The Hospital Volume - Outcome Relationship in Esophageal Cancer Surgery after the first wave of Centralization in the Dutch Upper Gastrointestinal Cancer Audit
*Daan M Voeten1, *Suzanne S Gisbertz1, *Jelle P Ruurda2, Lorenzo Ferri3, *Richard van Hillegersberg2, *Mark Ivo van Berge Henegouwen1
1Amsterdam UMC, Amsterdam, Netherlands2Utrecht University Medical Center, Utrecht, Netherlands3Montreal General Hospital, Montreal, QC, Canada
Previous literature demonstrating improved outcomes after esophagectomy in high-volume hospitals provoked the introduction of Dutch volume standards. After the threshold of at least 20 operations/year was adopted in 2011, the first centralization wave was completed by 2016. This study aimed to investigate the volume-outcome relationship in oncologic esophagectomy after centralization.
This nationwide cohort study included all esophagectomy patients registered in the Dutch Upper Gastrointestinal Cancer Audit after centralization (2016-2019). Annual hospital-volume was dichotomized around the national median into “low-volume”/“high-volume”. Multilevel multivariable logistic regression investigated the impact of hospital-volume on: overall complications, severe complications (≥Clavien-Dindo IIIa), 30-day/in-hospital mortality, surgical radicality(R0), lymph node yield, anastomotic leakage, pulmonary complications, pneumonia and the composite “textbook outcome” (R0 resection, ≥15 lymph nodes, LOS<21 days and no severe intra- or post-operative complication, readmission (to the ICU), or mortality).
In total, 3,259 esophagectomies were performed in 22 centers; 1,602 in 15 low-volume centers(<53 annual esophagectomies) and 1,657 in 7 high-volume centers(≥53). Lymph node yield was higher in high-volume hospitals. Anastomotic leakage rates were lower (OR:0.80, p=0.02) and the composite “textbook outcome” rates were higher (OR:1.25, p<0.01) in high-volume centers (Table1). There were no significant differences regarding the other outcomes.
This study shows that further improvement in outcomes after centralization for esophageal cancer surgery can be achieved in hospitals performing ≥53 cases annually, suggesting that increasing the threshold for centralization should be considered.
|Table 1. Multilevel multivariable logistic regression analyses of short-term surgical outcomes after oncologic esophageal cancer surgery compared between low and high-volume hospitals.|
|Annual esophagectomy hospital volume||Outcome incidence (%)||aORA||95% CIB||P-value|
|Overall postoperative complications (yes)||<53≥53||1031 (64.4%)1053 (63.5%)||10.97||0.83 – 1.13||0.71|
|Severe complicationsC(yes)||<53≥53||484 (30.2%)483 (29.1%)||10.93||0.79 – 1.10||0.40|
|30-day/in-hospital mortality(yes)||<53≥53||43 (2.7%)48 (2.9%)||11.08||0.71 – 1.64||0.72|
|Surgical radicallity (R0)(yes)||<53≥53||1505 (93.9%)1525 (92.0%)||10.98||0.70 – 1.39||0.93|
|Lymph node yield(>15)||<53≥53||1320 (82.4%)1509 (91.1%)||12.73||2.08 – 3.58||<0.01|
|Anastomotic leakage(yes)||<53≥53||319 (19.9%)280 (16.9%)||10.80||0.66 – 0.96||0.02|
|Pulmonary complication(yes)||<53≥53||514 (32.1%)516 (31.1%)||10.99||0.85 – 1.17||0.97|
|Pneumonia(yes)||<53≥53||328 (20.5%)347 (20.9%)||11.08||0.90 – 1.30||0.40|
|Textbook outcomeD(yes)||<53≥53||687 (42.9%)813 (49.1%)||11.25||1.07 – 1.46||<0.01|
|A adjusted Odds Ratio. Corrected for: sex, age, preoperative weight loss, BMI, Charlson Comorbidity Index, ASA-score, previous esophageal or gastric surgery, tumor location, histology, clinical Tumor stage, clinical Node stage, salvage surgery and year of surgery as random effect factor. When degrees of freedom were insufficient for correction for all possible confounders, only confounders leading to a 10% change in OR were included for analyses. Year of surgery as random effect was added to the model in case the log-likelihood ratio test showed a better fit compared to the original univariable model.B 95% Confidence intervalC Clavien-Dindo grade III or higherD Patients undergoing a radical, curative resection with at least 15 resected lymph nodes, without intra-operative complication, severe postoperative complicationE, reintervention, readmission (to the ICU), mortality and a length of hospital stay shorter than 21 days.|
Back to 2021 Abstracts