OUP user menu

Journal of Crohn's and Colitis: 10 (10)


Laurence J. Egan, Ireland

Associate Editors

Shomron Ben-Horin, IsraelSilvio Danese, ItalyPeter Lakatos, HungaryMiles Parkes, UKJesús Rivera-Nieves, USABritta Siegmund, GermanyGijs van den Brink, NLSéverine Vermeire, Belgium


Published on behalf of

Patients with Crohn's disease experience reduced general health and vitality in the chronic stage: Ten-year results from the IBSEN study

Marte Lie Høivik, Tomm Bernklev, Inger Camilla Solberg, Milada Cvancarova, Idar Lygren, Jørgen Jahnsen, Bjørn Moum
DOI: http://dx.doi.org/10.1016/j.crohns.2011.10.001 441-453 First published online: 1 May 2012


Background and aims: Data on the long-term effects of Crohn's disease (CD) on health-related quality of life (HRQoL) is scarce. We aimed to determine the HRQoL in CD patients 10 years after disease onset, to compare the results to the general population and to identify variables that could affect HRQoL.

Methods: CD patients from a population-based inception cohort (the IBSEN Study) met at a prescheduled ten-year follow-up. In addition to a structured interview, review of hospital records, clinical examination, laboratory tests and ileocolonoscopy, they completed a patient-reported questionnaire including the Short Form 36 (SF-36) and the Norwegian Inflammatory Bowel Disease Questionnaire (N-IBDQ). The SF-36 scores were compared to scores from the general population using one-sample t-tests. Standardized scores were calculated and interpreted according to Cohen's effect size index. The associations between relevant clinical and demographic factors and HRQoL were examined through linear regression analyses.

Results: Ninety-nine patients completed the HRQoL questionnaires (response rate 86%). Median age 39 years, 42% women. Compared to the general population the patients reported significantly lower SF-36 scores on the general health and vitality dimensions. IBDQ total scores were in line with scores of patients in remission. Except for current symptom severity no clinical parameters affected HRQoL scores. Work status and sick leave affected HRQoL negatively.

Conclusions: In this chronic stage of CD, reduced general health and vitality scores need attention while reductions in disease specific HRQoL seem to be less predominant.

  • Crohn's disease (CD)
  • Health-related quality of life (HRQoL)
  • Quality of life (QoL)
  • Population-based
  • Minimal clinically important difference (MCID)
  • Epidemiology

1 Introduction

Crohn's disease (CD) is a chronic inflammatory condition that can affect all segments of the alimentary tract. Disease onset occurs most often at a young age. The disease course is unpredictable; it can be complicated with intestinal strictures, fistulas and extraintestinal manifestations.1 There is no cure, and treatment may include both long term immunosuppressive medication and repeated surgery.2 However, overall mortality is not significantly increased compared to the general population.3 Consequently, patients often live with symptoms and uncertainty related to the disease for several decades. Knowledge about how patients perceive the impact of the disease is important for patients and health professionals. Health-related quality of life (HRQoL) is one patient-reported measure that can be used to examine these dimensions of a disease.

Previous studies have shown that HRQoL is reduced in CD patients compared to the general population 4,5 and that women have lower HRQoL scores than men.6,7 It has also been shown that disease distribution does not affect HRQoL, 8 whereas disease activity reduces HRQoL significantly.9,10 Both surgery and medication seem to improve HRQoL if disease activity is reduced.11 Work status appears to be an important determining factor for HRQoL.12

Recent reports have shown that the long-term prognosis in terms of clinical outcomes of CD is better than expected.13 However, there is a lack of population-based and long-term HRQoL studies on CD.14 Thus, we do not know how HRQoL is affected over time, which is an important aspect of the chronicity of the disease and of the description of the natural history of CD.

The IBSEN study (Inflammatory Bowel in South-Eastern Norway) is a large, population-based inception cohort study designed to follow newly diagnosed IBD patients prospectively to describe the natural course of the disease. In addition to comprehensive clinical data, the study focuses on HRQoL. Data from the study provide the opportunity to determine HRQoL in CD patients 10 years after disease onset.

2 Aim

The primary aim of this study was to determine HRQoL in CD patients after a ten-year disease course and to compare the results to the general population. Further, we aimed to identify demographic and clinical variables that could affect patients' HRQoL after a ten-year disease course.

