AIM: To establish the hospitalized prevalence of severe Crohn’s disease (CD) and ulcerative colitis (UC) in Wales from 1999 to 2007; and to investigate long-term mortality after hospitalization and associations with social deprivation and other socio-demographic factors.

METHODS: Record linkage of administrative inpatient and mortality data for 1467 and 1482 people hospitalised as emergencies for ≥ 3 d for CD and UC, respectively. The main outcome measures were hospitalized prevalence, mortality rates and standardized mortality ratios for up to 5 years follow-up after hospitalization.

RESULTS: Hospitalized prevalence was 50.1 per 100 000 population for CD and 50.6 for UC. The hospitalized prevalence of CD was significantly higher (P < 0.05) in females (57.4) than in males (42.2), and was highest in people aged 16-29 years, but the prevalence of UC was similar in males (51.0) and females (50.1), and increased continuously with age. The hospitalized prevalence of CD was slightly higher in the most deprived areas, but there was no association between social deprivation and hospitalized prevalence of UC. Mortality was 6.8% and 14.6% after 1 and 5 years follow-up for CD, and 9.2% and 20.8% after 1 and 5 years for UC. For both CD and UC, there was little discernible association between mortality and social deprivation, distance from hospital, urban/rural residence and geography.

CONCLUSION: CD and UC have distinct demographic profiles. The higher prevalence of hospitalized CD in more deprived areas may reflect higher prevalence and higher hospital dependency.

Inflammatory bowel disease (IBD), comprising mainly Crohn’s disease (CD) and ulcerative colitis (UC), can be one of the most severe gastrointestinal disorders with a significant health and economic burden[1]. Presenting commonly in adolescence, IBD often results in debilitating morbidity and frequent relapses that impact heavily on educational, social, professional and family life[1-3].

CD has increased in many countries in the past few decades, with highest prevalence often reported in western regions such as Northern and Western Europe and North America[3-7]. In the United Kingdom, the incidence of UC has shown greater stability over time[6,8], while CD has continued to rise in many but not in all regions[7,9-13]. Most studies have reported that CD is more common in females than in males[8,10,11,13-16], whereas for UC, higher risks have been reported in females[3,17,18] or in males[2,19], as well as similar risks for females and males[20,21].

In the United Kingdom, the prevalence of IBD is not known accurately, although based on regional studies, it is thought to affect approximately 150 000 people[1]. Of these, the majority have active IBD, with many having severe IBD that may lead to substantial inpatient stays. Our study focused on severe IBD, which we define as an emergency admission to hospital with a length of stay of ≥ 3 d, because we wanted to study people who were acutely ill with severe IBD. Inclusion of all patients admitted to hospital, regardless of length of stay, would have covered a wide spectrum of disease severity, from elective day case investigation of IBD and admissions for elective surgery, to emergency care of patients who were severely ill with IBD.

Little is known about the prevalence of severe IBD in the United Kingdom, how it varies geographically, and its relationship with social deprivation. Mortality following hospitalization for IBD is significant[22], but few studies have been based on long-term follow-up. Additionally, little is known about the effect on mortality of social deprivation and other socio-demographic factors such as area of residence and distance from hospital. These are important aspects of the epidemiology of IBD in the United Kingdom, which are important to understand in planning service provision and treatment.

The main objective of this study was to establish the prevalence of UC and CD that was severe enough to require admission to hospital for ≥ 3 d, and to report on subsequent mortality after up to 5 years follow-up. Wales has a population of 3 million, and a mixed urban and rural geography, with areas of significant social deprivation. Further objectives were to assess geographical variation in the hospitalized prevalence and mortality for CD and UC, and to investigate associations with social deprivation and other socio-demographic characteristics, including any association between mortality and distance of patients’ residence from hospital and urban or rural residence.

