Impact of Behçet’s syndrome on health-related quality of life: influence of the type and number of symptoms

Abstract

Objective. To assess the overall impact of Behçet’s syndrome (BS) on quality of life and the specific impact of the type and number of symptoms on the quality of life of adults with BS.

Methods. A questionnaire was mailed to the 641 adult members of the Behçet’s Syndrome Society in the UK. Participants provided information on socio-demographic characteristics, disease duration, current symptoms (mouth ulcers, genital ulcers, skin lesions, fatigue, joint problems, stomach/bowel problems, eye problems, pathergy reaction, headaches and other neurological problems), symptom control and quality of life (the EQ-5D index). Linear regression was used to test the associations of the type and number of symptoms with the EQ-5D index.

Results. Of the 447 members, 400 who returned the questionnaires had a confirmed diagnosis of BS. Of them, 362 had information on the variables selected for this analysis (76% females and 94% white British). The mean EQ-5D index was 0.47 (s.d. 0.38). Of the 10 symptoms assessed, joint problems had the strongest impact on quality of life, followed by neurological problems, pathergy reaction and stomach/bowel problems (adjusted coefficients of −0.15, −0.13, −0.11 and −0.18, respectively). Furthermore, the number of symptoms was significantly related to the EQ-5D index after adjustment for socio-demographic characteristics, disease duration and symptom control. The EQ-5D index decreased by −0.05 U for every additional symptom reported.

Conclusions. BS has a considerable impact on quality of life. Both the type and number of symptoms affect the quality of life of adults with BS.

Behçet’s syndrome, Quality of life, Signs and symptoms, Adults

Introduction

Behçet’s syndrome (BS) is a systemic inflammatory vasculitis of unknown aetiology, affecting vessels of different types, sizes and locations [1]. BS is mainly characterized by recurrent episodes of acute inflammation, including mucocutaneous manifestations (oral aphthous ulcers, genital ulcers and skin lesions) and gastrointestinal, musculoskeletal, neurological, ophthalmic and vascular involvement [1–3]. However, there is considerable variation in its clinical presentation between and within individuals [3]. The prevalence of BS in the UK is estimated to be 0.64 per 100 000 individuals [1, 2].

The burden of BS is usually confined to the initial years of its course, with most symptoms abating with the passage of time [2, 3]. Mucocutaneous manifestations are the most common symptoms in adults with BS while eye, vascular and neurological involvement are the most serious [2]. The complex patterns of symptoms in adults with BS can cause various degrees of activity limitation and restriction in everyday life. Although previous studies have shown the impact of BS on quality of life [4–13], most of them focused on the influence of one particular symptom, such as cutaneous manifestations [4], intestinal problems [5], ocular involvement [6], arthritis [8], oral ulcers [9, 13] and major depression [10]. The remaining studies looked at the relationship between disease activity and quality of life [7, 12] or compared the quality of life of BS patients with other groups [11]. No studies have explored the differential or cumulative impact of multiple symptoms on quality of life.

Health-related quality of life (HRQoL) can be assessed using either generic or specific measures. Generic HRQoL measures take into account numerous conditions, some occurring simultaneously, and thus collect information about wider effects of health status on daily living. The main advantage of generic HRQoL measures is that they allow comparison of various domains of quality of life for the condition being studied, as well as across populations, groups of patients and disease states [14]. One of the most widely used generic HRQoL measures is the EQ-5D, which is a simple, self-administered questionnaire that incorporates descriptions and valuations of different health states [15]. As such, EQ-5D provides both a health profile and an index for individuals or groups that allow clinical and economic evaluation of health interventions [15, 16]. Besides, EQ-5D can be used to quantify and compare the burden of different health conditions between groups. Therefore, the purpose of this study was to explore the impact of BS on HRQoL, as measured by the EQ-5D. A second objective was to assess the impact of the type and number of symptoms on the HRQoL of adults with BS.

