Development and Validation of a Nomogram for Predicting Post – Operative Poor Outcome in Abdominal EGS with Sepsis

0
0
0

Development and Validation of a Nomogram for Predicting Post-Operative In-Hospital Poor Outcome in Abdominal Related Emergency General Surgery Diseases Complicated with Sepsis

Emergency general surgery (EGS) encompasses procedures performed for acute, life-threatening conditions, often requiring immediate intervention. Patients undergoing EGS face significantly higher mortality risks compared to elective surgeries, with studies indicating an eightfold increase in mortality rates. Despite advances in surgical care, the absence of a robust risk stratification tool tailored to EGS patients—particularly those with abdominal pathologies complicated by sepsis—remains a critical gap. This study addresses this need by developing and validating a nomogram that integrates anatomical severity, physiological status, comorbidities, and surgical factors to predict post-operative in-hospital poor outcomes in this high-risk population.

Study Design and Patient Cohorts

The retrospective study enrolled 623 patients diagnosed with abdominal-related EGS diseases complicated by sepsis who underwent surgical intervention at a single tertiary care center in China. Patients were divided into development (January 2016–October 2019, n = 493) and validation (November 2019–October 2020, n = 130) cohorts. Exclusion criteria included pregnancy and incomplete medical records. The primary outcome—in-hospital poor outcome—was defined as either post-operative death or a hospital stay exceeding 15 days.

Data Collection and Predictor Variables

Clinical data were extracted from electronic health records, including pre-operative laboratory results, imaging findings, and intra-operative details. Predictor variables spanned five dimensions:

  1. Anatomical Severity: Assessed using the American Association for the Surgery of Trauma (AAST) grading system, which categorizes disease severity from Grade I (localized, organ-confined) to Grade V (widespread systemic involvement). Intra-operative findings were prioritized for grading when pre-operative imaging was inconsistent.
  2. Physiological Status: Evaluated using the Sequential Organ Failure Assessment (SOFA) score, Modified Early Warning Score (MEWS), and Acute Physiology and Chronic Health Evaluation II (APACHE II) score.
  3. Comorbidities: Chronic obstructive pulmonary disease, hypertension, diabetes, chronic kidney disease, and cardiovascular disease.
  4. Surgical Factors: Approach (open vs. endoscopic), procedure type, operation time, and prior abdominal surgery.
  5. Demographics: Patient age.

Statistical Analysis and Model Development

Univariate analysis identified variables associated with poor outcomes, followed by multivariable logistic regression using stepwise selection (entry P < 0.15, retention P < 0.10). The final model excluded operation time due to its limited impact on predictive accuracy, retaining four variables:

  • Age (OR: 1.018, 95% CI: 1.002–1.034, P = 0.027)
  • SOFA Score (OR: 1.322, 95% CI: 1.147–1.522, P < 0.001)
  • Surgical Approach: Open surgery correlated with a 5.6-fold higher risk of poor outcomes compared to endoscopic methods (P < 0.001).
  • AAST Grade: Higher grades predicted escalating risks: Grade IV (OR: 3.959, P = 0.008) and Grade V (OR: 10.591, P < 0.001) showed markedly worse prognoses.

The nomogram assigned weighted points to each variable, enabling clinicians to calculate individualized risk scores [Figure 1A].

Model Performance and Validation

The nomogram demonstrated robust discrimination, with a C-index of 0.869 in the development cohort and 0.867 in the validation cohort, indicating high predictive accuracy for poor outcomes. Calibration curves revealed strong agreement between predicted and observed outcomes in both cohorts [Figure 1D, E]. Decision curve analysis confirmed clinical utility, with net benefits exceeding traditional “treat-all” or “treat-none” strategies across probability thresholds of 10%–80% [Figure 1F, G].

Clinical and Epidemiological Insights

In the development cohort, 20.5% (n = 101) experienced poor outcomes: 58 deaths (57.4%) and 43 prolonged hospitalizations (42.6%). Common infection sources included appendicitis (33.5%), cholecystitis (24.3%), and small bowel pathologies (16.4%). The validation cohort mirrored these trends, with 23.1% (n = 30) poor outcomes (12 deaths, 18 prolonged stays).

Strengths and Innovations

This study represents the first effort to create a risk stratification tool for abdominal EGS patients in a resource-constrained setting. By integrating AAST anatomical grading with physiological and surgical factors, the nomogram addresses a critical gap in existing tools, which often overlook multidimensional risk determinants. Its graphical interface simplifies risk communication, aiding informed consent, triage decisions, and referrals to higher-care facilities. Pre-operative applicability is enhanced by moderate-to-strong concordance between imaging-based and intra-operative AAST grades, as demonstrated in prior studies.

Limitations and Future Directions

The single-center design and reliance on retrospective data limit generalizability. The model’s exclusion of post-operative complications—a key morbidity indicator—underscores the need for future iterations to incorporate morbidity endpoints. Expanding validation to multicenter cohorts and diverse healthcare settings will strengthen clinical adoption.

Conclusion

This nomogram provides a pragmatic, evidence-based tool for predicting poor outcomes in abdominal EGS patients with sepsis. By synthesizing anatomical, physiological, and surgical variables, it supports personalized risk assessment, enhances shared decision-making, and optimizes resource allocation in emergency surgical care.

doi.org/10.1097/CM9.0000000000001378

Was this helpful?

0 / 0