3 Materials and methods

3.1 Study design and study population

From January 1, 1990 to December 31, 1993, all newly diagnosed cases of IBD or possible IBD in four well-defined areas in south-eastern Norway (the counties of Oslo, Østfold, Telemark and Aust-Agder) were registered prospectively; 843 patients were included in this inception cohort. The cohort has been followed comprehensively for 10 years with scheduled follow-up visits at one, five and ten years (+/−1 year) after diagnosis. The organization of the cohort, the diagnostic criteria and the clinical follow-up are described in detail elsewhere.13,15,16 The present HRQoL study was conducted as a part of the ten-year follow-up study and included the subgroup of CD patients recruited between October 1, 1991 and December 31, 1993.

3.2 Data collection

The ten-year follow-up visit consisted of a patient-reported questionnaire that included two HRQoL questionnaires (the Norwegian Inflammatory Bowel Disease Questionnaire (N-IBDQ) and the Short Form 36 (SF-36)), a structured interview, a review of hospital records, a clinical examination, laboratory tests and when indicated, an ileocolonoscopy and a small-bowel enema. The patient-reported questionnaires were administered during the hospital visits and were completed before the clinical interview and examination. A study nurse checked the questionnaires to ensure that all questions had been answered.

3.3 HRQoL questionnaires

The IBDQ is a disease-specific multidimensional questionnaire that consists of 32 items scored on a seven-point Likert scale.17 Scores are given for each dimension or as a total score (32–224 points). Higher scores indicate a better HRQoL. The IBDQ has been translated to Norwegian (N-IBDQ) and validated in the IBSEN cohort.18 This validation process yielded a different factor structure than the original, with five dimensions instead of four: stool consistency and pattern (B1), emotional function (EF1), bowel pain and discomfort (B2), social function (SF) and worries (EF2). Nevertheless, the total score is the same in different versions of the questionnaire and can therefore be used to make comparisons between studies.

The SF-36 is a generic HRQoL questionnaire designed to assess functional status, well-being and general perception of health. The questionnaire has been shown to have high validity and reliability19–22 and is one of the most frequently used generic HRQoL instruments. The SF-36 consists of 36 items and one multi-item scale for each of eight conceptual domains. Higher scores indicate a better HRQoL. The domains are physical functioning (PF, 10 items), role limitation due to physical health (RP, 4 items), bodily pain (BP, 2 items), general health (GH, 5 items), vitality (VT, 4 items), social functioning (SF, 2 items), role limitations due to emotional problems (RE, 3 items) and mental health (MH, 5 items). A physical and a mental component summary score (PCS and MCS) can be computed, but these scores are conditioned by norm-based, country-specific conversion factors. The SF-36 has been translated and validated in Norwegian. The SF-36 has previously been demonstrated to have satisfactory psychometric properties in our IBD population.5 The Norwegian SF-36 population reference was described by Loge and Kaasa23 for a sample that was selected at random and retrieved from the National Population Register. The subjects received questionnaires by regular mail. The response rate was 67% (2323 persons). From this reference group we use overall scores, which were adjusted for age, gender and level of education,24 and crude, gender-stratified scores.23

3.4 Patient-reported symptoms and demographic data

Symptoms were patient-reported, with a recall period of 2 weeks prior to the follow-up visits, and were grouped into four categories: (1) no symptoms, (2) mild symptoms (do not interfere with everyday activities), (3) moderate symptoms (do interfere with everyday activities, may result in sick leave) or (4) severe symptoms (unable to carry out everyday activities, on sick leave or hospitalized). Other information from the patient questionnaires included current smoking status (defined as yes (more than one cigarette daily) or no); marital status (single, married/cohabitant, divorced, widower); educational status (grouped according to the Norwegian educational system: (1) second level, first stage (≤9 years), (2) second level, second stage (10–12 years), third level, college or university (> 12 years)); work status (patients were grouped into either working (including students and pupils) or not working (including unemployed, work disability, pensioners, housewives)); and sick leave (sometime during the last 6 months, yes or no).

