To evaluate the validity and responsiveness of a German-language version of the disease-specific Graves orbitopathy quality-of-life questionnaire (GO-QOL).
Prospective cross-sectional study.
At a multidisciplinary university orbital center, 310 consecutive unselected Graves orbitopathy outpatients answered the GO-QOL before undergoing complete ophthalmic and endocrine assessment. The main outcome measures were the GO-QOL and its 2 subscales, Visual Functioning and Appearance.
The QOL scores for the subscales Visual Functioning and Appearance were (mean ± SE) 72.5 ± 1.4 and 71.3 ± 1.5, respectively. Visual Functioning and Appearance were higher in mild (82.2 ± 2.2 and 86.0 ± 17.6) than in moderate to severe (66.6 ± 1.8, p < .001, 95% CI 15.6−2.9 and 65.5 ± 25.5, P < .001, 95% CI 15.1−26.0) and in sight-threatening Graves orbitopathy (41.9 ± 9.9, P < .001, 95% CI 25.4−55.3 and 58.5 ± 9.0, P < .001, 95% CI 15.0−39.4). Visual Functioning and Appearance were also lower in active (63.3 ± 2.2 and 64.5 ± 2.2) than in inactive stages (77.0 ± 1.9; P < .001, 95% CI 7.9−19.3 and 78.3 ± 1.8, P < .001, 95% CI 8.2−19.2). Visual Functioning was 81.6 ± 1.8 in patients without and 62.4 ± 2.0 in patients with diplopia ( P < .001, 95% CI 13.8−24.6). Appearance was lower in those receiving psychotherapy (64.7 ± 3.2) than in those without psychotherapy (74.6 ± 1.6, P = .005, 95% CI 3.0−16.7). Significant ceiling effects (≥15% at the highest value of the subscale) were observed for Appearance in 59 patients (19%) and for Visual Functioning in 85 patients (27%).
The German-language version of the GO-QOL shows evidence of validity in Graves orbitopathy and it usefully complements ophthalmic assessment in these patients.
Patients with Graves disease and related orbitopathy suffer from disfiguring proptosis and large lid fissures. Diplopia, orbital pain, lacrimation, corneal ulceration, and compression of the optic nerve may cause functional restrictions to the point of loss of vision. There is general agreement that patients’ perceptions of how they feel and how they are able to function in daily life should be included in the evaluation and monitoring of clinical parameters and treatment. Therefore, health-related quality of life in Graves orbitopathy has been evaluated using the Medical Outcomes Study 36- (or 24)-item Short-Form Health Status Survey. These studies emphasized the impact of Graves orbitopathy on both physical and psychological functioning as well as on patients’ quality of life (QOL). Also, anxiety and depression are prevalent in these patients. Together with a hyperthyroid state, blurred vision, stressful life events, depressive coping, and, foremost, anxiety have a major impact on QOL in Graves orbitopathy. Thus, the patients are not only physically ill, they also exhibit emotional distress. On the other hand, when using the National Eye Institute Visual Function Questionnaire (NEI VFQ-25) for measuring health-related quality of life in Graves orbitopathy, Bradley and associates concluded that this questionnaire did not thoroughly address several important issues for these patients.
A disease-specific questionnaire, GO-QOL (Graves orbitopathy quality-of-life questionnaire), has been developed and validated by members of the European Group on Graves Orbitopathy (EUGOGO), and has been proven to accurately measure well-being in Graves orbitopathy. The GO-QOL was originally developed in Dutch, and by now it has been translated into 6 languages. A forward and backward translation was performed by native speakers and the translations have been reviewed by 2 independent experts in the field of the disease, 1 ophthalmologist and 1 endocrinologist. The final translations were then reviewed by all members of the EUGOGO members and by the developers of the GO-QOL. All EUGOGO centers performed a pilot testing of the translated version at their weekly multidisciplinary orbital clinics. The different versions of the GO-QOL are available online on the EUGOGO website ( www.eugogo.eu ). In this prospective study, we aimed to evaluate the validity and responsiveness of the German version of the GO-QOL.
