The effect of uveitis and undiagnosed spondyloarthritis: a systematic review and meta-analysis

Delay diagnosis of spondyloarthritis (SpA) is associated with poor functional ability and quality of life. Uveitis is the most frequent extraarticular manifestation in SpA, and its prevalence increases with longer disease duration. This study examines the effect of uveitis on the disease activity and functional outcome of undiagnosed SpA. We reviewed published and unpublished studies. Data were pooled using the random-effects model; pooled means, and mean differences (MDs) were calculated. In the included 14 studies, disease activity, functional index, and inflammatory markers were measured in 2581 patients with SpA with uveitis and 13,972 without. The pooled mean delay in diagnosis of SpA with uveitis (6.08 years; 95% CI 4.77 to 7.38) was longer than those without (5.41 years; 95% CI 3.94 to 6.89). The Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score was the highest for a delay of 2–5 years (5.60, 95% CI 5.47 to 5.73) and the Bath Ankylosing Spondylitis Functional Index (BASFI) score was the lowest for a delay of < 2 years (2.92, 95% CI 2.48 to 3.37) and gradually increased to delay of > 10 years (4.17, 95% CI 2.93 to 5.41). Patients with SpA with uveitis had higher trend of Ankylosing Spondylitis Disease Activity Score (ASDAS)-CRP and BASDAI. The delay to diagnosis was longer in SpA with uveitis, and disease activity was often higher than those without uveitis. Early diagnosis of SpA with timely initiation of an appropriate management plan may reduce the adverse effects of the disease and improve functional ability.

In clinical practice, the functional ability of patients with SpA is evaluated using the Bath Ankylosing Spondylitis Functional Index (BASFI).Moreover, disease activity is quantified using two evaluation tools, namely the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) and Ankylosing Spondylitis Disease Activity Score (ASDAS) 16 .The BASDAI contains only subjective clinical items, whereas the ASDAS contains both subjective clinical items and objective laboratory measures including the erythrocyte sedimentation rate (ESR) and the C-reactive protein (CRP) level 1 .
Due to inconclusive emerging data regarding the disease activity and functional ability of patients with SpA with and without uveitis 7,9,15,[17][18][19][20][21] , this study aimed to provide further insight by conducting a meta-analysis of existing studies.Specifically, the study examined the disease activity and functional outcome of SpA, while also analyzing differences between patients with SpA with and without uveitis.

Characteristics of the study population
In the present study, we identified 689 articles from the online databases, among which 632 irrelevant articles were excluded.The remaining 57 articles were assessed for eligibility.Finally, 11 articles [6][7][8][9][10]14,15,[17][18][19]21 and 3 conference proceedings 20,22,23 including 16,553 patients with SpA were included in the meta-analysis. The PRISMA flo diagram (Fig. 1) for studies retrieved through the electronic search and the selection process for study inclusion.The quality of the included studies was assessed using the RoBANS tool 24 (Supplementary Fig. S1). In partiular, the studies included 2581 patients with SpA with uveitis and 13,972 patients with SpA without uveitis.Table 1 lists the detailed characteristics of each study.Among the included studies, three were conducted in Asian populations, including two studies in China 9,10 and one study in Taiwan 17 .Another 11 studies examined Caucasian populations from Europe, Latin America, and Turkey [6][7][8]14,15,[18][19][20][21][22][23] (Table 1).
The pooled mean values of delay in the diagnosis of SpA with and without uveitis were 6.08 (95% CI 4.77 to 7.38) years and 5.41 (95% CI 3.94 to 6.89) years, respectively.For continuous variables, mean differences (MDs) and 95% confidence intervals (CIs) were calculated.The delay in diagnosis was significantly longer in patients with uveitis than in those without uveitis (MD 1.04; 95% CI 0.28 to 1.80, p = 0.008; Table 3; Fig. 2).