Methods

This study adopted a cross-sectional postal survey design. Ethical approval was obtained from the Queen Mary Research Ethics Committee. The inclusion criteria were to be age ≥18 years and have a confirmed diagnosis of BS. The Behçet’s Syndrome Society (BSS) provided the names and addresses of potential participants. A large proportion of individuals with BS in the UK are members of the BSS. After excluding individuals aged <18 years, the total number of members was 641.

It was calculated that a minimum sample size of 312 individuals was required to estimate an s.d. of 0.45 in the EQ-5D index for adults with BS, with an absolute precision of 0.05 and 95% confidence level. This initial calculation was increased by 50% (n = 468) to compensate for possible non-response and incompleteness of data in the postal survey. It was decided to mail questionnaires to all adult members of the BSS (n = 641) because it was not possible to ascertain from institutional records which members had confirmed diagnosis of BS.

A cover letter, information sheet and pre-paid envelope were sent in the survey pack. One reminder was sent to increase the response rate. The survey questionnaire gathered information on participants’ socio-demographic (sex, age, ethnicity, marital status and education) and disease-related characteristics (age of diagnosis, current symptoms and symptom control). Ethnicity was self-assigned using an adaptation of the 2001 UK Census categories, which included 24 possible categories under five main ethnic groups (white, Asian, black, mixed or other). Education was indicated by the highest degree or qualification (no qualifications, secondary school, A levels, technical qualifications, first university degree or higher degree). Participants reported the BS symptoms they were currently experiencing by using a list of the 10 most common BS manifestations, namely mouth ulcers, genital ulcers, skin lesions, fatigue, joint problems, stomach/bowel problems, eye problems, pathergy reaction, headaches and other neurological problems [1–3]. They also reported the extent to which they consider their symptoms to be under control (no, sometimes, mostly or yes).

HRQoL was measured using the EQ-5D, which consists of five questions covering the following health domains: mobility, self-care, usual activity, pain and anxiety/depression. Participants were asked to choose their level of problems in each domain from three options: 1 (no problems), 2 (moderate problems) and 3 (extreme problems). These choices result in a five-digit score for each participant that reflects a unique health state. An index score can be derived for each health state to indicate its value or utility from the perspective of a country-specific general population [15, 16]. Scores in the UK EQ-5D value set range from −0.59 for the worst possible health status to 1.0 for perfect health, with 0 on the scale representing the state of being dead and negative values representing health states regarded as worse than being dead [17]. EQ-5D also includes a visual analogue scale (EQ VAS) from 0 for the worst imaginable health state to 100 for the best.

Statistical analysis

Due to small numbers in some response categories of covariates, marital status was reclassified as married (including remarried and co-habiting), single and separated/widowed (also including divorced); and education as no qualification, secondary school, A levels (including technical qualifications) and higher education (including first and higher university degrees). Besides, a summary score was calculated to indicate the number of symptoms reported by each participant, ranging between 0 and 10. This score represents the cumulative experience of symptoms.

Linear regression models were used to explore the association of the type and number of symptoms with the EQ-5D index, as the outcome measure resembled the normal distribution. First, the relationship between each type of symptom and the EQ-5D index was assessed in unadjusted, fully adjusted (sex, age, marital status, education, disease duration and symptom control) and mutually adjusted models (additionally for other types of symptoms). Secondly, the relationship between the number of symptoms and the EQ-5D index was assessed in unadjusted and fully adjusted models (sex, age, marital status, education, disease duration and symptom control). Statistical interactions (cross-products) of sex with the type and number of symptoms were examined by assessing their significance when added to their corresponding fully adjusted models one at a time.

Results

A total of 447 adult members of the BSS returned the questionnaires (70% response rate), of whom 400 had a confirmed diagnosis of BS. Of them, 362 provided usable data on the variables selected for this analysis. Nearly all the participants (94%) were white British and 76% were female (Table 1). The mean disease duration was 15.4 years (s.d. 11.3; range 0–53). Fatigue, joint problems and mouth ulcers were the most commonly reported symptoms (84, 81 and 77%, respectively). The total number of symptoms was 5.4 (s.d. 2.3; range 0–10), with only 2% of the participants reporting no symptoms.