Patients and Methods
In total, 310 consecutive unselected patients with Graves orbitopathy were prospectively included in this institutional study. At the multidisciplinary Orbital Center of the Johannes Gutenberg-University Medical Center in Mainz, Germany, each individual patient underwent complete ophthalmic and endocrine assessment. Prior to their medical investigation all patients answered the GO-QOL questionnaire. As proposed by EUGOGO, the GO-QOL is used at our institution daily for every patient with Graves orbitopathy as a separate diagnostic parameter, and it is not reserved for clinical trials. Additionally, all patients were asked whether they had psychotherapy because of Graves orbitopathy. Blood testing was performed on the same day.
To evaluate the validity and responsiveness of the German version of the GO-QOL, the data were assessed for significant correlations without hypothesis testing. More specifically, correlations between the 2 subscales of the German version of the GO-QOL and clinical findings were analyzed. We also looked for correlations of the GO-QOL subscales with the possible confounders age, sex, and smoking behavior. Furthermore, we investigated whether impaired quality of life was related to the need for psychotherapy. Although the GO-QOL was originally designed to measure quality-of-life impairment attributable to the eye disease only, correlations of thyroid function and of thyroid-related autoantibodies (TSH receptor autoantibodies) with the GO-QOL were tested, too.
We determined that the expected mean differences in GO-QOL scores between relevant subgroups were valid if they were at least 6 points. We chose 6 points as the minimal difference as Terwee and associates recommended 6 to 10 points as a minimal clinically important difference.
The Graves Orbitopathy Quality-of-Life Questionnaire
The GO-QOL consists of 2 subscales: 1 for visual functioning (Visual Functioning score; 8 questions referring to the limitations attributable to decreased visual acuity and/or diplopia) and 1 for appearance (Appearance score; 8 questions referring to the limitations in psychosocial functioning attributable to changes in appearance). The GO-QOL questions are scored as “severely limited” (1 point), “a little limited” (2 points), or “not limited at all” (3 points). After conversion to numerical values, the answers to questions 1 through 8 and 9 through 16 are added together to yield 2 raw scores ranging from 8 to 24 points. These raw scores are then transformed to 2 total scores from 0 to 100 using the following formula: total score = (raw score – #)/2 x # x 100, where # stands for the number of completed items. For both scores, higher scores indicate better QOL.
Disease activity and severity were evaluated according to the clinical activity score (CAS) and the NOSPECS (No signs or symptoms, Only signs, Soft tissue involvement, Proptosis, Extraocular muscle involvement, Corneal involvement, Sight loss) score and classified as recommended by EUGOGO. According to these recommendations, patients with a mild disease features only experience a minor impact on their daily life, which is insufficient to justify immunosuppressive or surgical therapy. These patients have only minor lid retraction (<2 mm), mild soft tissue involvement, proptosis less than 3 mm above normal for race and gender, intermittent or no diplopia, and conservatively treatable corneal exposure. In patients with moderately severe disease, Graves orbitopathy has sufficient impact on daily life to justify the risks of immunosuppression (active disease) or surgical intervention (inactive disease). These patients experience lid retraction of at least 2 mm, moderate or severe soft tissue involvement, and proptosis of at least 3 mm above normal adjusted for race and gender, and/or inconstant or constant diplopia. Patients with sight-threatening Graves orbitopathy suffer from optic neuropathy and/or corneal ulceration. In these cases, immediate intervention is urgently needed. Graves orbitopathy was classified as active if at least 3 of the following 7 symptoms were present: orbital pain at rest or in motion, swelling or redness of the eyelid or the conjunctiva, or swelling of the caruncle. Motility disorders were classified and scored according to the Gorman score as no diplopia (0 points), intermittent diplopia (eg, when tired; 1 point), inconstant diplopia (when looking in certain directions; 2 points), and constant diplopia (in primary position; 3 points). Hertel exophthalmometry was performed to quantify proptosis. Lid fissures were measured with a ruler in the primary gaze position.