Subgroup analysis
The heterogeneity among the included studies was obvious and we performed subgroup analysis for the mean difference of BASDAI score observed between SpA patients with and without uveitis.In the included studies, there were three study characteristics that may have affected the results.First, the delay diagnosis varied from 1.6 to 17.9 years.Second, the characteristics of participants varied in diagnostic criteria of SpA, sample size of patients and gender ratios.Third, the study design varied between cross-sectional and cohort studies.
In the subgroup analysis, the BASDAI score was significantly low in patients of SpA with uveitis and delay diagnosis less than 5 years (MD − 0.23; 95% CI − 0.40 to − 0.07, p = 0.005, I 2 = 0%) and it was high in study design as cross-sectional method (MD 0.69; 95% CI 0.23 to 1.16, p = 0.003, I 2 = 0%; Table 4; Fig. 3).This could be due to the disease activity of SpA with uveitis is higher than without.However, if diagnosed and treated within the first five years from symptom onset, the disease activity of SpA with uveitis can be lower than without.The nonsignificant result and heterogeneity was found in study design as cohort method (MD 0.03; 95% CI − 0.14 to 0.20, I 2 = 76%).A possible explanation could be that the cohort studies included multiple confounding factors.The score in sample size ≤ 500 patients was significant in high but with moderate heterogeneity (MD 0.62; 95% CI 0.10 to 1.13, p = 0.003, I 2 = 73%; Table 4; Fig. 3).In the six studies of subgroup of sample size ≤ 500 patients, there were three cross-sectional 7,17,18 and three cohort studies 8,14,20 , the heterogeneity may be attributed to differences in study design.There was a nonsignificant difference in delay diagnosis over 5 years (MD 0.15; 95% CI − 0.22 to 0.51).BASDAI scores also had nonsignificant differences in the groups of diagnostic criteria of SpA, the sample size of patients > 500 patients and gender ratios (Table 4; Supplementary Fig. S4).

Publication bias
The asymmetrical funnel plot indicated the publication bias in this meta-analysis (Fig. 4).We performed Egger's test to confirm the existence of bias but without significance (t value = 4.25, df = 13, p = 0.14).

