|
A correction has been published: N Engl J Med 2001;345(3):222.
|
| |||||||||||||||||||||||||||||||||||||||
Background A history of appendectomy is rare in patients with ulcerative colitis. This suggests a protective effect of appendectomy or that appendicitis and ulcerative colitis are alternative inflammatory responses. We sought to characterize this inverse relation further.
Methods We studied a cohort of 212,963 patients who underwent appendectomy before the age of 50 years between 1964 and 1993 and a cohort of matched controls who were identified from the Swedish Inpatient Register and the nationwide census. The cohort was followed until 1995 for any subsequent diagnosis of ulcerative colitis.
Results Patients who underwent appendectomy for appendicitis and mesenteric lymphadenitis had a low risk of ulcerative colitis (for patients with perforated appendicitis, the adjusted hazard ratio was 0.58 [95 percent confidence interval, 0.38 to 0.87]; for those with nonperforated appendicitis it was 0.76 [95 percent confidence interval, 0.65 to 0.90]; and for those with mesenteric lymphadenitis it was 0.57 [95 percent confidence interval, 0.36 to 0.89]). In contrast, patients who underwent appendectomy for nonspecific abdominal pain had the same risk of ulcerative colitis as the controls (adjusted hazard ratio, 1.06; 95 percent confidence interval, 0.74 to 1.52). For the patients who had appendicitis, an inverse relation with the risk of ulcerative colitis was found only for those who underwent surgery before the age of 20 years (P<0.001).
Conclusions Appendectomy for an inflammatory condition (appendicitis or lymphadenitis) but not for nonspecific abdominal pain is associated with a low risk of subsequent ulcerative colitis. This inverse relation is limited to patients who undergo surgery before the age of 20 years.
Another explanation is that appendicitis and ulcerative colitis are alternative inflammatory responses that are genetically or environmentally determined. That environmental factors may play a part is supported by a study that reported a lower incidence of appendectomy and a higher incidence of ulcerative colitis among Mormons in Britain and Ireland than in the general population, possibly because of differences in lifestyle.19
We sought to characterize further the relation between appendectomy and ulcerative colitis. We used data from a national data base in Sweden to examine the influence of the specific diagnosis and the age of the patient at the time of appendectomy on the relation between appendectomy and ulcerative colitis in a case–control study.
Methods
Study Population and Design
Since 1964, the Swedish National Board of Health and Welfare has compiled data on hospital discharges in its Inpatient Register. Besides a national registration number (that uniquely identifies every resident of Sweden), each record contains medical data, including surgical procedures performed (coded according to the Swedish Classification of Operations and Major Procedures20) and diagnoses at discharge (coded from 1964 to 1968 according to the 7th revision of the International Classification of Diseases [ICD-7],21 from 1969 to 1986 according to the 8th revision [ICD-8],22 and from 1987 to 1993 according to the 9th revision [ICD-9]23). In 1964 the register included only six counties, representing 20 percent of the Swedish population, but more counties were added successively. Since 1987 the register has included data from all Swedish hospitals. Since there is virtually no private hospital care in Sweden, our study was essentially population based.
The study protocol was approved by the Swedish Data Inspection Board, a Swedish federal agency that serves as an institutional review board for linkages of data bases. This process has been approved by the institutional review committees at the various hospitals in Sweden.
Patients Who Underwent Appendectomy
All patients in the Inpatient Register who had a discharge diagnosis indicating that an appendectomy (operation code 4510 or 4511) had been performed between 1964 and 1993 were identified (Table 1). Only patients who were under the age of 50 years at the time of surgery were included. Patients were divided into groups according to their discharge diagnoses, which included perforated appendicitis, nonperforated appendicitis, nonspecific abdominal pain, mesenteric lymphadenitis, and other diagnoses. Other diagnoses included salpingitis, gastroenteritis, ovarian cysts, inflammatory bowel disease, diverticulitis, other diseases of the appendix, and cholecystitis.
|
For each eligible patient with a history of appendectomy, one control patient without a history of appendectomy was matched according to age, sex, and township of residence. To establish the place of residence for all patients at the time of the surgery, we used data from the nationwide 1970, 1980, and 1990 censuses, and we chose the place of residence as that occupied closest to the time of the appendectomy. Consequently, for the patients who underwent appendectomy in the first half of each decade, the matching control was chosen from the census that preceded the operation. Eighty controls who died during the interval between the census and the operation were excluded from the analyses (Table 1).
