, continuous variables exhibiting skewed or non-normal distributions were summarized using the median [M] accompanied by the interquartile range (IQR), specifically detailing the 25th percentile (Q25) and the 75th percentile (Q75), which offers a more robust representation of central tendency and spread for such data. Categorical variables, including demographic factors, clinical characteristics, and binary outcomes like the presence or absence of fever, were presented as absolute frequencies (n) and relative percentages (%), allowing for a clear depiction of the distribution of these characteristics within the study population and its subgroups (training and validation sets). These descriptive statistical approaches were applied following the rigorous data quality control measures outlined previously (Section \u003cspan refid=\"Sec4\" class=\"InternalRef\"\u003e2.2\u003c/span\u003e), including the handling of missing data through multiple imputation and the management of outliers via the trimming method, ensuring that the presented summaries accurately reflect the characteristics of the final analytical cohort.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.6 Inferential Statistical Analysis\u003c/h2\u003e\u003cp\u003eComparative analyses between relevant groups (such as comparing baseline characteristics between the training and validation sets, or comparing predictor variables between febrile and afebrile parturients in univariate analysis, typically presented in the Results section) were conducted using appropriate inferential statistical tests selected based on the data type and distribution. For comparisons of continuous variables between two independent groups, the Independent Student's t-test was employed when the data were normally distributed and exhibited equal variances (homoscedasticity, typically assessed using Levene's test or similar\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e). In instances where normally distributed data displayed unequal variances, Welch's t-test, an adaptation robust to heteroscedasticity, was utilized\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. For continuous variables that did not follow a normal distribution, the non-parametric Mann-Whitney U test (also known as the Wilcoxon rank-sum test) was applied to compare the central tendencies of the two groups.\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e Comparisons involving categorical variables were performed using the Pearson's chi-squared (χ\u0026sup2;) test. In situations where the assumptions for the χ\u0026sup2; test were not met, specifically when expected cell frequencies were low (commonly defined as less than 5 in more than 20% of cells),\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e Fisher's exact test was employed as a more appropriate alternative, particularly for 2x2 contingency tables. Across all inferential statistical procedures undertaken in this study, a two-tailed probability value (P-value) of less than 0.05 was consistently adopted as the threshold for determining statistical significance. All statistical analyses, encompassing descriptive summaries and inferential comparisons, as well as the previously detailed model development and validation steps, were executed using the R statistical software environment, version 4.4.2 (R Foundation for Statistical Computing, Vienna, Austria).\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1 Study Cohort Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe final study population comprised 373 eligible parturients who received epidural labor analgesia during the specified study period. Following eligibility screening, these participants were randomly allocated, using a 7:3 ratio, into a training cohort (n = 261) designated for model development and an independent validation cohort (n = 112) reserved for testing the model\u0026apos;s generalizability. Within the total analyzed cohort, a notably high incidence of the primary outcome was observed, with 321 parturients (approximately 86.1%) developing epidural-related maternal fever (defined as maternal oral temperature \u0026gt; 37.0\u0026deg;C) during labor after epidural initiation. Crucially, a comparative statistical analysis of baseline demographic and pertinent clinical characteristics between the randomly formed training and validation cohorts revealed no statistically significant differences across all measured variables (all P \u0026gt; 0.05). This finding confirms the successful randomization process and ensures that the two cohorts were comparable at the outset, supporting the validity of using the validation cohort for unbiased assessment of the prediction model derived from the training cohort. Comprehensive details regarding the baseline characteristics of the study participants, potentially stratified by the occurrence of fever and explicitly comparing the training versus validation sets, are systematically presented (Table 1, Table 2).