Download pdf - Health‐related quality of life in patients with inflammatory bowel disease five years after the initial diagnosis

Health-related Quality of Life in Patients with Inflammatory BowelDisease Five Years after the Initial Diagnosis

T. Bernklev, J. Jahnsen, E. Aadland, J. Sauar, T. Schulz, I. Lygren, M. Henriksen, N. Stray,Ø. Kjellevold, M. Vatn, B. Moum & the IBSEN Study GroupMedical Dept., Rikshospitalet University Hospital, Oslo, Norway; Dept. of Internal Medicine, ØstfoldCentral Hospital, Fredrikstad, Norway; Dept. of Internal Medicine, Aker University Hospital, Oslo,Norway; Dept. of Internal Medicine, Telemark Central Hospital, Skien, Norway; Dept. of InternalMedicine, Aust-Agder Central Hospital, Arendal, Norway; Dept. of Gastroenterology, Division ofMedicine, Ulleval University Hospital, Oslo, Norway; Dept. of Internal Medicine, Diakonhjemmet,Oslo, Norway; Dept. of Internal Medicine, Telemark County Hospital, Rjukan, Norway; theInflammatory Bowel South-Eastern Norway (IBSEN) Group of Gastroenterologists

Bernklev T, Jahnsen J, Aadland E, Sauar J, Schulz T, Lygren I, Henriksen M, Stray N, Kjellevold Ø, VatnM, Moum B, the IBSEN Study Group. Health-related quality of life in patients with inflammatory boweldisease five years after the initial diagnosis. Scand J Gastroenterol 2004;39:365–373.

Background: Health-related quality of life (HRQOL) has become an important tool in evaluating patientsatisfaction in inflammatory bowel disease (IBD). So far, few prospective follow-up studies have beendone to identify variables that influence HRQOL. We aimed to identify demographic and clinicalvariables that influence HRQOL 5 years after diagnosis in patients with ulcerative colitis (UC) or Crohndisease (CD) included in a prospective follow-up study from 1990 to 1994 (the IBSEN study). Methods:All patients completed the Inflammatory Bowel Disease Questionnaire (IBDQ), a disease-specificquality-of-life questionnaire translated into Norwegian and validated. We present data from 497 patients(328 UC patients and 169 CD patients, mean age 43.3 years, 48% female). The impact of age, gender,smoking, symptom severity, disease distribution, rheumatic symptoms and surgery on IBD patients’HRQOL was analysed. Results: Women had a reduction in IBDQ total score of 10 points compared tomen, CD patients had a reduction of 7.5 compared to UC patients. The patients with moderate/severesymptoms had a 50 points lower score than the patients without symptoms. The patients with rheumaticsymptoms had a 10 points lower total score than the patients without these symptoms. All differenceswere statistically significant. The multiple regression analysis showed that symptom severity, rheumaticsymptoms and female gender were the strongest predictors of reduction in HRQOL for both diagnosisgroups. Conclusion: IBD symptoms, rheumatic symptoms and female gender have a significant influenceon patients’ HRQOL as measured by IBDQ. This was confirmed by the regression analysis.

Key words: Clinical significance; health-related quality of life (HRQOL); inflammatory bowel disease;inflammatory bowel disease questionnaire (IBDQ); IBD symptom severity; regression analysis

Tomm Bernklev, Solløkkasletta 10, NO-3233 Sandefjord, Norway (fax.�47 33 24 94 41, [email protected])

The inflammatory bowel diseases (IBD), ulcerativecolitis (UC) and Crohn disease (CD), are chronicgastrointestinal disorders with varying severity that

affect young and middle-aged people (1–6).Several different scoring systems are used to classify the

severity of IBD (7–10). However, these indices pay limitedattention to the patient’s subjective experiences. Moreover,they have limitations such as poor standardization, disregardfor rigorous validation and poor sensitivity for detectingfunctional disability and psychosocial problems (11). Dataabout patients’ own experiences, such as physical and socialfunctioning, mental health and working capacity, which arereferred to as health-related quality of life (HRQOL), arenecessary for clinicians and other healthcare providers whendeciding between different treatment options for IBD

patients. Investigating them has therefore become mandatoryin most clinical studies and they are often included as aseparate outcome measure in intervention trials.

There are several different questionnaires measuringHRQOL in IBD (12–15), and the Inflammatory BowelDisease Questionnaire (IBDQ) is the most commonly used(12, 13, 16). The psychometric properties of the IBDQ arewell described, and the questionnaire has now been translatedinto several languages, including Norwegian, and validated(17–23). The IBDQ has also been used in medical interven-tion trials (11, 24, 25) and for evaluating IBD patients whohave undergone surgery (26, 27). However, these clinical datahave mainly been collected from selected patient materials,and often from CD patients only. Thus, no HRQOL data areavailable from population-based cohorts of IBD patients

ORIGINAL ARTICLE

2004 Taylor & Francis DOI 10.1080/00365520310008386

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

followed prospectively over time. A recent paper hasaddressed some of these issues, but the patient data werecollected from a university hospital registry and the patientswere not followed prospectively (28).

