Acute diverticulitis is a common disease, responsible for an annual cost of $2.4 billion [1-3], including 300,000 admissions, and perhaps 3 times as many episodes in the outpatient setting [4]. Although most cases of uncomplicated diverticulitis are treated with antibiotics alone, diverticulitis remains one of the most common causes of colonic resection [5-7].
Historically, the threshold for recommending elective resection was after 2 episodes, but over the past 2 decades professional organizations have encouraged a more restrictive approach, moving away from counting episodes to determine indications for surgery [2].
Despite those recommendations, the incidence of elective operations for diverticulitis has continued to increase [8,9], and it is unclear whether patient and surgeon decision-making aligns with those recommendations.
A recent trial comparing elective surgery and conservative management for patients with recurrent diverticulitis revealed that surgery was associated with a significantly higher gastroenterological-specific quality of life (DIRECT trial) [10], emphasizing the importance of patient-specific outcomes in this disease.
Conventionally, surgeons have recommended elective surgery with the goal of preventing recurrence, progression, or – most importantly for patients – to prevent emergency surgery resulting in a colostomy [11,12]. Colostomies are associated with decreased quality of life, and both patients and surgeons have cited colostomy avoidance as a motivation for elective surgical intervention [13-16].
Assessment of the current risk of recurrence and colostomy, after surgical and nonsurgical treatment, has been challenging because of limitations in the available data. Previous large population-based studies evaluating diverticulitis recurrence have used health care claims that include only hospitalized patients.
Those studies overlook the dramatic shift toward outpatient care for diverticulitis over the past decade, including patients treated for recurrence as outpatients [17–19]. These limitations have prevented accurate risk assessment and preclude a data-driven approach to decision making.
Studies based on inpatient care have shown that approximately one-third of patients with diverticulitis will develop a recurrence requiring hospitalization after the primary episode [20-21].
Surgically treated patients are also at risk of recurrence [20], although the exact recurrence rate in that group has not been well defined. Colostomy rates are estimated to be between 5% and 7% after the first episode of diverticulitis [6], but elective surgery, by itself, has a small risk of colostomy, and the rate of “rescue colostomies” ( unplanned ostomy due to complications, including fistulization) with elective resection, is 1% to 3% [22-24]. What is unknown is whether medical treatment or surgery confers an increased risk of future colostomy.
To address the gap in evidence on the risk of recurrence and ostomy, and to support informed decision making, we compared the risk of recurrence and ostomy among patients treated for uncomplicated diverticulitis, using a national care claims database. including inpatient and outpatient claims, for millions of Americans. It was hypothesized that elective resection is associated with a reduced risk of disease and ostomy recurrence.
| Methods |
A time-to-event analysis for recurrence-free and ostomy-free survival was performed from a retrospective cohort of patients treated for uncomplicated diverticulitis, using the Thompson Reuters MarketScan Commercial Claims and Medicare Supplement Databases , between 1 January 2008 and December 31, 2014.
This data source includes claims and enrollment data for patients with employer-provided health insurance and their dependents, and Medicare claims for patients who continue to have private supplemental insurance in addition to being enrolled in Medicare.
Diagnoses and procedures are coded using the International Classification of Disease, 9th edition (ICD-9), Current Procedural Terminology (CPT), and Healthcare Common Procedures Coding System codes . This study was approved by the University of Washington Institutional Review Board .
> Cohort definition
Among all patients in the MarketScan database who were treated with inpatient admission for a primary diagnosis of uncomplicated diverticulitis (ICD-9 code 562.11), those who were subsequently treated with either medical therapy or elective surgery were selected. a second visit for uncomplicated diverticulitis (>6 weeks from the initial visit) (n = 26,471), and that analysis was repeated for patients treated at a third, fourth, and fifth visits (which included both inpatient and outpatients).
Uncomplicated diverticulitis was defined as diverticulitis without the concomitant presence of abscesses, peritonitis, intestinal perforation, colonic stenosis, colonic fistula, colonic hemorrhage, emergency colonic resection, or percutaneous drainage of an intra-abdominal abscess.