Table 1

Characteristics of the participants (n = 362)

Variable	n (%)	Mean (s.d.)	EQ-5D index Mean (s.d.)
Sex (missing = 3)
Men	85 (24)		0.54 (0.39)
Women	274 (76)		0.45 (0.37)
Age in years (missing = 12)		49.8 (12.4)
Ethnicity (missing = 1)
White British	339 (94)		0.47 (0.37)
White non-British	5 (1)		0.70 (0.06)
Asian	5 (1)		0.58 (0.40)
Other	12 (4)		0.25 (0.43)
Marital status (missing = 0)
Married	244 (67)		0.48 (0.37)
Single	45 (13)		0.50 (0.38)
Separated/widowed	73 (20)		0.40 (0.37)
Education (missing = 12)
No qualification	60 (17)		0.43 (0.35)
Secondary school	112 (33)		0.41 (0.37)
A levels	89 (25)		0.48 (0.40)
Higher education	89 (25)		0.56 (0.36)
Disease duration, years (missing = 22)		15.4 (11.3)
Current symptoms (missing = 2)
Mouth ulcers
No	83 (23)		0.58 (0.36)
Yes	277 (77)		0.43 (0.38)
Genital ulcers
No	194 (54)		0.56 (0.35)
Yes	166 (46)		0.36 (0.37)
Skin lesions
No	188 (52)		0.53 (0.36)
Yes	172 (48)		0.40 (0.38)
Fatigue
No	56 (16)		0.69 (0.33)
Yes	304 (84)		0.43 (0.37)
Joint problems
No	67 (19)		0.76 (0.29)
Yes	293 (81)		0.40 (0.36)
Stomach/bowel problems
No	163 (45)		0.59 (0.34)
Yes	197 (55)		0.37 (0.37)
Eye problems
No	203 (56)		0.54 (0.35)
Yes	157 (44)		0.37 (0.39)
Pathergy reaction
No	296 (82)		0.51 (0.36)
Yes	64 (18)		0.27 (0.38)
Headaches
No	161 (45)		0.58 (0.34)
Yes	199 (55)		0.38 (0.38)
Neurological problems
No	252 (70)		0.55 (0.34)
Yes	108 (30)		0.28 (0.38)
Total number of symptoms		5.4 (2.3)
Symptom control (missing = 5)
No	43 (12)		0.15 (0.39)
Sometimes	97 (27)		0.30 (0.35)
Mostly	151 (42)		0.56 (0.30)
Yes	66 (19)		0.73 (0.28)

The mean EQ-5D index was 0.47 (s.d. 0.38; range −0.59 to 1) compared with 0.86 (s.d. 0.23) in the UK general adult population. BS affected all the five domains of the EQ-5D descriptive system. Pain/discomfort (87%), usual activities (75%) and mobility (65%) were the most commonly affected domains in adults with BS. In addition, 60% of participants reported having at least moderate problems in the domain of anxiety/depression. Self-care was the least commonly affected domain (36%). The mean EQ VAS was 56.8 (s.d. 21.0; range 0–100).

The type of symptom had a significant impact on HRQoL, although the most prevalent symptoms were not those with the strongest impact on HRQoL (Table 2). Seven of the 10 symptoms assessed were associated with significant decrements in HRQoL (fully adjusted coefficients ranged between −0.08 for headaches and −0.22 for joint problems). However, when looking at all symptoms together, pathergy reaction as well as joint, neurological and stomach/bowel problems were the only symptoms having an independent impact on HRQoL. Of them, joint problems had the strongest impact on HRQoL (mutually adjusted coefficient of −0.15), followed by neurological problems (−0.13), pathergy reaction (−0.11) and stomach/bowel problems (−0.08), respectively. The interactions between sex and each particular symptom were not statistically significant (P > 0.213 for the 10 tests).