Follow-up
Patients who underwent appendectomy and controls were linked to the Inpatient Register for 1964 to 1995. Patients with a discharge diagnosis of ulcerative colitis or ulcerative proctitis (ICD-7 code 57220 or 57221, ICD-8 code 56310 or 56902, or ICD-9 code 556X) were identified. Some patients had received different discharge diagnoses of inflammatory bowel disease at different times. Reassignment of the diagnosis may occur as the disease progresses or if the histopathological diagnosis after surgery does not correspond to the diagnosis at the time of the first discharge. The most recent discharge diagnosis was considered the final diagnosis. The date of the first diagnosis of any inflammatory bowel disease was considered the time of the final diagnosis.
The follow-up for both groups of patients started one year after the date of the appendectomy and ended at the time of a discharge diagnosis of ulcerative colitis, on the date of death, or on December 31, 1995, whichever came first. Case patients and controls who had a diagnosis of ulcerative colitis before or within one year after appendectomy were considered to have preexisting cases and were excluded. Finally, each case patient or control whose matched counterpart had been excluded was also excluded; i.e., only complete pairs of case patients and matched controls were included in the analysis. Thus, the patients with a history of appendectomy who were matched to the 80 controls who died between the census and the time of the appendectomy were also excluded.
Statistical Analysis
The risk of ulcerative colitis after appendectomy was analyzed by calculation of the incidence-rate ratio (the ratio of the incidence rates of ulcerative colitis among the patients who underwent appendectomy to that among the controls). The incidence rate is the number of diagnosed cases of ulcerative colitis divided by the total number of person-years at risk during follow-up. The exact confidence intervals were calculated as described by Rothman.24
Multivariate analyses were performed with Cox proportional-hazards regression to adjust for any confounding effect of sex, age at the time of surgery, and year of surgery. The modification of the effect of appendectomy by age at the time of surgery was analyzed by the inclusion of terms for the interaction between diagnosis and age at the time of surgery in the models. To summarize the evolution over time of the risk of ulcerative colitis after appendectomy, we used Kaplan–Meier analysis to construct curves for the cumulative incidence of ulcerative colitis among the case patients and the controls, which were then compared with use of the log-rank test. Separate analyses were performed for the subgroups of case patients who underwent appendectomy for appendicitis before the age of 20 and at or after the age of 20, for mesenteric lymphadenitis, and for nonspecific abdominal pain and the corresponding controls. All tests of significance were two-sided. A P value of less than 0.05 was considered to indicate statistical significance.
Results
The characteristics of the patients in the case and control groups are presented in Table 2. After exclusions, there were 212,963 case patients and the same number of controls for analysis (Table 1). The follow-up encompassed more than 5 million person-years. During this period, 770 patients with a discharge diagnosis of ulcerative colitis were identified. One hundred three of these patients were later reclassified as having Crohn's disease. Twenty-eight patients with a first diagnosis of indeterminate colitis and 19 with a first diagnosis of Crohn's disease were reclassified as having ulcerative colitis. Thus, there were 714 patients with a final diagnosis of ulcerative colitis.
|
Patients who had an appendectomy had a significantly lower incidence of ulcerative colitis during follow-up than the controls (incidence-rate ratio, 0.74; 95 percent confidence interval, 0.64 to 0.86) (Table 3). This relation was found after appendectomy for appendicitis (incidence-rate ratio, 0.73; 95 percent confidence interval, 0.62 to 0.87) and for mesenteric lymphadenitis (incidence-rate ratio, 0.48; 95 percent confidence interval, 0.27 to 0.83). The relation was not found after appendectomy for nonspecific abdominal pain (incidence rate ratio, 1.34; 95 percent confidence interval, 0.77 to 2.38). The inverse relation of appendectomy to the risk of ulcerative colitis was observed only in patients who had undergone surgery before the age of 20 years.