The IBSEN (Inflammatory Bowel South-Eastern Norway)study group has followed an inception cohort of IBD patientsprospectively since 1990. The demographic and clinical data,both at inclusion and at follow-up, have been described indetail (4–6, 29). Five years after diagnosis, updated demo-graphic and clinical data were collected from all livingpatients, and questionnaires regarding HRQOL were admi-nistered during interviews with the patients in a cross-sectional design. The aim of the present study was todetermine HRQOL in patients with UC and CD, and toinvestigate the impact of demographic and clinical variableson HRQOL.

Materials and Methods

Study populationAll newly diagnosed cases of IBD, or possible IBD, between 1January 1990 and 31 December 1993 in four well-definedareas in southeastern Norway (the counties of Oslo, Østfold,Telemark and Aust-Agder) were registered by the localdepartments of gastroenterology. To ensure complete ascer-tainment, all 1236 general practitioners and clinicians at the14 hospitals in the four participating counties receivedinformation about symptoms consistent with IBD, and wereinvited to refer all potential cases of UC or CD to the localgastroenterological outpatient clinic. The information wasgiven prior to the start of the study in the form of three writtenreminders. In addition, information about the study waspresented at local meetings with the general practitioners, andall the practising gastroenterologists, internists, surgeons andpaediatricians in the relevant area were informed of the study.The total population of the geographical area was 966,427 on1 January 1992.

Endoscopy was chosen as the main instrument for diag-nosis and for determining the distribution of the disease in thecolon. A total of 618 cases of UC and indeterminate colitisand 225 cases of CD were diagnosed, which is consistent withthe high incidence reported in other studies in Norway andScandinavia in the 1970s and 1980s. The annual incidencewas similar throughout the study period, indicating completeascertainment (4, 5). At follow-up between 1 and 2 yearslater, the diagnosis was re-evaluated. Ninety-eight percent ofthe patients were available for follow-up. Four percentreceived an uncertain diagnosis at this point because thediagnosis of IBD could not be definitely confirmed (6).

Five years after the initial diagnosis, the patients wereinvited to make a clinical follow-up visit to their localhospitals. All these cases were reviewed by gastroenterolo-gists, and patients without a confirmed diagnosis of UC or CDwere then excluded, in addition to the patients who had died(30).

The remaining patients (454 with UC and 200 with CD)were invited to participate in the present HRQOL study. Ofthese, 533 fulfilled the inclusion criteria for definite IBD, gavewritten informed consent and completed the Norwegianversion of the Inflammatory Bowel Disease Questionnaire(N-IBDQ). These patients were also interviewed about theirdisease and were given a clinical examination by a gastro-enterologist.

Of these 533 patients, five were excluded because theywere below 18 years of age. Twenty-eight patients with UCwere excluded because they had undergone colectomy, and anadditional 3 patients were excluded because of non-compli-ance in completing the questionnaires. Thus, the IBDQ datafrom 497 patients (328 UC and 169 CD) were available foranalysis. Further demographic data are given in Table I.

Information on the other 121 patients (29 had moved awayfrom the area, 21 were suffering from concomitant seriousdisease or old age, and 71 patients did not respond to theinvitation) who had not attended the follow-up visit was basedon hospital records and telephone interviews.

Diagnostic criteria and classification of intestinal diseaseUniform methods and diagnostic criteria were used. Theinitial classification of UC and CD was based on symptomsconsistent with IBD for more than 4 weeks, excludinginfections and other acute or chronic non-IBD. The diagnosisof UC was based on the presence of at least three of thefollowing criteria: (1) a history of diarrhoea and/or blood/pusin the stools; (2) macroscopic appearance by endoscopy withcontinuous mucosal inflammation affecting the rectum incontinuity with some or all of the colon; (3) microscopicfeatures on biopsy compatible with UC; (4) no suspicion ofCD at small-bowel X-ray, ileoscopy or biopsy (4).

The diagnosis of CD was based on the presence of two ormore of the following established criteria (31): (1) typicalclinical features including abdominal pain, diarrhoea andweight loss; (2) macroscopic appearance at surgery orendoscopy: segmental, discontinuous and/or patchy lesionswith or without rectal involvement, discrete or aphthousulcerations, fissuring and penetrating lesions, cobblestone orstrictures; (3) radiological evidence of stenosis in the smallbowel, segmental colitis or findings of fistulae; (4) histo-logical evidence of transmural inflammation or epithelialgranulomas with giant cells.

The extent and localization of colonic disease were basedon endoscopical findings and characteristic histological signsof inflammation. When the extent of the disease changedduring follow up, the maximum extent of bowel involvementwas recorded. The patients with CD were classified accordingto disease localization; colitis, ilocolitis, ileitis and upper GI,and to the disease behaviour (32).

Data collectionAll patient data were collected between January 1995 andDecember 1999. The N-IBDQ was self-administered and was

Scand J Gastroenterol 2004 (4)

366 T. Bernklev et al.

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

answered by the patients at the hospital before they wereinterviewed and clinically examined by the gastroenterolo-gist. A standardized procedure was followed at all the centres,which allowed the patients to fill in the questionnaire alone inpeace and quiet. Before the patients left the clinic theinvestigator and a nurse checked the questionnaire to ensurethat all questions had been answered.

