- Email: [email protected]
Surgical management of esophageal diverticulum: a review of the Nationwide Inpatient Sample database
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Available online at www.sciencedirect.com
journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Surgical management of esophageal diverticulum: a review of the Nationwide Inpatient Sample database Michael T. Onwugbufor, BS,a Augustine C. Obirieze, MBBS, MPH,b Gezzer Ortega, MD, MPH,b Delenya Allen, MD, MPH,b Edward E. Cornwell III, MD, FACS,b and Terrence M. Fullum, MD, FACSa,b,c,* a
Howard University College of Medicine, Washington, District of Columbia Department of Surgery, Howard University College of Medicine, Washington, District of Columbia c Division of Minimally Invasive and Bariatric Surgery, Department of Surgery, Howard University College of Medicine, Washington, District of Columbia b
article info
abstract
Article history:
Background: Esophageal diverticulum is rare in the United States. The mainstay treatment
Received 5 January 2013
of symptomatic esophageal diverticulum is surgical correction. Much of the available
Received in revised form
information regarding esophageal diverticulum and its surgical management has been
6 May 2013
derived from small studies and institutional reviews. Our study objective was to investigate
Accepted 9 May 2013
the demographics, perioperative conditions, and predictors of outcomes after surgical
Available online 2 June 2013
treatment of acquired esophageal diverticulum using a nationally representative database. Methods: A retrospective review using the Nationwide Inpatient Sample database from
Keywords:
2000e2009 was performed for patients with acquired esophageal diverticulum. The patients
Esophagus
were stratified into Zenker’s diverticulum (ZD) or non-Zenker’s diverticulum (NZD)
Diverticulum
subgroups. The covariates retrieved included age, gender, ethnicity, insurance type, and
Zenker’s
Charlson comorbidity index. A multivariate analysis was performed to determine the
Epiphrenic
predictors of postoperative morbidity. Discharge-level weights were applied.
Diverticulectomy
Results: Overall, a total of 4253 patients met our inclusion criteria, 3197 (75%) with ZD and
Cricopharyngeal myotomy
1056 (25%) with NZD. In the ZD group, the mean age was 73 12.3 y, and most were men (55%) and white (67%). The mean length of stay was 5.82 8.08 d, and the mortality rate was 1.2%. The most common complication was septicemia or sepsis (2.0%). The black patients had higher odds of postoperative morbidity than the white patients (odds ratio [OR] 2.29, 95% confidence interval [CI] 1.02e5.17). The risk of overall postoperative morbidity was 52% greater for women (OR 1.52, 95% CI 1.01e2.29). An increasing Charlson comorbidity index was an independent predictor of morbidity. In the NZD group, the mean age was 69 13.9 y, and most were also men (51%) and white (63%). The mean length of stay was 8.13 10.56 d, and the mortality rate was 1.6%. The most common complication was air leak (3.1%). The black and Hispanic patients had higher odds of postoperative morbidity than the white patients (OR 1.97, 95% CI 1.05e3.72 and OR 2.37, 95% CI 1.06e5.30, respectively). An increasing Charlson comorbidity index was an independent predictor of
Presented at the Eighth Annual Academic Surgical Congress, February 5e7, 2013, New Orleans, Louisiana. * Corresponding author. Division of Minimally Invasive and Bariatric Surgery, Department of Surgery, Howard University Hospital, 2041 Georgia Avenue, Northwest, Tower Suite 4100-B, Washington, DC 20060. Tel.: þ1 202 865 1286; fax: þ1 202 865 3063. E-mail address: [email protected] (T.M. Fullum). 0022-4804/$ e see front matter ª 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2013.05.036
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
121
morbidity. Compared with laparoscopy, the risk of developing postoperative morbidity was higher with the thoracotomy procedure (OR 7.45, 95% CI 1.11e50.18). Conclusions: Using a nationally representative database, our study found that female gender, black race, and the presence of comorbidities were associated with increased postoperative morbidity among patients with ZD. Among the patients with NZD, black and Hispanic patients had worse postoperative morbidity than the white patients, and the presence of comorbidities was associated with increased postoperative morbidity. Thoracotomy for the correction of NZD was associated with increased postoperative morbidity compared with the laparoscopic approach. ª 2013 Elsevier Inc. All rights reserved.
1.
Introduction
Esophageal diverticulum is rare in the United States, with a prevalence of about 0.06%e4%, and it primarily affects the elderly [1]. A large proportion of patients with esophageal diverticula are asymptomatic and do not always require treatment [1,2]. However, the mainstay of treatment of symptomatic esophageal diverticulum has been surgical correction [1]. The most common symptoms that patients present with include dysphagia, regurgitation, halitosis, unintentional weight loss, chest pain, aspiration, and choking [1]. Zenker’s diverticulum (ZD; presenting in the proximal esophagus) is the most common form of esophageal diverticulum [3]. It has been shown to be a result of increased hypopharyngeal pressure during swallowing and dysfunction of the cricopharyngeus muscle and the upper esophageal sphincter [3,4]. Midesophageal diverticulum is frequently asymptomatic and not treated [1]. Traction has been proposed as the possible underlying mechanism causing it [5]. Finally, epiphrenic diverticulum is usually located in the distal 10 cm of the esophagus. Many investigators have suggested that a primary esophageal body motility disorder, such as diffuse esophageal spasm or a nonspecific esophageal motility disorder, underlies the development of the condition in most patients [5,6]. Some patients with epiphrenic diverticulum have also been shown to have lower esophageal sphincter abnormalities such as achalasia using manometry diagnostic studies [2,5]. Most of the information that has been collected regarding esophageal diverticulum and its surgical management has been gathered from small studies and institutional reviews. Therefore, our study evaluated the demographics, perioperative conditions, and surgical outcomes of acquired esophageal diverticulum in a larger population using a national database and determined the independent risk factors and predictors of the postoperative outcomes.
2.
