This figure shows 50% of recurrences occurred with the first year of operation, 75% within the first 5 years, and 25% after 5 years.
This figure shows a 27% mortality rate at a mean of 5.1 years after index repair.
eTable. Complete Listing of Patient, Operative, and Postoperative Factors
Customize your JAMA Network experience by selecting one or more topics from the list below.
Shankar DA, Itani KMF, O’Brien WJ, Sanchez VM. Factors Associated With Long-term Outcomes of Umbilical Hernia Repair. JAMA Surg. 2017;152(5):461–466. doi:10.1001/jamasurg.2016.5052
What factors lead to umbilical hernia recurrence?
This cohort study of 332 military veteran patients who underwent umbilical hernia repair found that ascites, liver disease, diabetes, obesity, and primary suture repair were significantly associated with increased rates of umbilical hernia recurrence. There was no significant difference in complication rates between mesh repair and primary suture repair.
Mesh repair should be considered for all patients undergoing umbilical hernia repair, given its lower recurrence rates and complication rates that are similar to that of primary suture repair.
Umbilical hernia repair is one of the most commonly performed general surgical procedures. However, there is little consensus about the factors that lead to umbilical hernia recurrence.
To better understand the factors associated with long-term umbilical hernia recurrence.
Design, Setting, and Participants
A retrospective cohort of 332 military veteran patients who underwent umbilical hernia repair was studied between January 1, 1998, and December 31, 2008, at the VA Boston Healthcare System. Recurrence and mortality outcomes were tracked from that period until June 1, 2014. Data were collected on patient characteristics, operative, and postoperative factors and univariate and multivariable analyses were used to assess which factors were significantly associated with umbilical hernia recurrence and mortality. All patients with primary umbilical hernia repair, with or without a concurrent unrelated procedure, were included in the study. Patients excluded were those who underwent umbilical hernia repair as a part of another major planned procedure with abdominal incisions. Data were collected from June 1, 2014, to November 1, 2015. Statistical analysis was performed from November 2, 2015, to April 1, 2016.
Main Outcomes and Measures
The primary study outcomes were umbilical hernia recurrence and death.
Of the 332 patients in this study, 321 (96.7%) were male, mean age was 58.4 years, and mean (SD) time of follow-up was 8.5 (4.1) years. The hernia recurrence rate was 6.0% (n = 20) at a mean 3.1 years after index repair (median, 1.0-year; range, 0.33-13 years). The primary suture repair recurrence rate was 9.8% (16 of 163 patients), and the mesh repair recurrence rate was 2.4% (4 of 169 patients). On univariate analysis, ascites (P = .02), liver disease (P = .02), diabetes (P = .04), and primary suture (nonmesh) repairs (P = .04) were significantly associated with increased recurrence rates. Patients who had a history of hernias (125 [39%]) were less likely to have umbilical hernia recurrences (χ21 = 4.65, P = .03). On multivariable regression analysis, obesity and ascites were associated with significantly increased odds ratios of recurrence of 3.3 (95% CI, 1.0-10.1) and 8.0 (95% CI, 1.8-34.4), respectively. Mesh repair was seen to decrease recurrence with odds of 0.28 (95% CI, 0.08-0.95). There was no significant difference in complication rates between mesh repair and primary suture repair. The survival rate was 73% (n = 242) at the end of the study. Factors associated with mortality were older age, smoking, liver disease, ascites, emergency or semiurgent repair, and need for intraoperative bowel resection.
Conclusions and Relevance
Ascites, liver disease, diabetes, obesity, and primary suture repair without mesh are associated with increased umbilical hernia recurrence rates. Elective umbilical hernia repair with mesh should be considered in patients with multiple comorbidities given that the use of mesh offers protection from recurrence without major morbidity.