Patients under 18 years of age and those with previous colon cancer or colonic operations were excluded ( n = 3120 excluded).
To accurately identify primary episodes of diverticulitis, patients were excluded if they did not have 1 full year of continuous enrollment, prior to primary diagnosis, that they were free of a diverticulitis diagnosis (n = 9163 excluded), or if they had discontinuous enrollment. between treatment and outcome (n = 2115 excluded).
The Charlson Comorbidity Index (CCI) adapted by Deyo [25] was calculated , and patients with immunosuppression diagnoses were identified: malignant disease (except non-melanoma skin cancer), lymphoma, human immunodeficiency virus, steroid use, transplant of solid organ, and other immunological disorders [26-28].
Differences between treatment groups were evaluated with chi-square tests for categorical data, and Student’s t- test and Kruskal-Wallis test for normally and non-normally distributed continuous variables, respectively.
> Types of treatment
Medical therapy was classified as a consultation with a physician with a diagnosis of diverticulitis and an associated hospitalization within 6 weeks of diagnosis (inpatient medical therapy), or a prescription for oral antibiotics to treat diverticulitis, within 1 week. of diagnosis (outpatient medical therapy).
An episode was defined as elective surgery if a colonic resection was performed within 6 weeks of the start of diverticulitis-related treatment for that episode.
This was intended to capture elective operations that may have been preceded by an outpatient clinic visit (e.g., a preoperative visit with a surgeon who billed for diverticulitis). Patients with potentially emergent operations in their second treatment episode were not included, excluding surgeries that occurred within 1 week of an emergency department visit [29].
> Results
Because numerous claims can be generated at the time of a single episode of diverticulitis, 6-week windows were assigned to capture all claims from a single episode of the disease [30]. Recurrence of diverticulitis was defined as any visit for treatment of diverticulitis (including inpatient, outpatient, complicated, uncomplicated, medical, or surgical), occurring more than the 6th week after the previous episode.
Outpatient recurrences were defined by a diagnosis of diverticulitis and a concomitant antibiotic prescription. Patients were followed for all subsequent treatment visits until disenrollment from their current health plan, or until the end of 2014, whichever came first. For patients who required an ostomy during primary elective surgical treatment, the time to event was considered one day.
The authors believe that patients would rate ileostomy (created to protect an anastomosis) and colostomy (created due to severe inflammation, perforation, or abscess) as equally unfavorable outcomes. Therefore, they measured the rate of any ostomy as a single outcome. They reported the proportion of patients with recurrence and ostomy at 1, 3, and 5 years, among those patients with continuous enrollment since treatment, respectively.
> Statistical analysis
A survival analysis was performed to determine the risk of diverticulitis recurrence and subsequent ostomy. Recurrence-free survival was summarized using Kaplan-Meier plots. Controlling for patient age, sex, CCI, immunosuppression, and type of health plan, a Cox regression analysis was performed to evaluate the hazards of recurrence and subsequent ostomy associated with either treatment.
The proportional hazards assumption was tested by plotting scaled Schoenfeld residuals over survival time. Data were missing only for health plan type in 2% of subjects, and the analysis revealed no significant differences in covariates for those with and without missing data; therefore, full case analysis was adopted.
> Sensitivity analysis
Due to the potential for bias in cohort and treatment selection, 3 sensitivity analyzes were performed.
1. First, because outcomes for patients after their second treatment visit may vary relative to patients treated at subsequent visits, we evaluated the rate and risk of recurrence and ostomy among patients treated for uncomplicated disease. , through medical or surgical therapy, at their third to fifth treatment visit (among those patients who had been treated medically for an uncomplicated illness in previous visits).
2. Second, because a 1-year washout period may not have precisely defined the initial disease, we also evaluated the rate and risk of recurrence and ostomy among patients with 3 and 5 years of enrollment. no disease before primary diagnosis.
