Indications for operation in suspected appendicitis and incidence of perforationBMJ 1994; 308 doi: https://doi.org/10.1136/bmj.308.6921.107 (Published 08 January 1994) Cite this as: BMJ 1994;308:107
- R Andersson,
- A Hugander,
- A Thulin,
- P O Nystrom,
- G Olaison
- Department of Surgery, Ryhov Hospital, S-551 85 Jonkoping, Sweden
- Department of Surgery, University Hospital, Linkoping, Sweden
- Correspondence to : Dr. Andersson.
Objective: To clarify poorly understood epidemiological features of appendicitis.
Design: Retrospective study of consecutive cases from a defined population and analysis of data from published studies.
Setting: County of Jonkoping, Sweden. 3029 patients who underwent operation in 1984-9 and 4717 patients from the county town who underwent operation in 1970-89, all for suspected appendicitis, plus 48 426 cases from six reported studies.
Main outcome measures: Incidences specific for age and sex and temporal trends of perforating and non-perforating appendicitis and removal of a normal appendix. Associations between diagnostic accuracy, rate of perforation, and incidences of removal of a normal appendix and of perforating and non-perforating appendicitis.
Results: The incidence of appendicitis was 116/100 000 inhabitants. Appendicitis was more common in male patients. The incidence of perforating appendicitis was independent of age, stable overtime, and uninfluenced by the rate of laparotomy, whereas the incidence of non-perforating appendicitis was age dependent, decreasing over time, and related to the diagnostic accuracy and rate of removal of a normal appendix.
Conclusions: Perforating and non-perforating appendicitis seem to be separate entities, and appendicitis that resolves spontaneously is common. This may have important implications for managing suspected appendicitis.
The incidence of appendicitis is declining and shows large geographical variations
A liberal attitude to exploration with an accompanying high rate of negative results (up to 50% in women) has been accepted in the hope of preventing perforation
In this study perforating and non-perforating appendicitis seemed to be different entities, and spontaneously resolving appendicitis was common
The readiness to explore influences the detection of resolving appendicitis and may explain variations in incidence of appendicitis and perforation rate
Perforation rate is useless as a measure of quality of management in suspected appendicitis
The many poorly understood epidemiological features of acute appendicitis include a widely varying incidence even within the same country,*RF 1-4* trends towards a falling incidence,*RF 5-7* and rising rate of perforation.8,9
We surveyed the epidemiology of appendicitis in terms of the patients' age and sex and the changes associated with time in a Swedish country. We also analysed how varying rates of exploratory laparotomy in suspected appendicitis influenced the incidence of perforating and non-perforating appendicitis.
Subjects and methods
Age And Sex Specific Incidences
From the computerised hospital register for Jonkoping county (population 302475 in 1987) we identified 3029 consecutive patients who underwent appendicectomy for suspected appendicitis in 1984-9. Patients who had incidental appendicectomies were excluded. The quality of the database is high, and errors in registration are avoided by continuous checks of extracts from the database with the original discharge letter.
A total of 1470 males and 1559 females had undergone an appendicectomy. Their median age was 21 years. Histological examination of excised tissue had been performed in 2509 (83%). An inconsistency between the discharge diagnosis and the results of the microscopic examination was found in 219 (9%) of the 2509 examined specimens. There were 179/1840 (10%) false positive diagnoses and 40/669 (6%) false negative diagnoses. The net error because of these false diagnoses was an overestimation of appendicitis in 139/2509 cases (6%).
In the 250 cases in which no histological examination of excised tissue took place the patients were slightly younger (mean (SD) age 25.1 (17.3)v 27.4 (17.9) years; P<0.007) and were more often male (52.7%v47.6%; P<0.05). There was, however, no difference in diagnostic accuracy (78% v 73%; P=0.15) and rate of perforation (16% in both groups) before the correction.
The final diagnosis corrected in accordance with the histological findings was appendicitis in 2100 cases, giving a diagnostic accuracy (ratio of inflamed appendixes to all appendicectomies) of 69%. The appendix was perforated in 330 cases (perforation rate 16%).