Umbilical hernias account for approximately 6% to 14% of all abdominal wall hernias in adults, and almost 90% of adult umbilical hernias are acquired.1,2 Despite umbilical hernia being common, there are no set surgical guidelines for its repair and there is no consensus on the best type of repair.2,3 Recurrence rates range from 1% to 43%,3-6 but the literature offers little consensus on factors that affect recurrence. Asolati et al1 found that type 2 diabetes, hyperlipidemia, and human immunodeficiency virus–positive status may be a factor in increased recurrence rates. Obesity (body mass index [BMI; calculated as weight in kilograms divided by height in meters squared]>30) has also been associated with higher rates of recurrence.7,8 A recent study9 found that patients with another known hernia at the time of repair were at increased risk of recurrence. In addition, studies have shown that hernia defects larger than 2 cm have increased rates of recurrence and that mesh repairs may decrease rates of recurrence.7,10 Still, few studies to date have assessed a wide range of patient and operative factors to determine predictors of umbilical hernia recurrence. In addition, to our knowledge, no study to date has had mean follow-up of patients with umbilical hernia repairs beyond 7 years. The primary aim of this study was to assess the long-term risk of umbilical hernia recurrences and the role that various patient characteristics and surgical techniques play.
To better understand the factors associated with umbilical hernia recurrence, we reviewed a 10-year cohort of veteran patients who underwent umbilical hernia repair at the VA (Veterans Affairs) Boston Healthcare System, Boston, Massachusetts.
We studied a retrospective cohort of 332 patients who underwent umbilical hernia repair at the VA Boston Healthcare System between January 1, 1998, and December 31, 2008, and we tracked recurrence and mortality outcomes from that period until June 1, 2014. The study was approved by the institutional review board of the VA Boston Healthcare System and met the criteria for patient consent exclusion. Data were collected from June 1, 2014, to November 1, 2015. Statistical analysis was performed from November 2, 2015, to April 1, 2016.
Using VistA Computerized Patient Record System (CPRS), version 18.104.22.168 (US Department of Veterans Affairs), we identified patients on the basis of Current Procedural Terminology codes for various umbilical hernia repairs. Patients included were those who underwent primary umbilical hernia repair (ie, open, laparoscopic, recurrent, emergent, or elective), 2 concurrent hernia repairs (ie, for inguinal hernia and umbilical hernia), and umbilical hernia repair with a concurrent unrelated procedure (eg, umbilical hernia repair and breast biopsy). Patients excluded were those who underwent umbilical hernia repair as a part of another major planned procedure with abdominal incisions (eg, umbilical hernia repair during a colectomy). The primary endpoints were recurrence (reoperation or clinical) and death. Patients were noted to have a hernia recurrence if it was reported during a physical examination and documented in the CPRS, if they had an operation for recurrence, or if it was indicated in imaging findings. The follow-up period was from the time of index repair to the last available record as of June 1, 2014, or death.
We collected patient data through medical record review of operative notes, general surgery notes, primary care notes, radiology reports, and other medical-surgical specialty notes. We studied 51 patient comorbidities, operative factors, and postoperative complications thought to be clinically significant in umbilical hernia recurrence. The variables and criteria used for each patient, operative, and postoperative factors are shown in the eTable in the Supplement.
We used logistic regression models to estimate the odds of recurrence and mortality as a function of demographic and clinical factors. The models included a time trend to control for improvements in medical knowledge and technology during the study period. We initially examined univariate relationships of all available variables and the 2 outcomes. Variables with a χ2 or Fisher exact test of P < .05 were eligible for inclusion in the multivariable models. We further narrowed the candidate variables to avoid multicollinearity (eg, liver disease and alcohol abuse). The final covariates in the recurrence model included age older than 60 years, chronic obstructive pulmonary disease, BMI greater than 30, ascites, mesh repair, and emergency or semiurgent operation. The mortality model included age as a continuous variable, smoking, liver disease, BMI greater than 35, large defect (≥2 cm), and emergency or semiurgent operation. We measured model discrimination using the C statistic. We also ran a sensitivity analysis using the same model specification, but with the redefined outcomes of 7-year mortality and 7-year recurrence. We performed all analyses in SAS, version 9.3 (SAS Institute Inc).
