AIM: To identify rates of post-discharge complications (PDC), associated risk factors, and their influence on early hospital outcomes after esophagectomy.

METHODS: We used the 2005-2013 American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database to identify patients ≥ 18 years of age who underwent an esophagectomy. These procedures were categorized into four operative approaches: transhiatal, Ivor-Lewis, 3-holes, and non-gastric conduit. We selected patient data based on clinical relevance to patients undergoing esophagectomy and compared demographic and clinical characteristics. The primary outcome was PDC, and secondary outcomes were hospital readmission and reoperation. The patients were then divided in 3 groups: no complication (Group 1), only pre-discharge complication (Group 2), and PDC patients (Group 3). A modified Poisson regression analysis was used to identify risk factors associated with developing post-discharge complication, and risk ratios were estimated.

RESULTS: 4483 total patients were identified, with 8.9% developing PDC within 30-d after esophagectomy. Patients who experienced complications post-discharge had a median initial hospital length of stay (LOS) of 9 d; however, PDC occurred on average 14 d following surgery. Patients with PDC had greater rates of wound infection (41.0% vs 19.3%, P < 0.001), venous thromboembolism (16.3% vs 8.9%, P < 0.001), and organ space surgical site infection (17.1% vs 11.0%, P = 0.001) than patients with pre-discharge complication. The readmission rate in our entire population was 12.8%. PDC patients were overwhelmingly more likely to have a reoperation (39.5% vs 22.4%, P < 0.001) and readmission (66.9% vs 6.6%, P < 0.001). BMI 25-29.9 and BMI ≥ 30 were associated with increased risk of PDC compared to normal BMI (18.5-25).

CONCLUSION: PDC after esophagectomy account for significant number of reoperations and readmissions. Efforts should be directed towards optimizing patient’s health pre-discharge, with possible prevention programs at discharge.

Esophagectomy is the mainstay treatment for localized esophageal carcinoma without medical contraindications[1]. It can also be indicated for certain benign conditions, including high-grade dysplasia in Barrett’s esophagus, caustic ingestion, reflux esophagitis complications, esophageal stricture, and esophageal neuromotor dysfunction (achalasia, spasm, scleroderma)[2]. Over 5000 esophagectomies are performed in the United States and United Kingdom annually[3]. Despite improved surgical techniques and intensive care unit therapy[4], post-esophagectomy mortality and morbidity rates remain suboptimal, ranging from 7%-28%[5,6] and 10%-27%[7], respectively. Moreover, readmission rates after esophagectomy range from 5%-25%[8,9], and the overall 5-year survival rate following esophagectomy range from 15%-40%[10,11].

These dismal outcomes after esophagectomy, compounded with national health policy changes by the Centers for Medicare and Medicaid Services (CMS) and the Affordable Care Act that emphasize reducing hospital readmission rates to improve health care quality, have spurred growing interest in investigating quality measures related to esophagectomy. Although several studies have explored esophagectomy complications and readmission[12-16] none to date have delved into risk factors associated specifically with post-discharge complications (PDC) after esophagectomy. Given that approximately a third of post-operative surgical complications occur post-discharge, and that PDC may differ from pre-discharge complications[17], we hypothesize that several clinical factors may increase risk for developing PDC after esophagectomy.

Using 2005-2013 data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP), we sought to identify the rate of PDC, their associated risk factors, and their influence on early hospital outcomes including readmission and reoperation. We believe that understanding these risk factors will provide additional insights to improve esophagectomy quality outcomes.

Although several studies have investigated postoperative complications and readmissions following esophagectomy, none to our knowledge have explored the distinct role of PDC on early hospital outcomes. This is the first study to use ACS-NSQIP to examine the rate of PDC, their associated risk factors, and their influence on early hospital outcomes after esophagectomy. ACS-NSQIP offers a unique opportunity to assess at a national, multi-institutional level these specific health quality measures that may be unavailable in other large, population-level databases. Our retrospective analysis demonstrates that PDC occur at a low rate but account for a significant number of reoperations and readmissions. Higher BMIs were associated with increased risk of PDC. PDC were different than pre-discharge complications and potentially preventable.

Of the patients with PDC, wound infection, pneumonia, and VTE were among common PDC that could serve as targeted areas for quality improvement. These PDC are consistent with prior studies examining perioperative complications after esophagectomy[21,22]. Interventions to reduce PDC, such as adopting best practices to prevent wound infection and VTE, may improve esophagectomy outcomes. In one study, selective anticoagulant thromboprophylaxis using a DVT risk factor index significantly decreased DVT rates after esophageal cancer surgery[23]. Another study showed that low molecular weight heparin prophylaxis resulted in a 72% reduction in DVT risk after general surgery[24]. These suggest that efforts aimed to reduce common complications after surgery may also be effective and applicable to esophagectomy and should be continued after discharge. Interestingly, PDC rates have remained similar over the studied years, revealing likely no significant changes in practice or indications.