3. Third, because patients treated medically as outpatients may have less severe disease than those treated medically as inpatients, we evaluated the rate and risk of recurrence and ostomy among those treated medically as inpatients and outpatients. separately, compared to elective surgery. All analyzes were performed with Stata software, v14.2 IC (StataCorp, College Station, TX).
| Results |
A total of 12,073 patients (mean age 56 years, 59% female, 10% immunosuppressed) were treated in a second consultation for uncomplicated diverticulitis. The median length of stay at the primary operation was 3 days (interquartile range [IQR]: 2-4 days).
At the second treatment visit, 81% of patients (n = 9832) were treated with medical therapy, while 19% (n = 2241) underwent elective surgery. Almost all patients underwent a computed tomography scan at the time of initial diagnosis or at the second episode of the disease (93.9% for medical therapy; 92.7% for surgical therapy).
The median time between the initial diagnosis and the second treatment visit was 199 days (IQR: 85-475 days; 257 days (IQR: 105-551) for the medical therapy group, and 88 days (IQR: 60-551) 143) for the elective surgery group).
Patients treated with elective surgery were younger, more frequently male, less commonly immunosuppressed, and a larger proportion had no comorbid conditions ( all P < 0.001).
There was no difference in the types of health plans between the treatment groups ( P = 0.13). Compared with patients who had no evidence of diverticulitis recurrence, those who recurred were slightly younger (54.6 vs 56.3 years; P < 0.001), and were less likely to have comorbidities (ICC ≥ 2, 17%). vs 21%; P < 0.001).
However, these groups did not differ in other factors. Similarly, patients who ultimately received an ostomy were only slightly more likely to have comorbidities (25% vs 19%; P = 0.04).
The median follow-up time was 537 days (IQR: 229-1042 days). Of all patients, 3844 (32%) had subsequent treatment consultations for diverticular recurrence (50% outpatient recurrences; 4% recurrences requiring emergency surgery).
Among patients with 1 year of follow-up, 32% of those treated medically and 12% after elective surgery had recurrence of diverticulitis. At 3 years, the recurrence rate was 51% after medical therapy and 12% after elective surgery (at 5 years, 61% and 15%).
After adjusting for differences between groups, patients treated with elective surgery had an 83% lower risk of diverticulitis recurrence compared with those treated with medical therapy (hazard ratio [HR], 0.17; 95% risk interval). confidence (CI): 0.15-0.20]. The median time to recurrence was shorter among patients treated medically (165 days; IQR: 76-396) than those treated surgically (272 days; IQR: 82 -612).
As after the second treatment visit, diverticulitis recurrence rates at one year were higher for medical treatment compared to elective surgery at the third visit (44% vs 9%; ORR: 0.18; 95 % CI: 0.14-0.23), in the fourth (53% vs 6%; TR: 0.09; 95% CI: 0.05-0.14), and in the fifth treatment consultation (54 % vs 7%; TR: 0.18; 95% CI: 0.09-0.34).
At 1 year, 1.6% of patients treated medically and 4.0% of those treated with elective surgery had an ostomy ( P < 0.001). Among patients with 3 years of follow-up, the ostomy rate did not appear higher in either group (1.8% with medical therapy vs. 2.9% after elective surgery; P = 0.09).
The adjusted risk of having an ostomy was more than twice as high in the elective surgery group compared with medical therapy (adjusted HR: 2.3; 95% CI: 1.8-3.0), but this model did not met the proportional hazards assumption. Importantly, ostomies in the elective surgery group were related to primary elective resection in all but 2 cases, which occurred within 6 weeks of primary treatment (76% within 1 week after elective surgery).
It could not be determined from the claims data whether these ostomies were performed prophylactically, due to anatomical challenges, or due to postoperative anastomotic leaks. In contrast, there was a continued distribution of ostomies after medical therapy.
A logistic regression model controlling for the covariates listed above revealed a 2.7 increased odds of ostomy after elective surgery at 1 year (95% CI: 1.9–3.8, proportional hazards not met).
The 1-year cumulative risk of ostomy remained low after treatment at the third visit (0.4% vs 3.9%; P < 0.001), at the fourth (0.5% vs 7.1%; P < 0.001 ), and at the fifth consultation (0.8% vs 2.3%; P = 0.43), for medical therapy and elective surgery, respectively.