There were 332 consecutive patients who underwent umbilical hernia repair between January 1, 1998, and December 31, 2008, and who fit our inclusion criteria. In our patient cohort, most were male (321 [96.7%]); the mean age was 58.4 years; the mean (SD) time of follow-up was 8.5 (4.1) years; and the mean BMI was 30.9 with 178 (54%) patients identified as obese, as defined by a BMI of 30 or greater. Among these patients, 242 (73%) had a smoking history and 70 (21%) had liver disease, as defined by cirrhosis, hepatitis, or ascites. In addition, 163 (49%) underwent primary suture repairs and 169 (51%) underwent mesh repairs. There was no significant difference in the underlying medical conditions (eg, diabetes, chronic obstructive pulmonary disease, ascites) between patients who underwent primary suture repair and those who underwent mesh repair. Of the hernias repaired, 292 (88%) were elective, 30 (9%) were an emergency, and 13 (4%) were semiurgent; 4 repairs (1.2%) were laparoscopic, and the remainder were open repairs. Eighty (24%) patients had concomitant operations at the time of their umbilical hernia repair, and 60 (74%) of these procedures were other hernia operations.
The recurrence rate was 6.0% (n = 20) at a mean 3.1 years after index repair (median, 1.0-year; range, 0.33-13 years). The primary suture repair recurrence rate was 9.8% (16 of 163 patients), and the mesh repair recurrence rate was 2.4% (4 of 169 patients). Postoperatively, 10 (50%) of the recurrences occurred within the first year, 15 (75%) occurred within the first 5 years, and 5 (25%) occurred after 5 years (Figure 1). Sixteen patients (80%) who had a recurrence underwent primary suture repairs, and 4 (20%) underwent mesh repairs. Within the primary suture repairs, 4 of the fascial defects were repaired with absorbable sutures, 11 were repaired with non–absorbable sutures, and 1 was repaired with unknown sutures. Within the mesh repairs, 3 were overlays and 1 was a combination underlay and overlay. Mesh location and type of suture could not be analyzed because of our small sample size.
On univariate analysis, ascites (Fisher exact test, P = .02), liver disease (Fisher exact test, P = .02), diabetes (χ21 = 4.17, P = .04), and primary suture (nonmesh) repairs (χ21 = 8.13, P = .04) were significantly associated with increased rates of recurrence (Table 1). Patients who had a history of hernias (125 [39%]) were less likely to have umbilical hernia recurrences (P = .03). Seventy-six (61%) patients with prior hernias received mesh as opposed to 87 (44%) patients without a history of hernia (P = .003). No postoperative complications significantly affected the rates of recurrence. Defect size did not affect the rates of recurrence; however, defect size was only recorded in 174 (54%) patients. On multivariable regression analysis, obesity and ascites were associated with significantly increased odds ratios (ORs) of recurrence—3.3 (95% CI, 1.0-10.1) and 8.0 (95% CI, 1.8-34.4), respectively (Table 2). Mesh repair was seen to decrease recurrence by 3.6 times with odds of recurrence of 0.28 (95% CI, 0.08-0.95).
Sixty-one (18%) patients experienced at least 1 complication within 30 days of repair. The most common complication was seroma (31 [9.6%]), followed by surgical-site infection (22 [6.9%]) and hematomas (8 [2.4%]). Of note, 3 of the 4 patients (75%) who needed bowel resection at the time of index repair developed a surgical-site infection. There were 2 cases of mesh infection and 3 cases of ascites leaks. Thirty-four patients (20%) who had a mesh repair had 1 or more complications, which was slightly higher, but not significantly, than the 27 patients (16%) with primary suture repair who had a complication after the procedure (P = .35).
The 30-day mortality rate was 0.3% (n = 1). This 1 death occurred on postoperative day 21 for unclear reasons.