Statistical analyses were performed using SPSS (Statistical Package for the Social Sciences, Version 17.0, Chicago, Illinois, USA), a commercially available software package. The Spearman correlation coefficient was calculated for correlations between 2 metric variables. To analyze correlations between the GO-QOL scores and metric variables, the Spearman correlation coefficient was calculated. Correlation coefficients >0.5 were defined to satisfactorily prove correlations. The χ 2 test was used to investigate dependencies between 2 categorical variables. Correlations between categorical and metric variables were calculated by the unpaired t test and means, standard errors (SE), and confidence intervals (CI) of the differences of the means were indicated. A significant result was indicated with P < .05; P values >.05 to .1 were considered a trend. Significant correlations were included in a multivariate analysis. More specifically, all variables that had a significant effect on the GO-QOL subscales within the univariate analyses were included in a stepwise linear regression model. The regression analysis was used to analyze the relationship between the independent variables (factors of influence on the GO-QOL, eg, disease severity or activity) and each dependent variable (the GO-QOL subscales Appearance or Visual Functioning) and to assess which of the independent variables may have no relationship to the dependent variable at all. Furthermore, the multivariate analysis was used to identify whether there were independent variables containing redundant information about the dependent variable; thus once 1 of them is known, the others are no longer informative. The regression model included factors that could confound the relationship between the clinical variables and the GO-QOL subscales, namely age, gender, and smoking behavior.
The fraction of patients scoring 0 or 100 on the 2 subscales of the GO-QOL was calculated to assess floor and ceiling effects. We chose to orientate on the definitions of significant ceiling effects by Bradley and associates as their requirements seemed to be similar to the ones of the present trial. These authors defined significant ceiling effects to be more than 15% and substantial ceiling effects to be more than 30% of responses at the maximum value for each subscale. On the other hand, significant and substantial floor effects were defined to be more than 15% and more than 30%, respectively, of the responses at the minimum value for the subscale.
Demographic data for the large group of consecutive and unselected patients with Graves orbitopathy are shown in Table 1 . More than half of the patients had diplopia and a moderately severe Graves orbitopathy, whereas an active disease was present in nearly 40% of the patient population. A relatively high percentage of subjects with optic neuropathy was noted, most probably because our institution is a reference orbital center attracting more severe cases. Also, the vast majority of the Graves patients were TSH receptor autoantibody positive. Most patients were undergoing anti-thyroid therapy with methimazole and were euthyroid. Finally, less than one-third of the Graves orbitopathy subjects smoked.
|Age, median (range), years||48.0 (13−81)|
|Active Graves orbitopathy||129||42|
|Moderate to severe||176||57|
|Proptosis, median (range), mm||19 (6.5−31)|
|Lid fissure width, median (range), mm||10 (5.5−16.5)|
|TSH receptor antibody–positive||260||70|
Disease-Specific Quality of Life
All 310 patients with Graves orbitopathy completed the GO-QOL (100% response rate). The frequencies of the responses to each question are shown in Table 2 . The majority of the patients reported limitations in their daily activities such as driving, indoor and outdoor activities, reading, watching TV, and hobbies, as well as impaired self-confidence.
