Outcomes of Esophageal Diverticulum Repair by Otolaryngologists and General Surgeons Over Time: A NSQIP Analysis

Abstract

Objective

To assess comorbidities, complication rates, and other metrics for open esophageal diverticula repair performed by either otolaryngologists or general surgeons over time.

Study Design

Retrospective database study.

Setting

National Surgery Quality Improvement Database (NSQIP).

Methods

The NSQIP database was queried for cases logged by either otolaryngologists (ENT) or general surgeons (GS) from 2005 to 2020 with CPT code 43130. Preoperative characteristics, postoperative outcomes (length of stay, complications, discharge destination), and operative time were compared between specialties and over time using SPSS and R, with significance set at 0.05.

Results

Preoperatively, GS patients were more likely to have histories of dyspnea (P = .004), pneumonia (P < .001), and >10% body weight loss within 6 months of surgery (P = .008). While there was no significant difference in rates of postoperative complications, patients undergoing repair by ENT had shorter hospital stays, were more likely to be discharged home, and their surgeries were more frequently outpatient and elective (all P < .001). Temporal analysis demonstrated a significant year‐over‐year decrease in length of stay across all cases (P < .001).

Conclusion

Our study suggests that the differences in length of hospital stay and discharge settings may be at least partially attributed to the fact that GS patients were admitted with more preoperative comorbidities. This could reflect hospital staffing, or endoscopic management of sicker patients by ENT. Closer comparison of postoperative management strategies may further elucidate the difference in outcomes.

Esophageal diverticula are outpouchings of esophageal mucosa at weak points in the muscular wall; they are commonly thought to occur due to abnormal motility and increased intraluminal pressure. ¹ , ² Diverticula are rare, with an estimated prevalence of 3.6%; however, they carry significant risks for dysphagia and aspiration. ³ Patients commonly present with symptoms of dysphagia, regurgitation, aspiration, cough, globus sensation, halitosis, weight loss, and voice hoarseness. ² , ⁴ , ⁵ , ⁶ Hypopharyngeal diverticula comprise over 70% of all esophageal diverticula. ⁷ An estimated 0.3% to 7% of pharyngoesophageal diverticula are associated with malignancy. ⁸ , ⁹ , ¹⁰ This carcinogenic transformation may be caused by repeated injury, irritation, and inflammation of the mucosa by food over time. ¹⁰ , ¹¹ One study found that diverticula patients who developed carcinoma had a long‐term history of untreated symptoms related to their diverticula, underscoring that early detection and management can shape outcomes. ⁸

Recommendations regarding management of small, asymptomatic diverticula are heterogenous; some studies recommend against intervention due to risk of iatrogenic injury, while others recommend repair to safeguard from aspiration and malnutrition. ² , ¹² , ¹³ , ¹⁴ One study comparing symptomatic diverticula patients who either received operative intervention or medical management found that while in the postoperative period surgery carried a significantly higher risk of morbidity due to anastomotic leak, there was no association with mortality and long‐term follow up showed patients who underwent open surgery exhibited significantly fewer symptoms. ¹²

Esophageal diverticula can be safely and effectively repaired with open surgery, or laser/stapler‐assisted endoscopy. ² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ The most common complications from both procedure types are due to leaks, which can occur in up to 15% of cases. ² Postoperative complications for open repairs of hypopharyngeal diverticula are reported to occur in 10.6% of cases ¹⁷ , higher than the complication rate reported for endoscopic repairs, 8.7%. ¹⁸ , ¹⁹ While both repairs carry similar risks of infection, abscess formation, and persistent diverticula, ⁶ , ¹⁷ , ²⁰ , ²¹ , ²² , ²³ mediastinitis and cervical emphysema are more common after endoscopic cases, while fistula formation and transient recurrent laryngeal nerve injury is more common in open cases. ¹⁸

One single‐institution study found that patients undergoing open repair experienced longer operative times, longer hospital stays, and required nasogastric tubes for longer periods of time. ²⁴ Another study concurred that endoscopic patients had shorter hospital stays and operative times, however, this study also found that endoscopic patients were more likely to require reoperation and less likely to be asymptomatic at 5 and 10 years postoperatively compared to patients who received open surgical repair. ²⁵ A similar long‐term study also found that the proportion of completely asymptomatic patients was significantly higher after open repair compared to endoscopic. ²⁰ Therefore, despite a higher complication rate and a more difficult postoperative course, open diverticulectomies continue to be widely employed due to improved long‐term outcomes.