Methods
We conducted a retrospective study using the Nationwide Inpatient Sample database from 2000e2009. The Nationwide Inpatient Sample is a part of the Healthcare Cost and Utilization Project, sponsored by the Agency for Healthcare Research and Quality. It is the largest all-payer inpatient care database publicly available in the United States. The data are from a 20%, stratified sample of discharges from US community hospitals, comprising 95% of the US population [7]. Using the
“International Classification of Diseases, Ninth Edition, Clinical Modifications” (ICD-9-CM) diagnosis codes, we identified patients 18 y with a primary diagnosis of acquired esophageal diverticulum (ICD-9-CM code 530.6). The study cohort was stratified into two subgroups according to the procedures (identified by ICD-9-CM procedure codes) the patient had undergone during the same admission to surgically address the esophageal diverticulum: Zenker’s diverticulum (ZD) and non-Zenker’s diverticulum (NZD). The ICD-9-CM procedures codes used to identify patients with ZD included those for cricopharyngeal myotomy, pharyngoesophageal diverticulectomy, and hypopharyngeal esophageal diverticulectomy. The patients with NZD were identified using the ICD-9-CM procedure codes for thoracoscopic diverticulectomy, laparoscopic diverticulectomy, esophagomyotomy, thoracotomy, and fundoplication. Associated preoperative symptoms and conditions, including dysphagia, regurgitation, halitosis, choking, coughing, weight loss, chest pain, esophageal cancer, and aspiration pneumonia were identified using the ICD-9-CM diagnosis codes (Table 1). We also recorded the postoperative complications, including air leaks, septicemia and/or sepsis, mediastinitis, mediastinal empyema with or without fistula, mediastinal abscess, postoperative wound infection, and postoperative hemorrhage using the ICD-9-CM diagnosis codes. These postoperative complications were selected in accordance with the most commonly reported and discussed postoperative complications in reports from previous institutional studies and reviews. Within our patient population, we identified as many ICD-9-CM codes matching these published complications as possible. The covariates retrieved include age, gender, race or ethnicity, Charlson comorbidity index, median household income quartile for the patient’s residential postal code, primary payer, hospital type (teaching versus nonteaching), hospital bed size (small, medium, or large), hospital location (rural versus urban), hospital ownership or control (government, private, or government, private collapsed, unclassified). For the study, our outcome of interest was the overall complication rate defined as having at least one of the postoperative complications previously listed. Univariate analyses on the covariates were conducted separately within the ZD and NZD subgroups. Multivariate logistic regression analyses and postanalysis Wald tests, performed separately within each of the two subgroups studied were used to investigate the predictors of postoperative morbidity, adjusting for patient age, gender, race or ethnicity, primary payer, median income in the residential postal code, Charlson comorbidity
122
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 1 e ICD-9-CM diagnosis and procedure codes used to identify selected pre- and postoperative conditions and procedures. Diagnosis and procedure Diagnosis Dysphagia Chest pain Weight loss Halitosis or choking Cough Aspiration pneumonia Esophageal cancer Mediastinitis Mediastinal abscess Mediastinal empyema with fistula Mediastinal empyema without fistula Air leaks Septicemia or sepsis Postoperative wound infection Postoperative hemorrhage Procedures Cricopharyngeal myotomy Pharyngoesophageal diverticulectomy Esophageal diverticulectomy Local excision of esophagus Endoscopic excision of esophagus Esophagomyotomy Thoracotomy Thoracoscopy Laparoscopy Fundoplication
ICD-9-CM codes 787.24 786.5 783.21 784.99 786.2 507.0 150.0e150.9 519.2 513.1 510.0 510.9 512.1e512.2 038.x, 995.91, 995.92 998.5, 998.59 998.11e998.12 29.31 29.32 42.31 42.32 42.33 42.7 34.0e34.02 34.21 54.21 44.66e44.67
index, discharge year, hospital region, bed size, location, teaching status, and ownership or control. The postanalysis Wald test was used to investigate the likelihood that a covariate (with all categories taken together) was an independent predictor of the development of postoperative morbidity. Discharge-level weights were applied in all analyses to obtain national estimates. The Hosmer-Lemeshow goodness-of-fit test was used to ascertain that the models fit. Statistical analyses were done using STATA/MP, version 11.0 (StataCorp, College Station, TX). Statistical significance was defined as P < 0.05.
3.
Results
Overall, a total of 4253 patients met our inclusion criteria, 3197 (75%) with ZD and 1056 (25%) with NZD. Among the patients with the ZD, the mean age was 73 12.3 y, with most patients aged 65 y (78%). Most of the patients with ZD were men (55%), white (67%), within the highest income quartile (35%), and underwent their procedures at a teaching hospital (57%). The mean length of stay in this group was 5.82 8.08 d, and the mortality rate was 1.22%. The most common preoperative symptoms were dysphagia (23.5%), aspiration pneumonia (7.4%), and weight loss (1.0%). The most common complication was septicemia or sepsis (1.97%). In the NZD group, the mean age was 69 13.9 y, with most patients aged 65 y (66%). Most were men (51%), white (63%), within the highest income quartile (33%), and had undergone their procedures at a teaching hospital (64%). The mean length
of stay was 8.13 10.56 d. The mortality rate was 1.61%. The most common preoperative symptoms were dysphagia (17.9%), aspiration pneumonia (5.7%), weight loss (1.2%), and chest pain (0.9%), and the most common complication was air leak (3.1%; Table 2). From 2000e2004, among the patients with NZD, 1.0%, 2.1%, 6.0%, 13.1%, and 77.9% had undergone thoracotomy, thoracoscopy, laparoscopy, fundoplication, and esophagomyotomy, respectively. From 2005e2009, 0.6%, 2.2%, 4.5%, 17.4%, and 75.3% had undergone thoracotomy, thoracoscopy, laparoscopy, fundoplication, and esophagomyotomy, respectively. On multivariate analysis, among the patients with ZD, black patients had more than twice the odds of morbidity compared with the white patients (odds ratio [OR] 2.29, 95% confidence interval [CI] 1.02e5.17). In addition, the risk of morbidity was 52% greater for women (OR 1.52, 95% CI 1.01e2.29). An increasing Charlson comorbidity index was an independent predictor of postoperative morbidity (Wald test P < 0.001; Table 3). Among the patients with NZD, black and Hispanic patients had higher odds of postoperative morbidity than white patients (OR 1.97, 95% CI 1.05e3.72 and OR 2.37, 95% CI 1.06e5.30, respectively). The independent predictors of postoperative morbidity were race or ethnicity (Wald test P ¼ 0.04), procedure type (Wald test P ¼ 0.01), and an increasing Charlson comorbidity index (Wald test P < 0.001). Finally, the risk of postoperative morbidity was greater after thoracotomy than after laparoscopy (OR 7.45, 95% CI 1.11e50.18; Table 4).
4.
Discussion
The main surgical procedures used for the treatment of ZD have been cricopharyngeal myotomy, diverticulectomy, and diverticulopexy, all of which can be used in combination with each other through a transcervical approach or endoscopically [1,3,4,8e10]. A retrospective review of 87 patients with ZD at a tertiary care institution [8] and another study of 11 patients with ZD by Constantin et al. [10] showed that cricopharyngeal myotomy is necessary for the treatment of ZD but might not be sufficient because it resulted in a poorer outcome in their studies compared with cricopharyngeal myotomy plus diverticulectomy or diverticulopexy. Regarding the invasive surgical approach for the treatment of ZD, two studies, by Lerut et al. [3] and Gutschow et al. [11], advocated the open surgical approach because of its superior outcome compared with the endoscopic approach, although a study of 297 patients with ZD by Bonavina et al. [12] showed no difference in outcomes for both approaches. Epiphrenic diverticulum can be surgically repaired using a transthoracic or laparoscopic approach and has almost always included the incorporation of a fundoplication in addition to diverticulectomy and/or esophagomyotomy [1,13]. Reznik et al. [14] assessed the pathophysiology-directed operative strategy of excision (diverticulectomy), repair of esophageal wall, and relief of functional and mechanical obstruction in 44 patients with epiphrenic diverticulum and showed it to be safe, with positive outcomes in most patients. A retrospective review by Varghese et al. [15] of 35 patients with epiphrenic diverticulum reported excellent long-term results using the
123
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 2 e Patient characteristics (2000e2009 Nationwide Inpatient Sample database; n [ 4253). Characteristic Mean age (y) Age group (n) <45 45e64 65e84 85 Gender (n) Male Female Unknown Race or ethnicity (n) White Black Hispanic Other or unknown Preoperative symptoms and conditions Dysphagia Chest pain Weight loss Cough Aspiration pneumonia Esophageal cancer Income quartile (n) Lowest Second Third Highest Unknown Primary payer (n) Private insurance Medicare Self-pay, no charge, or other Unknown Charlson comorbidity index (n) 0 1 2 Mean length of stay (d) Mortality (%) Postoperative complications (n) Mediastinitis Mediastinal abscess Mediastinal empyema Mediastinal empyema (with or without fistula) Air leaks Septicemia or postoperative sepsis Postoperative wound infection Postoperative hemorrhage Hospital teaching status (n) Nonteaching Teaching Hospital ownership or control (n) Government, nonfederal Private Government, private collapsed, uncategorized
ZD (n ¼ 3197)
Non-ZD (n ¼ 1056)
73.1 12.3
69.1 13.9
72 641 1948 536
(2.3) (20.1) (60.9) (16.8)
59 298 565 134
1747 (54.6) 1431 (44.8) 19 (0.6)
(5.6) (28.2) (53.5) (12.7)
539 (51.0) 516 (48.9) 1 (0.1)
2148 90 62 897 (n) 752 31 31 13 236 4
(67.2) (2.8) (1.9) (28.1) (23.5) (1.0) (1.0) (0.4) (7.4) (0.1)
189 9 13 2 60 3
(17.9) (0.9) (1.2) (0.2) (5.7) (0.3)
417 526 844 1122 288
(13.0) (16.5) (26.4) (35.1) (9.01)
177 174 254 349 102
(16.8) (16.5) (24.1) (33.1) (9.7)
761 2324 54 58
665 64 36 291
(23.8) (72.7) (1.7) (1.8)
1957 (61.2) 828 (25.9) 412 (12.9) 5.82 8.08 1.2
323 674 25 34
(63.0) (6.1) (3.4) (27.6)
(30.6) (63.8) (2.4) (3.2)
643 (60.9) 292 (27.7) 121 (11.5) 8.13 10.56 1.6
11 1 3 17
(0.3) (0.03) (0.1) (0.5)
5 4 2 14
(0.5) (0.4) (0.2) (1.3)
19 63 36 44
(0.6) (2.0) (1.13) (1.38)
33 30 20 13
(3.1) (2.8) (1.89) (1.23)
1373 (43.0) 1821 (57.0)
376 (35.6) 679 (64.3)
203 (6.5) 1205 (37.7) 1789 (56.0)
64 (6.1) 400 (37.9) 592 (56.1)
Data presented as mean standard deviation or n (%).
