Epidemiology and outcomes of non-infectious paediatric uveitis in the French Caribbean: a 15-year cohort study from a high-income health system in Martinique Authors: Arthur Felix Published: 2026 ABSTRACT AbstractBackgroundNon-infectious uveitis (NIU) are rare in children but might be associated with substantial visual impairment. Most NIU paediatric cohorts originate from Europe, North America, or Asia, with minimal representation of Caribbean populations. Our study aimed to describe the epidemiology, clinical features, aetiologies and outcomes of paediatric NIU in Martinique.FindingsWe conducted a monocentre, retrospective cohort study that included all children under the age of 18 diagnosed with NIU at the University Hospital of Martinique between 2010 and 2024. We identified cases through paediatric and ophthalmology registries, medical records, and national hospital databases. We excluded infectious uveitis and non-residents. Incidence estimates used population data. A total of 17 children with NIU were included in the study. 41% were male, with a mean age at diagnosis of 8.7 years (range: 2.5–16) and a median follow-up period of 7.6 years (range: 2.2–11.6). Eleven patients had anterior uveitis (64.7%), and 35.3% had panuveitis. Uveitis was unilateral in 64.7% of patients. The estimated incidence was 1.4 per 100,000 person-years (95% CI 0.83–2.27). The leading aetiology was juvenile idiopathic arthritis, accounting for 58.8% of cases. Systemic biologics were used in 17.6% of patients. Only one child experienced visual loss and required ocular surgery at the final follow-up.ConclusionsThe first Caribbean paediatric NIU cohort shows a similar etiological spectrum and good visual outcomes as reported in high-income countries.Keywords: Uveitis, Juvenile idiopathic arthritis, French West Indies, Caribbean Introduction Non-infectious uveitis (NIU) are a rare but potentially sight-threatening inflammatory disorder that frequently occurs alongside chronic autoimmune diseases, such as juvenile idiopathic arthritis (JIA), sarcoidosis, and idiopathic ocular inflammation [1]. Although paediatric NIU accounts for only 5–10% of all uveitis cases, it carries a high burden of visual morbidity due to its insidious onset, delayed diagnosis, and risk of long-term complications, such as cataracts, glaucoma, and macular oedema [2]. Although JIA-associated uveitis are the most frequent form of NIU in children in Europe and North America [3], significant geographical and ethnic variations in incidence, phenotype, and outcomes have been described worldwide [4]. Several autoimmune and inflammatory conditions, including lupus, sarcoidosis, and scleritis, show distinct epidemiological patterns and, in many cases, more severe disease courses in Afro-descendant (AD) and Caribbean populations compared to Caucasian populations [5, 6]. However, data on NIU among AD and Caribbean children remain extremely limited as most of the available studies from Latin America and the Caribbean are small and often comprise mixed adult-paediatric cohorts. These studies report heterogeneous aetiology distributions and variable access to ophthalmologic and rheumatologic care, as well as access to immunosuppressive treatment and biologics [4]. To date, there has been no comprehensive description of the epidemiology, clinical features, and outcomes of paediatric NIU in the Caribbean. It is unclear whether the low number of reported cases reflects a low true incidence rate, differences in genetic susceptibility or variability in diagnostic pathways. Martinique island, in the French Caribbean, comprises approximately 80,000 children under the age of 18. Although ethnicity is not officially recorded, more than 90% of the population is Afro-Caribbean [7]. Children in Martinique benefit from a universal healthcare system that provides access to advanced paediatric and ophthalmological care [7]. There is only one tertiary centre on the island accredited by the French Ministry of Health for paediatric ophthalmology and rheumatology. The main objective of our study was to describe the epidemiology, the characteristics and outcomes of children with NIU in Martinique between 2010 and 2024. Methods We conducted a monocentre retrospective, descriptive cohort study of all paediatric patients from Martinique diagnosed with NIU. We included children aged 0–17 years who were treated in the Martinique university hospital paediatric units between January 2010 and December 2024. We identified patients using registries from paediatricians and ophthalmologists, as well as data extracted from electronic hospital archives and the French Medicalization of Information Systems Program (DIM/PMSI), based on ICD-10 codes H20, H22, H30, and H44. Patients older than 18 years and/or non-resident were excluded. Patients with infectious uveitis were also excluded. After data extraction, the lists of patients were analysed for the relevance of the diagnoses, inclusion criteria and to eliminate duplicates. Data were collected by reviewing all available medical reports and follow-up. According to the SUN classification, senior ophthalmologists classify uveitis according to the segment of the eye affected [8]. All the children underwent an