A total of 28 patients with JIA, which included 25 cases referred from the general paediatric clinic, 2 cases diagnosed in paediatric ward and 1 case referred from a peripheral hospital were managed at the rheumatology unit. These accounted for 1.2% of the 2330 rheumatic disease cases managed over the study period. As shown in Table 1, 21 (75%) of our JIA cases were female and the mean age at presentation was 9.8 years. The mean duration of symptoms before diagnosis was 21.8 months. Constitutional symptoms such as intermittent fever, fatigue, weight loss and night sweat were documented in 15 patients. These 15 cases were made up of 7 with PJIA, 5 with systemic-onset JIA (SoJIA) and 3 with oligoarticular JIA (OJIA). PJIA constituted 14 (50%); OJIA, 9 (32.1%); and SoJIA, 5 (17.9%) of the JIA cases. Extended and persistent variants of OJIA were documented in 6 and 3 OJIA cases respectively. Extra-articular features were observed in 7 JIA patients (anterior uveitis in 2, pleural effusion in 1, generalized lymphadenopathy in 2, and hepatosplenomegaly in 2). Mortality was documented in 2 cases while 11 patients were lost to follow up.

We studied 28 cases of JIA which constituted 1.2% of the 2330 rheumatic cases diagnosed and managed in a public rheumatology unit over a 6-year period. The scarcity of paediatric rheumatologists in Africa is a challenge for the early detection and appropriate management of childhood-onset rheumatic diseases. There is currently no paediatric rheumatologist in Nigeria and suspected and confirmed cases of paediatric rheumatic diseases are managed by other paediatricians and adult-patient rheumatologists. Accordingly, the awareness and indices of suspicion for paediatric rheumatic diseases are very low among Nigerian clinicians. It is unknown, how large the annual number of missed paediatric rheumatic diseases is and what burden this might be contributing to the morbidity and mortality due to childhood non-communicable diseases.

The high mean age of onset of JIA observed in this study is similar to findings recorded in other African studies 678911. This finding contrasts with the lower age of onset documented among children from North America and Western Europe 14. Although, there is currently insufficient information to conclude that disease onset is generally delayed in the African patient; however, the complex interplay of genetics and environmental factors in the aetiology of JIA leaves this possibility open. In a study comparing phenotypic features of JIA between African American and non-Hispanic white, the median age of onset and median age of presentation were significantly higher among African American than the non-Hispanic whites 15. It is likely that a longer mean period since disease onset to diagnosis coupled with poor recall in many illiterate patients may be compounding this detail. The long road to diagnosis which is due to less assess to rheumatology care in our environment often takes the patient through several over-the-counter drug use and alternative medicine patronage.

The female-male ratio of 3:1 found in our study is larger than the range of 1-2.3:1 found in most other studies 678911161718. Furthermore, the closest ratio to ours of 2.3:1 was found in the only previous Nigerian study 6. The reasons for this may include the fact that polyarticular variant which is known to show female preponderance was the most important category identified in both Nigerian studies. Also, enthesitis-related JIA, a subtype that is more frequent in males was not found 19. Juvenile psoriatic arthritis and undifferentiated JIA were also absent. There is a recognized variation in the geographical, gender and age distributions of the seven ILAR JIA subcategories across the world. In developing countries, polyarthritis is the most common subtype while in Western countries, oligoarthritis is most frequently observed 2. SoJIA is the most predominant subtype recorded among Asian children 2. The distribution of juvenile enthesitis-related arthritis possibly reflects the patterns of occurrence of HLA-B27 around the world. As the prevalence of HLA-B27 is low in Africa 20, the enthesitis related arthritis is less commonly observed among black African children 911. Studies from across Africa have also shown the dominance of PJIA among the cases of JIA on the continent 6891011. The patterns observed among Africans have been demonstrated among African Americans. In a comparative study, Fitzpatrick et al 15 observed that African American children are more likely to have polyarticular subtypes compared with non-Hispanic white children. It was suggested that these racial variations in the distributions of the JIA subtypes may reflect vital roles of genetics in the pathogenesis of JIA 15.

A high proportion of our patients, irrespective of subtype, had prominent constitutional symptoms and elevated inflammatory markers. This finding is in agreement with the observation of Oyoo et al 9 who found that fever and elevated ESR were present in 73.5 % and 100% respectively of Kenyan patients. Similarly, in a multi-ethnic study of JIA by Consolaro et al 14, they observed that African patients with JIA had higher levels of disease activity and low frequencies of inactive diseases compared with North American and Western European patients. Although, we were unable to assess disease activity in the cases studied using validated composite tools, our patients might have had high baseline disease activities based on the levels of the inflammatory markers. Autoantibodies are not necessary for the diagnosis of JIA but they can be useful for excluding differential diagnoses of chronic arthritis in children. They are also important for sub-categorisation of PJIA into RF-positive and RF-negative subtypes as well as determining the risks of chronic anterior uveitis in those with ANA-positive OJIA 1.

Many of our patients had been on NSAIDs and in a few cases oral prednisolone before presentation. Suitable adjustments were made in these drugs and all patients were commenced on subcutaneous or oral methotrexate according to our local policy. Intra-articular steroids were administered in few cases. The only JIA-suitable biologic available in Nigeria is etanercept and only one in six patients eligible for it were able to afford it. Affordability of biologics and other emerging treatment modalities for JIA is a huge barrier to the optimal treatment of Nigerian patients. The only other biologic currently available in Nigeria, rituximab, has not been approved for JIA but anecdotal reports have documented its efficacy in refractory SoJIA 21.

Mortality in JIA has been shown to be higher than in the general population, and especially, higher among individuals with SoJIA than other subtypes 22. Davies et al 22 observed that the standardized mortality rate among 693 JIA cases was 7.3 (95% CI: 2.9-15) while the standardized mortality rate among 99 systemic JIA cases was 21.7 (95% CI: 5.9-55.4). The mortalities in our study were attributable to infection and suspected MAS. Previous studies have documented infection, MAS, and renal failure from amyloidosis as the major causes of death 2223. The Limitations of this study include the fact that the actual outcomes of these cases are likely to have been poorly appreciated due to the high proportion of patients lost to follow up. Also, our findings will have limited generalizability due to the few cases involved and the retrospective nature of the study.