transthoracic approach. Likewise, a similar review by Rosati et al. [16] of 20 patients with epiphrenic diverticulum who had undergone treatment with the laparoscopic approach
Table 3 e Adjusted ORs and 95% CIs for covariates used in multivariate analysis and P values of Wald test among patients with ZD. Variable Age group (y) <44 45e64 65e84 >85 Gender Male Female Race or ethnicity White Black Hispanic Other Charlson comorbidity index 0 1 2 Income quartile Lowest Second Third Highest Primary payer Private insurance Medicare Self-pay, no charge Hospital bed size Small Medium Large Hospital teaching status Nonteaching Teaching Hospital location Rural Urban Hospital control Government, nonfederal Private Government or private
OR
95% CI
1.65 0.60 0.73 Reference
0.46e5.97 0.23e1.52 0.44e1.22 Reference
Reference 1.52
Reference 1.01e2.29
Reference 2.29 0.66 0.91
Reference 1.02e5.17 0.08e5.31 0.56e1.49
Reference 2.02 2.61
Reference 1.29e3.17 1.53e4.44
Reference 0.76 0.86 0.68
Reference 0.36e1.61 0.45e1.66 0.34e1.38
Reference 1.20 1.37
Reference 0.52e2.77 0.33e5.69
Reference 2.84 3.09
Reference 0.99e8.17 1.13e8.42
Reference 0.83
Reference 0.47e1.44
Reference 1.77
Reference 0.66e4.76
Reference 1.19 2.08
Reference 0.45e3.16 0.68e6.36
P value (Wald test) 0.18
0.05
0.19
<0.001
0.80
0.89
0.09
0.44
0.09
0.25
reported satisfactory long-term outcomes. These studies have provided substantial comparative data regarding the various surgical techniques for the treatment of esophageal diverticulum. In our study, a review of a large national database, we classified the ZD group according to the unique primary types of procedures used for its surgical treatment. The NZD group was also classified according to the unique primary types of procedures used for the surgical treatment of epiphrenic diverticulum. Hence, most of the patients in the NZD group were assumed to mostly have had epiphrenic diverticulum. Similar to numerous studies [1e6,8e10], the most common symptom experienced by most of the patients in the present study was dysphagia. This symptom has been suggested to be the result of the abnormal swallowing mechanism and motility disorder associated with this pathologic entity [3e6]. Similar to other studies of ZD [4,9] and epiphrenic diverticulum [15],
124
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 4 e Adjusted ORs and 95% CIs for covariates used in multivariate analysis and P values of Wald test among patients with non-ZD. Variable Age group (y) <44 45e64 65e84 >85 Gender Male Female Race or ethnicity White Black Hispanic Other Procedure Laparoscopy Thoracoscopy Thoracotomy Charlson comorbidity index 0 1 2 Income quartile Lowest Second Third Highest Primary payer Private insurance Medicare Self-pay, no charge Hospital bed size Small Medium Large Hospital teaching status Nonteaching Teaching Hospital location Rural Urban Hospital control Government, nonfederal Private Government or private
Or
95% CI
1.16 0.72 0.76 Reference
0.42e3.19 0.37e1.41 0.51e1.14 Reference
Reference 1.23
Reference 0.90e1.67
Reference 1.97 2.37 1.10
Reference 1.05e3.72 1.06e5.30 0.75e1.62
Reference 3.42 7.45
Reference 0.77e15.09 1.11e50.18
Reference 1.60 2.30
Reference 1.13e2.28 1.49e3.54
Reference 1.02 0.87 0.88
Reference 0.59e1.77 0.52e1.47 0.53e1.47
Reference 1.04 1.29
Reference 0.60e1.80 0.44e3.78
Reference 1.21 1.46
Reference 0.65e2.27 0.82e2.60
Reference 0.91
Reference 0.58e1.41
Reference 1.76
Reference 0.86e3.58
Reference 1.88 1.75
Reference 0.79e4.46 0.66e4.64
P value of Wald test 0.42
0.27
0.04
0.01
<0.001
0.95
0.97
0.33
0.94
0.97
0.36
esophageal diverticulum mostly occurred in patients >65 y in the present study. According to a study by Thota and Richter [17], a possible explanation for the susceptibility of the older population is that morphologic changes in the pharynx and esophagus contribute to age-related changes in esophageal physiology [17]. Also, from our results, this pathologic entity appears to be more common within the white patient population. This is concordant with another study of 35 patients with epiphrenic diverticulum, in which 34 of the patients were white [15]. A suggestion has also been made of a genetic mechanism in the differences in the incidence of ZD related to race or ethnicity [10]. The length of stay was longer for the NZD group than for the ZD group. Moreover, leaks had also been reported to be the most common postoperative complication of
epiphrenic diverticulum in previous studies [6], just as was found in the present study. However, in the context of our study, the air leaks reflected the post-thoracotomy air leaks and iatrogenic pneumothorax secondary to the intraoperative manipulation of the lung parenchyma with traumatic laparoscopic or thoracoscopic instruments during access and repair of the thoracic esophagus. We believe that infectious mediastinal complications such as mediastinitis, mediastinal abscess, mediastinal empyema (with fistula), and mediastinal empyema (without fistula) were therefore consequences of suture or staple line leaks. In the results of the multivariate analysis of patients with ZD, the black patients had higher odds of developing postoperative complications than the white patients. Regarding gender, the women in the ZD group were more likely to develop postoperative complications than were the men. Furthermore, an increasing incidence of comorbidities was an independent predictor of postoperative morbidity. Because the mean age in the ZD group was >70 y, it could be inferred that most patients with ZD will present with significant comorbidities, increasing their overall risk of postoperative complications. In the results of the multivariate analysis of patients with NZD (epiphrenic), the black and Hispanic patients were more likely to develop postoperative complications than were the white patients. Race or ethnicity, procedure type, and increasing comorbidities were all independent predictors of postoperative morbidity. Furthermore, the increased odds of developing postoperative complications after thoracotomy for the treatment of NZD (epiphrenic diverticulum) compared with using a laparoscopic procedure is consistent with three recent reviews [1,6,18,19] that showed a higher complication rate and inferior outcomes after open surgery. Recently, minimally invasive approaches to epiphrenic diverticulum, such as laparoscopic and thoracoscopic technology, have been used by some investigators [15], although the transthoracic approach through a left thoracotomy remains the standard of care for epiphrenic diverticulum for most patients and surgeons [6]. However, even for expert thoracic surgeons, the left thoracotomy approach has been associated with a high morbidity rate, with leaks present in 21% of the patients [6]. Although some studies have reported that the minimally invasive procedures have also been associated with significant morbidity [15], several more recent studies have shown that the laparoscopic approach provides more excellent results [20e22]. Our study results have also supported this finding. With laparoscopy, good visualization of the distal esophagus can be accomplished, and staplers can be used to achieve diverticulectomy and the correction of any motility disorder. No large studies have yet compared the open and minimally invasive approaches for the treatment of epiphrenic diverticulum. In comparing the procedure choice and mortality between the early and latter periods of our study for NZD, we found no significant difference in the procedure choice and mortality between the early and latter parts of the present study. As a retrospective review using a national database, our study had limitations, including missing patient characteristics that might not have been entered into the database using the appropriate ICD-9-CM diagnosis and procedure codes. The use of administrative databases for population-based