3.5 Structured interview and patient records

Information on relapse and medication was collected in the structured interviews and through a review of patients' hospital records. Relapse was defined as “a change in the clinical condition that entailed more aggressive medical treatment or surgery” and was classified as relapse during the year preceding the ten-year follow-up (yes/no). Surgery was defined as any intra-abdominal procedure for active CD and was registered as surgery or no surgery during the preceding 10 years. The use of systemic corticosteroids during the year preceding the ten-year follow-up was classified as yes/no. The use of immunomodulators (azathioprine) was classified as yes/no during the previous 5 years. During the interview, patients were asked to categorize the clinical course of their disease from the time of diagnosis to the present ten-year follow-up according to predefined curves indicating; remission or mild severity of intestinal symptoms after initial high activity (C1), increase in the severity of intestinal symptoms after initial low activity (C2), chronic continuous activity (C3) or chronic intermittent activity (C4).

3.6 Laboratory tests

Hemoglobin (Hb) levels were categorized as anemia according to the definition from the World Health Organization (Hb < 12 g/dl in women and < 13 g/dl in men).25 C-reactive protein (CRP) levels were categorized as elevated when CRP was > 10 mg/l.

3.7 Assessment of disease phenotype

The patients were prospectively classified according to the Vienna Classification26 at the ten-year follow-up. The Vienna classification was the standard classification system at the time when the study was designed.

3.8 Minimal clinically important differences

As recommended in recent papers,27,28 we used the minimal clinically important difference (MCID), which is defined as the smallest difference in scores linked to a clinically relevant change, to assess the clinical relevance of our results. The SF-36 anchor-based MCIDs for CD have recently been published.29 To our knowledge, no MCIDs have been published for the IBDQ, but some publications have determined clinically relevant cut-off values.30,31

3.9 Statistics

Data were described using proportions for categorical variables and medians with ranges for continuous data. Comparisons of demographic and clinical data between groups were performed with Fisher's exact test for categorical and ANOVA for continuous variables.

All HRQoL dimensions had skewed distributions. We therefore applied non-parametric methods (Mann Whitney Wilcoxon, Kruskal Wallis and Spearman's rho, when appropriate) to compare the means for all dimensions on the SF-36 and the N-IBDQ between groups. However, we arrived at the same conclusions regarding statistical significant results also when using parametric methods (ANOVA, t-test and Pearson's correlation). Therefore, we chose to maintain standard parametric methods. This also makes it easier to compare our results to the literature.

Adjustment for age, gender and education was performed when appropriate using an analysis of co variation (ANCOVA). The results are presented as marginal means with 95% confidence intervals. The SF-36 scores of our patients were compared to normal population scores by transforming the original data to standard difference scores (s-scores = (mean patient score minus mean population score) divided by population standard deviation) and one sample t-test. S-scores were evaluated according to Cohen's effect size index32: < 0.2, no difference; 0.2 to 0.5, a slight difference; 0.5 to 0.8, a moderate difference; and > 0.8, a large difference. Multiple linear regression analysis (enter method) was used to assess the impact of relevant factors on HRQoL. Variables that were considered clinically relevant based on the literature and only those with a p-value < 0.2 in the univariate analyses (for the total N-IBDQ score or for more than four dimensions on the SF-36) were entered into multiple linear models. Due to multiple comparisons, the significance level was set to 1% in all analyses. All statistical analyses were performed with the Predictive Analytics Software PASW (version 18.0; IBM Corporation, Somers, New York, USA).

3.10 Ethical requirements

The regional ethics committee and the Norwegian Data Inspectorate approved the study. Confidentiality of patient identity and records was maintained using guidelines from the Norwegian Ministry of Health. The study was conducted in accordance with the Declaration of Helsinki.

4 Results

At the ten-year follow-up, 144 patients included in the IBSEN cohort between October 1, 1991 and December 31, 1993 had a definite CD diagnosis. Ten patients died, 19 were lost to follow-up (8 moved out of the area, 3 were untraceable, 1 could not meet because of other comorbidities and 7 were not willing to participate). The remaining 115 patients met at the prescheduled visit. Ninety-nine patients completed the HRQoL questionnaires (response rate 86.1%).

The patients who died were older than the patients who attended the hospital visit (median age at diagnosis 74 years [range 52 to 84 years] and 28 years [range 8 to 68 years], respectively). There was no age difference between the patients who were lost to follow-up (median age at diagnosis 32 years [range 14 to 76 years]) and the patients who attended the hospital visit. There was no difference with regard to gender, smoking, disease distribution or disease behavior (at the time of inclusion into the cohort) between the patients who had died, were lost to follow-up or had arrived at the follow-up (data not shown).