|
The relation of ulcerative colitis with the patient's diagnosis at the time of appendectomy, with adjustment for age, time period, and sex, was analyzed with use of Cox proportional-hazards regression. Significant inverse relations with ulcerative colitis were found for appendectomy for perforated appendicitis (adjusted hazard ratio, 0.58; 95 percent confidence interval, 0.38 to 0.87; P=0.01), for nonperforated appendicitis (adjusted hazard ratio, 0.76; 95 percent confidence interval, 0.65 to 0.90; P=0.001), and for mesenteric lymphadenitis (adjusted hazard ratio, 0.57; 95 percent confidence interval, 0.36 to 0.89; P=0.02). Appendectomy for nonspecific abdominal pain had no relation with ulcerative colitis (adjusted hazard ratio, 1.06; 95 percent confidence interval, 0.74 to 1.52; P=0.78). An age of less than 20 years at the time of surgery (adjusted hazard ratio, 0.65; 95 percent confidence interval, 0.56 to 0.76; P<0.001) and female sex (adjusted hazard ratio, 0.84; 95 percent confidence interval, 0.72 to 0.98; P=0.02) were significant covariates in the model.
Effect Modification by Age at the Time of Surgery
The inclusion of terms of interaction between diagnosis and age at the time of surgery produced significant results for the interaction of age and nonperforated appendicitis (P<0.001) but not for the interaction of age and perforated appendicitis (P=0.07), probably because of the smaller sample. The other interaction terms were not significant.
Separate multivariate analyses were performed for patients who had surgery before the age of 20 and those who had surgery at 20 or older to assess the relation between appendectomy and ulcerative colitis in the two age groups. Patients who underwent surgery for appendicitis before the age of 20 had a low risk of ulcerative colitis during follow-up, as compared with the controls (adjusted hazard ratio, 0.42; 95 percent confidence interval, 0.31 to 0.57; P<0.001). Patients who underwent surgery for appendicitis when they were 20 years of age or older had the same risk of ulcerative colitis as the controls (adjusted hazard ratio, 0.97; 95 percent confidence interval, 0.79 to 1.18; P=0.79).
Kaplan–Meier Analysis
Figure 1 and Figure 2 show the cumulative proportion of patients in whom ulcerative colitis developed after appendectomy performed for appendicitis before the age of 20 and at or after the age of 20, after appendectomy for mesenteric lymphadenitis, and after appendectomy for nonspecific abdominal pain, as well as the cumulative proportion of matched controls with ulcerative colitis. The decreased risk after appendectomy for appendicitis before the age of 20 (P<0.001) and after appendectomy for mesenteric lymphadenitis (P=0.005) persisted for up to 30 years. There was no significant difference between the controls and the patients who had surgery for appendicitis at the age of 20 or older (P=0.68) or the patients who had surgery for nonspecific abdominal pain (P=0.27).
|
|
Our findings confirm the previously reported inverse relation between appendectomy and ulcerative colitis. The incidence of ulcerative colitis was low after appendectomy for inflammatory conditions such as appendicitis and lymphadenitis, but no relation was found after appendectomy for nonspecific abdominal pain. These findings suggest that the inflammatory condition preceding the appendectomy, rather than the appendectomy itself, is inversely related to the subsequent development of ulcerative colitis.
The inverse relation suggests that appendicitis and mesenteric lymphadenitis are related to different pathogenic factors from those related to ulcerative colitis. One such factor is known for Crohn's disease. Crohn's disease is mediated by type 1 helper T cells, whereas ulcerative colitis is mediated by type 2 helper T cells. A weak positive relation has been reported between appendectomy and Crohn's disease.3,6,7,8 This may suggest that appendicitis is mediated by type 1 helper T cells, which would explain the inverse relation to ulcerative colitis. The pathogenesis of appendicitis, however, is not known.