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
outcome studies has been criticized because of the incidence of erroneous coding [23]. Because the Nationwide Inpatient Sample is an administrative database, the research data extrapolated might be subject to inaccurate and variable coding of symptoms, comorbidities, and procedures among physicians, hospitals, and coders, thereby leading to an underestimation of the presented data. Also, it can often be difficult to accurately distinguish postoperative complications from pre-existing conditions or comorbidities using the primary and secondary ICD-9-CM diagnosis codes. This is because an assessment of patient comorbidities and complications is established solely for administrative claims and could potentially be biased. Furthermore, post-thoracotomy complications that might not necessarily require a corrective procedure (e.g., mild air leak or mild pleural effusion) could potentially have been underreported. In addition, a few patients might have had midesophageal diverticulum and overlapped in the ZD and NZD groups, and a few patients with ZD might have undergone additional surgical procedures similar to those used for correcting NZD to address other secondary diagnoses within the same admission (e.g. Nissen fundoplication for GERD). Some of these patients could also have overlapped in the NZD group. Where possible, we tried to use very specific treatment procedure codes unique to ZD and epiphrenic diverticulum to stratify the patients into the two groups. Despite these limitations, our study represents a much larger sample size, providing more patient characteristics and greater statistical robustness and power in determining the risk factors and predictors of morbidity and mortality in patients with acquired esophageal diverticulum.
5.
Conclusions
Using a large nationally representative database, our study results have demonstrated that female gender, black race, and the presence of comorbidities are associated with increased postoperative morbidity among patients with ZD. Among patients with NZD, black and Hispanic patients had worse postoperative morbidity than white patients, and the presence of comorbidities was associated with increased postoperative morbidity. Thoracotomy for the correction of NZD was associated with increased postoperative morbidity compared with the laparoscopic approach. Additional investigation into these findings is warranted.
references
[1] Herbella FA, Patti MG. Modern pathophysiology and treatment of esophageal diverticula. Langenbecks Arch Surg 2012;397:29. [2] Nehra D, Lord RV, DeMeester TR, et al. Physiologic basis for the treatment of epiphrenic diverticulum. Ann Surg 2002; 235:346. [3] Lerut T, Coosemans W, Decaluwe´ H, et al. Zenker’s diverticulum. MMCTS 2009;2009. http://dx.doi.org/10.1510/ mmcts.2007.002881.
125
[4] Colombo-Benkmann M, Unruh V, Krieglstein C, Senninger N. Cricopharyngeal myotomy in the treatment of Zenker’s diverticulum. J Am Coll Surg 2003;196:370. [5] do Nascimento FA, Lemme EM, Costa MM. Esophageal diverticula: pathogenesis, clinical aspects, and natural history. Dysphagia 2006;21:198. [6] Soares R, Herbella FA, Prachand VN, Ferguson MK, Patti MG. Epiphrenic diverticulum of the esophagus: from pathophysiology to treatment. J Gastrointest Surg 2010; 14:2009. [7] Healthcare Cost and Utilization Project (HCUP). Healthcare Cost and Utilization Project Nationwide Inpatient Sample (NIS), 2009. Agency for Healthcare Research and Quality, Rockville (MD), http://www.hcup-us.ahrq.gov/nisoverview. jsp. [Accessed 15.03.12]. [8] Bonafede JP, Lavertu P, Wood BG, Eliachar I. Surgical outcome in 87 patients with Zenker’s diverticulum. Laryngoscope 2009;107:720. [9] Keck T, Rozsasi A, Gru¨n PM. Surgical treatment of hypopharyngeal diverticulum (Zenker’s diverticulum). Eur Arch Otorhinolaryngol 2010;267:587. [10] Constantin A, Mates IN, Predescu D, Hoara P, Achim FI, Constantinoiu S. Principles of surgical treatment of Zenker diverticulum. J Med Life 2012;5:92. [11] Gutschow CA, Hamoir M, Rombaux P, Otte JB, Goncette L, Collard JM. Management of pharyngoesophageal (Zenker’s) diverticulum: which technique? Ann Thorac Surg 2002;74:1677. [12] Bonavina L, Bona D, Abraham M, Saino G, Abate E. Longterm results of endosurgical and open surgical approach for Zenker diverticulum. World J Gastroenterol 2007;13: 2586. [13] Blom D. Epiphrenic diverticulum. Oper Tech Gen Surg 2006; 8:170. [14] Reznik SI, Rice TW, Murthy SC, Mason DP, AppersonHansen C, Blackstone EH. Assessment of a pathophysiologydirected treatment for symptomatic epiphrenic diverticulum. Dis Esophagus 2007;20:320. [15] Varghese TK Jr, Marshall B, Chang AC, Pickens A, Lau CL, Orringer MB. Surgical treatment of epiphrenic diverticula: a 30year experience. Ann Thorac Surg 2007;84:1801. [16] Rosati R, Fumagalli U, Elmore U, de Pascale S, Massaron S, Peracchia A. Long-term results of minimally invasive surgery for symptomatic epiphrenic diverticulum. Am J Surg 2011; 201:132. [17] Thota PN, Richter JE. Dysphagia in the elderly. Interdiscip Top Gerontol 2003;32:74. [18] Kilic A, Schuchert MJ, Awais O, Luketich JD, Landreneau RJ. Surgical management of epiphrenic diverticula in the minimally invasive era. JSLS 2009;13:160. [19] Zaninotto G, Portale G, Costantini M, Zanatta L, Salvador R, Ruol A. Therapeutic strategies for epiphrenic diverticula: systematic review. World J Surg 2011;35:1447. [20] Tedesco P, Fisichella PM, Way LW, Patti MG. Cause and treatment of epiphrenic diverticula. Am J Surg 2005; 190:891. [21] Melman L, Quinlan J, Robertson B, et al. Esophageal manometric characteristics and outcomes for laparoscopic esophageal diverticulectomy, myotomy, and partial fundoplication for epiphrenic diverticula. Surg Endosc 2009; 23:1337. [22] Klaus A, Hinder RA, Swain J, Achem SR. Management of epiphrenic diverticula. J Gastrointest Surg 2003;7:906. [23] Iezzoni LI, Foley SM, Daley J, Hughes J, Fisher ES, Heeren T. Comorbidities, complications, and coding bias: does the number of diagnosis codes matter in predicting in-hospital mortality? JAMA 1992;267:2197.
