Performance of the Emergency Surgery Score (ESS) Across Different Emergency General Surgery Procedures

Prognostic accuracy of emergency surgery score: a systematic review

PURPOSE

This systematic review aimed to summarize the literature regarding the prognostic accuracy of the emergency surgery score (ESS).

METHOD

PubMed, Embase, Web of Science, and Scopus were comprehensively searched by May 30, 2023. Two independent researchers performed the initial screening by reviewing the titles and abstracts of the non-duplicate records and selecting the full text of articles meeting our inclusion criteria. Finally, original studies that reported the prognostic accuracy of ESS in any emergency surgeries were included. Data from the included studies were extracted into a checklist designed based on the PRISMA guidelines. The area under the curve (AUC) was used to compare the prognostic accuracy of ESS in different settings.

RESULTS

Twenty-six studies met the inclusion criteria. ESS performed excellently in 30-day post-op mortality (AUC 0.84-0.89) and incidence of cardiac arrest (AUC 0.86-0.88) in emergency general surgeries. The AUC of ESS in overall 30-day morbidities varied from 0.72 to 0.82 in five cohort studies. In predicting the need for ICU admission, the study with the largest sample size reported the best sensitivity of ESS at 80% and the specificity at 85%. Moreover, an outstanding accuracy was observed for the prediction of 30-day sepsis/septic shock in emergency general surgeries (AUC 0.75-0.92).

CONCLUSION

Despite the acceptable prognostic accuracy of ESS in 30-day mortality, morbidities, and in-hospital ICU admission in different emergency surgeries, the high number of required variables and the high probability of missing data highlight the need for modifications to this scoring system.

Validation of the emergency surgery score (ESS) in a UK patient population and comparison with NELA scoring: a retrospective multicentre cohort study

INTRODUCTION

Accurate risk scoring in emergency general surgery (EGS) is vital for consent and resource allocation. The emergency surgery score (ESS) has been validated as a reliable preoperative predictor of postoperative outcomes in EGS but has been studied only in the US population. Our primary aim was to perform an external validation study of the ESS in a UK population. Our secondary aim was to compare the accuracy of ESS and National Emergency Laparotomy Audit (NELA) scores.

METHODS

We conducted an observational cohort study of adult patients undergoing emergency laparotomy over three years in two UK centres. ESS was calculated retrospectively. NELA scores and all other variables were obtained from the prospectively collected Emergency Laparotomy and Laparoscopic Scottish Audit (ELLSA) database. The primary and secondary outcomes were 30-day mortality and postoperative intensive care unit (ICU) admission, respectively.

RESULTS

A total of 609 patients were included. Median age was 65 years, 52.7% were female, the overall mortality was 9.9% and 23.8% were admitted to ICU. Both ESS and NELA were equally accurate in predicting 30-day mortality (c-statistic=0.78 (95% confidence interval (CI), 0.71-0.85) for ESS and c-statistic=0.83 (95% CI, 0.77-0.88) for NELA, p=0.196) and predicting postoperative ICU admission (c-statistic=0.76 (95% CI, 0.71-0.81) for ESS and 0.80 (95% CI, 0.76-0.85) for NELA, p=0.092).

CONCLUSIONS

In the UK population, ESS and NELA both predict 30-day mortality and ICU admission with no statistically significant difference but with higher c-statistics for NELA score. Both scores have certain advantages, with ESS being validated for a wider range of outcomes.

Effect of emergency general surgery on postoperative performance status in patients aged over 90 years

Background

Functional deterioration following emergency general surgery (EGS) poses a significant challenge in super-elderly patients. However, limited research has focused on assessing the deterioration in postoperative performance status (PS). This study aimed to investigate the impact of EGS on PS deterioration in super-elderly patients, and the extent to which deteriorated PS is recovered.

Methods

This historical cohort study comprised 77 super-elderly patients who underwent EGS between July 2015 and December 2020. Functional deterioration was evaluated by comparing preoperative and postoperative Eastern Cooperative Oncology Group Performance Status (ECOG-PS). The Emergency Surgical Score (ESS) was used as a risk-adjustment tool. Questionnaires were mailed to the patients and their families to assess post-discharge PS and obtain their impressions of EGS.

Results

Postoperative PS deteriorated in 35/77 patients (45.5 %). Significant differences were observed between the groups in terms of sex, serum C-reactive protein (CRP) levels, ESS scores, preoperative ECOG-PS, duration of operation, and major complications. Multivariate analysis of preoperative factors showed that ESS ≥7 (OR: 3.7, 95 % CI: 1.0-13), preoperative ECOG-PS ≤2 (OR: 5.9, 95 % CI: 1.7-21), and female sex (OR: 5.8, 95 % CI: 1.6-21) were associated with postoperative ECOG-PS deterioration. According to the questionnaire results, PS recovery post-discharge was observed in 6/36 (17 %) patients, and 34/36 (94 %) patients and their families expressed positive impressions of EGS.

Conclusions

EGS in super-elderly patients highly caused a deterioration in their PS, particularly in patients with maintained preoperative PS. PS hardly recovered; however, most patients and their families had positive impressions of the EGS.

Key message

We assessed the pre- and postoperative performance status of super-elderly patients who underwent emergency general surgery. Surgery caused a marked deterioration in patients' functional performance, which seldom recovered postoperatively.