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Identification of Predictors for Epidural-Related Maternal Fever\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe process of identifying significant factors associated with the development of epidural-related maternal fever was conducted using the data from the training cohort (n=261) through a structured two-phase analytical strategy designed to handle the multiplicity of clinical variables effectively. In the initial phase, Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis was employed as a robust feature selection technique. This method facilitated preliminary variable screening by identifying a subset of potentially influential predictors from a larger pool, while simultaneously mitigating multicollinearity and controlling for potential overfitting, thereby enhancing model stability. The LASSO procedure, including the determination of the optimal regularization parameter (lambda) via fivefold cross-validation, is illustrated graphically (Figure 2A-B). This analysis yielded seven candidate variables demonstrating non-zero coefficients, indicating their potential predictive value: maternal body weight upon admission, prior delivery history (parity), the total number of antenatal vaginal examinations performed, assessed uterine contraction intensity, peripheral white blood cell (WBC) count, the cumulative fentanyl dosage administered via the epidural, and baseline maternal temperature recorded before epidural placement. In the second phase, these seven LASSO-selected variables were incorporated into a multivariable logistic regression model to ascertain their independent predictive contributions while adjusting for each other. The analysis yielded the following adjusted odds ratios (ORs) with corresponding 95% confidence intervals (CIs) and P-values for each variable: body weight (OR=1.034, 95% CI: 0.991\u0026ndash;1.079, P=0.124), delivery history (OR=7.541, 95% CI: 1.408\u0026ndash;141.4, P=0.058), number of antenatal vaginal examinations (OR=2.860, 95% CI: 1.807\u0026ndash;4.768, P\u0026lt;0.001), uterine contraction intensity (OR=2.153, 95% CI: 0.896\u0026ndash;5.533, P=0.096), WBC Count (OR=1.196, 95% CI: 1.059\u0026ndash;1.356, P=0.004), fentanyl dosage (OR=0.731, 95% CI: 0.513\u0026ndash;1.031, P=0.076), and baseline temperature (OR=0.283, 95% CI: 0.023\u0026ndash;3.329, P=0.316). Based on the conventional statistical significance threshold (P\u0026lt;0.05), the number of antenatal vaginal examinations and the WBC count emerged as statistically significant independent predictors of epidural-related maternal fever within this multivariate model. The remaining variables, despite being selected by LASSO and included in the final model construction, did not demonstrate statistically significant independent associations at the P\u0026lt;0.05 level in this analysis. Complete, detailed results from the multivariable logistic regression analysis performed on the training cohort are systematically documented (Table 3).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3 Development of the Predictive Nomogram\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSubsequent to the identification of key predictive factors through statistical analysis of the training cohort data, a practical predictive tool was developed in the form of a nomogram. This nomogram was specifically constructed to facilitate individualized risk assessment for epidural-related maternal fever in parturients undergoing labor with epidural analgesia. The development process integrated the seven predictor variables that emerged from the combined LASSO screening and multivariable logistic regression analyses: maternal body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell (WBC) count, cumulative epidural fentanyl dosage, and baseline maternal temperature. The resulting nomogram provides a graphical representation of the multivariable logistic regression model, visually quantifying the weighted contribution of each of these distinct risk factors. By summing the points assigned to each predictor variable based on an individual parturient\u0026apos;s characteristics, clinicians can derive an estimated probability of developing intrapartum fever following epidural placement. The complete visual construct of this predictive nomogram, illustrating the points scale for each included variable and the corresponding total points to probability conversion axis, is presented for reference (Figure 3). This tool is intended to serve as a clinical decision-support aid for personalized risk stratification.