Available online at www.sciencedirect.com
journal homepage: www.JournalofSurgicalResearch.com
Association for Academic Surgery
Surgical management of esophageal diverticulum: a review of the Nationwide Inpatient Sample database Michael T. Onwugbufor, BS,a Augustine C. Obirieze, MBBS, MPH,b Gezzer Ortega, MD, MPH,b Delenya Allen, MD, MPH,b Edward E. Cornwell III, MD, FACS,b and Terrence M. Fullum, MD, FACSa,b,c,* a
Howard University College of Medicine, Washington, District of Columbia Department of Surgery, Howard University College of Medicine, Washington, District of Columbia c Division of Minimally Invasive and Bariatric Surgery, Department of Surgery, Howard University College of Medicine, Washington, District of Columbia b
article info
abstract
Article history:
Background: Esophageal diverticulum is rare in the United States. The mainstay treatment
Received 5 January 2013
of symptomatic esophageal diverticulum is surgical correction. Much of the available
Received in revised form
information regarding esophageal diverticulum and its surgical management has been
6 May 2013
derived from small studies and institutional reviews. Our study objective was to investigate
Accepted 9 May 2013
the demographics, perioperative conditions, and predictors of outcomes after surgical
Available online 2 June 2013
treatment of acquired esophageal diverticulum using a nationally representative database. Methods: A retrospective review using the Nationwide Inpatient Sample database from
Keywords:
2000e2009 was performed for patients with acquired esophageal diverticulum. The patients
Esophagus
were stratified into Zenker’s diverticulum (ZD) or non-Zenker’s diverticulum (NZD)
Diverticulum
subgroups. The covariates retrieved included age, gender, ethnicity, insurance type, and
Zenker’s
Charlson comorbidity index. A multivariate analysis was performed to determine the
Epiphrenic
predictors of postoperative morbidity. Discharge-level weights were applied.
Diverticulectomy
Results: Overall, a total of 4253 patients met our inclusion criteria, 3197 (75%) with ZD and
Cricopharyngeal myotomy
1056 (25%) with NZD. In the ZD group, the mean age was 73 12.3 y, and most were men (55%) and white (67%). The mean length of stay was 5.82 8.08 d, and the mortality rate was 1.2%. The most common complication was septicemia or sepsis (2.0%). The black patients had higher odds of postoperative morbidity than the white patients (odds ratio [OR] 2.29, 95% confidence interval [CI] 1.02e5.17). The risk of overall postoperative morbidity was 52% greater for women (OR 1.52, 95% CI 1.01e2.29). An increasing Charlson comorbidity index was an independent predictor of morbidity. In the NZD group, the mean age was 69 13.9 y, and most were also men (51%) and white (63%). The mean length of stay was 8.13 10.56 d, and the mortality rate was 1.6%. The most common complication was air leak (3.1%). The black and Hispanic patients had higher odds of postoperative morbidity than the white patients (OR 1.97, 95% CI 1.05e3.72 and OR 2.37, 95% CI 1.06e5.30, respectively). An increasing Charlson comorbidity index was an independent predictor of
Presented at the Eighth Annual Academic Surgical Congress, February 5e7, 2013, New Orleans, Louisiana. * Corresponding author. Division of Minimally Invasive and Bariatric Surgery, Department of Surgery, Howard University Hospital, 2041 Georgia Avenue, Northwest, Tower Suite 4100-B, Washington, DC 20060. Tel.: þ1 202 865 1286; fax: þ1 202 865 3063. E-mail address: [email protected] (T.M. Fullum). 0022-4804/$ e see front matter ª 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2013.05.036
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
121
morbidity. Compared with laparoscopy, the risk of developing postoperative morbidity was higher with the thoracotomy procedure (OR 7.45, 95% CI 1.11e50.18). Conclusions: Using a nationally representative database, our study found that female gender, black race, and the presence of comorbidities were associated with increased postoperative morbidity among patients with ZD. Among the patients with NZD, black and Hispanic patients had worse postoperative morbidity than the white patients, and the presence of comorbidities was associated with increased postoperative morbidity. Thoracotomy for the correction of NZD was associated with increased postoperative morbidity compared with the laparoscopic approach. ª 2013 Elsevier Inc. All rights reserved.
1.
Introduction
Esophageal diverticulum is rare in the United States, with a prevalence of about 0.06%e4%, and it primarily affects the elderly [1]. A large proportion of patients with esophageal diverticula are asymptomatic and do not always require treatment [1,2]. However, the mainstay of treatment of symptomatic esophageal diverticulum has been surgical correction [1]. The most common symptoms that patients present with include dysphagia, regurgitation, halitosis, unintentional weight loss, chest pain, aspiration, and choking [1]. Zenker’s diverticulum (ZD; presenting in the proximal esophagus) is the most common form of esophageal diverticulum [3]. It has been shown to be a result of increased hypopharyngeal pressure during swallowing and dysfunction of the cricopharyngeus muscle and the upper esophageal sphincter [3,4]. Midesophageal diverticulum is frequently asymptomatic and not treated [1]. Traction has been proposed as the possible underlying mechanism causing it [5]. Finally, epiphrenic diverticulum is usually located in the distal 10 cm of the esophagus. Many investigators have suggested that a primary esophageal body motility disorder, such as diffuse esophageal spasm or a nonspecific esophageal motility disorder, underlies the development of the condition in most patients [5,6]. Some patients with epiphrenic diverticulum have also been shown to have lower esophageal sphincter abnormalities such as achalasia using manometry diagnostic studies [2,5]. Most of the information that has been collected regarding esophageal diverticulum and its surgical management has been gathered from small studies and institutional reviews. Therefore, our study evaluated the demographics, perioperative conditions, and surgical outcomes of acquired esophageal diverticulum in a larger population using a national database and determined the independent risk factors and predictors of the postoperative outcomes.
2.