The total survival rate was 73% (n = 242) during the period of follow-up, as assessed by our medical record review from June 1, 2014, to November 1, 2015 (Figure 2). Ninety patients (27%) died at a mean of 5.1 years postoperatively (median, 5 years), despite the mean age of 58.4 years of our patient cohort. Among the deaths, 43 (48%) were of unknown causes, 18 (20%) were cancer related, 12 (13%) were secondary to liver disease, 6 (6.7%) were secondary to heart disease, 4 (4.4%) were secondary to renal disease, and 3 (3.3%) were secondary to sepsis. On univariate analysis (Table 1), older age (t = 6.19, P < .001), smoking (χ21 = 5.44, P = .02), liver disease (χ21 = 13.00, P < .001), ascites (χ21 = 49.64, P < .001), emergency or semiurgent repair (χ21 = 11.80, P < .001), and need for intraoperative bowel resection (Fisher exact test, P = .005) were significantly associated with increased long-term mortality rates. Twenty deaths occurred more than 30 days after emergency or semiurgent repair. These all occurred more than 30 days postoperatively. Twelve (60%) of these deaths were secondary to unknown causes and 25 (5%) were secondary to liver disease. On multivariable regression analysis, age (OR, 1.09; 95% CI, 1.06-1.13), history of smoking (OR, 2.27; 95% CI, 1.12-4.56), liver disease (OR, 5.60; 95% CI, 2.69-12.10), and emergency or semiurgent repair (OR, 2.18; 95% CI, 1.02-4.67) significantly increased the odds of long-term mortality (Table 3). Defects of 2 cm or greater had a protective effect on mortality with a 0.44 OR of death (95% CI, 0.23-0.86).
Umbilical hernias are a common problem encountered by general surgeons. Despite umbilical hernias being common, little has changed in their indications and methods of repair for many years. Several studies have attempted to analyze the factors associated with recurrence with short-term follow-up. In contrast, our study focused on surgical procedures during a 10-year period, and our mean follow-up time of 8.5 years is longer than that of any comparable study. Because of the small patient migration outside of the VA system and the ability to access the CPRS and Medicare mortality data, we have been able to capture outcomes that are difficult to study in other health care settings. Thus, our study sought to seek out the factors that might be associated with patients who are more likely to develop hernia recurrences.
Our finding of a 6.0% umbilical hernia recurrence rate is comparable to that of other studies. The primary suture repair recurrence rate was 9.8% and the mesh recurrence rate was 2.4%, which are also comparable to findings in the literature.3,11 Both univariate and multivariable analyses suggested that the use of mesh decreases the risk of developing a recurrence by 3.6 times. This information suggests that umbilical hernias should be repaired using mesh, especially if a patient has multiple comorbidities that are significantly associated with recurrence, such as obesity, diabetes, liver disease, and ascites. Surgeons commonly consider the size of the hernia defect when deciding whether to use mesh in a repair. Of the 167 patients who underwent repair with mesh, 64 (38%) did not have their defect size reported, 55 (33%) had a defect of 2 cm or greater, and 48 (29%) had a defect size of less than 2 cm. Because there was no significant difference between these groups, we are unable to conclude whether the size of defects should play a role in a surgeon’s decision to use mesh.
Most studies evaluating umbilical hernia recurrences have looked at short-term follow-up. Our median follow-up time was 8.5 years. In our study, 50% of recurrences occurred within the first year, 75% within the first 5 years, and 25% after 5 years of repair. This information implies that most studies reporting recurrence rates are not capturing up to 25% of recurrences that are occurring in the long term.
Interestingly, 193 (58%) patients who underwent an umbilical hernia repair had other hernias that either were repaired before the index repair or developed postoperatively. Therefore, we propose that umbilical hernias may be a type of “field defect,” and we support the idea that abnormal collagen metabolism could play a role in hernia development.9,10 We found that on univariate analysis, patients who had a history of other hernias (n = 125) were less likely to have umbilical hernia recurrences (P = .03). We speculate that surgeons might be more inclined to use mesh in patients with a history of other hernias. Seventy-six (61%) patients with other hernias received mesh, while only 87 (44%) patients without prior hernias received mesh (P = .003). The use of mesh in patients who are prone to develop hernia may protect them from recurrence.