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4 Assessment of Nomogram Discrimination\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe capacity of the developed nomogram to accurately differentiate between parturients who subsequently developed epidural-related maternal fever and those who remained afebrile was rigorously evaluated using Receiver Operating Characteristic (ROC) curve analysis in both the training and validation datasets. In the training cohort (n=261), the nomogram demonstrated excellent discriminative performance, yielding an Area Under the ROC Curve (AUC) of 0.847 (95% Confidence Interval [CI]: 0.778\u0026ndash;0.917). This high AUC value within the development dataset suggests a strong ability of the model to distinguish between the outcome groups based on the included predictors. When applied to the independent validation cohort (n=112), the model maintained good discriminative ability, achieving an AUC of 0.772 (95% CI: 0.649\u0026ndash;0.896). While indicating a slight decrease in performance compared to the training set, as is typically expected, this result confirms the model\u0026apos;s robust capacity for discriminating risk in a separate sample of parturients not used for its initial development. These findings consistently underscore the nomogram\u0026apos;s substantial ability to differentiate parturients at higher versus lower risk for developing epidural-related maternal fever. The corresponding ROC curves visually depicting these discrimination results for both the training and validation sets are provided (Figure 4).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.5 Assessment of Nomogram Calibration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe calibration of the predictive nomogram, which evaluates the concordance between the predicted probabilities of epidural-related maternal fever and the actual observed frequencies of the event, was assessed in both the training and validation cohorts. Visual inspection of the calibration plots revealed excellent agreement between the nomogram\u0026apos;s predictions and the observed outcomes across the spectrum of risk in the training set. This visual finding was corroborated by quantitative statistical testing; the Spiegelhalter Z test yielded a P-value of 0.974, indicating no statistically significant difference between the predicted probabilities and the observed frequencies, thus confirming good calibration within the data used for model development. Similarly, assessment in the independent validation cohort demonstrated strong calibration performance. The calibration plot for the validation set also showed close alignment between predicted and observed probabilities, further supported by a non-significant result from the Spiegelhalter Z test (P=0.109). Collectively, these calibration analyses performed on both datasets suggest that the nomogram provides accurate and reliable risk probability estimates for epidural-related maternal fever, demonstrating that the predicted risks align closely with the actual observed risks across different patient strata. The graphical calibration plots illustrating these findings for both cohorts are provided (Figure 5).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.6 Clinical Utility Assessment via Decision Curve Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo evaluate the potential clinical applicability and net benefit derived from using the developed nomogram to guide clinical decisions, Decision Curve Analysis (DCA) was conducted for both the training and validation datasets. DCA assesses the clinical consequences of adopting the prediction model across a range of threshold probabilities, comparing the net benefit of using the nomogram against default strategies, such as classifying all parturients as high-risk (treat all) or all as low-risk (treat none). In the training cohort, the DCA revealed that employing the nomogram to predict epidural-related maternal fever yielded a positive net benefit across a wide range of threshold probabilities, specifically from 1.5% up to 71%. This indicates that within this substantial risk threshold range, utilizing the nomogram for decision-making would be more advantageous than the default strategies. Importantly, the analysis performed on the independent validation cohort corroborated the clinical utility of the nomogram, demonstrating a positive net benefit over threshold probabilities ranging from 4.5% to 54.5%. These findings from the DCA confirm that the nomogram holds potential clinical value, offering improvements in net benefit across clinically relevant risk thresholds in both the development and validation settings. This suggests the model\u0026apos;s potential utility in optimizing clinical management strategies related to the risk of epidural-related maternal fever. The graphical representations of the DCA curves for both cohorts are displayed (Figure 6).