Table 2

Models for the association between different types of symptoms and the EQ-5D index (n = 322)

Symptoms	n (yes)	Unadjusted coefficient (95% CI)	Fully adjusted coefficienta (95% CI)	Mutually adjusted coefficientb (95% CI)
Mouth ulcers	245	−0.14 (−0.23, −0.04)**	−0.03 (−0.12, 0.05)	0.03 (−0.06, 0.12)
Genital ulcers	145	−0.18 (−0.26, −0.10)***	−0.10 (−0.18, −0.03)**	−0.06 (−0.14, 0.02)
Skin lesions	152	−0.13 (−0.21, −0.05)**	−0.06 (−0.13, 0.02)	0.01 (−0.07, 0.08)
Fatigue	270	−0.28 (−0.39, −0.17)***	−0.14 (−0.24, −0.04)**	−0.05 (−0.15, 0.05)
Joint problems	259	−0.35 (−0.45, −0.25)***	−0.22 (−0.31, −0.12)***	−0.15 (−0.25, −0.06)**
Stomach/bowel problems	176	−0.25 (−0.32, −0.17)***	−0.15 (−0.22, −0.07)***	−0.08 (−0.15, −0.01)*
Eye problems	138	−0.16 (−0.24, −0.08)***	−0.06 (−0.14, 0.02)	−0.01 (−0.08, 0.06)
Pathergy reaction	57	−0.24 (−0.35, −0.14)***	−0.17 (−0.26, −0.08)***	−0.11 (−0.20, −0.02)*
Headaches	178	−0.20 (−0.28, −0.12)***	−0.08 (−0.16, −0.004)*	0.00 (−0.07, 0.08)
Neurological problems	97	−0.25 (−0.34, −0.17)***	−0.17 (−0.25, −0.09)***	−0.13 (−0.21, −0.06)**

Linear regression was fitted. Regression coefficients reported represent differences in the EQ-5D index between patients with and without each specific symptom. ^aAdjusted for sex, age (continuous), marital status, education, disease duration and symptom control. ^bAdjusted for sex, age (continuous), marital status, education, disease duration, symptom control and other symptoms. *P < 0.05, **P < 0.01, ***P < 0.001.

The number of symptoms also had a significant impact on HRQoL (Table 3). There was a significant decrement in HRQoL for every additional symptom reported (fully adjusted coefficient of −0.05). For example, the cumulative negative effect of experiencing pathergy reaction, joint, neurological and stomach/bowel problems on HRQoL was significantly high (fully adjusted coefficient of −0.25). In addition, marital status and symptom control affected the HRQoL of adults with BS. Adults with BS who were separated or widowed reported poorer HRQoL compared with those married (fully adjusted coefficient of −0.09). The extent to which participants considered their symptoms to be under control positively affected their HRQoL (fully adjusted coefficients varied between 0.12 and 0.38). The interaction between sex and number of symptoms was not statistically significant (P = 0.660).

Table 3

Models for the association between number of symptoms and the EQ-5D index (n = 322)