aetiological assessment in the paediatric department : Hemogram and CRP; complete liver and kidney function tests; screening for tubulopathy or proteinuria; Quantiferon; complete infectious disease workup depending on aspect and topography; HLA-B27 and HLA-B51; complete lymphocyte phenotyping; antinuclear factor; angiotensin-converting enzyme and lysozyme; chest X-ray; accessory salivary gland biopsy; brain MRI with injection; lumbar puncture if posterior involvement. The diagnosis of JIA-associated uveitis was made according to the international criteria [9]. The paediatric inflammatory bowel diseases, Vogt–Koyanagi–Harada syndrome, and sarcoidosis were diagnosed according to the international classification [10–12]. The aetiological diagnoses were confirmed by both the paediatric rheumatology team and the paediatric ophthalmology team based on the results of the aetiological assessment and the clinical characteristics of the eye involvement. Data regarding the population were provided by the National Institute for Statistics and Economics (INSEE). The population of Martinique was stable throughout the study period, with 80,000 children aged < 18 years. All patients were treated according to French and European recommendations including all biologics without any shortage or unavailability [7]. In general terms, children were initially treated with glucocorticoid eye drops applied to the affected eye upon diagnosis. Depending on the inflammatory/visual response and/or the underlying cause, first-line treatment included systemic corticosteroids. Methotrexate and/or biologics were introduced according to clinical indications relating to inflammatory activity, visual response, glucocorticoid dependence or the underlying systemic disease. The follow-up period was defined as the time between the date of diagnosis and the last date of contact. We only performed descriptive analysis and, due to the small sample size, no analytical statistics were performed. Incidence rates were calculated by dividing the number of cases over the 2010–2024 period by the total person-years at risk (paediatric population × study duration) and expressed per 100,000 person-years, with 95% confidence intervals estimated using a Poisson distribution. The results were reported according to the STROBE methodology. The study summary was sent to all families to obtain their informed consent. Written informed consent was obtained from the parents or legal guardians of all participants, who were provided with a synopsis of the study. The Institutional Review Board of the University Hospital of Martinique approved the study under the number 2021/116. Results A total of 17 children with NIU were included in the study period. During the same period, five children with infectious uveitis (4/5 toxoplasmosis and 1/5 toxocariasis) were excluded (22%). Table 1 summarizes the baseline demographic and clinical characteristics. The median follow-up was 7.6 years (IQR: 2.2–11.6). The incidence observed in our population was 1.4 per 100,000 person-years (95% CI 0.83–2.27). Table 2 shows the individual characteristics of the patients who were included in the study. Table 1Clinical characteristics of children with non-infectious uveitis in MartiniqueVariableNumber of patients17Mean age (range min–max), years8.7 (2.5–16.0)Male sex, n (%)7 (41.2)Unilateral uveitis, n (%)11 (64.7)Bilateral uveitis, n (%)6 (35.3)Granulomatous, n (%)2 (11.8)Average visual acuity at diagnosis0.8 logMAR (IQR 0.5–1.0)Average visual acuity at last follow-up0.1 logMAR (IQR 0–0.2) SUN anatomical classification Anterior uveitis, n (%)11 (64.7)Intermediate uveitis, n (%)0 (0)Posterior uveitis, n (%)0 (0)Panuveitis, n (%)6 (35.3) Complications Cataract, n(%)1 (5.9)Glaucoma, n(%)0 (0)Ocular surgery (cataract), n (%)1 (5.9)Visual loss at last follow-up, n (%)1 (5.9) Treatments Glucocorticoid drops only, n(%)8 (47.1)Oral glucocorticoid < 4 months, (%)9 (52.9)Methotrexate, n(%)6 (35.3)Biologics (Anti-TNFa), n (%)3 (17.6)Median duration of treatment, months9 (range 3-125)Values were presented as mean (SD), range, or percentage, as appropriate. Uveitis classification follows the Standardization of Uveitis Nomenclature (SUN) criteria. Visual loss corresponds to permanent best-corrected visual acuity < 20/200 in at least one eye at last follow-up. Granulomatous was defined clinically as mutton-fat keratic precipitates and/or iris nodules. Best-corrected visual acuity was recorded and expressed in logMAR units for statistical analysis Clinical characteristics of children with non-infectious uveitis in Martinique Values were presented as mean (SD), range, or percentage, as appropriate. Uveitis classification follows the Standardization of Uveitis Nomenclature (SUN) criteria. Visual loss corresponds to permanent best-corrected visual acuity < 20/200 in at least one eye at last follow-up. Granulomatous was defined clinically as mutton-fat keratic precipitates and/or iris nodules. Best-corrected visual acuity was recorded and expressed in logMAR units for statistical analysis Table 2Individual patient characteristics (n = 17) Individual patient characteristics (n = 17) In terms of aetiology, JIA accounted for the majority of cases (58.8%), followed by