Aims To investigate the ability of the Behçet's disease ocular attack score 24 (BOS24) scoring system to predict visual acuity (VA) in patients with ocular Behçet's disease.
Design This is a retrospective study.
Methods We included 91 eyes of 50 patients with ocular Behçet's disease (33 males, 17 females) who were referred to our hospital between 1986 and 2008 with >5 years follow-up. Total BOS24 scores over a 5-year period, BOS24-5Y, were calculated as the sum of BOS24 scores for each attack over the 5-year study period for each eye. Change in VA was defined as change in best-corrected visual acuity (BCVA) from the first remission to the last remission at the end of the target period. Factors related to change in VA (age, gender, BCVA at the first remission, total number of immunosuppressive medications and total number of ocular attacks during the 5-year period and BOS24-5Y) were evaluated using a linear mixed model.
Results BCVA (logarithm of the minimal angle resolution) deteriorated from 0.16±0.30 (mean±SD) to 0.21±0.37 over the 5-year study period, but there was no statistical difference. The total number of ocular attacks during the 5-year period and BOS24-5Y scores were 10.0±7.9 and 36.8±40.8, respectively. Linear mixed-model analysis revealed that BOS24-5Y was the most important index for VA deterioration, followed by BCVA at the first remission.
Conclusions BOS24-5Y was found to be a significant positive prognostic index for VA deterioration in patients with ocular Behçet's disease.
Statistics from Altmetric.com
Behçet's disease (BD) is a chronic inflammatory disorder of unknown origin characterised by oral and genital ulcerations, intraocular inflammation, skin lesions and involvement of vascular, central nervous and gastrointestinal systems. The prevalence of ocular involvement in patients with BD has been reported to be 47%–75%.1 In a recent prospective multicentre survey of uveitis patients in Japan, BD was identified in 3.9% of all new patients with uveitis.2 Ocular involvement in BD is characterised by non-granulomatous intraocular inflammation and retinal vasculitis. Recurrent intraocular inflammation in patients with BD is typically intermittent and episodic and is referred to as inflammatory ocular attacks. Previous studies have indicated that recurrent inflammatory ocular attacks confer poor visual prognosis.3 ,4 In addition, ocular involvement, particularly posterior uveitis with retinal vasculitis, confers a poor visual prognosis.3 ,5
Recently, we reported a novel scoring system for uveitis related to BD, termed Behçet's disease ocular attack score 24 (BOS24).6 The BOS24 score, ranging from 0 to 24 points, consists of six parameters concerning ocular inflammatory symptoms. BOS24 was strongly related to the severity of the impression scores noted by a physician with a low level of interobserver variability.6 BOS24 may be useful in evaluating disease activity of ocular BD because total BOS24 scores over 6-month periods (BOS24-6M) as well as the severity of each ocular attack were shown to decrease following the initiation of infliximab therapy.6 However, BOS24 has not yet been evaluated as an index of visual outcome.
In the present study, we assessed the correlation between time-course changes in best-corrected visual acuity (BCVA) and total BOS24 scores over a 5-year period.
Patients and methods
Clinical records of patients with BD seen in the Department of Ophthalmology, The University of Tokyo Hospital from January 1986 to December 2008 were retrospectively reviewed from a database of patients with uveitis who were newly referred to the hospital.7
We included patients with ocular BD fulfilling the following criteria: (1) a diagnosis of BD confirmed at our hospital based on the criteria established by the Behçet's Disease Research Committee of Japan,8 (2) BCVA greater than 1.0 logarithm of the minimal angle resolution (logMAR) visual acuity (VA) at the first remission and (3) continuous follow-up for 5 years at our hospital. Unaffected eyes in unilateral cases were excluded from statistical analyses. In total, 91 eyes of 50 patients were enrolled in this study.
In our uveitis clinic, the patients with Behçet's uveitis were adequately educated to visit our hospital when the symptoms of ocular attacks appeared. We evaluated each attack using the BOS24 scoring system. It consists of 24 points summarised from six parameters of ocular inflammation symptoms, including anterior chamber cells (maximum 4 points), vitreous opacity (maximum 4 points), peripheral fundus lesions (maximum 8 points), posterior pole lesions (maximum 4 points), subfoveal lesions (maximum 2 points) and optic disc lesions (maximum 2 points). BOS24 was determined in eyes suffering from ocular inflammatory attacks on the day when the patient was examined by a doctor and judged to be the most severe period of inflammation during an exacerbation stage.6
Simultaneous bilateral attacks were counted as one attack for each eye and BOS24 was separately scored for each eye. A summation of BOS24 scores during the 5-year period, BOS24-5Y, was calculated as the sum of all attacks in the 5-year study period and used as an index of inflammatory activity over the given period.