For this study, we used hospital inpatient data from the Patient Episode Database for Wales (PEDW), which incorporates medical record linkage to mortality data from the National Health Service Administrative Register[23]. PEDW is a comprehensive administrative database that covers all inpatient and day-case admissions to every National Health Service (NHS) hospital in Wales, and has been used as the basis for previously published studies[24-27]. It covers 17 district general or teaching hospitals in eight NHS hospital trusts across 22 health boards, and serves the Welsh population of 3 million. We used record linkage of inpatient and mortality data because this enabled all repeat hospital admissions to be identified for the same patients, and enabled deaths following discharge from hospital to be identified, as well as in-hospital deaths.

The International Classification of Diseases, 10th Edition (ICD-10) codes used were: K50 for CDand K51 for UC. Patients with an emergency hospital admission of three or more consecutive days were classified as severe cases and included in the study. We selected only those admissions for which CD or UC was recorded as the principal diagnosis on the discharge record, and selected only the first admissions for each patient that met the study inclusion criteria during the 8-year study period from 1999 to 2007. We excluded all elective admissions, all day cases and inpatient admissions lasting < 3 d, all admissions subsequent to the first admission, and all admissions for which CD and UC were recorded on the discharge record as secondary diagnoses to other diseases such as colorectal malignancy.

Hospitalized prevalence rates for CD and UC were calculated per 100 000 resident population, using the numbers of severe hospitalized cases as the numerators and the Welsh residential population as the denominators. The direct method and the standard European population were used to standardize the prevalence rates. Mortality rates following hospital admission were calculated using the numbers of deaths (from all causes) as the numerators, and the total number of hospitalized cases, in the same time periods, as the denominators. Mortality rates were expressed as percentages and calculated at 1 and 5 years follow-up from the date of hospital admission.

To establish how mortality among the CD and UC patients compared with that in the general population, we used standardized mortality ratios (SMRs). These were calculated by applying the age- and sex-specific mortality rates in the general Welsh population to the numbers of study patients in the corresponding age- and sex-specific strata, to calculate the expected number of deaths in the study patients, and by comparing the observed deaths with the expected deaths.

Social deprivation scores were assigned to the patients’ place of residence on each hospital record by using residential postcodes at the Lower Super Output Area (LSOA) and the Welsh Index of Multiple Deprivation (WIMD) 2005[28]. The average population size of the 1896 LSOAs in Wales in 2005 was 1560. The WIMD 2005 was based on seven domains of social deprivation: income, housing, employment, environment, health, access to services, and education. The LSOAs were ranked according to their WIMD 2005 score and grouped into quintiles, with quintile I representing the 20% of least deprived LSOAs and quintile V the 20% of most deprived LSOAs in Wales.

Urban/rural residences were measured using five categories, based on settlement sizes that ranged from < 2500 to > 100 000 people. Distances from patients’ residential LSOAs (geographical center points) to their hospital of admission were calculated using ArcGIS Geographical Information System (GIS) software and were measured in five categories that ranged from < 3 to > 20 km. Other methods used included logistic regression analysis, χ² tests, and Pearson’s correlations.

From 1999 to 2007, there were 1467 patients hospitalized for ≥ 3 d following emergency admission for CD (mean ± SD, age: 45.8 ± 20.4 years) and 1482 for UC (mean ± SD, age: 52.0 ± 21.0 years). All the results that follow refer to these patients who meet the definition of severe IBD, although, for brevity, we omitted the repeated use of the term severe.

Hospitalized prevalence of CD and UC

The overall hospitalized prevalence rate per 100 000 population was 50.1 for CD and 50.6 for UC (Table 1). Geographically, across the 22 health boards in Wales (Table 1), the prevalence of Crohn’s varied from 31.5 per 100 000 (95% CI: 22.6-42.7) to 71.3 (59.3-84.9), whereas for UC, it varied from 38.2 per 100 000 (29.6-48.5) to 89.7 (66.6-118.3).