\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study successfully developed and internally validated the clinical nomogram specifically designed for the prediction of epidural-related maternal fever (ERMF) in parturients undergoing labor analgesia. By integrating seven readily accessible clinical variables \u0026ndash; maternal body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell (WBC) count, cumulative epidural fentanyl dosage, and baseline maternal temperature \u0026ndash; identified through a rigorous variable selection process involving LASSO regression and multivariable logistic analysis, the nomogram demonstrated commendable predictive performance. The model exhibited good discriminative capacity, effectively distinguishing between parturients who developed ERMF and those who did not, alongside reliable calibration, indicating accurate probability estimation across different risk strata. These findings, observed consistently in both the training and independent validation cohorts, represent a significant advancement towards providing a quantitative, individualized risk assessment tool to address a common and challenging clinical scenario in contemporary obstetric practice\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eThe inclusion of specific predictors within the final nomogram warrants interpretation regarding their potential mechanistic links to the pathophysiology of ERMF, which is predominantly considered a non-infectious inflammatory response.\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e The statistically significant association identified between an increased number of antenatal vaginal examinations and ERMF risk aligns with hypotheses suggesting that repeated examinations may increase bacterial translocation or induce localized inflammatory responses, even in the absence of overt clinical infection.\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e Similarly, the elevated WBC count, another significant predictor, likely reflects a heightened systemic inflammatory state, potentially pre-existing or developing in response to the stresses of labor and potentially amplified by epidural analgesia itself.\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e While maternal body weight and delivery history (likely reflecting nulliparity given the odds ratio trend) did not achieve conventional statistical significance in the final multivariable model, their selection by LASSO suggests predictive contribution; obesity is recognized as a chronic pro-inflammatory condition,\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e and nulliparity is often associated with longer labor durations and potentially different inflammatory signaling\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. The roles of uterine contraction intensity, fentanyl dosage, and baseline temperature appeared less pronounced or more complex in the final adjusted model, suggesting intricate interactions or the need for further investigation into their specific contributions to non-infectious inflammatory pathways during labor with epidural analgesia.\u003c/p\u003e\u003cp\u003eThe findings of this investigation resonate with, yet also extend, the existing body of literature concerning risk factors for intrapartum fever. Several factors incorporated into the nomogram, such as nulliparity, elevated maternal weight, increased number of vaginal examinations, and higher WBC counts, have been inconsistently reported as potential risk factors in previous studies focusing broadly on intrapartum fever or specifically on fever following epidural initiation\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. The current study provides further support for the relevance of these factors within the specific context of ERMF prediction. However, the primary contribution and novelty lie not merely in identifying individual risk factors, but in demonstrating their collective predictive utility through the development and validation of an integrated multivariable prediction model. To date, predictive tools combining multiple factors into a composite risk score for ERMF have been lacking\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. The construction of this nomogram, therefore, represents a crucial step forward, moving beyond associative analyses of single variables towards a synthesized, quantitative approach for personalized risk estimation, providing a framework potentially more robust than reliance on isolated clinical indicators.