The patients who completed the HRQoL questionnaires and those who did not were comparable with regard to age, smoking, educational status, marital status, disease distribution or disease behavior, surgery, relapse rates and use of corticosteroids and azathioprine (data not shown). However, in the group of patients who completed HRQoL questionnaires, significantly more patients were working (73% vs. 28%, p < 0.001), and more patients reported a less serious disease course (C1 and C2) (49% vs. 33%, p = 0.01) than in the group that did not complete the questionnaires.

Compared to the Norwegian reference population, the study sample was somewhat younger, had a higher level of education and included a larger proportion of men (Table 1).

View this table:
Table 1

Demographic characteristics of study population and general population.

Study population (n = 98)General populationa (n = 2323)
Age, yearsMedian (range)39 (19 to 78)45 (19–80)
Mean (SD)42 (13.7)
Gender% women42% (41)51%
Educational status (missing, n = 14)Second level, first stage4% (3)27%
Second level, second stage45% (38)45%
Third level, college/university51% (43)28%
Marital status (missing, n = 9)Single28% (25)20%
Married/cohabitant64% (57)70%
Divorced5% (4)5%
Widow(er)3% (3)5%

Proportions in percentage, n in parenthesis.

  • a Norweigan general population.23

The clinical characteristics of the included patients are shown in detail in Table 2.

View this table:
Table 2

Clinical characteristics of the study population.

Disease duration (months)
Median and range120 (108 to 140)
Mean and SD121 (7)
Age at diagnosis
A1 (< 40 years)75.5% (74)
A2 (> 40 years)24.5% (24)
L116.3% (16)
L227.6% (27)
L350.0% (49)
L46.1% (6)
B145.9% (45)
B232.7% (32)
B321.4% (21)
Current symptoms (missing, n = 5)
None44.1% (41)
Mild44.1% (41)
Moderate/severe11.8% (11)
Disease course since diagnosis
C144.9% (44)
C24.1% (4)
C313.3% (13)
C437.8% (37)
Relapse last year
Yes33.7% (33)
Surgery since diagnosis
Yes40.4% (40)
Corticosteroids last year
Yes23.5% (23)
5-ASA last 5 years
Yes64.3% (63)
Azathioprine last 5 years
Yes25.5% (25)
Hb g/dl (missing, n = 11)
Median and range14.2 (11.2 to 17)
Anemia9.3% (9)
CRP mg/l (missing, n = 12)
Median and range6 (0 to 42)
Elevated (CRP > 10 mg/l)23.7% (23)
Work status
Working/student72.4% (71)
Unemployed, disability or other27.6% (27)
Sick leave last 6 months (missing, n = 5)
Yes16.1% (15)
Smoking (missing, n = 13)
Yes32.3% (27)

Categorical variables are shown as percentages of total, n is shown in parenthesis.

    4.1 HRQoL scores

    Compared to the reference population, our CD patients had significantly lower SF-36 scores (adjusted for age, gender and education) on the general health and vitality dimensions (Table 3). Data on adjusted scores in the reference population were only available for patients younger than 75 years; therefore, patients aged over 74 years (n = 2) were excluded from these analyses. The other dimensional scores corresponded to the scores in the reference population. Women had significantly lower scores than men on the physical role and social function dimensions. On the social function dimension, the reduction was more than one MCID.

    View this table:
    Table 3

    Mean SF-36 dimensional scores with 95% confidence intervals. Only patients ≤ 74 years. MCIDs for comparison.

    General populationa,b (n = 2214)All patientsa (n = 96)Womenc (n = 40)Menc (n = 56)MCID
    PF8890 [85 to 95]89 [81 to 96]92 [87 to 97]8.7
    RP8072 [62 to 83]66 [50 to 82]82 [72 to 93]*18.4
    BP7671 [63 to 78]65 [53 to 77]79 [71 to 87]11.8
    GH7762 [55 to 69]§60 [49 to 71]64 [57 to 71]6.5
    VT6051 [45 to 58]§47 [37 to 58]58 [51 to 65]9.3
    SF8681 [74 to 88]75 [64 to 86]89 [82 to 96]*9.4
    RE8371 [60 to 82]64 [47 to 82]81 [70 to 93]15.6
    MH7974 [69 to 78]71 [64 to 79]77 [72 to 82]5.7

    MCID = minimal clinically important differences as determined by Coteur et al.29.

    • a Adjusted for age, gender and education.