All patients underwent complete ocular examination of both eyes at each visit including BCVA, slit-lamp biomicroscopy and binocular indirect ophthalmoscopy. BCVA was determined using a standard Landolt VA chart and then converted to the logMAR VA for statistical analysis. In this study, VA for counting fingers (CF) at a distance of 50 cm, CF at a distance of 5 cm, hand movements, light perception (LP) and no LP were designated as 0.01, 0.001, 0.0005, 0.0002 and 0.0001, respectively, as previously reported.9 The VA of patients with ocular BD considerably varies at each ocular attack and does not always correspond to the patients’ potential visual function. In order to exclude any possible temporary and acute influence of ocular attacks on VA, we adopted VA during remission in this study, as previously reported.4 ,5 ,9 A change in VA was defined as a change in BCVA from the first remission to the last remission at the end of the target period. As both eyes in some patients were included in this study, linear mixed-model analysis10 was employed to avoid pseudoreplication. In this model, a random effect was added to correct intraindividual data correlation. Clinical variables including age, gender, number of ocular attacks during the 5-year period, BOS24-5Y, BCVA at the first remission and total number of immunosuppressive medications were used as fixed-effect factors. Data are presented as mean±SD, unless stated otherwise. Probability (p) values of <0.05 were considered statistically significant. All statistical analyses were performed with SPSS for Windows V.16.0 (SPSS, Chicago, Illinois, USA).
All patients provided informed consent. This retrospective study was approved by the ethics committee of The University of Tokyo Hospital, and it conforms to provisions of the Declaration of Helsinki.
Clinical characteristics of all 50 patients are presented in table 1. Patients consisted of 33 males and 17 females, 24 had complete-type BD and 26 had incomplete-type BD. The mean age at the initial visit was 37.4±10.6 (range, 14–60) years. The time between onset of ocular disease and initial visit was 45.7±64.4 (range, 0–273) months. Bilateral ocular BD was observed in 49 patients (98%), whereas unilateral BD was observed in one patient (2%).
During the follow-up period, 47 patients received at least one of the following immunosuppressive medications: colchicine (n=39, 78%), ciclosporin (n=22, 44%), prednisolone (n=25, 50%), infliximab (n=6, 12%), cyclophosphamide (n=3, 6%), tacrolimus (n=2, 4%), azathioprine (n=2, 4%) and methotrexate (n=1, 2%). In total, 17 patients received monotherapy consisting of colchicine (n=11), prednisolone (n=4) or ciclosporin therapy (n=2), while the remaining 30 patients received combined therapy. At the initial visit, 86 eyes were phakia, 2 were aphakia and 3 were pseudophakia. Cataract surgery was performed on 22 out of 86 phakic eyes during the observation period. Glaucoma surgery was performed on nine eyes.
The number of ocular inflammatory attacks during the 5-year study period and BOS24-5Y were 10.0±7.9 and 36.8±40.8, respectively. The mean BCVA deteriorated from 0.16±0.30 to 0.21±0.37 after 5 years, but there was no statistical difference (paired t test).
To investigate the relationship between the number of ocular attacks during the 5-year study period and change in VA, simple linear regression analysis was performed (figure 1A). A slight but statistically significant relationship between the number of ocular attacks during the 5-year study period and change in VA was observed (R2=0.226, p<0.001). A statistically significant relationship between BOS24-5Y and change in VA was also found (figure 1B, R2=0.334, p<0.001).
To identify significant factors related to change in VA, we further performed linear mixed-model analysis (table 2). Linear mixed-model analysis revealed that BOS24-5Y had the largest effect on VA deterioration (p=0.001), followed by BCVA at the first remission (p=0.002) and age (p=0.009). However, the total number of ocular attacks, total number of immunosuppressive medications and patient gender were not found to affect VA.
In addition, to examine which aspect of ocular inflammation had the greatest effect on VA deterioration in BOS24-5Y, we sorted six parameters of BOS24-5Y into three indices: anterior chamber cells-5Y, vitreous opacity-5Y and posterior segment lesions-5Y. Posterior segment lesion scores were calculated as the sum of the scores of peripheral fundus lesions, posterior pole lesions, foveal lesions and optic disc lesions. Anterior chamber cells-5Y, vitreous opacity-5Y, posterior segment lesions-5Y, age and BCVA at the first remission were analysed by linear mixed-model analysis (table 3). Posterior segment lesions-5Y had the greatest effect on VA deterioration (p=0.002) in three indices of BOS24-5Y.
Relapsing and remitting uveitis is the typical form of ocular inflammation in BD, even if no treatment is administered. Thus, evaluation of ocular inflammatory activity in BD is challenging because severity can greatly vary between evaluation points.