Methods
We conducted a retrospective study using the Nationwide Inpatient Sample database from 2000e2009. The Nationwide Inpatient Sample is a part of the Healthcare Cost and Utilization Project, sponsored by the Agency for Healthcare Research and Quality. It is the largest all-payer inpatient care database publicly available in the United States. The data are from a 20%, stratified sample of discharges from US community hospitals, comprising 95% of the US population [7]. Using the
“International Classification of Diseases, Ninth Edition, Clinical Modifications” (ICD-9-CM) diagnosis codes, we identified patients 18 y with a primary diagnosis of acquired esophageal diverticulum (ICD-9-CM code 530.6). The study cohort was stratified into two subgroups according to the procedures (identified by ICD-9-CM procedure codes) the patient had undergone during the same admission to surgically address the esophageal diverticulum: Zenker’s diverticulum (ZD) and non-Zenker’s diverticulum (NZD). The ICD-9-CM procedures codes used to identify patients with ZD included those for cricopharyngeal myotomy, pharyngoesophageal diverticulectomy, and hypopharyngeal esophageal diverticulectomy. The patients with NZD were identified using the ICD-9-CM procedure codes for thoracoscopic diverticulectomy, laparoscopic diverticulectomy, esophagomyotomy, thoracotomy, and fundoplication. Associated preoperative symptoms and conditions, including dysphagia, regurgitation, halitosis, choking, coughing, weight loss, chest pain, esophageal cancer, and aspiration pneumonia were identified using the ICD-9-CM diagnosis codes (Table 1). We also recorded the postoperative complications, including air leaks, septicemia and/or sepsis, mediastinitis, mediastinal empyema with or without fistula, mediastinal abscess, postoperative wound infection, and postoperative hemorrhage using the ICD-9-CM diagnosis codes. These postoperative complications were selected in accordance with the most commonly reported and discussed postoperative complications in reports from previous institutional studies and reviews. Within our patient population, we identified as many ICD-9-CM codes matching these published complications as possible. The covariates retrieved include age, gender, race or ethnicity, Charlson comorbidity index, median household income quartile for the patient’s residential postal code, primary payer, hospital type (teaching versus nonteaching), hospital bed size (small, medium, or large), hospital location (rural versus urban), hospital ownership or control (government, private, or government, private collapsed, unclassified). For the study, our outcome of interest was the overall complication rate defined as having at least one of the postoperative complications previously listed. Univariate analyses on the covariates were conducted separately within the ZD and NZD subgroups. Multivariate logistic regression analyses and postanalysis Wald tests, performed separately within each of the two subgroups studied were used to investigate the predictors of postoperative morbidity, adjusting for patient age, gender, race or ethnicity, primary payer, median income in the residential postal code, Charlson comorbidity
122
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 1 e ICD-9-CM diagnosis and procedure codes used to identify selected pre- and postoperative conditions and procedures. Diagnosis and procedure Diagnosis Dysphagia Chest pain Weight loss Halitosis or choking Cough Aspiration pneumonia Esophageal cancer Mediastinitis Mediastinal abscess Mediastinal empyema with fistula Mediastinal empyema without fistula Air leaks Septicemia or sepsis Postoperative wound infection Postoperative hemorrhage Procedures Cricopharyngeal myotomy Pharyngoesophageal diverticulectomy Esophageal diverticulectomy Local excision of esophagus Endoscopic excision of esophagus Esophagomyotomy Thoracotomy Thoracoscopy Laparoscopy Fundoplication
ICD-9-CM codes 787.24 786.5 783.21 784.99 786.2 507.0 150.0e150.9 519.2 513.1 510.0 510.9 512.1e512.2 038.x, 995.91, 995.92 998.5, 998.59 998.11e998.12 29.31 29.32 42.31 42.32 42.33 42.7 34.0e34.02 34.21 54.21 44.66e44.67
index, discharge year, hospital region, bed size, location, teaching status, and ownership or control. The postanalysis Wald test was used to investigate the likelihood that a covariate (with all categories taken together) was an independent predictor of the development of postoperative morbidity. Discharge-level weights were applied in all analyses to obtain national estimates. The Hosmer-Lemeshow goodness-of-fit test was used to ascertain that the models fit. Statistical analyses were done using STATA/MP, version 11.0 (StataCorp, College Station, TX). Statistical significance was defined as P < 0.05.
3.
Results
Overall, a total of 4253 patients met our inclusion criteria, 3197 (75%) with ZD and 1056 (25%) with NZD. Among the patients with the ZD, the mean age was 73 12.3 y, with most patients aged 65 y (78%). Most of the patients with ZD were men (55%), white (67%), within the highest income quartile (35%), and underwent their procedures at a teaching hospital (57%). The mean length of stay in this group was 5.82 8.08 d, and the mortality rate was 1.22%. The most common preoperative symptoms were dysphagia (23.5%), aspiration pneumonia (7.4%), and weight loss (1.0%). The most common complication was septicemia or sepsis (1.97%). In the NZD group, the mean age was 69 13.9 y, with most patients aged 65 y (66%). Most were men (51%), white (63%), within the highest income quartile (33%), and had undergone their procedures at a teaching hospital (64%). The mean length
of stay was 8.13 10.56 d. The mortality rate was 1.61%. The most common preoperative symptoms were dysphagia (17.9%), aspiration pneumonia (5.7%), weight loss (1.2%), and chest pain (0.9%), and the most common complication was air leak (3.1%; Table 2). From 2000e2004, among the patients with NZD, 1.0%, 2.1%, 6.0%, 13.1%, and 77.9% had undergone thoracotomy, thoracoscopy, laparoscopy, fundoplication, and esophagomyotomy, respectively. From 2005e2009, 0.6%, 2.2%, 4.5%, 17.4%, and 75.3% had undergone thoracotomy, thoracoscopy, laparoscopy, fundoplication, and esophagomyotomy, respectively. On multivariate analysis, among the patients with ZD, black patients had more than twice the odds of morbidity compared with the white patients (odds ratio [OR] 2.29, 95% confidence interval [CI] 1.02e5.17). In addition, the risk of morbidity was 52% greater for women (OR 1.52, 95% CI 1.01e2.29). An increasing Charlson comorbidity index was an independent predictor of postoperative morbidity (Wald test P < 0.001; Table 3). Among the patients with NZD, black and Hispanic patients had higher odds of postoperative morbidity than white patients (OR 1.97, 95% CI 1.05e3.72 and OR 2.37, 95% CI 1.06e5.30, respectively). The independent predictors of postoperative morbidity were race or ethnicity (Wald test P ¼ 0.04), procedure type (Wald test P ¼ 0.01), and an increasing Charlson comorbidity index (Wald test P < 0.001). Finally, the risk of postoperative morbidity was greater after thoracotomy than after laparoscopy (OR 7.45, 95% CI 1.11e50.18; Table 4).
4.
Discussion
The main surgical procedures used for the treatment of ZD have been cricopharyngeal myotomy, diverticulectomy, and diverticulopexy, all of which can be used in combination with each other through a transcervical approach or endoscopically [1,3,4,8e10]. A retrospective review of 87 patients with ZD at a tertiary care institution [8] and another study of 11 patients with ZD by Constantin et al. [10] showed that cricopharyngeal myotomy is necessary for the treatment of ZD but might not be sufficient because it resulted in a poorer outcome in their studies compared with cricopharyngeal myotomy plus diverticulectomy or diverticulopexy. Regarding the invasive surgical approach for the treatment of ZD, two studies, by Lerut et al. [3] and Gutschow et al. [11], advocated the open surgical approach because of its superior outcome compared with the endoscopic approach, although a study of 297 patients with ZD by Bonavina et al. [12] showed no difference in outcomes for both approaches. Epiphrenic diverticulum can be surgically repaired using a transthoracic or laparoscopic approach and has almost always included the incorporation of a fundoplication in addition to diverticulectomy and/or esophagomyotomy [1,13]. Reznik et al. [14] assessed the pathophysiology-directed operative strategy of excision (diverticulectomy), repair of esophageal wall, and relief of functional and mechanical obstruction in 44 patients with epiphrenic diverticulum and showed it to be safe, with positive outcomes in most patients. A retrospective review by Varghese et al. [15] of 35 patients with epiphrenic diverticulum reported excellent long-term results using the
123
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 2 e Patient characteristics (2000e2009 Nationwide Inpatient Sample database; n [ 4253). Characteristic Mean age (y) Age group (n) <45 45e64 65e84 85 Gender (n) Male Female Unknown Race or ethnicity (n) White Black Hispanic Other or unknown Preoperative symptoms and conditions Dysphagia Chest pain Weight loss Cough Aspiration pneumonia Esophageal cancer Income quartile (n) Lowest Second Third Highest Unknown Primary payer (n) Private insurance Medicare Self-pay, no charge, or other Unknown Charlson comorbidity index (n) 0 1 2 Mean length of stay (d) Mortality (%) Postoperative complications (n) Mediastinitis Mediastinal abscess Mediastinal empyema Mediastinal empyema (with or without fistula) Air leaks Septicemia or postoperative sepsis Postoperative wound infection Postoperative hemorrhage Hospital teaching status (n) Nonteaching Teaching Hospital ownership or control (n) Government, nonfederal Private Government, private collapsed, uncategorized
ZD (n ¼ 3197)
Non-ZD (n ¼ 1056)
73.1 12.3
69.1 13.9
72 641 1948 536
(2.3) (20.1) (60.9) (16.8)
59 298 565 134
1747 (54.6) 1431 (44.8) 19 (0.6)
(5.6) (28.2) (53.5) (12.7)
539 (51.0) 516 (48.9) 1 (0.1)
2148 90 62 897 (n) 752 31 31 13 236 4
(67.2) (2.8) (1.9) (28.1) (23.5) (1.0) (1.0) (0.4) (7.4) (0.1)
189 9 13 2 60 3
(17.9) (0.9) (1.2) (0.2) (5.7) (0.3)
417 526 844 1122 288
(13.0) (16.5) (26.4) (35.1) (9.01)
177 174 254 349 102
(16.8) (16.5) (24.1) (33.1) (9.7)
761 2324 54 58
665 64 36 291
(23.8) (72.7) (1.7) (1.8)
1957 (61.2) 828 (25.9) 412 (12.9) 5.82 8.08 1.2
323 674 25 34
(63.0) (6.1) (3.4) (27.6)
(30.6) (63.8) (2.4) (3.2)
643 (60.9) 292 (27.7) 121 (11.5) 8.13 10.56 1.6
11 1 3 17
(0.3) (0.03) (0.1) (0.5)
5 4 2 14
(0.5) (0.4) (0.2) (1.3)
19 63 36 44
(0.6) (2.0) (1.13) (1.38)
33 30 20 13
(3.1) (2.8) (1.89) (1.23)
1373 (43.0) 1821 (57.0)
376 (35.6) 679 (64.3)
203 (6.5) 1205 (37.7) 1789 (56.0)
64 (6.1) 400 (37.9) 592 (56.1)
Data presented as mean standard deviation or n (%).