Health board	Map1	No. of LSOAs	Deprivation score (mean ± SD)	CD			UC
				Cases	Prevalence (per 100 000)	95% CI	Cases	Prevalence (per 100 000)	95% CI
Blaenau Gwent	1	47	33.8 ± 12.7	31	44.7	(30.4-63.5)	27	39.0	(25.7-56.7)
Bridgend	2	85	23.2 ± 13.0	41	31.5	(22.6-42.7)	62	47.6	(36.5-61.0)
Caerphilly	3	110	27.0 ± 14.1	97	57.0	(46.2-69.5)	91	53.5	(43.0-65.6)
Cardiff	4	203	21.3 ± 18.0	158	50.9	(43.3-59.5)	137	44.1	(37.1-52.2)
Carmarthenshire	5	112	22.0 ± 10.3	125	71.3	(59.3-84.9)	67	38.2	(29.6-48.5)
Ceredigion	6	47	15.0 ± 6.2	35	46.0	(32.0-63.9)	30	39.4	(26.6-56.2)
Conwy	7	71	18.0 ± 11.0	64	57.8	(44.5-73.8)	84	75.9	(60.6-94.0)
Denbighshire	8	58	20.4 ± 15.6	56	59.0	(44.6-76.7)	56	59.0	(44.6-76.7)
Flintshire	9	92	15.0 ± 10.2	74	49.6	(38.9-62.2)	77	51.6	(40.7-64.5)
Gwynedd	10	75	17.3 ± 7.4	63	53.7	(41.2-68.7)	70	59.6	(46.5-75.4)
Isle of Anglesey	11	44	20.7 ± 8.8	43	63.0	(45.6-84.9)	43	63.0	(45.6-84.9)
Merthyr Tydfil	12	36	37.2 ± 17.0	26	46.7	(30.5-68.4)	50	89.7	(66.6-118.3)
Monmouthshire	13	58	12.2 ± 6.0	37	42.9	(30.2-59.1)	44	51.0	(37.1-68.5)
Neath Port Talbot	14	91	28.2 ± 14.4	76	56.0	(44.1-70.0)	53	39.0	(29.2-51.0)
Newport	15	94	22.0 ± 15.6	73	52.4	(41.1-65.9)	71	51.0	(39.8-64.3)
Pembrokeshire	16	71	19.4 ± 8.9	44	38.0	(27.6-51.0)	47	40.6	(29.8-54.0)
Powys	17	80	14.3 ± 6.1	44	34.2	(24.9-46.0)	62	48.2	(37.0-61.8)
Rhondda Cynon Taff	18	152	29.2 ± 15.4	116	50.0	(41.3-60.0)	125	53.9	(44.9-64.2)
Swansea	19	147	22.6 ± 17.8	101	44.9	(36.6-54.6)	109	48.5	(39.8-58.5)
Torfaen	20	60	21.3 ± 10.8	46	50.6	(37.1-67.5)	51	56.1	(41.8-73.8)
Vale of Glamorgan	21	78	15.0 ± 10.5	49	40.7	(30.1-53.7)	74	61.4	(48.2-77.1)
Wrexham	22	85	19.9 ± 13.5	68	52.5	(40.8-66.5)	52	40.2	(30.0-52.6)
All Wales		1896	21.7 ± 14.3	1467	50.1	(47.5-52.7)	1482	50.6	(48.0-53.2)

¹For geographical details of each health board, see Figure 1. CD: Crohn's disease; UC: Ulcerative colitis; LSOA: Lower Super Output Area.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 WIMD 2005 quintiles for the 22 health boards and the 1896 geographical LSOAs in Wales. The key to the health board numbering is as follow: 1 Blaenau Gwent; 2 Bridgend; 3 Caerphilly; 4 Cardiff; 5 Carmarthenshire; 6 Ceredigion; 7 Conwy; 8 Denbighshire; 9 Flintshire; 10 Gwynedd; 11 Isle of Anglesey; 12 Merthyr Tydfil; 13 Monmouthshire; 14 Neath Port Talbot; 15 Newport; 16 Pembrokeshire; 17 Powys; 18 Rhondda Cynon Taff; 19 Swansea; 20 Torfaen; 21 The Vale of Glamorgan; 22 Wrexham.

