Background/aims Fuchs’ uveitis syndrome (FUS) is one of the frequently misdiagnosed uveitis entities, which is partly due to the absence of internationally recognised diagnostic criteria. This study was performed to develop and evaluate a set of revised diagnostic criteria for FUS.
Methods The clinical data of Chinese patients with FUS and patients with non-FUS were collected and analysed from a tertiary referral centre between April 2008 and December 2020. A total of 593 patients with FUS and 625 patients with non-FUS from northern China were enrolled for the development of diagnostic criteria for FUS. Three hundred and seventy-seven patients with FUS and 503 patients with non-FUS from southern China were used to validate the criteria. Clinical symptoms and ocular signs were collected from all patients with FUS and patients with non-FUS. Multivariate two-step cluster analysis, logistic regression and decision tree algorithms in combination with the clinical judgement of uveitis experts were used to revise diagnostic criteria for FUS.
Results Three essential findings including diffuse iris depigmentation, absence of posterior synechiae, mild inflammation in the anterior chamber at presentation and five associated findings including mostly unilateral involvement, cataract, vitreous opacities, absence of acute symptoms and characteristic iris nodules were used in the development of FUS diagnostic criteria. All essential findings were required for the diagnosis of FUS, and the diagnosis was further strengthened by the presence of associated findings.
Conclusion Revised diagnostic criteria for FUS were developed and validated by analysing data from Chinese patients and showed a high sensitivity (96.55%) and specificity (97.42%).
- diagnostic tests/investigation
- anterior chamber
Data availability statement
All data relevant to the study are included in the article or uploaded as supplemental information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Fuchs’ uveitis syndrome (FUS) is a chronic, mostly unilateral anterior uveitis, accounting for approximately 7% of Chinese patients with uveitis.1 It is one of the frequently misdiagnosed uveitis entities in the world.1–4 Patients with FUS generally have an insidious onset and low grade of disease activity, and often present with an asymptomatic mild inflammation.5 6 One of the important signs of FUS is heterochromia in white races.7 8 However, this sign is infrequently observed in patients with FUS from a yellow or black population, who have a dense melanin concentration in their iris.9–11 Although patients with FUS generally have a good prognosis, incorrect diagnosis may lead to unnecessary therapy and accelerate the development of cataract and secondary glaucoma.12–14 Therefore, a correct diagnosis is absolutely necessary for patients with FUS.
Throughout the history of reporting of FUS, the diagnosis of this syndrome has been a challenge due to its diverse and easily neglected manifestations. Kimura et al15 proposed the classical triad of signs (heterochromia, cyclitis and cataract), based on the clinical data from 750 American patients with uveitis (including 23 patients with FUS). La Hey et al2 made some modifications to the description made by Kimura and described a set of integrated diagnostic criteria which included the predominant clinical findings observed in 51 Dutch patients with FUS. The Standardisation of Uveitis Nomenclature (SUN) working group16 developed classification criteria for FUS according to the data from 1083 cases of anterior uveitis (including 146 FUS) by machine learning. The criteria stated above, especially the La Hey criteria have been widely used in clinical practice for a long time. However, there are still some potential limitations of these criteria that need further discussion. On the one hand, the three criteria were created from data mostly gathered from white patients. On the other hand, the sample size (23, 51 and 146 patients with FUS) used to develop these criteria was relatively small. Simply using the triad of signs proposed by Kimura is not enough to make a correct diagnosis. With regards to the La Heycriteria, over-emphasis on presenting with all essential findings and at least two associated findings may lead to an under-diagnosis of FUS. The classification criteria by the SUN working group emphasise specificity (exclusion of patients without the disease), which may lead to partial missed diagnoses. In this study, we therefore developed a set of revised diagnostic criteria (RDC) for FUS, based on the data obtained from a large number of Chinese patients and show that the chosen criteria have a high sensitivity and specificity.
A retrospective study was performed based on the clinical manifestations of 970 patients with FUS and 1128 patients with non-FUS referred to the uveitis centre of the First Affiliated Hospital of Chongqing Medical University during April 2008 to December 2020. The definition of FUS was made according to the La Hey criteria in combination with the clinical judgement of our uveitis specialists. Patients with other uveitis entities (online supplemental table 117–20), who presented with signs and symptoms of anterior uveitis, served as controls. Uveitis caused by infectious agents such as tuberculosis, syphilis and AIDS or associated with systemic disease including tubulointerstitial nephritis-associated uveitis syndrome and inflammatory bowel disease were all excluded. The diagnosis of the enrolled patients was made by more than two specialists from referring hospitals and then verified by uveitis specialists from our centre. In view of the potential impact of geographical factors on the diseases profile, we developed the RDC based on patients from northern China and then validated the novel set of criteria on a group of patients referred to us from southern China. The two parts of the country were demarcated by the Yangtze River. All patients were enrolled anonymously.