While endoscopic diverticula repair is predominantly performed by otolaryngologists, open repair can also be performed by general surgeons (GS). Sparse literature comparing outcomes across specialties exists, and these studies utilize small, single‐institution samples. The American College of Surgeon's National Surgical Quality Improvement Program (NSQIP) contains surgical case data from across the country, including postoperative complications for 30 days after surgery. This study utilizes NSQIP to compare pre‐operative comorbidities, health status, and postoperative medical and surgical complications between patients with esophageal diverticula repaired by either general surgeons or otolaryngologists. We also examined temporal trends in length of stay to provide greater context to our findings.

Materials and Methods

NSQIP was queried for cases performed by either otolaryngologists or general surgeons from 2005 to 2020 associated with the Current Procedural Terminology (CPT) code 43130, which describes “diverticulectomy of the hypopharynx or esophagus, with or without myotomy; cervical approach.” This code represents the open cervical approach, as endoscopic treatments were coded using 43199 prior to 2013 and 43180 starting in 2013. ²⁶ All cases included in the analysis were associated with either ICD‐9 code 530.6 or ICD‐10 code K22.5, both representing a cervical approach to “diverticulum of the esophagus, acquired.” These may include Zenker's Diverticulum (ZD), Killian‐Jamieson Diverticulum (KJD), or Laimers Diverticulum (LD).

Preoperative comorbidities and postoperative outcomes (including complications within 30 days, length of stay, discharge disposition, and operative time) were defined as a priori variables of interest. Chi‐squared analyses were used to compare demographics and preoperative characteristics including history of severe COPD, congestive heart failure, dialysis, disseminated cancer, dyspnea, hypertension, independent functional health status, pneumonia, sepsis, chronic steroid use, prolonged course on ventilator, and recent weight loss prior to surgery. Rates of postoperative medical and surgical complications including wound disruptions, surgical site infections, pneumonia, sepsis, unplanned intubation, DVT/thrombophlebitis, cardiac arrest, death within 30 days of surgery, unplanned reoperation, unplanned readmission, and hospitalization longer than 30 days were also compared.

The variable “discharge destination” was converted to a binary, where discharge to home included patients discharged to either home or a facility that was home, and discharge to facility included patients who expired in the hospital or were discharged to hospice, rehabilitation, acute care, skilled and unskilled nursing facilities, and other transitional care units. Functional health status was also binarized, independent or not, where independent patients could complete acts of daily living without any assistance. American Society of Anesthesiologists (ASA) class was also binarized, patients in class 1 and 2 were grouped as low ASA and 3 and 4 were grouped as high ASA. Length of hospital stay and operative time were compared using Mann‐Whitney U tests. Finally, we performed exploratory post hoc analyses examining temporal trends. Length of stay (LOS) was right‐skewed, therefore log transformation was applied prior to regression analyses.

IBM SPSS software (version 29.0.2.0) and R (4.4.1) were used to perform all statistical analysis. Statistical significance for all tests was set at 0.05. Data visualization was performed with Microsoft Excel v16.74. A nonhuman research determination was approved by the institutional review board at Rutgers University, New Brunswick, New Jersey (Pro2025000277).

Results

Of the 644 cases performed with CPT code 43130 from 2005 to 2020, 284 (44.1%) were performed by GS and 360 (55.9%) by ENT. Patients did not vary significantly in gender and age distributions in either group (Table 1). Analysis of preoperative comorbidities showed that GS patients were more likely to have a high ASA score (P < .046) and a history of dyspnea at rest or on exertion (13.4% vs 6.6%, P = .004) compared to ENT patients (Figure 1). GS patients were also more likely to have lost more than 10% of their total body weight in the 6 months preceding surgery (4.9% vs 1.4%, P = .008). There were no significant differences in rates of other preoperative comorbidities, including congestive heart failure, COPD, diabetes, dialysis, disseminated cancer, hypertension, and steroid use for chronic conditions (all P > .05) (Table 2).

Table 1.