|Seriously Limited n (%)||A Little Limited n (%)||Not Limited n (%)||Missing Value n (%)|
|Visual functioning subscale|
|1) Bicycling||50 (16.1)||81 (26.1)||170 (54.8)||9 (2.9) a|
|2) Driving||53 (17.1)||95 (30.6)||147 (47.4)||15 (4.8) a|
|3) Moving around the house||11 (3.5)||76 (24.5)||223 (71.9)||0|
|4) Walking outdoors||16 (5.2)||128 (41.3)||166 (53.5)||0|
|5) Reading||32 (10.3)||150 (48.8)||128 (41.3)||0|
|6) Watching TV||23 (7.4)||152 (49.0)||135 (43.5)||0|
|7) Hobby or pastime||33 (10.6)||118 (38.1)||159 (51.3)||0|
|8) Hindered from something that you wanted to do||33 (10.6)||103 (33.2)||174 (56.1)||0|
|Yes, Very Much n (%)||Yes, A Little n (%)||No, Not At All n (%)||Missing Value n (%)|
|9) Changed appearance||122 (39.4)||115 (37.1)||73 (23.5)||0|
|10) Being stared at in the streets||34 (11.0)||86 (27.7)||190 (61.3)||0|
|11) People react unpleasantly||15 (4.8)||78 (25.2)||217 (70.0)||0|
|12) Influence on self-confidence||67 (21.6)||101 (32.6)||142 (45.8)||0|
|13) Socially isolated||8 (2.6)||29 (9.4)||273 (88.1)||0|
|14) Influence on making friends||30 (9.7)||53 (17.1)||227 (73.2)||0|
|15) Appear less often on photos||67 (21.6)||44 (14.2)||199 (64.2)||0|
|16) Mask changes in appearance||44 (14.2)||82 (26.5)||184 (59.4)||0|
The GO-QOL scores for the subscales Visual Functioning and Appearance were (mean ± SE) 72.5 ± 1.4 and 71.3 ± 1.5, respectively.
Visual Functioning Subscale
The GO-QOL subscale Visual Functioning negatively correlated with the disease severity score (r = −0.44, P < .001) and the activity score (r = -0.42, P < .001) as well as with motility disorders (r = -0.39, P = .001). Visual functioning was 82.2 ± 2.2 in patients with mild vs in patients with moderate to severe Graves orbitopathy (66.6 ± 1.8, P < .001, 95% CI 15.6−2.9) and patients with a sight-threatening disease (41.9 ± 9.9, P < .001, 95% CI 25.4−55.39). The difference of the Visual Functioning score between patients with moderate to severe Graves orbitopathy and with sight-threatening Graves orbitopathy was also statistically significant ( P = .002, 95% CI 9.5−40.0). Patients with an active disease had a markedly lower Visual Functioning score (63.3 ± 2.2) than those with an inactive disease (77.0 ± 1.9, P < .001, 95% CI 7.9−19.3). Also, Visual Functioning score closely correlated with diplopia: Visual Functioning was 81.6 ± 1.8 in patients without and 62.4 ± 2.0 in patients with diplopia ( P < .001, 95% CI 13.8−24.6). Specifically, Visual Functioning was 62.6 ± 3.9 in patients with intermittent diplopia, 67.4 ± 2.6 in patients with inconstant diplopia, and 50.7 ± 4.1 in patients with constant diplopia. The differences in the Visual Functioning score between patients with intermittent and constant diplopia ( P = .040, 95% CI 0.6−23.3) and between those with inconstant and constant diplopia ( P = .001, 95% CI 7.1−26.2), although not between patients with intermittent and inconstant diplopia ( P = .312, 95% CI −14.1−4.5), were statistically significant. No correlations between Visual Functioning and thyroid function or TSH receptor autoantibodies were found. There was a weak but statistically significant negative correlation of the Visual Functioning score with age (r = −0.24, P < .001), whereas the score did not correlate with age or with smoking behavior.
The multivariate analysis of clinical activity, severity, diplopia, and the possible confounders age, sex, and smoking behavior with respect to visual functioning was performed using a linear regression model. Within stepwise regression, sex and smoking behavior were excluded from analyses, whereas disease severity alone was sufficient to predict visual functioning (regression coefficient ± SE = −4.79 ± 0.57, P < .001).
Significant ceiling effects (15% or more of responses at the highest value of the subscale) were observed for Visual Functioning as 85 patients (27%) responded at the ceiling. No significant floor effects were noted, although 5 patients (2%) scored at the floor of Visual Functioning.