Assessment of HRQOL.IBDQ is a disease-specific quality-of-life questionnaire (12). It consists of 32 questions, with atotal score and 4 underlying dimensions; Bowel Function,Emotional Function, Social Function and Systemic Function.The responses are graded on a 7-point Likert scale from 7 (nota problem) to 1 (a very serious problem), giving a possibletotal score range of 224 to 32. The higher the score the betterthe HRQOL. This questionnaire has recently been translatedinto Norwegian (N-IBDQ), validated and published. TheN-IBDQ has five underlying dimensions in contrast to thefour dimensions in the original IBDQ: Bowel Function-I(Stool consistency and pattern, B1), Bowel Function-II(Bowel pain and discomfort, B2), Emotional Function-I(E1), Emotional Function-II (Worries, E2) and Social Func-tion (SF). The differences in dimensional scores between thetwo versions have been discussed in detail in a previouspublication (23)

Clinical condition questionnaireThe patients were asked about their clinical condition at thevisit. The recall period was the previous 2 weeks. The fourpossible answers were: no symptoms, mild symptoms (do notinterfere with everyday activities), moderate symptoms (dointerfere with everyday activities, may result in sick leave)

and severe symptoms (unable to carry out everyday activities,on sick leave or hospitalized).

Other background informationAt the consultation with the gastroenterologist, additionalinformation was recorded, including smoking habits, symp-tom severity, disease distribution according to the Viennaclassification, history of surgery (CD) and rheumatic symp-toms.

Statistical methodsAll descriptive data are given as means with standarddeviations; median and range are given when appropriate.Box plots are used to visualize symptom categories. Acorrelation analysis (Pearson’s) was made between symptomseverity and N-IBDQ scores. Comparisons of quality of lifebetween UC and CD groups were performed using ANCOVA(analyse of covariance) estimating marginal mean scores with95% confidence intervals adjusted for age and gender, asthese covariates are known to influence HRQOL question-naires.

To adjust for multiple comparisons we used Bonferroni’smethod.

In a multiple linear regression analysis (enter method), weinvestigated the impact of demographic and clinical data(independent variables) on the N-IBDQ total and dimensionscores (dependent variables). The analysis was performed forUC and CD separately.

The significance level was set to 5%; all tests were two-sided. All the statistical analyses were performed with SPSSVersion 11.0 (SPSS, Chicago, Ill., USA) for Windows.

Table I. Demographic and clinical data. Actual numbers of patients in each category with percentages in parentheses

n = 6541 n = 5332 n = 4973

UC (n = 454) CD (n = 200) UC (n = 361) CD (n = 172) UC (n = 328) CD (n = 169)

Age (mean � s) 44.9 (�16.4) 37.9 (�15.9) 45.1 (�15.3) 38.3 (�15.6) 45.7 (�14.9) 38.6 (�15.5)% women 47.3 50 46.2 49.4 47.6 50.9Smokers

Yes 57 (13) 81 (41) 43 (12) 71 (41) 39 (12) 71 (42)No 395 (86.6) 115 (57) 314 (87) 97 (56) 288 (87.7) 95 (56)Not known 2 (0.4) 4 (2) 4 (1) 4 (2) 1 (0.3) 3 (2)

SymptomsNone 258 (57) 87 (44) 203 (56) 80 (46) 183 (56) 79 (47)Mild 153 (34) 82 (41) 128 (35) 68 (40) 122 (37) 68 (40)Moderate/severe 34 (7) 31 (15) 30 (9) 24 (14) 23 (7) 22 (13)Not known 9 (2) 0 0 0 0 0

RelapseYes 316 (70) 150 (75) 251 (70) 125 (73) 231 (70) 123 (73)No 138 (30) 50 (25) 110 (30) 47 (27) 97 (30) 46 (27)

Extra-intestinal manifestationsYes 171 (38) 93 (47) 137 (38) 80 (47) 124 (38) 80 (47)No 235 (52) 82 (41) 182 (50) 72 (42) 169 (51) 70 (42)Not known 48 (10) 25 (12) 42 (12) 19 (11) 35 (11) 19 (11)

1All patients with verified IBD diagnosis after 5 years.2All patients who attended the 5-year follow-up and consented to participate in this study.3All patients entered in the database.s= standard deviation.


Health-related Quality of Life in IBD 367

Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

Ethical requirementsThe study was performed in accordance with the principlesset out in the Helsinki Declaration, and was approved by theregional ethics committee. Permission was also obtained fromthe Norwegian Data Registry.

Results

Gender, age and diagnostic groupThe mean IBDQ score for women and men, adjusted for ageand diagnostic group, is presented in Table II. There was astatistically significant difference for gender, both in totalscore and within the dimensions B2, E1 and SF. The analysisof diagnostic groups adjusted for age and gender found areduction in total score of 7.4 points for CD patientscompared with UC patients. The same difference wasobserved for the dimension scores except for the dimension‘bowel pain and discomfort (B2)’ (Table III).

When gender and diagnostic group were combined, women

with CD had a statistically significant reduction in IBDQscore compared with UC women and UC and CD men (TableIV).