Variables	n	Unadjusted coefficient (95% CI)	Fully-adjusted coefficient (95% CI)
Sex
Men	79	0 [Reference]	0 [Reference]
Women	243	−0.09 (−0.18, 0.01)	−0.02 (−0.10, 0.06)
Age (in 10 years)	322	−0.04 (−0.07, −0.003)*	−0.02 (−0.05, 0.02)
Marital status
Married	216	0 [Reference]	0 [Reference]
Single	39	−0.01 (−0.14, 0.12)	−0.05 (−0.16, 0.06)
Divorced/widowed	67	−0.11 (−0.21, −0.003)*	−0.09 (−0.17, −0.004)*
Education
No qualification	57	0 [Reference]	0 [Reference]
Secondary school	101	−0.02 (−0.14, 0.10)	−0.04 (−0.14, 0.07)
A levels	83	0.05 (−0.07, 0.18)	0.05 (−0.06, 0.15)
Higher education	81	0.14 (0.01, 0.27)*	0.10 (−0.01, 0.21)
Disease duration (in 10 years)	322	−0.001 (−0.04, 0.04)	−0.01 (−0.04, 0.03)
Number of symptoms	322	−0.08 (−0.09, −0.06)***	−0.05 (−0.07, −0.03)***
Symptom control
No	37	0 [Reference]	0 [Reference]
Sometimes	88	0.13 (0.01, 0.26)*	0.12 (−0.001, 0.24)
Mostly	141	0.39 (0.27, 0.51)***	0.30 (0.18, 0.42)***
Yes	56	0.58 (0.44, 0.71)***	0.38 (0.23, 0.52)***

Linear regression was fitted. Regression coefficients reported represent differences in the EQ-5D index between patients with and without each specific symptom. *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

This is the first study to use a generic HRQoL measure in adults with BS in the UK. It demonstrates clearly the negative impact of BS on quality of life, at least to the extent that the latter is measured by the EQ-5D. Adults with BS who responded to the postal survey were characterized by a long-lasting illness with multiple symptoms. Participants experienced high levels of pain and discomfort as well as moderate to extreme problems with everyday functioning. More importantly, adults with BS reported poorer HRQoL than the reference sample of the general adult population and groups with other chronic conditions in the UK, as reported in previous studies using the EQ-5D index (Table 4). Deterioration in HRQoL of adults with BS was such that the EQ-5D index decreased by 45% compared with the general population [17]. Furthermore, BS had one of the worst impacts on HRQoL in the UK population [18–24], only comparable to the impact of multiple sclerosis [25] and arthritis [26, 27]. Therefore, this study shows that adults with BS appear to live in a utility state that is equivalent or somewhat worse than many other prevalent chronic conditions.

Table 4

EQ-5D index for the general adult population and different groups of patients in the UK

Study	Group	Female, %; age, mean (s.d.), range, years	EQ-5D index
			n	Mean (s.d.)
Harper et al. [18]	COPD	51; 62 (10.3), 35+	125	0.52 (0.16)
Brazier et al. [27]	OA of the knee	>67; 64	107	0.46 (0.17)
Kind et al. [17]	General population	54; 18+	3392	0.86 (0.23)
Myers and Wilks [19]	Chronic fatigue syndrome	67; 39.2 (12.7), 15–74	85	0.56 (0.35)
Walters et al. [20]	Venous leg ulcers	67; median age: 75	233	0.57 (0.18)
Haywood et al. [21]	AS	26; 42.1 (12.6), 18–75	342	0.53 (0.35)
Akehurst et al. [22]	Irritable bowel syndrome	86; 47.1 (11.9)	157	0.68 (0.25)
Currie et al. [23]	Ischaemic heart disease	Not reported	789	0.64 (0.29)
	Crohn's disease	Not reported	73	0.69 (0.29)
	Ulcerative colitis	Not reported	61	0.79 (0.24)
Matza et al. [24]	Type 2 diabetes	35; 55.7 (10.3), 30–75	130	0.75 (0.30)
McCrone et al. [25]	Multiple sclerosis	72; 54.5 (11.4), 18–89	1923	0.41 (0.34)
Harrison et al. [26]	Long-standing arthritis	68; 60.6 (11.2), 18+	466	0.59 (0.22)
	Severe arthritis	76; 60.2 (11.7)	188	0.55 (0.27)
	Very early arthritis	69; 55.5 (15.0), 18+	182	0.46 (0.31)
	Active arthritis	75; 57.9 (12.2)	223	0.34 (0.33)
Current study	BS	76; 49.8 (12.4), 19–81	362	0.47 (0.38)

COPD: chronic obstructive pulmonary disease.