The GO-QOL subscale Appearance closely correlated with proptosis, and with lid fissure width. Patients with a proptosis up to 20 mm (Group 1) had a mean Appearance score of 77.4 ± 1.7. Graves orbitopathy patients with a proptosis from 20.5 to 24 mm (Group 2) and with values over 24 mm (Group 3) had mean Appearance scores of 66.9 ± 2.6 and 50.8 ± 5.3. The differences of the Appearance scores between Groups 1 and 2 (P = .001, 95% CI 4.3−16.6), Groups 1 and 3 (P < .001, 95% CI 16.3−36.9), and Groups 2 and 3 (P = .005, 95% CI 5.0−27.3) were statistically significant.
Patients with lid fissures up to 10 mm had higher Appearance scores (79.1 ± 2.1) than those with larger lid fissures (62.2 ± 3.2, P = .001, 95% CI 9.7−24.0). Appearance also correlated with clinical activity and severity, and in a trend with TSH receptor autoantibodies. Patients with an inactive orbitopathy had a higher Appearance score (78.3 ± 1.8) as compared to those with active disease (64.5 ± 2.2, P < .001, 95% CI 8.2−19.2). Also, Appearance was markedly higher in patients with mild Graves orbitopathy (86.0 ± 17.6) as compared to those with moderately severe (65.5 ± 25.5) and/or sight-threatening stages (58.5 ± 9.0). The differences of Appearance were significantly different between patients with mild and with moderate to severe disease ( P < .001, 95% CI 15.1−26.0), as well as between patients with mild and sight-threatening disease ( P < .001, 95% CI 15.0−39.4), but not between patients with moderate to severe and sight-threatening Graves orbitopathy ( P = .383, 95% CI −8.8−22.7). Although there was a trend for TSH receptor autoantibody–negative patients to have higher Appearance scores (76.7 ± 2.3) than antibody-positive patients (70.7 ± 1.8), this was not statistically significant ( P = .054, 95% CI −0.1−12.1). However, Appearance did not correlate with thyroid function. The Appearance score correlated with gender and with smoking behavior but not with age. Female patients had lower Appearance scores (70.9 ± 1.5) than male patients (80.9 ± 3.1, P = .01, 95% CI 2.5−17.5) and smokers had lower Appearance scores (68.1 ± 1.6) than nonsmokers (75.3 ± 1.9, P = .025, 95% CI 0.9−13.6).
The multivariate analysis of TSH receptor autoantibodies, clinical activity and severity, proptosis, and lid fissure width, and of the possible confounders age, sex, and smoking behavior, with respect to Appearance was performed using a linear regression model. Within stepwise regression smoking, age, and TSH receptor autoantibodies were excluded from analysis. Proptosis alone was sufficient to predict Appearance (regression coefficient b ± standard error SE (b) = −2.55 ± 0.44, P < .001).
Significant ceiling effects (15% or more of responses at the highest value of the subscale) were observed for Appearance as 59 of 310 patients (19%) responded at the ceiling. No significant floor effects were noted as 2 patients (13%) scored at the floor of Appearance only.
More than one-fifth (66/310, 21.3%) of the patients received psychotherapy. Appearance score was lower in those receiving psychotherapy (64.7 ± 3.2) than in those without psychotherapy (74.6 ± 1.6, P = .005, 95% CI 3.0−16.7).
Validity of Differences Between the Relevant Subgroups
Mean differences in GO-QOL scores between relevant subgroups were found to be valid if they were at least 6 points. The differences of the Appearance and Visual Functioning scores between these groups are summarized in Table 3 . For both scores, valid differences were found between patients with active vs inactive Graves orbitopathy, as well as between patients with mild, moderate to severe, and sight-threatening stages of the disease. The difference in the Visual Functioning score was valid between patients without and with diplopia, as well as between patients with inconstant vs constant diplopia. Regarding the Appearance score, the differences were also valid between TSH receptor autoantibody–positive and TSH receptor autoantibody–negative patients and between patients with and without psychotherapy. The differences in the Appearance scores between female patients and male patients (difference: 10 points) and between smokers and nonsmokers (difference: 7.2 points) were also found to be valid.