Demographic characteristics and medical history (symptoms and signs) of all patients were collected anonymously at the patients’ initial visit, and also included previous clinical records from referring hospitals (table 1).21 The same uveitis specialist (PY) inquired about the medical history and examined all patients. In our study, elevated intraocular pressure (IOP) was defined as an IOP ≥21 mm Hg. Mild inflammation in the anterior chamber at presentation was defined as the presence of stellate or medium-sized keratic precipitates (KPs) and/or minimal cells and flare in the aqueous (1+ or 2+) at initial visit at our centre or from the clinical records of the referring hospitals.22 Previous studies reported that the KPs in patients with FUS are characterised by a stellate appearance or a medium size, which are distributed diffusely or centrally.1 4 23 Diffuse iris depigmentation was defined as a uniform shedding of pigment in the iris. In patients with FUS, slight iris depigmentation usually occurs at the pupillary collar and is accompanied by effacement of the iris crypts. With the gradually increasing loss of pigment, heterochromia (striking iris depigmentation) may eventually occur. Different degrees of diffuse iris depigmentation are shown in figure 1. These changes in the iris are easily detected with a slit-lamp biomicroscope with a narrow slit beam. In normal eyes, a translucent zone is rarely observed and the slit-beam band has a smooth appearance. Whereas in case of iris depigmentation, the transparent zone gradually becomes larger and the slit-beam band becomes curved (figure 2). Iris nodules in patients with FUS generally presented as small white nodules at the pupillary margin (Koeppe nodules) or white dots scattered on the anterior surface of the iris (Busacca nodules).2 24 Koeppe nodules and Busacca nodules observed in FUS are quite different from those observed in granulomatous uveitis (online supplemental figure 1). They are usually smaller in size and definitely show a fluffy appearance.1 25 Vitreous opacities could be assessed with a slit-lamp biomicroscope or by B-scan ultrasonography.
Developing the RDC for FUS
A total of 593 patients with FUS and 625 patients with non-FUS from northern China were used as a ‘development set’ for establishing the RDC. Complete clinical data were obtained for each patient and evaluated by using a data mining method (multivariate two-step cluster analysis,26 binary logistic regression27 and decision tree28) to establish the RDC. Description of the development of the RDC is given in detail in the online supplemental file 1.
Evaluating the RDC for FUS
An evaluation set was used to validate the developed diagnostic efficacy of the RDC for FUS, and included 880 patients with uveitis (377 FUS and 503 non-FUS) from southern China. The sensitivities, specificities and area under the receiver operating characteristic (ROC) curves among the RDC, the La Hey criteria and the SUN classification criteria were compared using this evaluation set.
Quantitative variables were compared using the Student’s t-test. Categorical variables were recorded as frequencies or percentages and compared with the χ2 test. The sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), negative likelihood ratio (NLR) and regression coefficient (B points) of the variables were recruited after binary logistic regression analysis. Sensitivities, specificities and ROC curves with calculations of area under the receiver operating characteristic curves (AUCs) were estimated among the RDC, the La Hey criteria and the SUN classification criteria. AUCs were computed for each finding with 95% CIs by the exact Clopper-Pearson method. SPSS V.24.0 (IBM) was used to analyse all the aforementioned methods. Python scikit-learn package V.0.18.2 was used to conduct decision tree classification. A p value <0.05 was considered significant.
Development of RDC for FUS
In the development set, the data of 593 patients with FUS and 625 control patients from northern China were analysed to develop the RDC for FUS. The demographic data and clinical manifestations in the patients with FUS and control patients are shown in table 1. Almost all patients with FUS showed varying degrees of diffuse iris depigmentation without posterior synechiae. There were significant differences in all variables between the FUS and control groups (p<0.01, table 1) except the mean disease duration, gender distribution, the number of Han Chinese and elevated intraocular pressure. A total of 16 variables with a p<0.01 after using Student’s t-test or the χ2 test were used as candidate factors for the diagnostic model.