Both the type and number of symptoms had a negative effect on HRQoL. Musculoskeletal, neurological and gastrointestinal involvement and skin hyperreactivity contributed considerably and independently to deterioration in HRQoL. In contrast, highly prevalent symptoms such as mouth ulcers and fatigue, although debilitating, had a small relative impact on HRQoL compared with other symptoms. There are two possible explanations for this finding: first, some symptoms may be more severe than others, particularly those associated with organ systemic involvement [1, 2]; secondly, adults with BS may have learned how to cope with frequent symptoms during the course of their condition, which in turn become less distressing with every recurrence. In addition, the accumulation of symptoms in adults with BS was related to further deteriorations in HRQoL suggesting a cause–effect relationship. Every increase in the number of symptoms leads to a utility loss of 0.05 U on the EQ-5D index, which can be considered as meaningful based on current estimates for the minimally important difference in the EQ-5D index among different groups of patients [28, 29]. These findings suggest that practitioners should not overlook the type and number of symptoms in the treatment plan of patients with BS. Furthermore, these findings highlight the need for collaborative multidisciplinary health care teams to provide appropriate medical care to BS patients. As expected, marital status and symptom control affected the HRQoL of adults with BS. Further research should elucidate the buffer effect of marriage on the impact of BS on quality of life. Socio-psychological factors such as social support from a special person may help in coping with symptoms of BS.

The present findings should be interpreted taking into account the strengths and limitations of the study. The strengths were the higher number of adults with BS compared with previous studies; the relatively high response rate for a postal survey; and the use of a standardized internationally recognized and widely used HRQoL measure. Although the EQ-5D has been criticized as having insufficient dimensions, insensitive scales and excessively low utility values for certain health states [30], our findings support its validity among adults with BS as it was able to discriminate between adults with different types and numbers of symptoms. This study also has some limitations. First, the cross-sectional nature of the data restricts the ability to make inferences regarding the direction of the associations. Nevertheless, it is less reasonable to believe that domains of quality of life affected symptoms rather than the association being the other way around. Secondly, the study did not include a control group for comparison of HRQoL levels. However, the EQ-5D allowed comparisons with the general population and other groups where the same instrument was used [17–27]. The general population is a good control group because the prevalence of BS is low and it will therefore contribute very little to the average impact on quality of life in the general population. Thirdly, as we used a postal questionnaire for data collection, the diagnosis of BS was self-reported. However, the information obtained from participants was based on a previous clinical examination, and it is unlikely that this resulted in some misclassification. Fourthly, this study focused on the frequency of symptoms but not on their severity, as the latter would have required a longer and less attractive questionnaire for a postal survey. However, this precluded us from assessing the impact on quality of life associated with the extent of organ system involvement. Further studies should explore associations between involvement in different organ systems and quality of life. Fifthly, the study sample consisted mainly of white British and women. Although the proportion of white British in the sample (94%) was similar to the 2001 UK Census (92%), women were overrepresented. Although this uneven sex distribution could have influenced our findings, we controlled for a comprehensive set of socio-demographic confounders and found additionally that the impact of the type and number of symptoms on quality of life did not differ by sex. Therefore, the present findings represent valid relationships for the variables of interest, but may not be generalizable to all the BSS members.

In conclusion, BS has a considerable impact on HRQoL, as measured by the EQ-5D. Adults with BS reported high levels of pain/discomfort, and problems with mobility and usual activities. The impact of BS is substantial when compared with the general adult population, and groups with other serious chronic conditions, such as multiple sclerosis and arthritis, in the UK. This study also demonstrates the significant burden of single symptoms (especially musculoskeletal, neurological and gastrointestinal problems and skin hyperreactivity) and the even further burden of cumulative symptoms in adults with BS in terms of impact on HRQoL.

Acknowledgements

Funding: This research was partially supported by the BSS.

Disclosure statement: The authors have declared no conflicts of interest.