Our modified Poisson regression analysis demonstrates several factors that may increase the risk of PDC and therefore enable us to stratify higher-risk patients for perioperative risk assessment. These include patients with BMI 25.0-29.9 and BMI ≥ 30, compared to normal BMI 18.5-25.0. Although obesity is associated with higher incidence of medical comorbidities like hypertension, diabetes, and cardiovascular disease[25], the association of BMI and complications after esophagectomy is conflicting. Several studies have suggested that overweight and obese patients may have increased risk for complications after esophagectomy, such as longer operative times[26] and greater risks for anastomotic leaks[27], respiratory complications[27], and surgical site infections[28]. However, other studies have shown that obese patients do not have more postoperative complications and longer hospital stays compared to non-obese patients[14,29,30]. Several routine measures implemented in the hospital setting to decrease the most common post-operative complications (i.e., early mobilization, incentive spirometry and pulmonary toilet, DVT prophylaxis, daily dressing changes) are not continued after discharge, and the lack of preventive actions may impact obese patients more than non-obese. As such, appropriate interventions for higher BMI patients may be beneficial and may include enhanced perioperative management of patient comorbidities, preoperative risk stratification, additional patient education, continuation of pulmonary toilet, exercise programs, and DVT prophylaxis after discharge and earlier follow-up. For example, one study has shown that performing preoperative risk analysis on pulmonary function and general status when selecting patients for transthoracic esophagectomy reduced postoperative morbidity rates[31]. Interventions targeting higher BMI patients undergoing esophagectomy may therefore reduce PDC rates.

We decided to separate the patients with pre-discharge complications to keep our groups as homogeneous as possible for fair comparison. Since these patients have a longer hospital stay (median LOS 14 d), a 30-d follow-up may not be long enough to record PDC. It was interesting, however, to see that the most common types of PDC are different than those occurring in the initial post-operative period. This information is helpful to the physician to identify patients at risk of PDC and to plan preventative measures.

The overall readmission rate in our study is 12.8%, consistent with another study’s rate of 18.6%[12]. From our study, 39.5% of PDC patients underwent reoperation (2012-2013) compared to 22.4% of pre-discharge complication patients. This suggests that reoperation may have been indicated due to complications that developed after discharge and subsequent delays in management. Even more striking is our finding that 66.9% of PDC patients were readmitted, compared to 6.6% of pre-discharge complication patients. Recent changes in health policy and reimbursements have brought issues of health care costs to the forefront, resulting in some hospital administrations advocating for earlier discharge of patients. Readmission has become a major focus from health care quality and cost-savings standpoints. However, the use of readmission as a quality metric is still debatable[32,33]. In the case of esophagectomy, efforts to reduce costs by promoting earlier discharge may increase readmission[16], reoperation, and PDC rates; hospital readmission after esophagectomy for esophageal cancer is also associated with poor survival[12]. One study has shown that postoperative complications lead to greater readmission rates after colon resection[34]. In bariatric surgery patients, PDC account for a substantial source of patient morbidity and readmissions[17]. Rather than emphasizing shorter hospital lengths of stay, putting more energy into preventing complications after esophagectomy, and optimizing patient health prior to discharge instead may lead to improved surgical quality outcomes and reduced hospital spending.

Despite the advantages of ACS-NSQIP, the database does pose some limitations to our study. ACS-NSQIP does not contain consistent information on tumor histology, margins, stage, surgical history, neoadjuvant chemotherapy (within 30 d preoperatively), and neoadjuvant radiation (within the last 90 d) that could otherwise provide greater context in interpreting our results. Several complications common to esophagectomy, such as anastomotic leaks, chyle leak, and delayed gastric emptying, are not captured in the database. ACS-NSQIP also identifies 30-d postoperative readmission rather than 30-d post-discharge readmissions, which may introduce immortal person-time bias and lead to shorter number of follow-up days for patients with longer hospital stays. Although PDC could lead to greater mortality, because ACS-NSQIP does not capture data beyond 30-d postoperatively, further studies examining long-term outcomes including mortality is warranted. Readmission data is available only from 2011-2013, but the large patient sample population within that cohort should prevent significant alterations in our conclusions. It is also uncertain whether our findings are applicable to all hospital settings, as hospital participation is voluntary and comprised primarily of academic, tertiary-care centers. Hospital size, setting, patient volume, teaching status, and individual surgeon experience also cannot be adjusted for.

In summary, PDC after esophagectomy account for a significant number of reoperations and readmissions. Adopting best practices to reduce common PDC like VTE and infection, and performing interventions for higher-risk individuals such as those with high BMI, should therefore be considered.

The authors would like to thank Mr. Edwin Lewis for his generous support of Dr. Lidor’s Department of Surgery Research Fund.