Symptom severityThe patients’ self-reported symptoms relating to IBD duringthe last 14 days before the follow-up visit and thecorresponding IBDQ total scores are presented in Fig. 1. Asexpected, severe IBD symptoms were associated with asignificant reduction in IBDQ score.

A corresponding correlation analysis (Pearson’s r) for eachdiagnostic group gave r values of �0.31 to �0.52 for UCpatients and �0.42 to �0.61 for CD patients, indicating aslightly higher correlation between symptom severity andHRQOL in CD patients compared with UC patients.

Disease distribution and behaviourPatients were classified according to disease distribution asfollows: in UC; proctitis, procto-sigmoiditis, left-sided colitis

Table II. IBDQ scores by gender. Scores are given as estimated marginal mean scores (95% confidence intervals) adjusted for age anddiagnostic group

Women (n = 239) Men (n = 258) P valuea

B1 42.06 (41.19–42.92) 43.15 (42.30–44.00) 0.077B2 26.00 (25.21–26.79) 28.27 (27.49–29.04) �0.001E1 55.74 (54.16–57.32) 60.53 (58.98–62.08) �0.001E2 24.30 (23.77–24.82) 24.98 (24.47–25.50) 0.067SF 24.40 (23.80–25.00) 25.54 (24.95–26.13) 0.008Total score 178.15 (174.45–181.85) 188.43 (184.80–192.06) �0.001

aANCOVA, P values based on the F distribution, adjustment for multiple comparisons: Bonferroni.

Table III. IBDQ scores by diagnostic group. Scores are given as estimated marginal mean scores (95% confidence intervals) adjusted for ageand gender

UC (n = 328) CD (n = 169) P valuea

B1 43.27 (42.55–43.98) 41.94 (40.94–42.94) 0.037B2 27.61 (26.96–28.26) 26.65 (25.74–27.56) 0.097E1 59.33 (58.03–60.63) 56.95 (55.13–58.77) 0.040E2 25.07 (24.63–25.50) 24.22 (23.61–24.82) 0.027SF 25.78 (25.29–26.28) 24.15 (23.46–24.84) �0.001Total score 186.99 (183.94–190.03) 179.59 (175.33–183.86) 0.006


Table IV. IBDQ scores by gender and diagnostic group. Scores are given as estimated marginal mean scores (95% confidence intervals)adjusted for age

Group n B1 B2 E1 E2 SF Total score

Women, UC 154 42.83(41.74–43.93)

26.67(25.74–27.60)

57.44(55.61–59.27)

24.87(24.22–25.52)

25.28(24.56–26.01)

182.83(178.35–187.32)

Women, CD 85 41.27(39.82–42.72)

25.28(23.85–26.71)

54.01(51.27–56.75)

23.74(22.81–24.67)

23.57(22.40–24.74)

173.45(166.90–180.00)

Men, UC 174 43.71(42.83–44.60)

28.62(27.79–29.45)

61.27(59.62–62.92)

25.24(24.70–25.77)

26.21(25.73–26.70)

191.17(187.54–194.80)

Men, CD 84 42.58(41.12–44.05)

27.89(26.54–29.25)

59.78(56.91–62.65)

24.74(23.87–25.62)

24.89(23.70–26.08)

185.68(178.98–192.37)

P valuea 0.043 �0.001 �0.001 0.042 �0.001 �0.001




Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

and extensive colitis: in CD; colitis, ileo-colitis, ileitis andupper GI. In addition to disease localization, patients with CDwere classified according to age at diagnosis (�40 years, �40years) and disease behaviour (non-stricturing/non-penetrat-ing, stricturing, penetrating). There were no statisticallysignificant differences in N-IBDQ scores according to diseasedistribution within the categories, and total scores were 184–188 for UC and 178–184 for CD. However, in CD patients 40years or above with complicated disease, the N-IBDQ totalscore was reduced from 180 to 165 in fistulizing patients andto 145 in patients with stenosis. These reductions arestatistically and clinically significant.

Rheumatic symptomsA rheumatologist examined all the patients and registered allrheumatic symptoms either present at the examination orreported by the patients in their medical history (33). In total,81 (25%) of the patients with UC reported rheumaticsymptoms; the corresponding figure for CD patients was 50(30%). The N-IBDQ total scores are presented in Table V. UCpatients with rheumatic symptoms reported a significantlylower score compared to patients without such symptoms.

SurgeryThe impact of surgery on HRQOL applies to CD patients

only, since colectomized UC patients were excluded from thestudy. Fifty-five out of 169 CD patients (32.5%) underwentsurgery between inclusion and 5-year follow-up. But whenage and gender were adjusted for, no difference in total scorebetween the two CD groups was found (N-IBDQ: 179 versus180).

Smoking habitsA total of 110 patients (22.3%) were current smokers, 58% ofwhom were women. Of the CD patients, 71 (43%) werecurrent smokers, while the corresponding figure for UCpatients was 39 (12%). Women were more often smokers thanmen in both UC and CD (CD women/CD men: 51% versus34%; UC women/UC men: 14% versus 10%) groups.

The smoking CD patients had a statistically lower IBDQtotal score than non-smokers (175 versus 183, P = 0.029).However, no significant difference was found in IBDQ totalscore between smokers and non-smokers in the UC group.