    • b CI or SD not available in the literature.24

    • c Adjusted for age and education.

    • § Statistical difference between general population and all patients, p-value ≤ 0.01.

    • * Statistical difference between men and women, p-value ≤ 0.01.

    Comparison to the reference population using s-scores stratified by gender (Fig. 1) revealed that women with CD reported small to moderate negative effects on all but the physical function dimension (with the most pronounced effects on the physical role, general health and vitality dimensions) compared to women in the reference population. Men with CD reported small negative effects on three out of eight dimensions (general health, vitality and mental health) compared to men in the reference population.

    Figure 1

    SF-36, S-scores stratified by gender. Text: S-scores = (mean patient score minus mean population score) divided by population standard deviation. Cohen's effect size index: < 0.2 no difference, 0.2 to 0.5 small difference, 0.6 to 0.8 moderate difference and > 0.8 large difference. PF: physical function, RP: role physical, BP: bodily pain, GH: general health, VT: vitality, SF: social function, RE: role emotional, MH: mental health.

    Patients who reported current symptoms had significantly reduced scores on seven out of eight dimensions, and there was a trend toward reduced scores on the physical function dimension. (Table 4). Work status (not working) was associated with reduced scores on the general health and physical role dimensions. Use of corticosteroids was associated with a trend toward a reduction on the physical function and the bodily pain dimensions. Increased CRP was associated with a reduced bodily pain score and a trend toward reduced scores on the physical function and general health dimensions. All these absolute differences were above the corresponding MCIDs. Clinical variables such as phenotype based on the Vienna classification (location and behavior), relapse, surgery, use of azathioprine and anemia did not have statistically significant effects on SF-36 scores nor did patient-reported ten-year disease course, smoking or sick leave (Appendix 1).

    View this table:
    Table 4

    SF-36 dimensional scores adjusted for age, gender and education.

    SymptomNo94 [87 to 101]88 [76 to 100]89 [81 to 97]70 [61 to 80]65 [57 to 73]94 [86 to 101]85 [70 to 100]82 [75 to 88]
    Mild83 [76 to 91]60 [46 to 73]58 [48 to 67]58 [48 to 69]41 [31 to 50]76 [67 to 85]56 [39 to 74]67 [60 to 75]
    Sev82 [70 to 93]27 [7 to 48]32 [18 to 46]43 [27 to 59]20 [8 to 34]57 [43 to 71]42 [16 to 68]59 [48 to 70]
    p-Value0.017< 0.001< 0.0010.013< 0.001< 0.0010.003< 0.001
    CorticosteroidNo90 [85 to 95]74 [64 to 94]74 [66 to 82]63 [56 to 70]53 [46 to 60]83 [76 to 90]71 [59 to 82]75 [70 to 80]
    Yes80 [72 to 88]62 [45 to 80]61 [48 to 74]55 [43 to 67]48 [36 to 60]74 [62 to 86]79 [59 to 99]70 [62 to 79]
    CRPNormal92 [86 to 97]76 [65 to 89]76 [68 to 84]64 [55 to 72]54 [46 to 62]83 [75 to 90]69 [56 to 82]74 [69 to 80]
    Elevated83 [75 to 91]60 [45 to 76]56 [44 to 68]51 [39 to 63]47 [35 to 58]76 [65 to 87]82 [63 to 101]75 [66 to 83]
    916201377− 13− 1
    WorkYes91 [86 to 96]76 [66 to 86]71 [64 to 79]65 [58 to 72]53 [45 to 60]81 [74 to 88]69 [57 to 81]74 [69 to 79]
    No83 [74 to 92]45 [27 to 62]64 [51 to 78]45 [33 to 57]47 [34 to 60]80 [68 to 92]80 [60 to 101]70 [61 to 79]

    Scores adjusted for age, gender and education in ANCOVA analysis. Adjusted mean and 95% confidence intervals. Only variables with significant differences (p-value ≤ 0.01) or trends (p-value ≤ 0.05) in more than one dimension are shown in this table.

    PF: physical function, RP: role physical, BP: bodily pain, GH: general health, VT: vitality, SF: social function, RE: role emotional, MH: mental health.

    ∆ Delta value between categories.

    MCID = minimal clinically important difference as determined by Coteur et al.29.