transthoracic approach. Likewise, a similar review by Rosati et al. [16] of 20 patients with epiphrenic diverticulum who had undergone treatment with the laparoscopic approach
Table 3 e Adjusted ORs and 95% CIs for covariates used in multivariate analysis and P values of Wald test among patients with ZD. Variable Age group (y) <44 45e64 65e84 >85 Gender Male Female Race or ethnicity White Black Hispanic Other Charlson comorbidity index 0 1 2 Income quartile Lowest Second Third Highest Primary payer Private insurance Medicare Self-pay, no charge Hospital bed size Small Medium Large Hospital teaching status Nonteaching Teaching Hospital location Rural Urban Hospital control Government, nonfederal Private Government or private
OR
95% CI
1.65 0.60 0.73 Reference
0.46e5.97 0.23e1.52 0.44e1.22 Reference
Reference 1.52
Reference 1.01e2.29
Reference 2.29 0.66 0.91
Reference 1.02e5.17 0.08e5.31 0.56e1.49
Reference 2.02 2.61
Reference 1.29e3.17 1.53e4.44
Reference 0.76 0.86 0.68
Reference 0.36e1.61 0.45e1.66 0.34e1.38
Reference 1.20 1.37
Reference 0.52e2.77 0.33e5.69
Reference 2.84 3.09
Reference 0.99e8.17 1.13e8.42
Reference 0.83
Reference 0.47e1.44
Reference 1.77
Reference 0.66e4.76
Reference 1.19 2.08
Reference 0.45e3.16 0.68e6.36
P value (Wald test) 0.18
0.05
0.19
<0.001
0.80
0.89
0.09
0.44
0.09
0.25
reported satisfactory long-term outcomes. These studies have provided substantial comparative data regarding the various surgical techniques for the treatment of esophageal diverticulum. In our study, a review of a large national database, we classified the ZD group according to the unique primary types of procedures used for its surgical treatment. The NZD group was also classified according to the unique primary types of procedures used for the surgical treatment of epiphrenic diverticulum. Hence, most of the patients in the NZD group were assumed to mostly have had epiphrenic diverticulum. Similar to numerous studies [1e6,8e10], the most common symptom experienced by most of the patients in the present study was dysphagia. This symptom has been suggested to be the result of the abnormal swallowing mechanism and motility disorder associated with this pathologic entity [3e6]. Similar to other studies of ZD [4,9] and epiphrenic diverticulum [15],
124
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
Table 4 e Adjusted ORs and 95% CIs for covariates used in multivariate analysis and P values of Wald test among patients with non-ZD. Variable Age group (y) <44 45e64 65e84 >85 Gender Male Female Race or ethnicity White Black Hispanic Other Procedure Laparoscopy Thoracoscopy Thoracotomy Charlson comorbidity index 0 1 2 Income quartile Lowest Second Third Highest Primary payer Private insurance Medicare Self-pay, no charge Hospital bed size Small Medium Large Hospital teaching status Nonteaching Teaching Hospital location Rural Urban Hospital control Government, nonfederal Private Government or private
Or
95% CI
1.16 0.72 0.76 Reference
0.42e3.19 0.37e1.41 0.51e1.14 Reference
Reference 1.23
Reference 0.90e1.67
Reference 1.97 2.37 1.10
Reference 1.05e3.72 1.06e5.30 0.75e1.62
Reference 3.42 7.45
Reference 0.77e15.09 1.11e50.18
Reference 1.60 2.30
Reference 1.13e2.28 1.49e3.54
Reference 1.02 0.87 0.88
Reference 0.59e1.77 0.52e1.47 0.53e1.47
Reference 1.04 1.29
Reference 0.60e1.80 0.44e3.78
Reference 1.21 1.46
Reference 0.65e2.27 0.82e2.60
Reference 0.91
Reference 0.58e1.41
Reference 1.76
Reference 0.86e3.58
Reference 1.88 1.75
Reference 0.79e4.46 0.66e4.64
P value of Wald test 0.42
0.27
0.04
0.01
<0.001
0.95
0.97
0.33
0.94
0.97
0.36
esophageal diverticulum mostly occurred in patients >65 y in the present study. According to a study by Thota and Richter [17], a possible explanation for the susceptibility of the older population is that morphologic changes in the pharynx and esophagus contribute to age-related changes in esophageal physiology [17]. Also, from our results, this pathologic entity appears to be more common within the white patient population. This is concordant with another study of 35 patients with epiphrenic diverticulum, in which 34 of the patients were white [15]. A suggestion has also been made of a genetic mechanism in the differences in the incidence of ZD related to race or ethnicity [10]. The length of stay was longer for the NZD group than for the ZD group. Moreover, leaks had also been reported to be the most common postoperative complication of
epiphrenic diverticulum in previous studies [6], just as was found in the present study. However, in the context of our study, the air leaks reflected the post-thoracotomy air leaks and iatrogenic pneumothorax secondary to the intraoperative manipulation of the lung parenchyma with traumatic laparoscopic or thoracoscopic instruments during access and repair of the thoracic esophagus. We believe that infectious mediastinal complications such as mediastinitis, mediastinal abscess, mediastinal empyema (with fistula), and mediastinal empyema (without fistula) were therefore consequences of suture or staple line leaks. In the results of the multivariate analysis of patients with ZD, the black patients had higher odds of developing postoperative complications than the white patients. Regarding gender, the women in the ZD group were more likely to develop postoperative complications than were the men. Furthermore, an increasing incidence of comorbidities was an independent predictor of postoperative morbidity. Because the mean age in the ZD group was >70 y, it could be inferred that most patients with ZD will present with significant comorbidities, increasing their overall risk of postoperative complications. In the results of the multivariate analysis of patients with NZD (epiphrenic), the black and Hispanic patients were more likely to develop postoperative complications than were the white patients. Race or ethnicity, procedure type, and increasing comorbidities were all independent predictors of postoperative morbidity. Furthermore, the increased odds of developing postoperative complications after thoracotomy for the treatment of NZD (epiphrenic diverticulum) compared with using a laparoscopic procedure is consistent with three recent reviews [1,6,18,19] that showed a higher complication rate and inferior outcomes after open surgery. Recently, minimally invasive approaches to epiphrenic diverticulum, such as laparoscopic and thoracoscopic technology, have been used by some investigators [15], although the transthoracic approach through a left thoracotomy remains the standard of care for epiphrenic diverticulum for most patients and surgeons [6]. However, even for expert thoracic surgeons, the left thoracotomy approach has been associated with a high morbidity rate, with leaks present in 21% of the patients [6]. Although some studies have reported that the minimally invasive procedures have also been associated with significant morbidity [15], several more recent studies have shown that the laparoscopic approach provides more excellent results [20e22]. Our study results have also supported this finding. With laparoscopy, good visualization of the distal esophagus can be accomplished, and staplers can be used to achieve diverticulectomy and the correction of any motility disorder. No large studies have yet compared the open and minimally invasive approaches for the treatment of epiphrenic diverticulum. In comparing the procedure choice and mortality between the early and latter periods of our study for NZD, we found no significant difference in the procedure choice and mortality between the early and latter parts of the present study. As a retrospective review using a national database, our study had limitations, including missing patient characteristics that might not have been entered into the database using the appropriate ICD-9-CM diagnosis and procedure codes. The use of administrative databases for population-based