Patient Demographics

	ENT (n = 360)	GS (n = 284)	P‐value
Sex			.957
Female	144 (32%)	113 (40%)
Male	216 (68%)	171 (60%)
Race			<.001
American Indian/ Alaskan Native	4 (1%)	1 (0.4%)
Asian	1 (0.2%)	1 (0.4%)
Black/African American	13 (3.6%)	7 (2.5%)
Other	2 (0.4%)	24 (8.5%)
Unknown	42 (14.8%)	29 (10%)
White	298 (80.3%)	222 (78.2%)
Age			.339
<45	16 (4.4%)	7 (2.5%)
46‐60	63 (17.5%)	41 (14.4%)
60‐74	134 (37.2%)	108 (38%)
75+	147 (50.7%)	128 (45.1%)

Bold value indicates statistical significance.

Figure 1.

Comparison of statistically significant pre‐operative comorbidities for patients receiving open diverticula repair by either otolaryngologists or general surgeons.

Table 2.

Preoperative Comorbidities, Complications, and Outcomes Across Specialties

	ENT (n = 360)	GS (n = 284)	P‐value
Preoperative comorbidities
ASA classification			.046
Low (ASA 1 or 2)	187 (51.9%)	125 (44.1%)
High (ASA 3 or 4)	173 (48.1%)	159 (55.9%)
Congestive heart failure within 30 days prior to surgery	5 (1.4%)	3 (1.1%)	1.00
COPD history (severe)	24 (6.7%)	27 (9.5%)	.185
Diabetes requiring oral agents, other injectables, or insulin	35 (9.7%)	30 (10.6%)	.725
On dialysis	0 (0.0%)	1 (0.4%)	.441
Disseminated cancer	1 (0.3%)	1 (0.4%)	1.000
Dyspnea	24 (6.7%)	38 (13.4%)	.004
Elective surgery	294 (81.9%)	168 (59.2%)	<.001
Emergency surgery	1 (0.3%)	1 (0.4%)	1.000
“Independent” functional health status	350 (97.5%)	269 (94.7%)	.092
Hypertension requiring medications	173 (48.1%)	158 (55.6%)	.056
Sepsis	0 (0.0%)	0 (0.0%)	1.00
Septic shock	0 (0.0%)	0 (0.0%)	1.00
Reported steroid use for a chronic condition	11 (3.1%)	17 (6.0%)	.070
>10% loss body weight in last 6 months	5 (1.4%)	14 (4.9%)	.008*
On ventilator for >48 hours at time of surgery	0 (0.0%)	1 (0.4%)	1.00
Surgical complications
Wound disruptions	0 (0.0%)	2 (0.7%)	.194
Superficial SSI	4 (1.1%)	3 (1.1%)	1.00
Deep incisional SSI	4 (1.1%)	4 (1.4%)	1.00
Organ space SSI	6 (1.7%)	6 (2.1%)	.678
Medical complications
Cardiac arrest requiring CPR	0 (0.0%)	1 (0.4%)	.441
CVA/stroke with neurological deficit	1 (0.3%)	0 (0.0%)	1.000
DVT/thrombophlebitis	1 (0.3%)	1 (0.4%)	1.00
Pneumonia	5 (1.4%)	6 (2.1%)	.482
Sepsis	6 (1.7%)	4 (1.4%)	1.00
Septic shock	0 (0.0%)	1 (0.4%)	.441
Urinary tract infections	2 (0.6%)	2 (0.7%)	1.00
On ventilator >48 hours, failure to wean	2 (0.6%)	1 (0.4%)	1.00
Unplanned intubation	2 (0.6%)	5 (1.8%)	.268
Other outcomes
Death within 30 days	1 (0.3%)	3 (1.1%)	.462
Days from operation to discharge	2.49	3.08	.007
Mean length of total hospital stay (days)	2.60	3.53	<.001
Mean total operation time (minutes)	108.32	96.71	.048
Still in the hospital >30 days	0 (0%)	0 (0%)	1.00
Reoperation within 30 days	17 (4.7%)	21 (7.4%)	.153
Discharge to home	290 (80.6%)	168 (59.2%)	<.001
Outpatient surgery	151 (41.9%)	68 (23.9%)	<.001

Bold values indicate statistical significance.