\u003c/p\u003e\u003cp\u003eThe practical implications of this validated nomogram warrant careful consideration for potential clinical application. If externally validated, this tool could serve as a valuable adjunct to clinical judgment at the point of care, potentially calculated using readily available data shortly after epidural placement or during the course of labor. Its primary utility lies in providing an objective, individualized estimation of ERMF risk, thereby enabling clinicians to prospectively identify parturients situated at the higher end of the risk spectrum. Such identification could facilitate tailored patient counseling regarding the likelihood of developing a non-infectious fever and inform decisions about heightened surveillance, such as more frequent temperature monitoring. Furthermore, in parturients predicted to be at low risk who subsequently develop a borderline fever, the nomogram score might contribute, alongside the overall clinical picture, to a more confident deferral of extensive infectious workups, thereby potentially reducing unnecessary interventions, associated costs, and maternal anxiety\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. It must be emphasized, however, that the nomogram serves as a risk stratification instrument and decision-support aid; it does not supplant clinical acumen or obviate the need for appropriate investigation of maternal fever, particularly when clinical signs suggestive of infection are present.\u003c/p\u003e\u003cp\u003eThis investigation possesses several methodological strengths that bolster the credibility of its findings. Foremost is the novelty of developing and validating the nomogram specifically tailored for the prediction of ERMF, addressing a recognized gap in clinical prediction tools for this common scenario. The statistical approach employed was robust, utilizing LASSO regression for objective feature selection to handle potential multicollinearity and reduce model complexity, followed by standard multivariable logistic regression for model finalization. The study benefited from the inclusion of a range of pertinent and generally accessible clinical predictors. Crucially, the internal validation process employed a randomly allocated, independent cohort, providing a rigorous assessment of the model's generalizability beyond the training data.\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e Moreover, the evaluation of the nomogram's performance was comprehensive, extending beyond standard measures of discrimination (AUC) to incorporate critical assessments of calibration (agreement between predicted and observed risk via calibration plots and statistical testing) and potential clinical utility (quantified through Decision Curve Analysis), offering a more holistic perspective on the model's potential value in practice.\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e\u003cp\u003eDespite these strengths, several limitations inherent to the study design and execution must be acknowledged. The retrospective nature of the data collection introduces susceptibility to potential information bias, documentation inconsistencies, and the influence of unmeasured confounding variables that could affect the observed associations.\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e Conducted at a single tertiary care center in Fujian, China, the findings' generalizability may be limited, as patient demographics, baseline risks, specific clinical practices, and epidural analgesia protocols might differ significantly in other institutions or geographical regions. The definition of ERMF relied on a specific oral temperature threshold, which may not capture all relevant inflammatory responses or perfectly align with definitions used elsewhere. Furthermore, some predictor variables, such as uterine contraction intensity, may be subject to subjective assessment or measurement variability. Consequently, the critical next step requires external validation of this nomogram in larger, diverse, multi-center prospective cohorts to confirm its predictive accuracy and transportability across different clinical settings and populations.\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e,\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e Future research should also focus on prospectively evaluating the nomogram's real-world impact on clinical decision-making, resource utilization, and patient outcomes, alongside further investigations into the complex pathophysiology of ERMF to potentially refine predictive models or identify targets for intervention.\u003c/p\u003e\u003cp\u003eThis study successfully developed and internally validated a novel predictive nomogram incorporating seven readily accessible clinical and obstetric factors for estimating the individualized risk of epidural-related maternal fever in parturients. The nomogram demonstrated good discriminative ability and reliable calibration in both training and validation cohorts, suggesting its potential utility as a practical, quantitative decision-support tool. By enabling enhanced risk stratification, this validated model offers a promising approach to aid clinicians in identifying individuals at higher risk for ERMF, potentially facilitating more tailored patient monitoring, informed counseling, and rationalized clinical management strategies within contemporary obstetric anesthesia practice, pending further external validation.