Our findings also confirm a previous report of an influence of age at the time of surgery on the inverse relation.6 This influence was found only after surgery for appendicitis, with an inverse relation to ulcerative colitis after surgery for appendicitis in childhood and adolescence and no relation after surgery for appendicitis in adulthood. This finding suggests an age-dependent difference in the pathogenesis of appendicitis. Such a difference would be consistent with the strong age-dependent variation in the incidence of appendicitis, which peaks during adolescence.17
The chief strengths of our study are its size and its design as a cohort study. Most previous studies analyzed the frequency of appendectomy in patients with preexisting and newly diagnosed cases of ulcerative colitis as compared with the frequency in matched control patients at orthopedic or dermatology outpatient clinics. Thus, there was a risk of selection bias and recall bias. This problem is also suggested by the unusually high proportion of controls who had undergone appendectomy, which reached 17 to 25 percent in some of these studies,5,6,8 as compared with an estimated lifetime risk of 7 percent.25
Our study also has potential weaknesses. The specificity of a diagnosis of appendicitis is somewhat less secure than that of a diagnosis of a noninflamed appendix, since in the former case the surgeon may overdiagnose macroscopic pathologic changes. In a study of the Inpatient Register of Jönköping county in central Sweden, a region that is also included in the national register, 10 percent of patients had been given a false positive diagnosis of appendicitis and 6 percent had been given a false negative diagnosis.17 These figures are probably representative of the whole register. The differentiation between mesenteric lymphadenitis and nonspecific abdominal pain is also imprecise. These errors, however, would only lead us to underestimate the differences we observed.
Epidemiologic studies have shown that appendicitis that resolves spontaneously is common.17 A diagnosis of appendicitis may therefore be related to care-seeking behavior and other underlying socioeconomic factors. This potential source of selection bias, however, does not apply to patients with perforated appendicitis.
The specificity of the diagnoses of ulcerative colitis in the Swedish Inpatient Register was assessed in one previous study and was found to be about 90 percent.18 We did not include patients with distal colitis or mild inflammation that did not necessitate hospital care. The inverse relation therefore applies only to patients with extensive colitis or severe inflammation. There is, however, no reason to believe that the underreporting of patients with mild ulcerative colitis should be different for the case patients and the controls. Among the controls, there was a cumulative incidence of 150 newly diagnosed cases of ulcerative colitis per 100,000 persons during follow-up. After we included the preexisting cases, the cumulative incidence was just under 300 per 100,000 persons. This is very close to the prevalence found in epidemiologic studies in Sweden during this period.26 Thus, most cases of ulcerative colitis were probably identified.
Recent studies have reported an increased risk of appendectomy among the children of parents who smoke.27,28 Smoking may therefore act as a confounder. Unfortunately, information about smoking was not available to us. In previous studies, the inverse relation between appendectomy and ulcerative colitis remained after adjustment for smoking status.3,9,10
In conclusion, we confirmed an inverse relation between appendectomy and ulcerative colitis. We found that this inverse relation was confined to appendectomy in young patients with inflammation in the appendix or mesenteric lymph nodes. We found no such relation for patients who had an appendectomy for nonspecific abdominal pain or for older patients who had surgery for appendicitis.
Supported by grants from the Scientific Committee at Ryhov County Hospital, the Health Research Council in southeastern Sweden, and the Bengt Ihre Foundation.
Source Information
From the Department of Surgery, County Hospital Ryhov, Jönköping (R.E.A.); the Division of Surgery, University Hospital, Linköping (G.O.); the Department of Gastroenterology, Örebro Medical Center Hospital, Örebro (C.T.); and the Department of Medical Epidemiology, Karolinska Institute, Stockholm (A.E.) — all in Sweden; and the Department of Epidemiology, Harvard School of Public Health, Boston (A.E.).
Address reprint requests to Dr. Andersson at the Department of Surgery, Ryhov Hospital, S-551 85, Jönköping, Sweden, or at roland.andersson{at}ryhov.ltjkpg.se.
References
| |||||||||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