For UC, there was a significant association between hospitalized prevalence and age (P < 0.05, Table 2). The hospitalized prevalence of UC increased with age, and was highest for males and females aged > 65 years (103 and 93.4 per 100 000, respectively). There was no significant association between the hospitalized prevalence of CD and age. The highest age-specific prevalence of CD was in the 16-29 years age group, for both males (61.4) and females (76.0). Overall, the hospitalized prevalence of CD was significantly higher (P < 0.05) in females (57.4) than in males (42.2), whereas the hospitalized prevalence of UC was very similar in females (50.1) and males (51.0, Table 2).

		CD			UC
		Cases	Prevalence (per 100 000)	95% CI	Cases	Prevalence (per 100 000)	95% CI
Males	All	601	42.2	(38.9-45.8)	726	51.0	(47.4-54.9)
	0-15	25	8.4	(5.5-12.5)	20	6.7	(4.1-10.4)
	16-29	151	61.4	(52.0-72.0)	115	46.7	(38.6-56.1)
	30-44	141	47.6	(40.1-56.1)	168	56.7	(48.5-66.0)
	45-64	163	44.5	(38.0-51.9)	199	54.4	(47.1-62.5)
	65+	119	54.5	(45.2-65.3)	224	103	(89.7-117)
Females	All	866	57.4	(53.7-61.4)	756	50.1	(46.6-53.8)
	0-15	21	7.5	(4.6-11.4)	11	3.9	(2.0-7.0)
	16-29	185	76.0	(65.5-87.8)	114	46.9	(38.6-56.3)
	30-44	232	74.3	(65.1-84.5)	166	53.2	(45.4-61.9)
	45-64	222	58.9	(51.4-67.2)	189	50.1	(43.3-57.8)
	65+	205	69.4	(60.2-79.6)	276	93.4	(82.7-105)
All Wales		1467	50.1	(47.5-52.7)	1482	50.6	(48.0-53.2)

The ages of three Crohn’s patients, two males and one female, was unknown.

Figure 1 shows the geographical distribution of the social deprivation quintiles to the 1896 LSOAs across Wales. The hospitalized prevalence of CD was associated significantly with social deprivation at the LSOA level (Pearson’s correlation = 0.07, P = 0.004), although the association was not strong. The prevalence of CD tended to be higher in more deprived areas, although there were not quite significant differences between the most and least affluent quintiles (Figure 2). For UC, there was no significant association between prevalence and social deprivation at the LSOA level, or across any of the five social deprivation quintiles (Figure 2).

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Hospitalized prevalence of severe Crohn’s disease and ulcerative colitis according to social deprivation, 1999-2007. I: Least deprived social deprivation quintile; V: Most deprived. Vertical bars represent 95% CI. These prevalence rates are standardised for age group and gender.

The main aims of this record linkage study were to establish the extent and prognosis of severe CD and UC in Wales, and how they are associated with social deprivation, geography and other socio-demographic aspects. It is thought that IBD affects approximately 150 000 people in the United Kingdom, with corresponding prevalence rates of 55-140 per 100 000 for CD, 160-240 per 100 000 for UC, and approximately 13 300 combined new cases diagnosed each year[1]. We found a hospitalized prevalence rate of 50.1 per 100 000 population for severe CD and 50.6 for severe UC. If these prevalence rates for Wales were applied to the whole of the United Kingdom, this would suggest a total of approximately 62 000 people hospitalized with severe IBD over the 9-year study period (30 000 for CD and 32 000 for UC).

The main strength of this study was its size, which covered almost 3000 patients who were hospitalized for IBD in a geographically defined population of nearly 3 million over a 8-year period. Further strengths are that it was based on systematic record linkage, which enabled long-term mortality follow-up of 5 years, by monitoring deaths that occurred in and following discharge from hospital. The study also used geographical measures of social deprivation to measure social inequality comprehensively on a national scale, as well as other novel prognostic/demographic risk factors, such as distance from home to hospital and urban or rural residences.