In the sensitivity analysis, no influence of the duration of continuous enrollment before the initial episode of diverticulitis (washout period) on the risk of recurrence or the risk of ostomy, depending on the treatment group, was identified. Nor were any differences identified in the results due to the type of medical therapy (inpatient vs. outpatient), or in patients with immunosuppression.
| Discussion |
Diverticulitis is a major reason for colostomy, a treatment outcome ranked among the worst non-fatal complications by both patients and physicians [31,32].
After an episode of diverticulitis there is a lifelong risk of disease recurrence [6]; Therefore, when counseling patients, surgeons have historically recommended an elective resection of the colon to prevent recurrence and/or a future more serious episode, requiring an emergency colostomy. Although rates of emergency surgery for diverticulitis are highest at initial presentation [5,6,33], the risk of an ostomy after subsequent episodes has not been well described.
Most administrative data sets are restricted to inpatient care, which has made it difficult to describe recurrence rates that occur in the outpatient setting. In this study, we addressed that limitation and showed that, while patients treated with elective resection have a considerably lower risk of recurrence compared to those treated medically, as many as 6% of patients will have a disease recurrence within 1 year after elective surgery, and the rate is as high as 15% at 5 years.
Contrary to the authors’ hypothesis, the rate of subsequent ostomy is low for both treatments, but is higher among surgically treated patients.
Likewise, in this study, almost all ostomies were performed at or shortly after resection, rather than after subsequent surgical emergencies. To the authors’ knowledge, this is the first time that diverticulitis recurrence and long-term risk of ostomy have been evaluated from a disease burden perspective for both inpatients and outpatients.
The findings have important implications when counseling patients about the choice of elective surgery to prevent the impact of future episodes of diverticulitis or the risk of an ostomy.
Observational studies in hospitalized patients with diverticulitis suggest that up to one-third of patients with acute diverticulitis will have a recurrence [20,21,34]. In the present study, estimates of recurrence after medical therapy alone (32% at 1 year, 61% at 5 years) are likely somewhat higher than previously reported, capturing episodes managed in the outpatient setting [35 ].
It was previously thought that relapsed diverticulitis represented a more virulent form of the disease and was therefore more likely to fail medical treatment [36]. As a result, elective surgery was recommended to prevent recurrences after a second episode. But that belief has been changing [35] and guidelines now discourage the “2 strikes, you’re out” approach, favoring individualization based on patients’ concerns [2].
For many patients whose concerns are related to fear of the ostomy, an elective colonic resection is often presented as a long-term cure for recurrence and avoidance of colostomy. Despite that perspective, as many as 25% of patients have persistence of symptoms after resection [20]. A small number of studies have described that the risk of recurrent diverticulitis persists after resection in 5.8%–8.7% of cases [30,35,37,38].
Persistent symptoms are often attributed to inadequate resection, and these recurrences are thought to be due to inflammation of newly formed or previously unresected diverticula after a colosigmoid anastomosis [30].
The present study confirmed this recurrence rate across a national sample, reporting a 1-year recurrence rate of 6% after elective surgery at the second visit (and 7%–9% after surgery at subsequent visits). ) For patients with 3 to 5 years of follow-up (the minority in this cohort), recurrence after surgery reached 12% to 15%.
The risk of a colostomy is a major concern in many patients [31], and one factor motivating the choice between medical and surgical therapy is the fear of future emergency surgery resulting in a colostomy.
Although such fear was a driver of older recommendations for early surgical intervention [11,12], previous analyzes have reported that only a small number of patients will ever require an emergency colostomy after an initial episode (5.5% ) [6], and approximately 18 patients may need to undergo elective resection for diverticulitis to prevent a single emergency surgery [2,6].
To date, the risk of an ostomy has been attributed to subsequent emergency operations, but when a patient undergoes elective resection for diverticulitis, there is also the risk of an ostomy during the preventive operation. Approximately 2% to 5% of colonic anastomoses suffer disruption requiring reintervention [39], some of which will require a rescue colostomy. Additionally, surgeons sometimes use protective temporary ostomies at the time of elective resection.