Use of artificial intelligence for nonlinear benchmarking of surgical care

BACKGROUND

Existent methodologies for benchmarking the quality of surgical care are linear and fail to capture the complex interactions of preoperative variables. We sought to leverage novel nonlinear artificial intelligence methodologies to benchmark emergency surgical care.

METHODS

Using a nonlinear but interpretable artificial intelligence methodology called optimal classification trees, first, the overall observed mortality rate at the index hospital's emergency surgery population (index cohort) was compared to the risk-adjusted expected mortality rate calculated by the optimal classification trees from the American College of Surgeons National Surgical Quality Improvement Program database (benchmark cohort). Second, the artificial intelligence optimal classification trees created different "nodes" of care representing specific patient phenotypes defined by the artificial intelligence optimal classification trees without human interference to optimize prediction. These nodes capture multiple iterative risk-adjusted comparisons, permitting the identification of specific areas of excellence and areas for improvement.

RESULTS

The index and benchmark cohorts included 1,600 and 637,086 patients, respectively. The observed and risk-adjusted expected mortality rates of the index cohort calculated by optimal classification trees were similar (8.06% [95% confidence interval: 6.8-9.5] vs 7.53%, respectively, P = .42). Two areas of excellence and 4 for improvement were identified. For example, the index cohort had lower-than-expected mortality when patients were older than 75 and in respiratory failure and septic shock preoperatively but higher-than-expected mortality when patients had respiratory failure preoperatively and were thrombocytopenic, with an international normalized ratio ≤1.7.

CONCLUSION

We used artificial intelligence methodology to benchmark the quality of emergency surgical care. Such nonlinear and interpretable methods promise a more comprehensive evaluation and a deeper dive into areas of excellence versus suboptimal care.

Emergency thoracic surgery patients have worse risk-adjusted outcomes than non-emergency patients

BACKGROUND

Outcomes for patients undergoing emergency thoracic operations have not been well described. This study was designed to compare postoperative outcomes among patients undergoing emergency versus nonemergency thoracic operations.

METHODS

We retrospectively analyzed the American College of Surgeons National Surgical Quality Improvement Program database (2005-2018). We identified patients who underwent emergency thoracic operations using current procedural technology codes. Patients were then sorted into 1 of 4 cohorts: lung and chest wall, hiatal hernia, esophagus, and pericardium. Emergency versus nonemergency outcomes were compared. Univariate logistic regression was performed with "emergency status" as the independent variable and 30-day postoperative outcomes as the dependent variables. Multiple logistic regression models were performed to control for preoperative factors.

RESULTS

Of 90,398 thoracic operations analyzed, 4,044 (4.5%) were emergency. Common emergency operations were pericardial window (n = 580, 10.2%), laparoscopic hiatal hernia repair (n = 366, 8.9%), thoracoscopic partial lung decortication (n = 334, 8.1%), thoracoscopic wedge resection (n = 301, 7.3%), thoracoscopic total lung decortication (n = 256, 6.2%), and open repair of hiatal hernia without mesh (n = 254, 6.2%). In all 4 cohorts, 30-day postoperative complications occurred more frequently after emergency surgery. After controlling for patient characteristics, 8 complications were more frequent after emergency lung and chest wall surgery, 5 complications were more frequent after emergency hiatal hernia surgery, and 3 complications were more frequent after emergency pericardium surgery. Risk-adjusted complications were not different after emergency esophageal surgery.

CONCLUSION

Patients undergoing emergency thoracic operations have worse risk-adjusted outcomes than those undergoing nonemergency thoracic operations. Subset analysis is needed to determine what factors contribute to increased adverse outcomes in specific patient populations.

Revision of the AAST grading scale for acute cholecystitis with comparison to physiologic measures of severity

BACKGROUND

Grading systems for acute cholecystitis are essential to compare outcomes, improve quality, and advance research. The American Association for the Surgery of Trauma (AAST) grading system for acute cholecystitis was only moderately discriminant when predicting multiple outcomes and underperformed the Tokyo guidelines and Parkland grade. We hypothesized that through additional expert consensus, the predictive capacity of the AAST anatomic grading system could be improved.

METHODS

A modified Delphi approach was used to revise the AAST grading system. Changes were made to improve distribution of patients across grades, and additional key clinical variables were introduced. The revised version was assessed using prospectively collected data from an AAST multicenter study. Patient distribution across grades was assessed, and the revised grading system was evaluated based on predictive capacity using area under receiver operating characteristic curves for conversion from laparoscopic to an open procedure, use of a surgical "bail-out" procedure, bile leak, major complications, and discharge home. A preoperative AAST grade was defined based on preoperative, clinical, and radiologic data, and the Parkland grade was also substituted for the operative component of the AAST grade.

RESULTS

Using prospectively collected data on 861 patients with acute cholecystitis the revised version of the AAST grade has an improved distribution across all grades, both the overall grade and across each subscale. A higher AAST grade predicted each of the outcomes assessed (all p ≤ 0.01). The revised AAST grade outperformed the original AAST grade for predicting operative outcomes and discharge disposition. Despite this improvement, the AAST grade did not outperform the Parkland grade or the Emergency Surgery Score.

CONCLUSION

The revised AAST grade and the preoperative AAST grade demonstrated improved discrimination; however, a purely anatomic grade based on chart review is unlikely to predict outcomes without addition of physiologic variables. Follow-up validation will be necessary.

LEVEL OF EVIDENCE

Diagnostic Test or Criteria, Level IV.