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthical Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis retrospective study was approved by the Institutional Ethics Committee of The First Affiliated Hospital of Fujian Medical University (Approval No. MRCTA, ECFAH of FMU [2025] 508) and adhered to the principles of the Declaration of Helsinki. Patient consent handling for this analysis complied with the approved protocol. All the patients have been informed and signed informed consent before the experiments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConsent for publication was obtained from the participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData and Code Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIndividual participant datasets are not publicly available due to patient privacy and ethical restrictions. De-identified summary data may be accessible from the corresponding author upon reasonable request, contingent upon institutional approval and data use agreements. Analyses employed standard functions in R software (v4.4.2), including the \u0026apos;rms\u0026apos; package; specific code is available upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Fujian Medical University Research and Development Fund Project (2023QH1114) and the Youth Scientific Research Project of Fujian Provincial Health Commission (2024QNA039).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization：X.L., L.L., T.C.\u003c/p\u003e\n\u003cp\u003eMethodology: X.L., L.L., T.C.\u003c/p\u003e\n\u003cp\u003eSoftware: T.C.\u003c/p\u003e\n\u003cp\u003eFormal Analysis: T.C., Q.Z.\u003c/p\u003e\n\u003cp\u003eData Curation: X.L., L.D., X.Z., Y.L., L.H.\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Original Draft: T.C., X.L.\u003c/p\u003e\n\u003cp\u003eWriting \u0026ndash; Review \u0026amp; Editing: X.L., L.L., T.C.\u003c/p\u003e\n\u003cp\u003eSupervision, Project Administration: X.L., T.C., Q.Z.\u003c/p\u003e\n\u003cp\u003eAll authors have read and approved the final manuscript.\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eGratitude is expressed to the participating parturients and the staff of the Departments of Anesthesiology and Obstetrics at The First Affiliated Hospital of Fujian Medical University for their support in data acquisition and institutional assistance.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eAmerican College of Obstetricians and Gynecologists. ACOG Practice Bulletin No. 209: Obstetric Analgesia and Anesthesia. Obstet Gynecol. 2019 Mar;133(3):e208-e225. DOI: 10.1097/AOG.0000000000003132.\u003c/li\u003e\n\u003cli\u003eSharpe EE, Arendt KW. Epidural Labor Analgesia and Maternal Fever. Clin Obstet Gynecol. 2017 Sep;60(3):365-374. DOI: 10.1097/GRF.0000000000000270. \u003c/li\u003e\n\u003cli\u003eSelina Patel, Sarah Ciechanowicz, Yair J Blumenfeld, Pervez Sultan. Epidural-related maternal fever: incidence, pathophysiology, outcomes, and management. Am J Obstet Gynecol. 2023 May;228(5S):S1283-S1304.e1. DOI: 10.1016/j.ajog.2022.06.026.\u003c/li\u003e\n\u003cli\u003eGoetzl L. Epidural analgesia and maternal fever: a clinical and research update. Curr Opin Anaesthesiol. 2012 Jun;25(3):292-9. DOI: 10.1097/ACO.0b013e3283530d7c.\u003c/li\u003e\n\u003cli\u003eKatherine W Arendt, B Scott Segal. The association between epidural labor analgesia and maternal fever. Clin Perinatol. 2013 Sep;40(3):385-98. DOI: 10.1016/j.clp.2013.06.002.\u003c/li\u003e\n\u003cli\u003eMillicent Anim-Somuah, Rebecca Md Smyth, Allan M Cyna, Anna Cuthbert. Epidural versus non-epidural or no analgesia for pain management in labour. Cochrane Database Syst Rev. 2018 May 21;5(5):CD000331. DOI: 10.1002/14651858.CD000331.pub4.\u003c/li\u003e\n\u003cli\u003eLieberman E, Lang JM, Frigoletto F Jr, Richardson DK, Ringer SA, Cohen A. Epidural analgesia, intrapartum fever, and neonatal sepsis evaluation. Pediatrics. 1997 Mar;99(3):415-9. DOI: 10.1542/peds.99.3.415.\u003c/li\u003e\n\u003cli\u003eSarah Morton, Justin Kua, Christopher J Mullington. Epidural analgesia, intrapartum hyperthermia, and neonatal brain injury: a systematic review and meta-analysis. Br J Anaesth. 2021 Feb;126(2):500-515. DOI: 10.1016/j.bja.2020.09.046.\u003c/li\u003e\n\u003cli\u003eMoons KGM, Royston P, Vergouwe Y, Grobbee DE, Altman DG. Prognosis and prognostic research: what, why, and how? BMJ. 2009 Feb 23;338:b375. DOI: 10.1136/bmj.b375..