An abdominal pathological condition other than acute appendicitis was the cause of symptoms in 382 (13%) cases, of which 105 (4%) required an operation. The most common cause was lymphadenitis mesenterica (193 (6%) cases). A gynaecological disorder was found in 64(4%) female patients. Further details of the patients have been published elsewhere.10
The incidence of perforating and non-perforating appendicitis and of removal of a normal appendix were calculated in relation to the patients' age and sex. Because of the age related rise in the proportion of people who underwent appendicectomy the population at risk of acquiring appendicitis was adjusted by subtracting the estimated number of people who had previously undergone the operation according to the cumulative age specific incidence of appendicectomy, assuming a constant rate of appendicectomy over time.
To study temporal trends we analysed the results of 4717 appendicectomies performed in the town of Jonkoping (population 108235 in 1987) in 1970-89. We chose this town because in other parts of the county the hospital register for the period was incomplete. The incidence of perforating and non-perforating appendicitis and of removal of a normal appendix, standardised for age and sex to the demographic state in 1987, were calculated for each of the four component periods of five years. No correction of discharge disgnosis was made as results from histological examinations were not available before 1984.
Association Between Laparotomy Rate And Appendicitis Rate
The influence of the rate of laparotomy on incidences of perforating and non-perforating appendicitis was analysed by reviewing previously published studies from defined populations. Diagnostic accuracy and rates of negative results from laparotomy were used as indicators of the attitude to surgical exploration in suspected appendicitis as the numbers and the proportion of operations that yielded negative results are directly related to readiness to explore.
A search of Index Medicus back to 1970 found six relevant studies.*RF 7 11-16* With our data these studies included 53 143 patients from 15 defined geographical areas who had undergone laparotomy. As some series represented more than one time period from the same area there were in total 23 sets of data from different times and places.
The X(super2) test was used for analysis of categorical data and the t test for normally distributed continuous variables. Relative risks were calculated with Taylor series 95% confidence intervals. Correlation analysis was performed with Pearson coefficient and P values of <0.05 were considered significant. Calculations were done with EPI INFO17 and STATGRAPHICS.
Relation Of Incidence Of Appendicitis To Age And Sex
From 1984 to 1989 the incidence of appendicectomy per 100 000 people in Jonkoping county was 167 and that of appendicitis 116. Appendicectomy was equally common in both sexes, whereas the incidence of appendicitis had a male to female ratio of 129:102 (relative risk 1.27; 95% confidence interval 1.16 to 1.38; P<0.0001). The incidence of perforating appendicitis was 18/100 000, with a male to female ratio of 23:14 (1.71; 1.37 to 2.14; P<0.0001). The corresponding figure for non-perforating appendicitis was 97 per 100 000, with a male to female ratio of 106:89 (1.20; 1.09 to 1.32;P<0.0002).
The incidence of appendicitis was strongly age dependent, peaking at 10-14 years. This variation, however, was mainly among cases of non-perforating appendicitis, and the incidence of perforating disease was almost stable at all ages (fig 1).
The rate of removal of a normal appendix was 51 per 100 000, with a male to female ratio of 34:69 (0.50;0.43 to 0.57; P<0.00001), reflecting the lower diagnostic accuracy in female patients (60% v 79%).
Temporal Trends, 1970-89
The incidence of appendicectomy in the town of Jonkoping fell by 29% (from 256 to 182 per 100 000, P<0.0001) between 1970-4 and 1985-9 (table I). The corresponding figure for appendicitis was 24% (from 173 to 132 per 100 000; P<0.0001). A trend towards more conservative management was indicated by a 40% drop in the number of operations to remove a normal appendix (from 84 to 50 per 100 000 population; P<0.0001) and a rise in diagnostic accuracy from 67% to 73% (P<0.01). Concomitantly the incidence of non-perforating appendicitis fell by 27% from 152 to 111 per 100 000 (P<0.00001), while that of perforating appendicitis remained stable at around 21 per 100 000. As a result the perforation rate rose from 12% to 16% (P<0.05).
Influence Of Management Policy
Among the 23 sets of data from defined places and times there were wide variations in the incidence of appendicectomy and appendicitis. The variation, however, was mainly in the incidence of non-perforating appendicitis and removal of a normal appendix, whereas the incidence of perforating appendicitis was less variable. For non-perforating appendicitis the highest rate (235 per 100 000) was reported from the Republic of Ireland11 and the lowest (11 per 100 000) from Melanesia,12 but the rate of perforating appendicitis in the Republic of Ireland was only double that in Melanesia (20 v 9 per 100 000 population).
The influence of the readiness to explore in suspected appendicitis, expressed as diagnostic accuracy and incidence of removal of a normal appendix was studied by correlation analysis (table II). The extreme results from Melanesia and the Republic of Ireland were excluded from this analysis.