The identified ostomy rate of 4% at 1 year after elective resection is consistent with the findings of other authors [22-24], but the consideration of the very low ostomy rate in the medical group (1.6%) , can influence the decision-making of those who seek to avoid this result.
The difference in ostomy risk between groups may not be clinically significant, but patients should be advised that, although the risk of ostomy is low (5%), it is unlikely to improve with elective resection and, in In reality, it is greater when considering the risk of ostomy with the elective procedure.
In this study, both treatment groups have been restricted by excluding patients with possible indications for colon resection according to current guidelines (i.e., abscess, fistula, stricture, hemorrhage, perforation, peritonitis, or emergency presentation ). However, being an observational study of uncomplicated diverticulitis, these groups likely differed in the timing and frequency of symptoms.
Indeed, the time between the first (initial) episode and the second differed between patients treated with elective surgery (median 88 days; IQR: 60-143 days), and those treated with medical therapy (median 257; IQR: 105-143 days). 551).
That suggests that patients treated with elective surgery may be more likely to experience "latent diverticulitis," or some form of more virulent or burdensome disease, manifesting as earlier surgery than the medically treated cohort. This hypothesis warrants further study, given that current guidelines do not evaluate that subgroup of patients, and when intervention is most appropriate for their disease [2].
Studies using administrative data have important limitations. Diagnoses applied to claims can be moved from one consultation to another within the patient’s medical history.
To manage this uncertainty about the true episodes of the disease and the severity of any recurrence (such as the presence or absence of complicating factors, such as abscesses or fistulas), it was specified that all visits for diverticulitis should be concurrent with at least 1 form of treatment (hospitalization, operation, or outpatient prescription of antibiotics).
Another challenge to comparing outcomes between treatments in observational cohorts is the potential for differences in disease severity between treatment groups (i.e., unmeasured confounding by indication), and a lack of certainty about the elective or emergent nature of a colonic resection.
A doctor may recommend surgery or antibiotics based on symptom burden, disease severity, or his or her own specialty, but those factors are not captured by administrative databases. This study was carefully restricted to patients with uncomplicated diverticulitis (excluding abscesses, peritonitis, etc.), to decrease the heterogeneity of disease severity. Likewise, it is unknown whether some episodes of relapse instead represent persistent latent disease.
It may be that there are surgery factors not measured in the surgical cohort, such as persistent inflammation on CT scan, persistent leukocytosis on repeat labs, or persistent symptoms leading to more visits or longer stays, that warrant characterization in a cohort with data detailed clinics. A data set spanning a greater number of years would allow for longer periods of removal and follow-up. For example, future research should include measuring the proportion of patients who have their ostomy closed.
Finally, the MarketScan database provides claims from Americans with private health insurance and does not include Medicaid, or the uninsured population; therefore, it is not completely generalizable. Likewise, by restricting themselves to patients with continuous enrollment, the authors selected a cohort with stable, full-time employment, who, consequently, may be healthier than other groups.
Additionally, because patients appear to be frequently dropped from private insurance, estimates of outcomes at 3 and 5 years after an event may be imprecise due to low follow-up rates (24%-28% and 4% -8%, respectively).
Despite that limitation, to the authors’ knowledge, this study is the largest published to date on inpatient and outpatient claims for diverticulitis, and provides valuable information not available through other data sources.
| Conclusions |
For patients with a history of uncomplicated diverticulitis, surgeons have previously recommended surgery to prevent recurrence and decrease the likelihood of a future ostomy.
This study shows that, while elective surgery clearly decreases the risk of disease recurrence, 6% to 15% of patients treated with elective surgery will have a recurrence 5 years after surgery.
These findings suggest that the risk of a future ostomy is low regardless of treatment choice, but may currently be higher in patients with elective surgery, when the risk of an ostomy related to elective resection is considered.
These findings challenge the appropriateness of elective surgery for ostomy prevention and should be useful to inform decision making.