Multiple linear regression analysisIn the multiple linear regression analysis, IBD symptoms(severity), rheumatic symptoms and female gender were themost frequently occurring independent variables for both UCpatients and CD patients. The variable that contributed most

Fig. 1. IBDQ total score and symptom category at 5-year follow-up adjusted for age and gender. The differences are statistically significant(P� 0.001a) for both UC (left panel) and CD (right panel) patients. aANCOVA, P values based on the F distribution, adjustment for multiplecomparisons: Bonferroni.

Table V. Rheumatic symptoms and corresponding N-IBDQ total score. Scores are given as estimated marginal mean scores (95% confidenceintervals) adjusted for age and gender (patient number in parentheses)

Symptoms No symptoms P*

UC 180.5 (174.8–186.2) (n = 81) 189.1 (185.8–192.4) (n = 247) 0.011CD 173.0 (164.3–181.6) (n = 50) 182.7 (177.2–188.2) (n = 119) N.S. (0.062)

a ANCOVA, P values based on the F distribution, adjustment for multiple comparisons: Bonferroni.



Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

to explaining the variance in both diagnostic groups was IBDsymptoms (Table VI).

Discussion

This study is the first to describe HRQOL in a population-based cohort of IBD patients followed prospectively for 5years. Since the introduction of the IBDQ (12, 13), this hasbecome the standard for evaluating HRQOL in IBD patients,and several papers have reported HRQOL scores fromselected IBD populations in different clinical situations (24–27).

A total of 121 patients with verified IBD did not meet up atthe hospital and could not be included in this investigation.However, as can be seen from Table I, they did not differsignificantly from those included with regard to age, gender,smoking habits and disease symptoms and severity. Theinclusion of these patients would not have altered our mainfindings or conclusions regarding HRQOL. We thus believethat our findings regarding HRQOL are representative forIBD patients in general.

We found that the IBDQ total score was significantlyreduced by 10 points in women compared to men. Kim et al.(19) found that women had a total score of 6–8 points lowerthan men in their study. Irvine et al. (34) investigated theinfluence of different variables on the IBDQ score in IBDpatients included in an international multi-centre trial, andfound that female gender was a significant predictor for a poorHRQOL. Our results support this finding.

Differences in HRQOL scores between men and women

are frequently reported in the literature, both in normalpopulations and in patient groups. Problems relating toinfertility, pregnancy, sexual activity and social coping mayhelp explain these findings (35–38). However, the clinicalsignificance of our finding regarding a gender differencecould be questioned.

We also found a difference in N-IBDQ scores between thetwo diagnostic groups. Both total and dimension scores werelower in CD patients, indicating that they had a lowerHRQOL than UC patients. We observed a significantreduction in CD patients in the two emotional dimensions.This finding is in accordance with that of Nordin et al. (28),who reported more anxiety and depression in CD patients thanin UC patients. The same findings have also been reported byothers (39, 40). This probably reflects the fact that CD is amore severe systemic disease than UC, which is usuallylimited to colonic mucosal inflammation alone.

Women with CD had a statistically and clinically sig-nificant reduction in IBDQ score compared to the othergroups. There may be a number of explanations for thisfinding. These patients were slightly younger (39.6 yearsversus 43.3 years) and reported more severe symptoms relatedto the disease at the time of the visit than the other groups(60% versus 40%).

The relationship between smoking and IBD has arousedmuch interest and it is now well established that smoking isnegatively associated with UC and positively associated withCD. Smoking also has opposite effects on the clinical courseof the two diseases, with a possible beneficial effect in UC anda detrimental effect in CD. In our study, smoking CD patients

Table VI. In a multiple linear regression analysis (enter method) we investigated the effect of demographic and clinical data (independentvariables) on the N-IBDQ total and dimensional scores (dependent variables) for each diagnostic group separately. All P values equal to orless than 0.05 are given. Adjusted r2 for each dimension is also given

B1 B2 E1 E2 SF Tot

UCAge 0.045Sex 0.026 0.023Symptom severity �0.001 �0.001 �0.001 �0.001 �0.001 �0.001Disease distributionRheumatic symptoms 0.002 0.039 0.009SmokingAdjusted r2 0.266 0.175 0.131 0.146 0.160 0.229

CDAgeSex 0.031 0.024 0.035Symptom severity �0.001 �0.001 �0.001 �0.001 �0.001 �0.001Disease localizationDisease behaviour 0.010Rheumatic symptoms 0.008 0.043SmokingSurgeryAdjusted r2 0.330 0.326 0.218 0.211 0.300 0.361

B1 = Bowel Function-I (stool consistency and pattern).B2 = Bowel Function-II (bowel pain and discomfort).E1 = Emotional Function-I.E2 = Emotional Function-II (Worries).SF = Social Function.Tot = Total score.



Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

had a significantly lower IBDQ score than non-smokers.There was also a significant difference between smoking andnon-smoking women. The difference between smokers andnon-smokers among the UC patients was of the samemagnitude, but was not statistically significant. Russel andco-workers (41) found a relationship between smoking andHRQOL in both UC and CD patients, but we only found asignificant difference between smoking and non-smokingwomen with CD in the IBDQ total score.