The multivariate two-step cluster analysis was used to analyse these 16 candidate variables and identified only two distinct clusters that discriminated patients with FUS (n=610) from patients with non-FUS (n=608). This analysis showed that diffuse iris depigmentation was a significant determinant of phenotype clustering, ranking the highest among the clustering variables. A total of seven variables with ‘Importance’ >0.05 are showed in online supplemental figure 2). The 16 candidate variables were also analysed by a stepwise binary logistic regression with forward selection. Six variables showed statistical significance (p<0.05). Detailed information (OR, sensitivity, specificity, PPV, NPV, PLR, NLR and regression coefficient) of these six variables is shown in table 2. After analysing the results of cluster analysis and logistic regression, a total of eight variables (diffuse iris depigmentation, mild inflammation in the anterior chamber at presentation, unilateral involvement, cataract, absence of posterior synechiae, redness, vitreous opacities and iris nodules) were chosen to develop the RDC for FUS.
The decision tree model was subsequently developed to explain the relationship between the eight variables and disease diagnosis, and was based on recursive partitioning analysis. As shown in online supplemental figure 3, ‘diffuse iris depigmentation’ was positioned at the top of the tree model, followed by ‘mild inflammation in the anterior chamber at presentation’ and ‘absence of posterior synechiae’. A prediction result was attached to each terminal node. When the input of one of these three variables mentioned above was ‘0’, the prediction result was taken as the output of a terminal node (‘Non-FUS’). This decision tree model showed that the presence of these three signs was required for the diagnosis of FUS and included diffuse iris depigmentation, mild inflammation in the anterior chamber at presentation and absence of posterior synechiae. Three variables (unilateral involvement, cataract and vitreous opacities) had no effect on the results of the final classification (‘FUS’) and were included as associated findings. The other two variables (redness and iris nodules) were not shown in the decision model but may be useful for the prediction of FUS and were also included as associated findings. According to the clinical judgement of uveitis experts, ‘redness’ could be summarised as absence of acute symptoms, and characteristic iris nodules might be more suitable than iris nodules for FUS. All essential findings were required for the diagnosis of FUS, and the diagnosis was further strengthened by the presence of associated findings. The RDC for FUS can be seen in table 3. The differences among the RDC, the La Hey criteria and the SUN classification criteria are shown in table 4.
Validation of the RDC for FUS
In the validation set, we used patients with FUS (n=377) and control patients with uveitis (n=503) from southern China (online supplemental table 1). Sensitivities, specificities and ROC curves with calculations of AUCs were compared among our RDC, the La Hey criteria and the SUN classification criteria (online supplemental table 2). The sensitivity of the RDC was 96.55%, considerably higher than that seen in the SUN classification criteria and the La Hey criteria (84.08% and 77.72%), which proved that our RDC may be the optimal choice to facilitate the inclusion of FUS. The specificity of our RDC (97.42%) was lower than the SUN classification criteria (99.40%), and the same as the La Hey criteria. The ROC curves of the three criteria are shown in online supplemental figure 4. The AUC of the RDC was 0.970 (95% CI, 0.957 to 0.983), which is the highest among the three criteria, suggesting that our RDC is superior for the diagnosis of FUS in a Chinese population. There were statistical differences (p<0.01) between our RDC and the other two criteria.
In this study we developed a novel set of simple diagnostic criteria to diagnose FUS in Chinese patients with uveitis. We believe that our criteria may also be used in other ethnic populations with heavily pigmented irises. One of the main criteria is the observation of a diffuse iris depigmentation, which can be seen by careful examination of the iris with a narrow beam on a slit-lamp biomicroscope. We used novel techniques to develop our revised criteria, which included a combination of cluster analysis, logistic regression and decision tree methods.2 26 29 The RDC established with these three data mining tools included three essential findings and five associated findings and showed a higher sensitivity and larger AUC in comparison to the SUN classification criteria and the La Hey criteria.2
FUS is a phenotypic expression of several probable etiologic causes of inflammation which differentiates it from diseases that have defined etiologies like cytomegalovirus anterior uveitis or other herpesvirus-induced anterior uveitis. Until now, the aetiology and underlying mechanism of FUS are not yet clear. Previous studies have shown that Toxoplasma gondii and Rubella virus may be possible agents that are involved in the pathogenesis of FUS, but as yet are not used as a diagnostic criterion.30 31 The diagnosis of FUS is a clinical one. It is not necessary to perform real time PCR analysis in the diagnosis. However, it is definitely necessary to do PCR and aqueous humour analysis in study of the pathogenesis of this disease if samples can be obtained during cataract surgery.