ENT had significantly shorter average length of hospital stays (2.6 days vs 3.53 days, P < .001) and were more likely to be discharged home than GS patients (80.6% vs 59.2%, P < .001) (Figures 2 and 3). ENT surgeries were also more likely to be elective (81.7% vs 59.1%, P < .001) and take place in the outpatient setting (41.9% vs 23.9%, P < .001) (Figure 2).

Figure 2.

Comparison of postoperative outcomes in for patients receiving open diverticula repair by either otolaryngologists or general surgeons.

Figure 3.

Comparison of mean length of total hospital stay (days) between patients operated on by general surgery and otolaryngology for open diverticula repair.

There was no significant difference in rates of wound disruptions, surgical site infections, pneumonia, sepsis, unplanned intubation, cardiac arrest, stroke, DVT/thrombophlebitis, urinary tract infections, or failure to wean off a ventilator within 48 hours of surgery between groups (Table 2). GS patients had shorter average operative times compared to ENT patients (96.71 vs 108.32 minutes, P = .048, Figure 4). There was no significant difference in 30‐day mortality between ENT and GS patients (0.3% vs 1.1%, P = .462) (Table 2).

Figure 4.

Comparison of mean total operating time (minutes) between patients operated on by General surgery or Otolaryngology for open diverticula repair.

Across all cases, the most common postoperative complications were superficial (1.1%), deep (1.2%), and organ space (1.9%) surgical site infections, and pneumonia (1.7%); the overall complication rate was 10.7% for open diverticula repair. 9.2% of ENT patients experienced any kind of medical or surgical complication, compared to 12.7% of GS patients (P = .153).

ENT case volume increased over time, rising at a steeper rate compared to GS cases on regression (P = .003, Figure 5). ENT case volume also increased substantially after the introduction of the endoscopic code, 43180, in 2015 (33.0 vs 16.2 cases/year, P = .002), while GS volumes remained stable (17.7 vs 17.8 cases/year, P = .97).

Figure 5.

Frequency of reported open cervical repairs over time.

Mean LOS also varied over time. Median LOS across all years was 3.01 days (SD 4.20 days, Table 3) and was typically 2 to 3 days from 2005 to 2010, and 1 to 2 days from 2011 to 2020. A small number of patients in all years experienced prolonged hospitalizations up to 35 days (Figure 6). On log‐transformed linear regression, year of admission was significantly associated with a decrease in LOS, corresponding to a 1.5% reduction in LOS per year (B = −0.015, 95% CI −0.028 to −0.001, P = .033).

Table 3.

Summary Statistics for Length of Stay Over Time

Year	Median	Range	SD
2005	2	1‐3	1.414
2006	3	1‐27	10.139
2007	3	0‐9	2.281
2008	2	0‐10	2.23
2009	2	0‐25	5.406
2010	2	0‐24	3.216
2011	1	0‐25	7.396
2012	1	0‐22	4.165
2013	2	0‐31	4.853
2014	2	0‐23	4.281
2015	2	0‐35	5.817
2016	1	0‐15	2.636
2017	2	0‐18	3.771
2018	1	0‐18	3.069
2019	1	0‐11	2.193
2020	1	0‐7	1.441

Figure 6.

Box plots of length of stay by year. ASA, American Anesthesiology Association Class.

Discussion

Open diverticulectomy is an effective treatment of esophageal diverticula, with lower rates of recurrence compared to endoscopic surgery. ² While open surgery for esophageal diverticula is routinely performed by both GS and ENT, no existing studies compare outcomes between the specialties. Therefore, we sought to compare preoperative comorbidities and patient outcomes following diverticulectomies performed by ENT and GS, acknowledging identifiable temporal trends over the study period.

Preoperatively, patients who were operated on by GS were more likely to have a high ASA class (3 or 4), dyspnea at rest or on exertion, or have recent unintentional weight loss. These changes could be partially due to management differences between specialties. A study from Finland comparing treatment for Zenker's across 4 specialties found that otolaryngologists were the second most likely to treat the condition surgically (62.7%), while general surgeons were the least likely specialty (28.5%) to treat surgically, frequently opting for conservative medical management (71.5%). ²⁷ If general surgeons treat the condition less frequently, the patients they do choose to operate on may have more advanced disease or co‐morbidity burden.