Our study had a low complication rate of 18%, despite the 50% of patients who underwent repair with mesh. If seromas are excluded, the total complication rate was only 11.7%, which is comparable to the rate in the literature.12,13 In 332 patients, we saw a low surgical-site infection rate of 6.6% (n = 22) as well as a low mesh infection rate of 0.9% (n = 2). There were no cases of mesh causing fistulas in our long-term follow-up. There was no significant difference in the complication rate (of which most are seromas) between those who had mesh repair and those who did not. In view of its low risks and high benefits, mesh repair should be considered in most cases of umbilical hernia repair, regardless of defect size.
Our study had a 27% mortality rate, at a mean of 5.1 years after index repair, for reasons that are unclear. Because the mean age of our patients was 58.4 years, this rate suggests that 27% of patients were dying by age 63 years. Expectedly, age, smoking, and liver disease increased the odds of mortality in this patient cohort. There was only 1 death within 30 days, for a 30-day mortality of 0.3%.
A surprising finding was that patients who underwent emergency or semi-urgent repair had a 2.2 times increased odds of death. Twenty deaths occurred more than 30 days after emergency or semi-urgent repair, and 12 (60%) of these deaths were secondary to unknown causes and 5 (25%) were secondary to liver disease. Although there is a trend toward higher mortality rates in patients who underwent emergency repair, it is difficult to interpret whether the deaths were related to the emergency or to underlying medical conditions given that the etiology of the majority of these deaths is unknown.
A major limitation of this study was that patients had variable follow-up periods. For example, patients studied in 1998 had 16 years of follow-up, while those who had operations in 2008 had only 6 years of follow-up. To assess the severity of this bias, we ran sensitivity analyses that included only patients who strictly met a 7-year follow-up period. We found that there was minimal bias in our multivariable mortality model, with no changes in the sign or significance of covariates. In our recurrence model, however, the important covariates lost significance because of the smaller sample size used in the sensitivity analysis, and we could not retain the major findings of our study while applying the model.
There were several other limitations to the study. First, we were able to note recurrences only if they were described in the CPRS. Second, we were unable to accurately account for the size of the hernias because of inconsistent recording in the operative notes. Third, we studied a cohort composed of mostly male veterans, so these results may not be applicable to other populations. Fourth, the causes of death of the majority of the deceased patients in our cohort were unknown. Therefore, it was difficult to know if their deaths were related to operative complications.
Umbilical hernia repairs in our cohort study of 332 mostly male military veterans reveal that ascites, liver disease, diabetes, obesity, and primary suture repair without mesh are significantly associated with an increased rate of recurrence. Mesh repairs should be considered in these patients, if not in all patients, given that there was no significant difference in the complication rates between mesh repair and primary suture repair. In this cohort, there was a 27% long-term death rate, at a median of 5 years, from index repair operation for reasons that are unclear. Future research looking into the mortality rates in a matched cohort comparing patients with umbilical hernia may help elucidate whether the death rates in our cohort are different from the rates in the general population.
Corresponding Author: Vivian M. Sanchez, MD, Department of Surgery, Boston University School of Medicine, 1400 VFW Pkwy, SS112, West Roxbury, MA 02132 (email@example.com).
Accepted for Publication: October 28, 2016.
Published Online: January 25, 2017. doi:10.1001/jamasurg.2016.5052
Author Contributions: Dr Sanchez and Ms Shankar had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Shankar, Itani, Sanchez.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Shankar, Sanchez.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: O'Brien, Sanchez.
Administrative, technical, or material support: Itani, O'Brien, Sanchez.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was funded by the VA Healthcare System.
Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: The authors acknowledge the VA Healthcare System for funding the design and conduct of this study and the interpretation of the data.
Create a personal account or sign in to: