Dr J. Y. Kang, Department of Gastroenterology, St George's Hospital, Blackshaw Road, London SW17 0QT, UK. E-mail: firstname.lastname@example.org
Background: Diverticular disease emerged as a common problem in Western countries over the course of the 20th century.
Aims: To determine the time trends in diverticular disease for hospital admissions in England between 1989/1990 and 1999/2000 and population mortality rates from 1979 to 1999.
Methods: Hospital Episode Statistics were obtained from the Department of Health and mortality data from the Office for National Statistics.
Results: Between 1989/1990 and 1999/2000, annual age-standardized hospital admission rates for diverticular disease increased by 16% for males (from 20.1 to 23.2 per 100 000) and 12% for females (from 28.6 to 31.9 per 100 000). Female rates were significantly higher than male rates throughout the study period. The proportions of admissions with an operation increased by 16% for males (from 22.9% to 24.1%) and 14% for females (from 19.7% to 22.3%). Older patients were less likely to undergo operation than younger patients. In-patient case fatality rates and population mortality rates remained unchanged.
Conclusions: Admission rates for diverticular disease increased over the study period. The proportion of patients who underwent operation increased, but in-patient and population mortality rates remained unchanged. With an ageing population, diverticular disease will become an increasingly important clinical problem in England.
Although acquired diverticular disease was first described as far back as 1700, it remained an uncommon problem until the turn of the 20th century.1,2 Over the following few decades, mortality rates rose several-fold and the condition became increasingly detected during life, at least in Western countries.1–3 However, these time trends may not necessarily be secondary to a marked and continual rise in the real prevalence of diverticular disease. An ageing population, increasing use of barium enema examinations and changing fashions in death certification could be contributory factors.4 The epidemiological, pathological and clinical aspects of diverticular disease have been well studied. However, relatively little is known about the time trends since the 1980s.3 In this study, we analysed hospital admission, operation and case fatality rates for diverticular disease in England from 1989/1990 to 1999/2000, and population mortality rates from 1979 to 1999.
Two main sources of data were used for this study: Hospital Episode Statistics5 and mortality statistics for England.
The Hospital Episode Statistics database, administered by the Statistics Division at the Department of Health, provides information on in-patient care delivered by National Health Service hospitals in England from April 1987 onwards. In the 1999/2000 financial year, the Hospital Episode Statistics system collected nearly 12 million records detailing episodes of in-patient treatment delivered in National Health Service hospitals in England. Data for the financial years 1998/1999 and 1999/2000 have not yet been adjusted to account for shortfalls in the number of records submitted or for missing or invalid clinical information. Although private hospitals are not covered, the Hospital Episode Statistics database includes private patients treated in National Health Service hospitals.
The data recorded include consultant episodes, defined as ‘the period during which an admitted patient is under the care of a particular medical consultant within a hospital provider’. Each consultant episode generates one record. Data extracted for the purpose of this study are finished consultant episodes, each ended by internal transfer within the hospital, discharge or death. As patients can have more than one episode during an admission, only first episodes were used in this study as a proxy for the number of admissions. The International Classification of Diseases Ninth Revision (ICD9)6 was used until April 1995, and ICD107 from this date onwards.
Hospital admission rates
Our study examines the time trends in hospital admissions for diverticular disease. We identified all first consultant episodes in which the main diagnosis was colonic diverticular disease, diverticulitis and diverticulosis (ICD9 562.1 and ICD10 K57.2–57.9) for the financial years 1989/1990–1999/2000. Day case admissions were excluded.
We obtained admission rates per 100 000 population by age and sex for each year in the study period by dividing the annual number of admissions for the diagnosis by the England mid-year population estimates. As the admissions data are presented in financial years and the population estimates in calendar years, we used the population estimates for the calendar year overlapping 9 months of any given financial year to calculate the admission rates. For example, to calculate the hospital admissions rate in 1989/1990, we divided the number of admissions in 1989/1990 by the 1989 mid-year population estimates.
Age-standardized rates allow us to assess the true change in hospital admission rates over time, and the difference between the sexes, by adjusting for differing age distributions both between the sexes and over the period 1989/1990–1999/2000. We calculated age-standardized rates by applying the age-specific rates, by 5-year age groups up to age 84 years and then 85 years and over, to the European standard population.
We identified all hospital admissions in which a surgical operation, excluding endoscopic procedures (G14–G19, G43–G47, G54–G57, G64–G67, G79–G82, J24–J25, J38–J45, H20–H30),8 was performed. As previously, we omitted day case admissions. We then calculated the percentage of admissions in which an operation was performed by age and sex. We also calculated the age-standardized percentage of admissions including an operation by applying the age-specific rates in the same age groups as previously. However, to calculate the operation age-standardized rates, we used the age distribution of hospital admissions for diverticular disease to determine the age distribution of the standard population.
Case fatality rates
We identified all hospital admissions for diverticular disease ended by death as opposed to discharge or internal transfer. We calculated case fatality rates by age and sex as the percentage of admissions ending in death. Age-standardized rates were then calculated using the same standard population as for the operation rates.
Population mortality rates
We obtained mortality statistics for England from the Office for National Statistics. These are recorded in calendar years and, since January 1979, have been coded using ICD9.6 Deaths due to diverticular disease from 1979 to 1999 were identified for both males and females and presented as age-standardized rates per 1 000 000 persons.
Trends over time
For each series of data, we obtained 95% confidence intervals for the age-standardized rates in every analysis year. Subsequently, for data series that exhibited an upward or downward trend throughout the study period, we examined whether the confidence intervals in the first and last analysis years overlapped. If the confidence intervals did not overlap, we concluded that there was a statistically significant increase (or decrease) in the rates between the beginning and end of the study period.
Between 1989/1990 and 1999/2000, the age-standardized hospital admission rate for diverticular disease in England increased significantly from 25.1 to 28.2 per 100 000 population (Table 1). The admission rate amongst females was higher than that amongst males throughout the study period, and increased by 12%, from 28.6 per 100 000 in 1989/1990 to 31.9 per 100 000 in 1999/2000. The male admission rate increased by 16%, from 20.1 to 23.2 per 100 000. The male rate increased most over the final 3 years of the study period.
Table 1. Age-standardized hospital admission rates for diverticular disease of the colon per 100 000 population, by sex, 1989/1990–1999/2000
Age-specific hospital admission rates for diverticular disease were very low below an age of 35 years (less than 2 per 100 000), but increased dramatically with age to 250.9 and 314.2 per 100 000 at ≥ 85 years for men and women, respectively (Figure 1). The female age-specific admission rate was greater than the male rate for age groups of 45 years and above.
The age-standardized percentage of admissions for diverticular disease with an operation increased significantly between 1989/1990 and 1999/2000, by 16.3% for males and by 13.5% for females (Figure 2). In 1999/2000, 26.6% of men and 22.3% of women admitted for diverticular disease underwent an operation, and this percentage was higher in men than in women in each year of the study period.
Younger patients were more likely than older patients to undergo an operation following admission for diverticular disease. In 1999/2000, the percentage of admissions with an operation decreased with age from 54.2% in the 25–34-year age group to 13.4% in the ≥ 85-year age group (Figure 3).
Case fatality rates
In contrast with the hospital admission rate and operation rate for diverticular disease, the age-standardized case fatality rate did not increase significantly between 1989/1990 and 1999/2000 (Figure 4). Overall rates fell from 3.2% in 1989/1990 to 2.6% in 1991/1992 and then increased, reaching 3.4% in 1999/2000. From 1994/1995 onwards, female case fatality rates were marginally higher than male rates.
Population mortality rates
Although there were fluctuations, age-standardized mortality rates for diverticular disease did not change significantly for either sex from 1979 to 1999 (Figure 5). Female mortality rates were consistently higher than male rates.
There are several inherent problems in assessing the frequency of diverticular disease. Many subjects are asymptomatic. Symptoms, when they occur, are neither sensitive nor specific. There are no pathognomonic physical signs. The diagnosis of diverticular disease in life is therefore dependent on imaging studies, i.e. barium enema or colonoscopy, or is made at surgery or autopsy. Necropsy studies tend to include older individuals and are usually undertaken by interested pathologists who are hence more likely to identify the condition. Colonic imaging is usually undertaken only in symptomatic individuals, and hence imaging studies tend to overstate the disease prevalence.
The frequency of diverticular disease increases with age. Although it is uncommon below the age of 40 years, more than 50% of subjects above the age of 80 years are found to be affected at necropsy.4,9,10 The reported prevalence in barium enema studies ranges from 2% to 45%, depending on the patient population and age groups studied. About 5% of patients above the age of 40 years have diverticular disease, rising to one-third for individuals above the age of 60 years,2 and two-thirds for those aged over 85 years.11 Although many authorities feel that diverticular disease has increased in frequency over recent decades,2 the effect of confounding factors, such as increasing awareness of the condition and more frequent use of diagnostic investigations, cannot be discounted. The prevalence of diverticular disease in the general population is unknown. In a small series of symptom-free volunteer subjects in Oxford, 19% of individuals between the ages of 40 and 59 years, 29% of those aged between 60 and 79 years, and 42% of subjects aged above 80 years were found to be affected.12 A male preponderance of between 1.6 : 1 and 3 : 1 was reported in studies up to the 1930s, but subsequent reports from the 1950s and 1960s have shown higher prevalence rates amongst females.13
A recent task force convened by the American Gastroenterological Association confirmed that diverticular disease is a major clinical problem. In 1998, there were 2.2 million cases in the USA,14 and total health care costs came to $2358 million. Diverticular disease was fifth in the list of digestive diseases, after gastro-oesophageal reflux disease, gall-bladder disease, colorectal cancer and peptic ulcer disease. It was ahead of chronic liver disease, irritable bowel syndrome and inflammatory bowel disease.14
There are ethnic and geographical variations in diverticular disease, the condition being less frequent in underdeveloped countries.4 Westerners living in the East are affected more than the indigenous population. In a study involving four teaching hospitals in Scotland, Singapore, Nigeria and Fiji, the annual admission rate for colonic diverticulitis was 12.88 per 100 000 population in Scotland compared with 0.17 per 100 000 population in Nigeria.15 In Singapore, the annual admission rate was 0.13 per 100 000 population for indigenous people and 5.41 per 100 000 population for Europeans. Similarly, in Fiji, annual admission rates were 0.34 per 100 000 for Indian Fijians, 0.21 per 100 000 for non-Indian Fijians and 7.62 per 100 000 for Europeans.
Diverticular disease in Western countries affects the sigmoid and distal descending colon predominantly, whereas right-sided disease is more common in Asia.16 Painter and Burkitt suggested that left-sided diverticular disease in Western populations is linked to the decreased consumption of dietary fibre.17 This leads to constipation and increases the intraluminal pressure in the sigmoid colon, segmentation and herniation of the mucosa through the muscle in areas of weakness.1 In contrast, caecal diverticula are traditionally thought to be congenital in origin, with all muscular coats in their walls.9
The most common complication is acute diverticulitis, presenting with abdominal pain, sometimes associated with perforation, intestinal obstruction, fistula and abscess formation. Lower gastrointestinal tract bleeding from diverticula is not associated with acute underlying inflammation, but rather with erosion of the submucosal blood vessels at the neck of a diverticulum.1
Although the lifetime prevalence of diverticulitis has been stated to be 10–25% amongst subjects with diverticular disease,1 available data are limited. Kyle and Davidson compared hospital admission rates for colonic diverticulitis in north-east Scotland during 1958–1961 and 1968–1971 and showed an increase from 12.8 to 23.5 per 100 000 population per year.18 However, as the operation and mortality rates declined, the increased admission rates could simply have resulted from the admission of patients with milder disease. Makela et al. found that the incidence of admissions for perforated sigmoid diverticular disease in Finland increased from 2.4 per 100 000 population in 1986 to 3.8 in 2000.19 The overall resection rate was 90% and the mortality rate was 9%. Hart et al., in Norwich, UK, also reported an incidence rate of perforated diverticular disease of 4.0 per 100 000 per year.20
Our data indicate that hospital admission rates in England for diverticular disease increased in the last decade of the 20th century. Although female admission rates were consistently higher than male rates, admission rates for males increased more than those for females. The proportion of male patients undergoing an operation was about 15% higher than that for females, although operation rates increased for both sexes over the period of the study. Hospital case fatality rates were slightly higher in females from 1994/1995 onwards, but overall remained unchanged over the study period. Mortality rates for diverticular disease remained stable in England from 1979 to 1999. Rates of approximately 22 deaths per 1 000 000 women per year and 13 deaths per 1 000 000 men per year approximate to the US figures of 1.74 and 0.76 deaths per 100 000 per year for women and men, respectively.14
Nomenclature is a problem. We have used the ICD codes for colonic diverticular disease, diverticulosis and diverticulitis. These diagnoses are often used interchangeably, especially by non-specialists. Depending on the hospital, the diagnostic coding could be made by the consultant, junior doctor or coding clerk. With the use of Hospital Episode Statistics, incomplete recording of admissions is another potential problem.21 However, as there has not been any significant change in either the recording process or disease classification over the study period, the completeness of recording should have remained broadly unaltered and our study of time trends should be valid.
Factors other than disease incidence can affect the number of hospital admissions. However, the increase in admission rates demonstrated in our study is unlikely to be due to a lower threshold for admission, as operation rates have increased, whereas case fatality rates have remained stable. A further possibility is that more diagnostic studies have been performed in recent years, and hence patients with abdominal symptoms may have been more likely to be diagnosed with diverticular disease. However, no major change in the use of barium studies and colonoscopy occurred during the decade in question. Repeat admissions, e.g. for staged surgery, would be considered as separate admissions for the purpose of this study, but the frequency of staged procedures would not have changed greatly over the study period. Indeed, the advent of computerized tomography and radiological drainage would, if anything, have resulted in fewer multi-stage procedures being performed.1 Our data on time trends are age-standardized, and hence not caused by an ageing population.
As there is an association between the use of aspirin and non-steroidal anti-inflammatory drugs on the one hand, and perforated diverticular disease on the other,22 the increasing use of low-dose aspirin for cardiovascular prophylaxis23 may be one reason for the rising incidence of perforated diverticular disease. However, our study is ecological, involving populations rather than individual patients, and so conclusions cannot be drawn on causality.24
In conclusion, admission rates for diverticular disease increased steadily in England over the last decade of the 20th century, especially amongst males. As it is predominantly a disease of the elderly, diverticular disease will become an increasingly important health burden in this country.
The authors would like to thank the Hospital Episode Statistics team at the Department of Health for providing the admissions data and the mortality team at the Office for National Statistics for the mortality data. Drs A. R. Hart and C. R. Morris of the University of East Anglia gave helpful advice. The findings of this study were presented at the British Society of Gastroenterology Annual Meeting in March 2002.