inflammatory bowel disease (IBD) (11.8%), Vogt–Koyanagi–Harada syndrome (5.9%), sarcoidosis (5.9%), and idiopathic cases (11.8%) (Tables 1 and 2). All the JIA patients had oligoarticular forms and positive ANA. All cases of uveitis associated with JIA were diagnosed within the first six months of ophthalmological surveillance. All of these patients had articular symptoms (mostly isolated lower limb arthritis) prior to being diagnosed with uveitis. None of them were taking methotrexate or biologics at the time of uveitis diagnosis. All patients required topical or systemic glucocorticoid, and 17.6% received systemic biologics, mostly anti-TNFα. Only one patient (5.9%) reported visual loss or ocular surgery at last follow-up. Discussion This retrospective study presents the first paediatric cohort of NIU in the Caribbean, focusing on predominantly AD children from Martinique. Despite the high prevalence of autoimmune and inflammatory diseases in Caribbean population [5], to our knowledge, no previous study has specifically documented the epidemiology, clinical phenotypes, and outcomes of paediatric uveitis in this population. Our tertiary centre provided an opportunity to characterize uveitis in an underrepresented population in the literature without the confounding effects of barriers to care, as Martinique is a Caribbean island with a high-income country health system [7]. By contrast, most published paediatric uveitis cohorts originate from North America, Europe, or Asia with little representation of Caribbean children [4]. The lack of data has historically limited our understanding of potential variations in disease expression, access to treatment, and long-term prognosis related to ethnicity [4]. Therefore, we believe that our study fills a knowledge gap by offering insight into disease patterns in a well-resourced Caribbean setting. The clinical presentation of uveitis in this cohort was characterised by mostly anterior uveitis, with more than one-third of cases involving panuveitis. JIA was the leading cause, accounting for nearly 58.8% of cases, which is consistent with most studies from Europe and North America [4, 13, 14]. The retrospective design limited the detailed characterisation of ocular inflammatory features, such as vitritis, snowballs, retinal vasculitis or infiltrates, which were not systematically recorded or quantified in ophthalmologists’ reports. Consequently, the SUN anatomical categories [8] were used for classification, but a cautious interpretation of inflammatory features is required. Granulomatous uveitis and sarcoidosis were notably uncommon in our cohort, affecting only one patient. Despite the higher incidence of sarcoidosis among individuals of African descent being well-established, studies conducted in Afro-Caribbean populations report lower incidence rates [15]. Granulomatous uveitis appears to be more prevalent in reports from developing countries, where infectious aetiologies (e.g., tuberculosis, toxoplasmosis, and toxocariasis) often predominate [4, 13, 14, 16]. Over the study period, the ratio of NIU to infectious uveitis was approximately 3:1. The low prevalence of infectious uveitis reported here is more consistent with what is reported in high-income countries [4] than in emerging countries, where infectious uveitis is more common (40–60% depending on the region) [4, 14, 16–18]. This difference could be indicative of an epidemiological transition in high-income countries with universal healthcare systems. Therefore, our findings could also highlight the distinct aetiological spectrum of uveitis in high-income settings compared to low- or middle-income countries. Overall, visual outcomes were favourable, with visual acuity improving from moderate impairment at diagnosis to near-normal values at the final follow-up. This improvement is clinically significant in paediatric NIU, where persistent inflammation and delayed treatment can result in permanent visual impairment [3]. The low incidence of cataracts, the absence of glaucoma and the limited need for ocular surgery suggest that severe structural complications were uncommon in this cohort. However, these findings should be interpreted with caution given the small sample size. As relapse episodes were neither prospectively defined nor systematically recorded throughout the study period, their frequency could not be analysed. Therefore, treatment escalation was only interpreted as a description of treatment exposure and therapeutic intensity, rather than as a surrogate marker of relapse frequency or disease control. Several similarities and differences emerge when comparing our results with those of other paediatric uveitis cohorts worldwide. The incidence observed in our population (approximately 1.4 per 100,000 person-years) falls within the lower range of estimates reported in population-based studies from Finland, Norway, and South Korea, in which incidences range from 4 to 14 per 100,000 [4, 19, 20] . However, the reported incidence reflects a hospital-based estimate rather than a strictly population-based rate. However, given that Martinique has a single tertiary referral centre for paediatric ophthalmology and rheumatology, to which all suspected cases are theoretically