Previously, ocular inflammatory activity was evaluated on the basis of the frequency of ocular inflammatory attacks,9 ,11–14 localisation of ocular inflammatory attacks, including anterior uveitis, posterior uveitis and panuveitis14 and the presence of signs of severe ocular inflammation (hypopyon or inflammation involving the retina, macula or optic disc).14 In addition, several authors have proposed methods for evaluating ocular inflammation using a posterior uveitis score,15 laser flare-cell photometry16 or a fluorescein angiography scoring system.17
The BOS24 scoring system has several advantages over the evaluation methods described above. First, compared with laser flare-cell photometry or fluorescein angiography, the BOS24 scoring system is simple and can be applied in all clinical settings. Second, this scoring system enables the evaluation of both anterior segment inflammation and posterior segment inflammation. Third, this score can be used to evaluate ocular inflammation over a target period through the summation of BOS24 scores, in addition to assessing individual ocular attack. The sum of BOS24 scores reflects both individual attack severity and the total number of ocular attacks. Finally, the BOS24 scoring system can be retrospectively employed using medical records. Despite these advantages, BOS24 has not yet been examined as an index of visual outcome.
In the present study, linear mixed-model analysis revealed that BOS24-5Y was the most important index of VA deterioration in ocular BD. Interestingly, the total number of ocular attacks was not a statistically significant index of VA deterioration in linear mixed-model analysis. Conversely, simple linear regression analysis revealed that both BOS24-5Y and the total number of ocular attacks were associated with VA deterioration. This may be explained by the fact that the BOS24-5Y score provides information regarding the total number of ocular attacks and individual attack severity. Therefore, BOS24-5Y may provide more accurate prognostic information than the number of ocular attacks alone. The significant results of this study demonstrate the utility of the BOS24 scoring system as a positive prognostic index for VA deterioration in patients with ocular BD.
Other important findings of this study were as follows: the score of posterior segment lesions was the most important index of VA deterioration, and anterior chamber cells and vitreous opacity were not found to influence VA. Anterior chamber cells and vitreous opacity are characterised by exacerbations and periods of remission; therefore, it is reasonable that those are not critical determinants of VA prognosis. Conversely, it has been reported that recurrent attacks of posterior segment inflammation lead to severe retinal damage and visual loss.3 ,5 ,18–20 The present study corroborates these previous reports.
In the present study, we assessed the correlation between change in VA and gender or age. It has been reported that uveitis is more severe and the risk of useful vision loss is higher in males than in females.5 ,18 ,20 However, no difference in VA was found between the genders in this study. This may be explained by the fact that gender only affects VA outcomes through disease severity, adequately incorporated into the BOS24 score. Patient bias may also have contributed because this study was performed in a university hospital where patients with greater disease severity may have been included regardless of gender.
It has been reported that early BD onset is associated with BD severity,21 and several papers have assessed the correlation between age and visual outcome. Demiroğlu et al reported that an age of ≤32 years is a risk factor for visual impairment;22 however, other authors reported that age is not linked to VA.3 ,5 In the present study, age was a positive index of VA deterioration in linear mixed-model analysis. Further analysis revealed no significant correlation between age and visual change, as determined by the Spearman's rank correlation coefficient (p=0.12), similar to previous reports.3 ,5 Although early onset of BD is associated with disease severity,21 it is adequately incorporated into the BOS24 score. In other words, these results indicate that older patients are more likely to develop reduced VA than younger patients in cases where the ocular inflammation is comparable. Presumably, older patients are more likely to develop cataracts and loss of retinal function as a result of recurrent ocular attacks leading to visual impairment.
Linear mixed-model analysis also revealed that BCVA at the first remission had an effect on VA deterioration (p=0.002). Eyes with good baseline VA became exacerbated to a greater extent than eyes with poor baseline VA. Eyes with poor baseline VA could not deteriorate any further. On the other hand, the total number of immunosuppressive medications used in the observation period did not affect VA change. Presumably, immunosuppressive medications influence visual outcome through a reduction in the number and/or severity of ocular attacks, which is well reflected in the BOS24 scoring system.
We thought that cataract surgery during the observation period may affect visual change. Cataract surgery was performed on 22 out of 86 phakic eyes during the observation period. Further analysis limited to 86 phakic eyes revealed no significant correlation between cataract surgery and visual change, as determined by linear mixed-model analysis (p=0.50, table 4). There are two possible explanations for this result. First, there were some cases whose VA did not improve because of pre-existent macular atrophy and/or chorioretinal atrophy. Second, some cases suffered from severe ocular inflammation attacks after cataract surgery.
However, evaluation using the BOS24 scoring system has several limitations. First, scoring was not possible in patients who did not attend clinic during an ocular attack or who had invisible ocular fundus lesions as a result of vitreous opacity. Second, although the duration of ocular inflammation attack could be another factor for visual change, we cannot evaluate the exact duration of attacks retrospectively. Third, the selection process was not independent of bias because of retrospective design. Further prospective studies are required to confirm the results of this study.
In conclusion, BOS24-5Y is a more reliable prognostic index than total number of ocular attacks for VA deterioration in patients with ocular BD. The BOS24 scoring system allows precise evaluation of the ocular BD severity.
Contributors All the authors meet the criteria for authorship.
Competing interests None declared.
Ethics approval The ethics committee of The University of Tokyo Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.