Regarding extra-intestinal diseases, the question of interestis the impact of rheumatic complaints on HRQOL. Given thelarge study population, concomitant diseases other than therheumatic ones will probably be equally distributed insubgroups of patients, and therefore not influence the HRQOLbetween patient groups.

We have demonstrated that symptom severity, femalegender, diagnostic group (CD) and rheumatic symptoms alllead to a statistically significant reduction in the IBDQ score,both within each dimension and in the total score. Theregression analysis showed that symptom severity, rheumaticsymptoms and female gender were the strongest predictors ofreduction in HRQOL for both diagnostic groups.

Although the IBDQ total score in this IBD population waslower than that of a normal population (42), it is still relativelyhigh, indicating that most patients had a satisfactory healthstatus despite having a chronic disease. However, it isimportant to identify patients who report a significantreduction in their HRQOL in order to focus on their specialneeds, which should include medical follow-up and socialsupport.

Assessing the clinical impact of statistically significantdifferences is a general problem in HRQOL research, and anumber of methods for overcoming it have been proposed.One suggestion has been to define differences of more than 10points on a 0–100 point scale (e.g. SF-36) as clinicallysignificant (43), but this method has its disadvantages, sincethe means and standard deviations vary considerably and thescales do not relate linearly to external criteria. Also, in somecases, differences smaller than 10 points may be clinicallyimportant (44).

Lydick & Epstein describe two different methods (45), thedistribution-based interpretation and the anchor-based inter-pretation. With the first method we found that a 7-point changein IBDQ total score was clinically significant, and with thesecond that differences of 13 points in IBDQ total score wereclinically significant. Juniper et al. (46) have determined theminimal important difference (MID) for the Asthma Qualityof Life Questionnaire, which also uses a 7-point Likert scale.They concluded that the MID score per item was close to 0.5(range 0.42–0.58). A MID score of 0.5 for the N-IBDQcorresponds to a change of 16 points in the total score, whichhas also been found by Irvine et al. (16). However, as Juniperet al. write in their conclusion; their findings are limited towithin patient changes and a MID of 0.5 may not be correctfor discriminating between patient groups.

Norman et al. (47) conducted a systematic review of theliterature to identify studies that computed an MID. Theyconcluded that the threshold of discrimination for changes inHRQOL for chronic diseases appears to be approximatelyhalf a standard deviation. Compared to our findings, thisindicates that a change of 14 points in total score is thethreshold for MID in our population.

Establishing a fixed value for clinical importance, whetherit is 16 points, 14 points or 7 points, is difficult, since therelevant differences vary with patient populations. A clini-cally important difference has to be established for eachpatient group under consideration and the questionnaires haveto be validated specifically for each patient group. Inepidemiological studies, however, even smaller changesmay be clinically important, and our results indicate that adifference of 7–8 points in the IBDQ total score may be ofclinical importance in our patient population. Since the clini-cal interpretation of changes in score is important for thefuture use of this HRQOL questionnaire, this issue should beaddressed in future studies.

Conclusion

We assessed the HRQOL 5 years after diagnosis in apopulation-based cohort of IBD patients followed prospec-tively. A relatively high IBDQ total score was found in mostpatients, indicating a satisfactory HRQOL. However, wefound that CD women had a clinically significant reduction inscore compared with UC women and CD and UC men.Several demographic and disease-related variables like age,gender, smoking, symptom severity, disease distribution,rheumatic symptoms and surgery were investigated in thecontext of the IBDQ score, and the presence of most of themwas found to result in a clinically significant reduction in thepatient’s HRQOL.

A multiple linear regression model showed that symptomseverity, rheumatic symptoms and female gender were thestrongest predictors of reduction in HRQOL for bothdiagnostic groups.

Acknowledgements

The following members of the Inflammatory Bowel South-Eastern Norway (IBSEN) Study Group of Gastroenter-ologists: Erik Aubert, SØF, Sarpsborg; Taran Søberg andthe late Magnus Melsom, SØF, Moss; Per Tolas, SØF,Halden; Peter Utzon, Lovisenberg Hospital, and BorgarFlaaten, Notodden Hospital, are all thanked for participatingin the study. Permission to use the original IBDQ was kindlygiven by Professor G. Guyatt at McMaster University,Canada. This project was financed with the aid of EXTRAfunds from the Norwegian Foundation for Health andRehabilitation.



Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

References

1. Grand RJ, Homer DR. Approaches to inflammatory boweldisease in childhood and adolescence. Pediatr Clin N Am 1975;22:835.

2. Hendriksen C, Kreiner S, Binder V. Long-term prognosis inulcerative colitis: based on results from a regional patient groupfrom the county of Copenhagen. Gut 1985;26:158–63.

3. Munkholm P, Langholz E, Davidsen M, Kreiner S, Binder V.Incidence and prevalence of Crohn’s disease in the county ofCopenhagen 1962–87: a 6-fold increase in incidence. Scand JGastroenterol 27:6091992.