referred, under-ascertainment is likely to be limited. Nevertheless, some underestimation cannot be ruled out due to the retrospective design, particularly for mild cases or undiagnosed patients. Cohorts from the United States and South America generally report higher proportions of bilateral disease and often include older children [4, 18, 21]. In contrast, our population displayed a mixed pattern, with a relatively young mean age and predominantly unilateral involvement. Cohorts from Asia often have a high prevalence of idiopathic uveitis, which differs from our JIA majority population [4, 22]. However, our incidence estimate was hospital-based rather than strictly population-based, so comparisons with population-based studies should be treated with caution. The proportion of patients receiving anti-TNF biologics was slightly lower, but comparable to the results of Western countries studies [2, 14, 22]. Overall, our findings were similar to those of paediatric cohorts in Western countries, suggesting the impact of healthcare infrastructure and early screening on the spectrum and severity of uveitis and visual outcomes rather than ethnicity; however, this result should be tempered by the low number of patients in our cohort. This study has several strengths. First, it is the largest series and the only one specifically describing Caribbean children with NIU in a high-income setting. This provides unique insight into an underreported population in the literature. Second, care was delivered within a universal, free healthcare system that provided ready access to paediatric ophthalmologists, rheumatologists, imaging services, and biologics. This could help to reduce socioeconomic and structural biases, providing a more accurate picture of the expression of disease. Third, systematic data extraction and cross-verification across ophthalmology and rheumatology records ensured robust case identification and minimized missing data. Fourth, the study included a comprehensive evaluation of anatomical classification, aetiology, and treatment patterns (Table 1). However, our study has several limitations. Due to its retrospective design, certain biases could not be avoided, especially during data collection. It is possible that some cases of NIU were excluded from our study due to incomplete historical documentation. The single-island sample and small cohort size limited external generalizability and precluded formal subgroup analyses. Although informative, the incidence estimate may have been underestimated due to the biases inherent in retrospective studies of rare diseases over long periods. Furthermore, since this cohort exclusively represents NIU, we cannot extrapolate the overall burden and epidemiology of uveitis, including infectious aetiologies, from these data. Our high-income, universal system is unique in this region, which might limit the ability to extrapolate to other Caribbean countries [7].There was also potential for recruitment bias because our methodology primarily identified patients requiring hospitalization. However, most patients with NIU benefit from at least one hospitalization or day hospital stay for diagnostic investigations. Despite these limitations, our monocentre study will contribute to the knowledge of these diseases in children, given the rarity of NIU in children and the lack of previous descriptions of Caribbean patients. Conclusion Overall, this study is the first to provide a detailed description of non-infectious paediatric uveitis in Caribbean children within a high-income, universal healthcare system. The etiological profile, which was dominated by JIA, and the favourable access to immunomodulatory treatment, resulted in favourable outcomes for children. The incidence was slightly lower than in the previously published main series. Acknowledgements None. Abbreviations Afro-descendant Anti–tumour necrosis factor alpha Inflammatory bowel disease International Classification of Diseases, 10th revision National Institute for Statistics and Economic Studies Juvenile idiopathic arthritis Non-infectious uveitis Author contributions AF were responsible for data extraction, data analysis, results interpretation, wrote the draft of the manuscript. All authors had responsibility for the decision to submit the study for publication. Funding This work did not receive any financial assistance. Data availability All data generated or analysed during this study are included in this published article and tables. Declarations Ethics approval and consent to participate This study was conducted in accordance with the Declaration of Helsinki and its later amendments. Ethical approval was obtained from the Institutional Review Board of the University Hospital of Martinique approved the study under the number 2021/116. Written informed consent was obtained from the parents or legal guardians of all participants, who were provided with a synopsis of the study. Consent for publication Written informed consent to publication was obtained from the parents or legal guardians of all participants. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement All data generated or analysed during this study are included in this published article and tables.