j o u r n a l o f s u r g i c a l r e s e a r c h 1 8 4 ( 2 0 1 3 ) 1 2 0 e1 2 5
outcome studies has been criticized because of the incidence of erroneous coding [23]. Because the Nationwide Inpatient Sample is an administrative database, the research data extrapolated might be subject to inaccurate and variable coding of symptoms, comorbidities, and procedures among physicians, hospitals, and coders, thereby leading to an underestimation of the presented data. Also, it can often be difficult to accurately distinguish postoperative complications from pre-existing conditions or comorbidities using the primary and secondary ICD-9-CM diagnosis codes. This is because an assessment of patient comorbidities and complications is established solely for administrative claims and could potentially be biased. Furthermore, post-thoracotomy complications that might not necessarily require a corrective procedure (e.g., mild air leak or mild pleural effusion) could potentially have been underreported. In addition, a few patients might have had midesophageal diverticulum and overlapped in the ZD and NZD groups, and a few patients with ZD might have undergone additional surgical procedures similar to those used for correcting NZD to address other secondary diagnoses within the same admission (e.g. Nissen fundoplication for GERD). Some of these patients could also have overlapped in the NZD group. Where possible, we tried to use very specific treatment procedure codes unique to ZD and epiphrenic diverticulum to stratify the patients into the two groups. Despite these limitations, our study represents a much larger sample size, providing more patient characteristics and greater statistical robustness and power in determining the risk factors and predictors of morbidity and mortality in patients with acquired esophageal diverticulum.
5.
Conclusions
Using a large nationally representative database, our study results have demonstrated that female gender, black race, and the presence of comorbidities are associated with increased postoperative morbidity among patients with ZD. Among patients with NZD, black and Hispanic patients had worse postoperative morbidity than white patients, and the presence of comorbidities was associated with increased postoperative morbidity. Thoracotomy for the correction of NZD was associated with increased postoperative morbidity compared with the laparoscopic approach. Additional investigation into these findings is warranted.
references
[1] Herbella FA, Patti MG. Modern pathophysiology and treatment of esophageal diverticula. Langenbecks Arch Surg 2012;397:29. [2] Nehra D, Lord RV, DeMeester TR, et al. Physiologic basis for the treatment of epiphrenic diverticulum. Ann Surg 2002; 235:346. [3] Lerut T, Coosemans W, Decaluwe´ H, et al. Zenker’s diverticulum. MMCTS 2009;2009. http://dx.doi.org/10.1510/ mmcts.2007.002881.
125
[4] Colombo-Benkmann M, Unruh V, Krieglstein C, Senninger N. Cricopharyngeal myotomy in the treatment of Zenker’s diverticulum. J Am Coll Surg 2003;196:370. [5] do Nascimento FA, Lemme EM, Costa MM. Esophageal diverticula: pathogenesis, clinical aspects, and natural history. Dysphagia 2006;21:198. [6] Soares R, Herbella FA, Prachand VN, Ferguson MK, Patti MG. Epiphrenic diverticulum of the esophagus: from pathophysiology to treatment. J Gastrointest Surg 2010; 14:2009. [7] Healthcare Cost and Utilization Project (HCUP). Healthcare Cost and Utilization Project Nationwide Inpatient Sample (NIS), 2009. Agency for Healthcare Research and Quality, Rockville (MD), http://www.hcup-us.ahrq.gov/nisoverview. jsp. [Accessed 15.03.12]. [8] Bonafede JP, Lavertu P, Wood BG, Eliachar I. Surgical outcome in 87 patients with Zenker’s diverticulum. Laryngoscope 2009;107:720. [9] Keck T, Rozsasi A, Gru¨n PM. Surgical treatment of hypopharyngeal diverticulum (Zenker’s diverticulum). Eur Arch Otorhinolaryngol 2010;267:587. [10] Constantin A, Mates IN, Predescu D, Hoara P, Achim FI, Constantinoiu S. Principles of surgical treatment of Zenker diverticulum. J Med Life 2012;5:92. [11] Gutschow CA, Hamoir M, Rombaux P, Otte JB, Goncette L, Collard JM. Management of pharyngoesophageal (Zenker’s) diverticulum: which technique? Ann Thorac Surg 2002;74:1677. [12] Bonavina L, Bona D, Abraham M, Saino G, Abate E. Longterm results of endosurgical and open surgical approach for Zenker diverticulum. World J Gastroenterol 2007;13: 2586. [13] Blom D. Epiphrenic diverticulum. Oper Tech Gen Surg 2006; 8:170. [14] Reznik SI, Rice TW, Murthy SC, Mason DP, AppersonHansen C, Blackstone EH. Assessment of a pathophysiologydirected treatment for symptomatic epiphrenic diverticulum. Dis Esophagus 2007;20:320. [15] Varghese TK Jr, Marshall B, Chang AC, Pickens A, Lau CL, Orringer MB. Surgical treatment of epiphrenic diverticula: a 30year experience. Ann Thorac Surg 2007;84:1801. [16] Rosati R, Fumagalli U, Elmore U, de Pascale S, Massaron S, Peracchia A. Long-term results of minimally invasive surgery for symptomatic epiphrenic diverticulum. Am J Surg 2011; 201:132. [17] Thota PN, Richter JE. Dysphagia in the elderly. Interdiscip Top Gerontol 2003;32:74. [18] Kilic A, Schuchert MJ, Awais O, Luketich JD, Landreneau RJ. Surgical management of epiphrenic diverticula in the minimally invasive era. JSLS 2009;13:160. [19] Zaninotto G, Portale G, Costantini M, Zanatta L, Salvador R, Ruol A. Therapeutic strategies for epiphrenic diverticula: systematic review. World J Surg 2011;35:1447. [20] Tedesco P, Fisichella PM, Way LW, Patti MG. Cause and treatment of epiphrenic diverticula. Am J Surg 2005; 190:891. [21] Melman L, Quinlan J, Robertson B, et al. Esophageal manometric characteristics and outcomes for laparoscopic esophageal diverticulectomy, myotomy, and partial fundoplication for epiphrenic diverticula. Surg Endosc 2009; 23:1337. [22] Klaus A, Hinder RA, Swain J, Achem SR. Management of epiphrenic diverticula. J Gastrointest Surg 2003;7:906. [23] Iezzoni LI, Foley SM, Daley J, Hughes J, Fisher ES, Heeren T. Comorbidities, complications, and coding bias: does the number of diagnosis codes matter in predicting in-hospital mortality? JAMA 1992;267:2197.