Our findings could also reflect GS surgeries being performed within the inpatient setting rather than as elective procedures. This may be reflective of larger numbers of general surgeons on staff relative to otolaryngologists and the growing shift in otolaryngology toward ambulatory surgical centers for elective surgeries amidst financial incentives and safety advancements. ²⁸ , ²⁹ , ³⁰ Another important consideration is that general surgeons exclusively perform open diverticulectomy, whereas otolaryngologists offer both open and endoscopic approaches. As a result, patients with greater medical comorbidities likely undergo endoscopic rather than open repair with ENT surgeons, influencing the observed patient selection and outcomes.

GS patients in our study had longer hospital stays and were less likely to be discharged home. One likely contributor to this finding is the difference in preoperative ASA class between the 2 specialties. One study utilizing NSQIP to assess predictive factors for complications following open diverticula repair found that functional health status, systemic sepsis, dyspnea, and recent weight loss were not significantly different between patients who experienced complications or longer hospital stays and those who did not. ¹⁷ This same study found that age, ASA class, and smoking status were independent predictors of adverse events. ¹⁷

The observed differences in surgical outcomes between specialties may also be due to characteristics of the diverticula themselves or variations in postoperative management. Softeland et al compared endoscopic ZD repair by gastrointestinal surgeons or otolaryngologists in Norway and reported that ENT patients were more likely to have larger diverticula, delayed diet advancement, experience longer hospital stays, were less likely to receive prophylactic antibiotics, more likely to receive intraoperative dexamethasone, and more likely to have postoperative infections compared to GS patients. ⁴

While Softeland et al's findings suggest potential specialty‐related differences, key distinctions limit their applicability to our cohort. Their analysis involved endoscopic repairs, and the study was conducted in a different country, where perioperative practices, discharge criteria, and antibiotic/steroid protocols differ substantially from those in the United States. These differences may account for the divergence from our findings, as our US‐based data demonstrated shorter hospital stays and similar complication rates among ENT patients compared to GS patients.

Postoperative management and timing of care for ZD is also variable and may contribute to differences in outcomes. Most patients receive a clear liquid diet after surgery, and then advance to thickened liquids or soft foods over 1 to 2 weeks to allow the mucosa time to heal. ⁷ , ³¹ , ³² A survey from the American Bronchoesophagological Association (ABEA) reported that most (64.7%‐92.3% depending on approach) patients undergoing treatment of ZD by otolaryngologists began peroral diets on POD 1, but did not assess general surgery ZD repairs. ³³ Although studies directly examining general surgeon's dietary practices in ZD repair are lacking, research in other surgical areas suggests a more conservative approach. For example, Chough et al found that patients who underwent colorectal surgery with general surgeons were less likely to have an oral diet ordered on POD 0 or 1 compared to those treated by colorectal surgeons. ³⁴ Another single‐institution study of elective intestinal surgery reported that patients treated by general surgeons were less likely to receive POD1 oral feeding compared. ³⁵ Although these findings come from different contexts, they raise the possibility that general surgeons may also have later postoperative feeding in diverticula repairs compared to otolaryngologists.

Given that our GS cohort was more acutely ill at baseline, postoperative practices could affect length of stay and discharge destinations between the 2 specialties. Prospective data from the POUCH Collaborative similarly found that complications were strongly associated with both prolonged hospitalization and dietary restriction, with over 20% of patients with complications discharged NPO with a nasogastric tube, compared to only 1% of those without complications. ³⁶ A study conducting a more direct comparison of postoperative management strategies between specialties, particularly assessing the rate of diet advancement, may further explain the difference in outcomes we observed.

Despite these outcome differences, there were no differences in the overall rates of medical or surgical complications between specialties. The most common complications within 30 days of repair for both specialties included surgical site infections and pneumonia, with an overall complication rate of 10.7%. Other studies report similar trends in overall complication types and rates. ¹⁷ , ³⁷ In our study, 9.2% of ENT patients and 12.7% of GS patients experienced any kind of medical or surgical complication measured. Similar to Rizzetto et al's study, ⁶ our findings suggest that there is benefit to discussing endoscopic surgery rather than a transcervical approach for diverticula patients with chronic systemic disease. This could shorten the recovery period and minimize postoperative complications for higher‐risk patients.