The incidence of perforating appendicitis was unrelated to diagnostic accuracy and rates of removal of a normal appendix. The figures for non-perforating appendicitis, however, correlated strongly with those for removal of a normal appendix (r=0.85; P<0.0001; fig 2) and were inversely related to diagnostic accuracy (r= minus 0.59; P=0.005; fig 3).
The perforation rate was unrelated to diagnostic accuracy and to the incidence of perforating appendicitis but was inversely related to the rate of non-perforating appendicitis (r=minus 0.62; P=0.003) and removal of a normal appendix (r=minus 0.52; P=0.02).
In this study the incidence of perforating appendicitis was independent of time, place, patients' age, and rates of exploratory laparotomy in suspected appendicitis. The contrast with the widely varying incidence of non-perforating appendicitis supports Luckman's proposition that perforating and non-perforating appendicitis are separate entities,16 as do earlier observations of higher rate of obstruction and faecaliths in gangrenous and perforating appendices than in phlegmonous appendixes.18
The natural course of non-perforating appendicitis is not known. Many surgeons regard acute appendicitis as a progressive inflammation,19 but spontaneous resolution may occur(20-22) and microscopy of excised appendixes may show signs of healed inflammation.23 Recurrent appendicitis has also been reported24
Changes In Management
A high rate of exploratory laparotomy in suspected appendicitis increases the number of confirmed cases,25 presumably by adding cases of self limiting inflammation which otherwise would have escaped detection. Inflammation without symptoms, which is seen in up to 35% of incidentally removed appendixes,26 may also be erroneously diagnosed at laparotomy as appendicitis. The observed incidence of non-perforating appendicitis will therefore be influenced by a willingness to perform exploratory laparotomy in cases of suspected appendicitis.
This hypothesis is confirmed by our findings of a strong association between the incidence of non-perforating appendicitis and rate of laparotomy with a stable incidence of perforating appendicitis and indicates that appendicitis commonly resolves. The declining registered incidence of appendicitis in the town of Jonkoping and the observed geographical variations may therefore be explained by differences in the management of patients with suspected appendicitis. The increase in diagnostic accuracy in the town of Jonkoping was not the result of any new diagnostic methods.
The diagnosis of appendicitis is usually overestimated clinically at operation (in 6% according to our study). Histological confirmation, which is usually mandatory, was lacking in 17% of the cohort from Jonkoping county for 1984-9. These patients did not represent any selected subgroup as the characteristics were almost identical to the remaining cases. The error induced by this group is therefore only marginal (less than 1%).
Changes in the clinical diagnosis of appendicitis at operation could be an alternative explanation to the decreasing incidence of appendicitis as clinical diagnosis of appendicitis without histological confirmation was used in the temporal trend analysis. This trend has, however, continued after 1984 and after histological confirmation and is consistent with the unanimous reports of a decreasing incidence of appendicitis from other countries.*RF 5-7*
Removal of a normal appendix was more common in women. A lower diagnostic accuracy in women is a common finding in most studies. Gynaecological diseases may mimic appendicitis because of the proximity to the female pelvic organs, but this is probably not the only explanation.10
Goals Of Treatment
The ultimate goal in treating suspected appendicitis is to minimise the number of unnecessary laparotomies without increasing the incidence of perforated appendixes. Diagnostic accuracy may be increased by a conservative attitude to explore in uncertain cases.27 This strategy is criticised, however, for giving an increased perforation rate. Our study shows that a seemingly increased perforation rate can be explained by misclassification of spontaneously resolving cases. Reported rises in perforation rates with duration of symptoms may also be explained by selection bias due to spontaneous resolution of milder inflammation.15 The perforation rate is therefore worthless as a measure of quality in the management of appendicitis.
That appendicitis commonly resolves implies that future diagnostic and therapeutic policies should aim at early detection and operation in patients with perforating appendicitis or progressive inflammation and at active observation and investigation of alternative diagnoses in other patients. Studies on clinical signs and laboratory findings in relation to the degree of inflammation or perforation have shown promising results.28 29 The role of explorative laparoscopy in this context is as yet unclear but may increase the yield of alternative diagnoses in prolonged abdominal pain.30
We thank Mrs R McVeigh Kallqvist for linguistic revision. The study was supported by a grant from the committee for research and development, Jonkoping County Council.