4. Moum B, Vatn MH, Ekbom A, Aadland E, Fausa O, Lygren I, etal. Incidence of ulcerative colitis and indeterminate colitis in fourcounties of southeastern Norway, 1990–93. A prospectivepopulation-based study. Scand J Gastroenterol 1996;31:362–6.

5. Moum B, Vatn MH, Ekbom A, Aadland E, Fausa O, Lygren I, etal. Incidence of Crohn’s disease in four counties of southeasternNorway, 1990–93. A prospective population-based study. ScandJ Gastroenterol 1996;31:355–61.

6. Moum B, Ekbom A, Vatn MH, Aadland E, Sauar J, Lygren I, etal. Inflammatory bowel disease: re-evaluation of the diagnosis ina prospective population based study in south eastern Norway.Gut 1997;40:328–32.

7. Best WR, Becktel JM, Singleton JW, Kern F. Development of aCrohn’s Disease Activity Index; National Cooperative Crohn’sDisease Study. Gastroenterology 1976;70:439–44.

8. Harvey RF, Bradshaw JM. A simple index of Crohn’s diseaseactivity. Lancet 1980;1:514.

9. Myren J, Bouchier IAD, Watkinson G, Softley A, Clamp SE, deDombal FT. The OMGE Multinational inflammatory boweldisease survey 1976–1982. A further report on 2657 cases. ScandJ Gastroenterol 1984;19 Suppl 95:1–27.

10. van Hees PAM, van Elteren PH, van Lier HJJ, van TongerenJHM. An index of inflammatory activity in patients with Crohn’sdisease. Gut 1989;21:279–86.

11. Irvine EJ. Quality of Life: measurement in inflammatory boweldisease. Scand J Gastroenterol 1993;28 Suppl 199:36–9.

12. Guyatt G, Mitchell A, Irvine EJ, Singer J, Williams N, GoodacreR, et al. A new measure of health status for clinical trials ininflammatory bowel disease. Gastroenterology 1989;96:804–10.

13. Irvine EJ, Feagan B, Rochon J, Archambault A, Fedorak RN,Groll A, et al. Quality of life: a valid and reliable measure oftherapeutic efficacy in the treatment of inflammatory boweldisease. Canadian Crohn’s Relapse Prevention Trial StudyGroup. Gastroenterology 1994;106:287–96.

14. Love JR, Irvine EJ, Fedorak RN. Quality of life in inflammatorybowel disease. J Clin Gastroenterol 1992;14:15–9.

15. Drossman DA, Patrick DL, Mitchell CM, Zagami EW, Appel-baum MI. Health related quality of life in inflammatory boweldisease: functional status and patients worries and concerns. DigDis Sci 1989;34:1379–86.

16. Irvin EJ. Development and subsequent refinement of theinflammatory bowel disease questionnaire: a quality-of-lifeinstrument for adult patients with inflammatory bowel disease.J Ped Gastroenterol Nutr 1999;28:S23–7.

17. Russel MGVM, Pastoor CJ, Brandon S, Rijken J, Engels LGJB,van der Heijde DMFM, et al. Validation of the Dutch translationof the inflammatory bowel disease questionnaire (IBDQ): ahealth-related quality of life questionnaire in inflammatorybowel disease. Digestion 1997;58:282–8.

18. Han SW, McColl E, Steen N, Barton JR, Welfare MR. TheInflammatory Bowel Disease Questionnaire: a valid and reliablemeasure in ulcerative colitis patients in the North East ofEngland. Scand J Gastroenterol 1998;33:961–6.

19. Kim WH, Cho YS, Yoo HM, Park IS, Park EC, Lim JG. Qualityof life in Korean patients with inflammatory bowel diseases:ulcerative colitis, Crohn’s disease and intestinal Behccet’sdisease. Int J Colorec Dis 1999;14:52–7.

20. Cheung WY, Garratt AM, Russell IT, Williams JG. The UKIBDQ—A British version of the inflammatory bowel diseasequestionnaire: development and validation. J Clin Epidemiol2000;53:297–306.

21. Hjortswang H, Jarnerot G, Curman B, Sandberg-Gertzen H, TyskC, Blomberg B, et al. Validation of the inflammatory boweldisease questionnaire in Swedish patients with ulcerative colitis.Scand J Gastroenterol 2001;36:77–85.

22. Pallis AG, Vlachonikolis IG, Mouzas IA. Assessing health-related quality of life in patients with inflammatory boweldisease, in Crete, Greece. BMC Gastroenterol 2002;2:1. Epub2002 Jan 10.

23. Bernklev T, Moum B, Moum T, the Inflammatory Bowel South-Eastern Norway (IBSEN) Group of Gastroenterologists. Qualityof life in patients with inflammatory bowel disease: translation,data quality, scaling assumptions, validity, reliability andsensitivity to change of the Norwegian version of IBDQ. ScandJ Gastroenterol 2002;37:1164–74.

24. Feagan BG, Rochon J, Fedorak RN, Irvine EJ, Wild G,Sutherland L, et al, for the North American Crohn’s StudyGroup: Methotrexate for the treatment of Crohn’s disease. NEngl J Med 1995;332:292–7.