Clinical diagnostic criteria differ from classification criteria, although both try to reduce misdiagnosis, classification criteria typically emphasise specificity when trade-off is required, whereas diagnostic criteria emphasise sensitivity to ensure inclusion of a homogeneous group of patients with characteristic clinical features.32 The SUN classification criteria for FUS were proposed in 2020 and performed satisfactorily using an internal data set from the SUN working group.16 In our cohort, the SUN classification criteria also showed a high specificity, but its sensitivity was significantly lower than our RDC, which may due to an overemphasis on the presentation of unilateral involvement in patients with FUS. Previous studies showed that the number of patients with FUS with binocular involvement accounts for 10%–21% from a yellow or black population.2 10 Mostly unilateral involvement was included as associated findings in our RDC. In our patients with bilateral involvement, stellate or medium-sized KPs distributed diffusely or centrally, mild anterior chamber inflammation and absence of posterior synechiae in combination with diffuse iris depigmentation were of great important features to make the diagnosis.
There are some similar characteristics between our RDC and the La Hey criteria such as the presence of diffuse iris depigmentation (diffuse iris stromal atrophy) and absence of posterior synechiae. It should be noted that our set of criteria was developed by analysing data from 970 patients with FUS and 1128 controls with uveitis. Some of the La Hey criteria, such as heterochromia was not frequently observed in our set of patients but could be incorporated into a composite set of criteria for non-Chinese individuals. More studies are needed to address this issue. As mentioned above, we would like to stress the analysis of iris depigmentation by observation of the translucent zone generated in the iris by using a slit-lamp biomicroscope with a narrow slit-beam. Other uveitis entities like the Posner-Schlossman syndrome may show diffuse iris depigmentation, but this depigmentation is heterogeneous and uneven. Iris depigmentation can also be detected in patients with herpetic uveitis, but here the depigmentation is approximately sectoral and locally distributed at the lesion site. Therefore, the finding of diffuse iris depigmentation is of great importance and quite specific for the diagnosis of FUS.
Our revised criteria propose the concept of ‘mild inflammation in the anterior chamber at presentation’ that defined as the presence of mild inflammation at initial visit or from the clinical records of the referring hospitals. In practice, many patients had already been treated with anti-inflammatory drugs by local hospitals and mild inflammation may thus be absent at visit. Therefore, it is extremely important to trace back to the previous clinical records. Vitreous opacities were considered as an important associated finding because they were observed in a vast majority of the patients with FUS.1 3 Characteristic iris nodules were included in our RDC and this sign emphasised the small size and a fluffy appearance of the nodules. Elevated intraocular pressure was not mentioned in our criteria because this variable lost statistical significance due to the presence of patients with Posner-Schlossman syndrome in our control group. Chorioretinal lesions were also not included in our RDC because this sign was not presented in 104 patients with FUS as previously reported1 and those enrolled in this study. A further comparison of our RDC with the La Hey criteria and the SUN classification criteria shows that our criteria are simpler and superior, since we have confined the diagnosis to three essential criteria and do mention five associated findings as a support of the diagnosis.
We would like to point some limitations of our study. First of all, our findings were obtained from a group of Chinese patients with uveitis and although we believe that it can be generalised to other heavily iris-pigmented populations, this should of course be verified in non-Chinese patients. Another limitation is the fact that consensus among uveitis specialists on diagnosis is modest at best, with great individual variation in the absence of a ‘gold standard’ for diagnosis of the disease. Our uveitis specialist’s opinion plays a strong and somewhat self-referential role in defining criteria. Further validation of our criteria would need a prospective analysis in consecutive patients with uveitis from other ethnic populations.
In conclusion, we developed RDC for FUS using data from 593 patients with FUS and 625 cases with non-FUS uveitis from northern China and validated the developed criteria using another set of patients with uveitis (377 FUS and 503 non-FUS) from southern China. Higher sensitivity and larger AUC were found in our RDC in comparison to the SUN classification criteria and the La Hey criteria. Further validation of the newly developed criteria would need a prospective analysis in other racial or ethnic populations.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplemental information.
The study followed the principles of the Declaration of Helsinki and was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (No. 2020027).
Contributors PY and WZ conceived and designed the study. WZ, ZC, HZ, YZ and ZZ collected the clinical data. WZ and GS analysed and interpreted the data. WZ and ZC wrote the initial draft. PY, WZ, ZC and AK judged data interpretation and edited the manuscript. PY supervised the study. All authors provided a final review and approved the manuscript before submission.
Funding This work was supported by Chongqing Chief Medical Scientist Project (2018), Chongqing Outstanding Scientists Project (2019), Chongqing Science & Technology Platform and Base Construction Program (cstc2014pt-sy10002) and Chongqing Key Laboratory of Ophthalmology (CSTC, 2008CA5003).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
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.