\u003c/li\u003e\n\u003cli\u003eVinod P Balachandran, Mithat Gonen, J Joshua Smith, Ronald P DeMatteo. Nomograms in oncology: more than meets the eye. Lancet Oncol. 2015 May;16(5):e172-80. DOI: 10.1016/S1470-2045(14)71116-7.\u003c/li\u003e\n\u003cli\u003eErik von Elm, Douglas G Altman, Matthias Egger, Stuart J Pocock, Peter C G\u0026oslash;tzsche, Jan P Vandenbroucke; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Bmj-british Medical Journal. 2007 Oct 20;335(7624):806-8. DOI: 10.1136/bmj.39335.541782.AD.\u003c/li\u003e\n\u003cli\u003eDoyle DJ, Goyal A, Garmon EH. American Society of Anesthesiologists Classification. [Updated 2023 Jul 3]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK441940/. \u003c/li\u003e\n\u003cli\u003eSterne JAC, White IR, Carlin JB, Spratt M, Royston P, Kenward MG, Wood AM, Carpenter JR. Multiple imputation for missing data in epidemiological and clinical research: potential and pitfalls. BMJ. 2009 Jun 29;338:b2393. DOI: 10.1136/bmj.b2393.\u003c/li\u003e\n\u003cli\u003eCollins GS, Reitsma JB, Altman DG, Moons KGM. Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): The TRIPOD Statement. Bmj-british Medical Journal. 2015 Jan 7:350:g7594. DOI: 10.1136/bmj.g7594.\u003c/li\u003e\n\u003cli\u003eAlireza Daneshvar, Golalizadeh Mousa. Regression shrinkage and selection via least quantile shrinkage and selection operator. PLoS One, 2023 Feb 16;18(2):e0266267. DOI: 10.1371/journal.pone.0266267\u003c/li\u003e\n\u003cli\u003eR Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. 2024. URL https://www.R-project.org/.\u003c/li\u003e\n\u003cli\u003eSandro Sperandei. Understanding logistic regression analysis. Biochem Med. 2014 Feb 15;24(1):12-8. DOI: 10.11613/BM.2014.003.\u003c/li\u003e\n\u003cli\u003eHanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982 Apr;143(1):29-36. DOI: 10.1148/radiology.143.1.7063747.\u003c/li\u003e\n\u003cli\u003eSteyerberg EW, Vickers AJ, Cook NR, Gerds T, Gonen M, Obuchowski N, Pencina MJ, Kattan MW. Assessing the performance of prediction models: a framework for traditional and novel measures. Epidemiology. 2010 Jan;21(1):128-38. DOI: 10.1097/EDE.0b013e3181c30fb2.\u003c/li\u003e\n\u003cli\u003eSpiegelhalter DJ. Probabilistic prediction in patient management and clinical trials. Stat Med. 1986 Jul-Sep;5(5):421-33. DOI: 10.1002/sim.4780050506.\u003c/li\u003e\n\u003cli\u003eVickers AJ, Elkin EB. Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making. 2006 Nov-Dec;26(6):565-74. DOI: 10.1177/0272989X06295361.\u003c/li\u003e\n\u003cli\u003eMar\u0026iacute;a J Blanca, Jaume Arnau, F J Garc\u0026iacute;a-Castro, Rafael Alarc\u0026oacute;n, Roser Bono. Non-normal Data in Repeated Measures ANOVA: Impact on Type I Error and Power. Psicothema. 2023 Feb;35(1):21-29. DOI: 10.7334/psicothema2022.292.\u003c/li\u003e\n\u003cli\u003eLevene H. Robust tests for equality of variances. In: Olkin I, Ghurye SG, Hoeffding W, Madow WG, Mann HB, editors. Contributions to Probability and Statistics: Essays in Honor of Harold Hotelling. Stanford (CA): Stanford University Press; 1960. p. 278-92.\u003c/li\u003e\n\u003cli\u003eB L WELCH. The generalisation of student\u0026apos;s problems when several different population variances are involved Biometrika. 1947;34(1/2):28-35. DOI: 10.1093/biomet/34.1-2.28.\u003c/li\u003e\n\u003cli\u003eJoshua N Pritikin, Timothy R Brick, Michael C Neale. Multivariate normal maximum likelihood with both ordinal and continuous variables, and data missing at random. Behav Res Methods. 2018 Apr;50(2):490-500. DOI: 10.3758/s13428-017-1011-6.\u003c/li\u003e\n\u003cli\u003eShuhua Chang, Deli Li, Yongcheng Qi. Pearson\u0026apos;s goodness-of-fit tests for sparse distributions. J Appl Stat. 2021 Dec 30;50(5):1078-1093. DOI: 10.1080/02664763.2021.2017413.\u003c/li\u003e\n\u003cli\u003eGluck O, Mizrachi Y, Ganer Herman H, Bar J, Kovo M, Weiner E. The correlation between the number of vaginal examinations during active labor and febrile morbidity, a retrospective cohort study. BMC Pregnancy Childbirth. 2020 Aug 28;20(1):495. DOI: 10.1186/s12884-020-02925-9.\u003c/li\u003e\n\u003cli\u003eEllulu MS, Patimah I, Khaza\u0026rsquo;ai H, Rahmat A, Abed Y. Obesity and inflammation: the linking mechanism and the complications. Arch Med Sci.2017 Jun;13(4):851-863. DOI: 10.5114/aoms.2016.58928.\u003c/li\u003e\n\u003cli\u003eJan P Vandenbroucke 1, Erik von Elm, Douglas G Altman, Peter C G\u0026oslash;tzsche, Cynthia D Mulrow, Stuart J Pocock, Charles Poole, James J Schlesselman, Matthias Egger; STROBE Initiative. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): explanation and elaboration. PLoS Med. 2007 Oct 16;4(10):e297. DOI: 10.1371/journal.pmed.0040297.