We defined severe IBD as emergency hospitalization for IBD, as the principal reason for admission, which lasted ≥ 3 d. This was to distinguish severe illness from relatively minor illness and from day-case investigations. We accept that the definition may also have included some patients who were admitted with long stays for other serious illnesses, for example, myocardial infarction or pneumonia, who also had IBD. However, we sought to minimize this by restricting case selection to people in whom IBD itself was coded as the principal reason for admission.

The main limitation of record linkage studies such as this is the lack of detailed information about the history, severity and management of the illness. It was not possible to ascertain the extent of disease from endoscopic evidence or detailed pathology information from our data sources. Our hospitalized prevalence rates referred specifically to CD and UC patients with an inpatient stay of ≥ 3 d following emergency admission. Although length of stay is often affected by factors such as age, comorbidity, and effectiveness of self care and social support, we believe that a length of stay of ≥ 3 d after emergency admission usually indicates severe disease that requires intensive medical or surgical treatment rather than admission for diagnosis or assessment.

For both males and females, we found a continuous rise in the hospitalized prevalence of UC as age increased. For CD, we found highest prevalence among patients aged 16-29 years, with a second peak in prevalence in males aged > 65 years and females aged ≥ 60 years. We also found a higher prevalence of CD in females than in males, but a similar prevalence of UC among males and females, which is consistent with other studies[8,10,11,13,14,16,19,29].

IBD has been linked with higher socioeconomic groups in some studies, particularly for CD[30,31], but also for UC[31], although other studies have reported no association[32-34]. Instead, we found slightly higher hospitalized prevalence of CD in more socially deprived locations, although there were not quite significant differences across the social deprivation quintiles. As our study focused on more severe cases, possible reasons for this include higher levels of health-care dependency as well as higher levels of smoking and more comorbidity among patients from more deprived social backgrounds. For UC, we found no significant association between hospitalized prevalence and social deprivation. The differing demographic profiles between CD and UC, and the stronger evidence of an association between CD and social deprivation suggests that environmental factors play a stronger part in the etiology of CD, but attempts to establish environmental factors have proved inconclusive and sometimes controversial in the past[31,33-37].

Overall mortality was 6.8% at 1 year and 14.6% at 5 years for CD, compared with 9.2% at 1 year and 20.8% at 5 years for UC. At 5 years, mortality was high among older patients (49% and 54% for CD and UC, respectively, among patients aged ≥ 65 years). The higher mortality for UC is consistent with other studies[16,22,38], although there is a lack of large studies with long-term follow-up for people hospitalized with IBD. Mortality at 1 year follow-up was five times higher than in the general population for CD and four times higher for UC. At 5 years, mortality was increased two fold for CD and UC. An earlier study has reported SMRs of 3.2 and 2.4 after 3 years follow-up for people hospitalized with severe CD and UC, respectively, in England[22], and further demonstrates that people hospitalized with severe IBD are often at increased risk of mortality, which continues in the long term. Elective colectomy may well be the most effective means of reducing these excess risks of mortality[22].

Although there was some variation in prognosis geographically across health boards, it was not significant for either CD or UC. We also found little discernible association between social deprivation quintiles, distance travelled to hospital and rural versus urban residence, and subsequent mortality at either 1 or 5 years. As IBD often relapses and requires regular monitoring and surveillance, it is possible that patients who live remotely from outpatient clinics and hospital care may have poorer outcomes. Our findings for mortality outcomes do not support such concerns, although recent health service developments in the United Kingdom that encourage greater management and treatment of IBD in primary care may partially offset the distances travelled by patients.

We found that severe CD and UC had quite distinct demographic profiles in Wales. CD and UC also showed slightly differing patterns of association with social deprivation. The higher prevalence of hospitalized CD in more deprived areas may reflect higher smoking rates, more comorbidity and higher levels of hospital dependency, as well as higher prevalence.