      In the linear regression analysis, the only variables that fulfilled the established criteria for inclusion were age, current symptoms and work status (Table 5). Older age had a negative effect on the physical function dimension. There was a trend toward a negative effect on physical role and bodily pain as well as a trend toward a positive effect on mental health. Current symptoms affected all but the physical function, while work status (not working) affected physical function, physical role and general health and had a trend toward negative effects on the bodily pain and vitality dimensions.

      View this table:
      Table 5

      Linear regression model for SF-36 dimensional scores and the IBDQ total score.

      Older age (per year)− 0.6 [− 0.8 to − 0.3]*− 0.5 [− 0.9 to 0]§− 0.3 [− 0.7 to − 0.1]§nsnsnsns0.3 [0 to 0.5]§ns
      Current symptoms− 6 [− 11 to − 1]§− 30 [− 38 to − 22]*− 28 [− 34 to − 22]*− 14 [− 20 to − 8]*− 18 [− 24 to − 13]*− 21 [− 27 to − 15]*− 22 [− 33 to − 11]*− 11 [− 16 to − 6]*− 30 [− 36 to − 24]*
      Not working− 16 [− 25 to − 8]*− 33 [− 48 to − 19]*− 11 [− 21 to − 1]§− 25 [− 35 to − 14]*− 11 [− 21 to − 13]§nsnsnsns

      Linear regression model fitted to estimate the effect of selected variables on SF-36 dimensional scores and the IBDQ total score. The presented results are estimated β's with 95% confidence intervals.

      ns = non significant.

      PF: physical function, RP: role physical, BP: bodily pain, GH: general health, VT: vitality, SF: social function, RE: role emotional, MH: mental health, IBDQ: Inflammatory Bowel Disease Questionnaire total score.

      • * p-Value ≤ 0.01.

      • § p-Value ≤ 0.05 (trend).

      Overall, the N-IBDQ total score was 183 (95% CI: 178 to 189) (Table 6). We found no gender differences. Current symptoms and sick leave were associated with reduced N-IBDQ scores; no symptoms 202 (196 to 207), mild symptoms 178 (173 to 184) and moderate to severe symptoms 134 (123 to 145), p < 0.001; no sick leave 187 (181 to 193), sick leave 168 (154 to 181), p = 0.01 (age- and gender-adjusted analyses). No other clinical or demographic factors had an effect on the total N-IBDQ scores (Appendix 2). In the multiple linear regression analyses, only current symptoms were associated with reduced N-IBDQ scores (Table 5).

      View this table:
      Table 6

      Mean IBDQ dimensional scores with 95% confidence intervals.

      All patientsa (n = 99)Womenb (n = 42)Menb (n = 57)
      IBDQ total183 [178 to 189]181 [173 to 189]186 [179 to 193]
      B143 [42 to 44]43 [41 to 45]44 [42 to 45]
      E157 [54 to 59]55 [52 to 59]58 [55 to 62]
      SF25 [24 to 26]25 [24 to 27]26 [24 to 27]
      B227 [26 to 28]27 [25 to 28]27 [26 to 29]
      E225 [24 to 26]25 [24 to 26]25 [24 to 26]

      IBDQ: Inflammatory Bowel Disease Questionnaire.

      B1: stool consistency and pattern, E1: emotional function.

      SF: social function, B2: bowel pain and discomfort, E2: worries.

      • a Adjusted for age and gender.

      • b Adjusted for age.

      5 Discussion

      In this study, we determined the HRQoL in a population-based cohort of CD patients who have had the disease for 10 years. The patients were from demographically uniform areas and had equal access to health care. The study was conducted prior to the introduction of anti-tumor necrosis factor (TNF) inhibitors, and none of the patients had received anti-TNF in their ten-year disease course. We therefore consider this inception cohort to represent, the natural spectrum of disease presentation 10 years after disease onset. We believe that this study adds important information to the description of the natural course and outcomes in CD patients. The patients displayed reduced general health and vitality scores, while the other SF-36 dimensional scores were comparable to those of the general population. The overall disease-specific HRQoL (N-IBDQ) score was above a level seen in patients in remission.30 Only current symptoms, work status and sick leave had a significant negative effect on the HRQoL.