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 1 to 3 are available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Epidural-Related Maternal Fever, Nomogram, Prediction Model, Risk Stratification, Labor Epidural Analgesia","lastPublishedDoi":"10.21203/rs.3.rs-6854966/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6854966/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eObjectives\u003c/strong\u003e: Epidural-related maternal fever (ERMF) is a frequent complication following labor analgesia, posing diagnostic challenges due to its overlap with infectious causes. Reliable tools for predicting individualized risk are currently lacking. This study aimed to develop and validate a clinical nomogram to predict the risk of ERMF in parturients undergoing epidural labor analgesia.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods\u003c/strong\u003e: A retrospective cohort study was conducted involving parturients receiving epidural analgesia from January 2022 to December 2023 at a single university-affiliated hospital. Eligible participants (n=373) were randomly allocated using a 7:3 ratio into training and validation datasets. Potential predictor variables were screened using the LASSO regression, and significant predictors were identified via multivariable logistic regression analysis within the training set. The nomogram's performance was rigorously evaluated using the Area Under the Receiver Operating Characteristic curve (AUC) for discrimination, calibration plots with associated statistical tests for agreement, and Decision Curve Analysis (DCA) for clinical utility.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults\u003c/strong\u003e: ERMF was observed in 321 (86.1%) parturients included in the study. Seven independent predictors were incorporated into the final nomogram: body weight, delivery history, number of antenatal vaginal examinations, uterine contraction intensity, white blood cell count, epidural fentanyl dosage, and baseline temperature. The nomogram demonstrated good discriminative performance with an AUC of 0.847 (95% CI: 0.778–0.917) in the training set and 0.772 (95% CI: 0.649–0.896) in the validation set. Calibration analyses indicated excellent agreement between predicted probabilities and observed ERMF frequencies in both datasets (P \u0026gt; 0.05). DCA confirmed positive net clinical benefit across wide and clinically relevant ranges of threshold probabilities (1.5%-71% in training; 4.5%-54.5% in validation).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDiscussion\u003c/strong\u003e: This study successfully developed and internally validated the first nomogram designed to predict individualized risk for ERMF. By integrating readily available clinical factors, the nomogram demonstrates robust predictive accuracy and calibration, alongside potential clinical utility. It represents a valuable tool to support enhanced risk stratification and inform clinical management decisions for parturients receiving epidural labor analgesia, pending further external validation.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Nomogram for Predicting Epidural-Related Maternal Fever Following Labor Analgesia: A Retrospective Cohort Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-14 06:21:30","doi":"10.21203/rs.3.rs-6854966/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"reviewerAgreed","content":"32517474450518198639599610787402326337","date":"2026-05-12T00:52:54+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-07-28T23:19:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"199495660570777692862274411121854842054","date":"2025-07-08T12:39:54+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-07-06T05:40:59+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-06-12T13:23:55+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-06-12T01:36:46+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-06-12T01:34:41+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Pregnancy and Childbirth","date":"2025-06-09T13:23:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-pregnancy-and-childbirth","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"prch","sideBox":"Learn more about [BMC Pregnancy and Childbirth](http://bmcpregnancychildbirth.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/prch/default.aspx","title":"BMC Pregnancy and Childbirth","twitterHandle":"@BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4ad2a5af-b7a6-4801-9848-0fce310beb0a","owner":[],"postedDate":"July 14th, 2025","published":true,"recentEditorialEvents":[{"type":"reviewerAgreed","content":"32517474450518198639599610787402326337","date":"2026-05-12T00:52:54+00:00","index":87,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2025-07-14T06:21:30+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-14 06:21:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6854966","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6854966","identity":"rs-6854966","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}