Discussion
In a previous meta-analysis, the mean delay in the diagnosis of SpA was 6.7 (95% CI 6.2 to 7.2) years 11 , and in the present study, that of SpA with and without uveitis was 6.08 and 5.41 years, respectively.The delay in the diagnosis of SpA significantly differed between patients with and without uveitis (Table 3).Uveitis can be the first presenting symptom of SpA, and studies have reported a prevalence of 40-50% of previously undiagnosed SpA in patients presenting with uveitis 3,13 .Wendling et al. 6 reported that uveitis occurred before the first symptom of inflammatory back pain in 37% and simultaneously in 18% of cases.The delay in the diagnosis of SpA is associated with poor functional impairment, rapid radiographic progression, poor quality of life, and decreased treatment response 11 .In the present study, the pooled mean BASDAI score was the highest for a delay of 2-5 years (5.60, 95% CI 5.47 to 5.73).By contrast, the pooled mean BASFI score was the lowest for a delay of < 2 years (2.92, 95% CI 2.48 to 3.37), and its severity gradually increased to a delay of > 10 years (4.17, 95% CI 2.93 to 5.41; Table 3).Early recognition can lead to the early initiation of the most appropriate and effective therapy.Early diagnosis of SpA may reduce the adverse effects of the disease and improve functional ability.In this study, we found that patients with SpA with uveitis had a prolonged delay in diagnosis, which was highly associated with higher BASFI scores and poorer prognostic outcomes (Table 3).Awareness regarding uveitis is crucial considering its role in the diagnostic process of SpA for treatment choices and health-related quality of life 4 .Backache or arthralgia were not the primary symptom of patients with SpA with uveitis during their presentation to an ophthalmologist.The degree of pain and disability, as reflected by BASDAI and BASFI scores, was often noted to be mild 13 .In the present study, the BASDAI score was significant lower in patients of SpA with uveitis with delay diagnosis less than 5 years (MD − 0.23; 95% CI − 0.40 to − 0.07).Previous studies have examined disease activity by using different assessment systems.Approximately 40-65% of patients with low BASDAI scores (< 4) had high disease activity, as measured using the ASDAS (≥ 2.1) [25][26][27] .In this study, we found a higher disease   activity, as measured using the ASDAS (MD 0.18; 95% CI 0.11 to 0.25), in patients with SpA with uveitis than in those without.Measuring the BASDAI rather than the ASDAS is routine in clinical practice.However, some patients of uveitis especially early undiagnosed SpA suspected to have high disease activity despite low BASDAI scores should be examined using the ASDAS.
In the present study, we found no significant difference in the ESR between patients with and without uveitis (MD − 0.80; 95% CI − 4.62 to 3.03 mm/h).However, the CRP level tended to be higher in patients with uveitis (MD 1.48; 95% CI − 0.17 to 3.14 mg/L).Furthermore, we performed a subgroup analysis of ASDAS-CRP and ASDAS-ESR.Although the ASDAS might be affected by infectious or inflammatory conditions, it is regarded as more objective than the BASDAI.The ASDAS-CRP was significantly higher in patients with SpA with uveitis (MD 0.17; 95% CI 0.10 to 0.24).However, ASDAS-ESR did not differ between patients with and without uveitis (MD 0.03; 95% CI − 0.65 to 0.72; Table 3).
Uveitis may be a key feature leading to SpA diagnosis 6,12,17 .The nonspecific and often subtle symptoms of inflammatory back pain make early diagnosis and subsequent treatment challenging.In the subgroup analysis, the BASDAI score was significant higher in SpA patients with uveitis in study design as cross-sectional method (MD, 0.69; 95% CI 0.23 to 1.16).Patients with SpA with uveitis exhibited a marked reduction in the activities of daily living compared with those without uveitis (47.1% vs. 23.5%) 17,28.However, in this study, no significant difference in BASFI scores was noted (MD − 0.12; 95% CI − 0.46 to 0.21).Because the delay in diagnosis was significantly longer in patients with uveitis, we further analyzed the pooled mean BASFI score.The lowest in delay diagnosis of < 2 years was 2.92 (95% CI 2.48 to 3.37), and the scores gradually increased to delay in diagnosis of > 10 years (4.17, 95% CI 2.93 to 5.41).Prolonged delay in diagnosis is common among patients with SpA and the occurrence of uveitis may be the reason for their first interaction with medical care, the occurrence of uveitis presents a unique opportunity for identifying such patients with undiagnosed SpA 29 .Therefore, awareness regarding uveitis among ophthalmologists and primary care physicians is vital considering its role in the diagnostic process and the selection of appropriate treatment for improving health-related quality of life 4 .
This study has some limitations that should be addressed.First, in our included studies, the disease activity indices of BASDAI, ASDAS and BASFI all require patients to answer the questionnaire and it is quite challenging for physician and patients to determine and make the assessment.Most of the SpA-related uveitis was diagnosed by ophthalmologists, however some of the history of uveitis was patients' self-reported and this may reduce the accuracy.This may cause variations in findings among different geographical regions.Moreover, in terms of baseline characteristics, the included studies did not provide the age of uveitis diagnosis.We found the BASDAI was significant lower in patients of SpA with uveitis in delay diagnosis less than 5 years (MD − 0.23; 95% CI − 0.40 to − 0.07).Uveitis may occur before SpA symptoms or diagnosis 6,12,17 .The reason for a longer delay in diagnosis of SpA in patients with uveitis may be due to the mild symptoms of SpA.Additional cohort studies focusing on the age at diagnosis in patients of SpA and uveitis should be conducted.
In the present study, we found that patients with SpA with uveitis had a longer delay in diagnosis than did those without.Uveitis is the most frequent EAM in SpA 3 and may affect disease activity, quality of life, and treatment decisions.In our meta-analysis, we found a significant difference in ASDAS-CRP and BASDAI between

Figure 1 .
Figure 1.PRISMA flow diagram for studies retrieved through the electronic search and the selection processes.

Figure 3 .
Figure 3. Forest plot of the subgroup analysis of BASDAI in patients of spondyloarthritis with and without uveitis.(a) Delay diagnosis less than 5 years, (b) Study Design: Cross-sectional, (c) Sample Size <500 patients.

Table 1 .
Characteristics of the included studies.Delay delay time of diagnosis, NA not available, NY 1984 modified New York criteria for ankylosing spondylitis, ESSG European Spondylarthropathy Study Group criteria for SPA, ASAS Assessment of Spondylarthritis International Society classification Criteria for SPA, CASPAR Classification Criteria for Psoriatic Arthritis, Kahn Classification Criteria for SAPHO syndrome, AS Ankylosing Spondylitis, SPA Spondylarthritis, PsA Psoriatic Arthritis, SAPHO syndrome synovitis acne pustulosis hyperostosis osteitis syndrome.

Table 3 .
The pooled mean and mean difference of clinical data in spondyloarthritis patients with and without uveitis of the included studies.

Table 4 .
Subgroup analysis of mean difference of BASDAI score in spondyloarthritis patients with and without uveitis of the included studies.*Number of study.