      The SF-36 scores were comparable to those of the reference population, except for the general health and vitality dimensions. The reduction in general health scores was noteworthy because the absolute value was more than twice the published MCID for this dimension, and the effect size was moderate in both genders. A reduction in the general health dimension has been described in population-based studies of patients in different disease states4,5 and in studies on patients in clinical remission.33 In our study, the patients without a relapse in the year before follow-up also had reduced scores on the general health dimension compared to the reference population (63 [55 to 71] vs. 77 points, respectively). The questions in the general health dimension focus on perceptions of one's own health compared to others and expectations of future health. With a long history of a relapsing disease, it is not surprising that our patients have low scores on this dimension even when they do not experience a relapse. This finding is in agreement with results from both qualitative and quantitative studies in which concerns about the future and the uncertain nature of the disease were found to be important limitations for CD patients.10,34,35

      The reduction in vitality scores is numerically smaller than our finding for general health, but the reduction equals one MCID and has a small effect on the gender-stratified S-scores (with a larger negative effect for women than for men) (Fig. 1). There was no statistically significant gender difference on the adjusted vitality scores. The absolute difference, however, was above one MCID. A type II error might cause the lack of statistical significance. A reduction in the vitality dimension has been linked to the presence of fatigue.36 Low vitality scores are more predominant in CD than in UC patients in other studies.37 Two recent studies, one hospital-based38 and one population-based,39 have shown that chronic fatigue is more common in CD patients than in UC patients. The reductions in vitality scores might therefore be a reflection of increased fatigue in this population.

      The bodily pain dimension was not reduced in the overall group. Women had a 13-point-lower score than men, which is above the MCID of 11.8 points, but this difference did not reach statistical significance at the 1% level (p = 0.035). Bodily pain was also reduced by more than one MCID in patients who reported current symptoms and use of corticosteroids and by almost two times the MCID for patients reporting being on sick leave (but this difference was not statistically significant). This finding suggests that pain is an important limitation for these patients. It has been shown that pain anxiety and pain catastrophizing are pronounced in CD patients40 and that fear of pain is an important concern for them.35 Our results support these observations.

      One study has shown that corticosteroids reduce HRQoL scores, while immunomodulators are associated with a better HRQoL.41 We found none of the above trends in our study. This could be due to a type II error.

      Phenotype based on the Vienna classification did not seem to affect either generic or disease-specific HRQoL in our study. This finding is in line with other studies8 and is in accordance with the fact that phenotype is unimportant to the patient's perception of the disease, even though it might be important with regard to a clinical prognosis.

      We found no effect of surgery on HRQoL. Previous studies have shown contradictory results with regard to surgery and HRQoL: some have shown improved HRQoL after surgery,42 others have found reduced HRQoL after surgery.10,41 These contradictory results are likely caused by different time spans between surgery and the HRQoL measurement as well as a lack of standardization in the trials mentioned. Our study was not designed to investigate HRQOL before and after surgery, and the time span between surgery and the HRQOL measurement differed within the patient group. These factors likely contributed to our findings.

      It is well known that HRQoL in CD is reduced in patients with active disease.9,40,43,44 However, the definition and the mode of measurement of disease activity vary between subjective, patient-reported questions and clinician- and laboratory-dependent complex indexes. Additionally, some of the complex “objective” indexes are likely affected by subjective interpretation.45 Maunder et al. have shown that in UC, a self-report symptom index was as effective as the St. Marx index (clinically and endoscopy-based) in predicting inactive or active disease.46 We did not include an activity index, but we included one patient-reported question on current symptoms (IBD symptoms in the last 14 days). As expected, patients who responded “no current symptoms” had SF-36 scores similar to, and even above, the reference population, and they had N-IBDQ scores above the scores described in CD patients in remission.30 Patients who reported current symptoms had the lowest HRQoL scores on both the SF-36 and the N-IBDQ. One could hypothesize that this simple question is as effective as any disease activity index to differentiate between patients in remission and relapse.

      It has been shown that CRP level correlates well with both endoscopic activity and the Crohn's disease activity index (CDAI).47 An increased CRP level was only associated with a reduced bodily pain score and a trend toward reduced scores in the physical function and general health dimensions. We did not find a linear association between CRP and HRQoL scores (data not shown). This result might reflect the fact that none of the patients had a substantially increased CRP level (the highest being 42 mg/l). We do not know whether other confounding factors could cause the lack of association between CRP levels and HRQoL in our study.

      Anemia has been linked to a reduced HRQoL both in CD48,49 and in other chronic conditions.50 In our study, anemia was not associated with a reduced HRQoL; however, less than 10% of our patients had anemia, and no patient had Hb levels below 11 g/dl. Thus, the degree of anemia was very limited, and the clinical relevance was likely to be small.