Notably, ENT procedures increased substantially over the study period, while GS did not. Case volumes before and after 2015 were also examined to understand the impact of the introduction of code 43180, which describes transoral repair of Zenker's diverticulum. Prior to this, some endoscopic cases were likely billed under the open code, raising the possibility of miscoding. ³⁸ , ³⁹ However, after 2015, ENT volumes for the open code still increased significantly, whereas GS volume did not significantly change, suggesting that introduction of 43180 was not associated with reduced use of the open diverticulectomy code (43130). Thus, miscoding alone is unlikely to explain the observed trends, which may instead reflect increasing diverticulum repair volume or evolving referral patterns favoring subspecialists.

We also observed a decrease in median length of stay during the study period, at an approximate 1.5% reduction in LOS per year. While NSQIP does not capture operative approach or instrumentation used, the temporal decline likely at least partially reflects adoption of minimally invasive tools such as staplers, lasers, or energy‐based devices, improving time to discharge. ⁴⁰

The temporal improvement in LOS also likely reflects evolving perioperative management patient selection strategies. It is possible that earlier in the dataset, more complex or acutely ill patients may have undergone open repair, whereas in later years, these patients may have underwent endoscopic repair with appropriate coding, making healthier patients more likely to undergo open repair and reducing overall LOS within the group.

Our study was limited in comparing long‐term outcomes, as NSQIP follows patients for only 30 PODs. Additionally, there are other procedure‐specific complications such as transient recurrent laryngeal nerve injury, persistent diverticula, and cervical emphysema that are not captured by NSQIP, but which may represent important differentiating factors between specialties. We also could not compare data including the size of diverticula, preoperative patient optimization, and postoperative medical management protocols. Differences in protocols, particularly diet advancement and discharge criteria, likely contribute to the outcome disparities observed and represent targets for standardization across specialties. Because CPT code 43130 broadly encompasses open, cervical repairs of a diverticulum of the hypopharynx or esophagus, it lacks specificity for Zenker's versus other pharyngoesophageal diverticula. As such, our results reflect open diverticular repairs, which limits procedure‐specific interpretation. Our temporal analyses are limited as NSQIP data does not provide information on institutional practices, operative approach, or coding changes over time, all of which may influence observed trends. Participation in NSQIP is voluntary and may select for institutions with more resources and higher surgical volume. Despite limitations, NSQIP has rigorous requirements for data entry and provides standardized data on a large, diverse, national patient sample to provide reliable insight on postoperative outcomes for diverticula patients by specialty.

Conclusion

While there was no difference in rates of postoperative complications, patients undergoing GS repair had longer hospital stays and were less likely to be discharged home compared to patients operated on by ENT. However, GS surgeries were more likely to take place nonelectively, in the inpatient setting. This could reflect hospital staffing differences or trends toward outpatient or trends towards outpatient and endoscopic surgeries within otolaryngology. Our findings suggest that the difference in length of hospital stay and discharge settings may be at least partially attributed to the fact that GS patients present with greater pre‐operative comorbidities and higher ASA classes, as well as likely variation in postoperative diet practices between specialties. These results underscore the importance of considering endoscopic surgery as an alternative to an open approach for higher‐risk patients, especially given the comparable complication rates between surgical approaches. Importantly, our temporal analysis demonstrated a significant overall decrease in length of stay year over year, highlighting improvements in perioperative management and evolving surgical practice patterns across the study period. Ultimately, aligning postoperative protocols between specialties and leveraging evolving surgical techniques may improve efficiency while maintaining safety for patients undergoing esophageal diverticula repair.

Author Contributions

Nour Abdel‐Azim, design, analysis, writing, reviewing; Emma Thompson, design, analysis, writing, reviewing; Om Tripathi, literature review, writing, reviewing; Kenneth Yan, design, analysis, writing, reviewing.

Disclosures

Competing interests

None.

Funding source

None.

Data Availability Statement

The data that support the findings of this study are available from the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP). Access to these data is subject to approval by the American College of Surgeons and is available to eligible researchers through application at https://www.facs.org/quality-programs/data-and-registries/acs-nsqip/.