25. Irvine EJ. Effects of budesonide therapy on quality of life inactive Crohn’s disease. Res Clin Forums 1996;18:77–85.

26. Casellas F, Lopez-Vivancos J, Badia X, Vilaseca J, MalageladaJR. Impact of surgery for Crohn’s disease on health-relatedquality of life. Am J Gastroenterol 2000;95:177–82.

27. Broering DC, Eisenberger CF, Koch A, Bloechle C, Knoefel WT,Izbicki JR. Quality of life after surgical therapy of small bowelstenosis in Crohn’s disease. Dig Surg 2001;18:124–30.

28. Nordin K, Pahlman L, Larsson K, Sundberg-Hjelm M, Loof L.Health-related quality of life and psychological distress in apopulation-based sample of Swedish patients with inflammatorybowel disease. Scand J Gastroenterol 2002;37:450–7.

29. Moum B, Ekbom A, Vatn MH, Aadland E, Sauar J, Lygren I, etal. Clinical course during the first year after diagnosis inulcerative colitis and Crohn’s disease. Results of a large pros-pective population-based study in southeastern Norway 1990–1993. Scand J Gastroenterol 1997;32:1005–12.

30. Jahnsen J, Moum B, Schulz T, Sauar J, Lygren I, Stray N, et al.Inflammatory bowel disease, disease course and status 5 yearsafter diagnosis (the IBSEN study). Gastroenterology 2000;118:A1005.

31. Binder V, Both H, Hansen PK, Hendriksen C, Kreiner S, Torp-Pedersen K. Incidence and prevalence of ulcerative colitis andCrohn’s disease in the county of Copenhagen, 1962 to 1978.Gastroenterology 1982;83:563–8.

32. Gasche C, Scholmeric J, Brynskov J, D’Haens G, Hanauer SB,Irvine EJ, et al. A simple classification of Crohn’s disease: reportof the Working Party for the World Congresses of Gastroenter-ology, Vienna 1998. Inflamm Bowel Dis 2000;6:8–15.

33. Palm Ø, Moum B, Jahnsen J, Gran JT. The prevalence andincidence of peripheral arthritis in patients with inflammatorybowel disease, a prospective population-based study (the IBSENstudy). Rheumatology 2001;40:1256–61.

34. Irvine EJ, Grace E, Kerr GD, Lyrenas E, Bolin T, O’Morain C, etal. Non-disease related factors affecting health related quality oflife (HRQOL) in inflammatory bowel disease (IBD). AbstractDDW, 16–22 May 1998, New Orleans, LA.

35. Glise H, Wiklund I. Health related quality of life and gastro-intestinal disease. J Gastroenterol Hepatol 2002;17 Suppl:S73–85.

36. Wenger NK. Quality of life issues in hypertension: consequencesof diagnosis and considerations in management. Am Heart J1988;116:628–32.

37. Vandenburg MJ, Evans SJW, Kelly BJ, Bradshaw F, CurrieWJC, Cooper WJD. Factors affecting the reporting of symptomsby hypertensive patients. Br J Clin Pharmacol 1984; 10: 189–95.

38. van Wijk CM, Kolk AM. Sex differences in physical symptoms:the contribution of symptom perception theory. Soc Sci Med1997;45:231–46.

39. Drossman D, Leserman J, Mitchell M, Zhiming M, Zagamin E,Patrick D. Health status and health care use in persons withinflammatory bowel disease: a national sample. Dig Dis Sci1991;36:1746–55.

40. Helzer JE, Stillings WA, Chammas S, Norland CC, Alpers DH. A



Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.

controlled study of the association between ulcerative colitis andpsychiatric diagnoses. Dig Dis Sci 1982;27:513–8.

41. Russel MG, Nieman FH, Bergers JM, Stockbrugger RW, theSouth Limburg IBD study group: Cigarette smoking and qualityof life in patients with inflammatory bowel disease. Eur JGastroenterol hepatol 1996;8:1075–81.

42. Irvine EJ, Donnelly M. Quality of life comparably impaired inpatients with ulcerative colitis and irritable bowel syndrome.Gastroenterology 1994;106:A517.

43. Hjermstad MJ, Fayers PM, Bjordal K, Kaasa S. Health-relatedquality of life in the general Norwegian population assessed bythe European Organization for Research and Treatment of

Cancer Core Quality-of-Life Questionnaire: the QLQ = C30(�3). J Clin Oncol 1998;16:1188–96.

44. Testa MA, Simonson DC. Assessment of quality of lifeoutcomes. N Engl J Med 1996;334:835–40.

45. Lydick EG, Epstein RS. Clinical significance of quality of lifedata. In: Spilker B, editor. Quality of life and pharmaco-economics in clinical trials, 2nd edn, 1996.

46. Junipher EF, Guyatt GH, Willan A, Griffith LE. Determining aminimal important change in a disease-specific quality of lifequestionnaire. J Clin Epidemiol 1994;47:81–7.

47. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changesin health-related quality of life. The remarkable universality ofhalf a standard deviation. Med Care 2003;41:582–92.

Received 18 August 2003Accepted 24 November 2003



Scan

d J

Gas

troe

nter

ol D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Mel

bour

ne o

n 11

/16/

14Fo

r pe

rson

al u

se o

nly.