      There is a lack of well-documented MCIDs in HRQoL research on IBD. The MCIDs for the SF-36 we used in our study29 were computed with an anchor-based approach, which implies that calculation of the changes should be based on a well-defined and relevant change in another outcome (clinical or patient-based) — the anchor. A change of 16 points on the IBDQ total score and a change of 50 points on the CDAI were used as anchors for the SF-36. The clinical relevance of these two scales is debated, and many researchers suggest that higher numbers (up to 32 for IBDQ and 100 for the CDAI) are needed to identify a clinically relevant change in these scales. The MCIDs might therefore be underestimated and consequently the clinical importance of some our results might be overestimated.

      IBDQ cut-off values for remission and clinical improvement have been computed in a trial with CD patients treated with infliximab.30 Remission was defined as a CDAI score > 150 and the corresponding IBDQ-total score was 168. In our study, patients with no current symptoms had a total IBDQ score of 202 (196 to 207) and patients with no relapse during the last year had a score of 185 (179 to 192). A 168 point cut-off for remission therefore seems somewhat low in the light of our results. This is also supported by an observational study of UC patients where an IBDQ total score of 205 corresponded to patient-defined remission.31

      Among the responders in our study, a higher proportion of patients reported to be working and having disease course C1 or C2 (less serious disease courses) than among the non responders. This could have biased our results. However, disease course does not seem to be associated with changes in HRQoL in our results thus this has not likely affected our findings substantially. A too low proportion of people not working among our respondents might have led to a slight overestimation of HRQoL scores mainly in the role physical and the general health dimensions of the SF-36.

      Because it has been shown that both somatic and psychiatric co-morbidity can affect HRQoL,48,51 more information on these issues could have strengthened our study. However, our study population is relatively young, so somatic comorbidity is expected to be limited. If psychiatric comorbidity was a major problem in our study population, we would have anticipated a substantial reduction in the mental health dimension of the SF-36, but this was not the case.

      6 Conclusion

      In this population-based cohort study of CD patients with a ten-year disease course, we found greater reductions in the SF-36 scores than in the N-IBDQ scores. The SF-36 dimensions of general health and vitality were significantly reduced compared to the general population. All other SF-36 dimensions were comparable to those of the general population, and the N-IBDQ scores were comparable to the scores of patients in remission. Except for current symptoms, we found no clinical variables that affected HRQoL scores. Work status (not working) and being on sick leave affected HRQoL negatively. Our results suggest that in this chronic stage of CD, general symptoms might be more limiting than disease-specific symptoms. We consider this to be an important finding when treating these patients.

      Conflict of interest



      We thank the following members of the IBSEN Study Group: Morten Vatn, Akershus University Hospital, Magne Henriksen; Østfold Hospital, Øyvind Palm; Oslo University Hospital, Njaal Stray; Diakonhjemmet Hospital, Jostein Sauar; Telemark Hospital, Stein Dahler; Befell Hospital, Øystein Kjellevold; Telemark Hospital Finn Strøm; Lovisenberg Diakonale Hospital, Ole Høie; Sørlandet Hospital.

      Permission to use the IBDQ was kindly given by Professor Gordon Guyatt at McMaster University, Canada.

      Permission to use the SF-36 was given by Professor Stein Kaasa, St. Olavs University Hospital, Trondheim, Norway.

      The authors have contributed to the following:

      The conception and design of the study: BM, ICS, IL, JJ and TB

      Acquisition of data: BM, ICS, IL and JJ

      Interpreting data: MLH, BM and TB

      Statistical analysis: MLH and MC

      Drafting the manuscript: MLH

      Revising the manuscript for intellectual content: MLH, TB, ICS, MC, IL, JJ and BM

      Final approvement of the submitted manuscript: MLH, TB, ICS, MC, IL, JJ and BM


      SF-36 dimensional scores adjusted for age, gender and education.

      View this table:


      IBDQ total score, adjusted for age and gender.

      View this table:


      • ☆☆ Conference presentation: Parts of the results have been presented at: ECCO conference Innsbruck February 2007, poster UEGW Stockholm October 2011 (abstract accepted for poster presentation).

      • Funding: Research grant from South-Eastern Norway Regional Health Authority.


      View Abstract