Introduction. 2010 marked a century since Perthes' disease was first described, but the aetiology and mechanism remain unknown. Worldwide, the incidence of Perthes' disease varies widely, yet this may be through differences in study design, population denominators or case ascertainment. It is suggested that differential exposure to adverse socioeconomic circumstances may be a key precipitant, although this remains the subject of debate. This work draws on several epidemiological studies that have sought to develop the understanding of Perthes' disease by examining a case register from Merseyside, discharge data from Scotland and the world's largest community disease register. A systematic review was performed to ensure a robustness and homogeneity between published studies in order to allow meaningful comparisons. Methods. Studies were based on data from the Merseyside Perthes' Disease Register (1976–2008), the UK General Practice Research Database (GPRD, 1990–2008) and hospital discharge data for Scotland (2000–2009). Temporal trends and geographic patterns were analysed and the relationship to deprivation investigated. A systematic review of the published literature was used to explore international variations in incidence up to December 2010 focusing upon the influence of race and latitude. Analyses were conducted using Poisson regression. Results. Systematic Review: 21 studies were included which described 27 populations in 16 countries, observing 124 million person years. The annual incidence ranged from 0.2 – 19.1 per 100,000 0–14 year-olds. Race was a key determinant with East Asians least affected and Whites most affected (East Asian IRR 1.0 (Ref), South Asian IRR 2.9 (2.4, 3.5), White IRR 8.8 (8.2, 9.6)). Latitude was a strong predictor of disease, even after adjusting for race. Each 10 degree increase in latitude resulted in the incidence rate increasing by a factor of 1.44 (1.30, 1.58). GPRD Study. There was a dramatic decline in Perthes' disease incidence within the UK, with annual incidence rates falling from 12.2 to 5.7 cases/100,000 0–14 year-olds over the 19 year study period (p<0.001). There was marked geographic variation in incidence with incidence rates in Scotland more than twice those in London (10–39 (95%CI 8.05 – 13.2) vs 4.6 (95% CI 3.4 – 6.1) per 100.000 0–14 year-olds). A more rapid decline in incidence was apparent in the Northern regions compared to Southern regions. The most deprived quintile had the highest disease incidence (rate ratio 1.49 (95% CI 1.10 – 2.04)) and, with the exception of London, regional incidence showed a strong linear relationship to regional deprivation score (p<0.01). Merseyside. There was a dramatic decline in Perthes' disease incidence within Liverpool with rates falling from 14.2 to 7.7 cases/100,000 0–14 year-olds over the 34-year study period (p<0.001). Incidence rates similarly halved within the nearby region of Knowsley (p=0.01) but remained largely static in the more affluent regions of Sefton where the incidence has remained around 7.2 cases/100,000 0–14 year-olds (p=0.73). The association with socioeconomic deprivation is striking with over three times the incidence in the most deprived quintile of multiple deprivation compared to the most affluent quintile of IMD (11.5 vs 3.8 cases/100.000 0–14 year-olds (p<0.001). The incidence, by ward region, was strongly correlated to the ward index of deprivation (p<0.001) (IRR 1.014 (1.007 – 1.021)). Scotland. Hospital discharge rates due to Perthes' disease fell annually by 5.6% (95% CI 2.4% – 8.8% p<0.001) between 2000–2010. Given that there has been no significant change in treatment practice during this period and the results of the above studies, this is likely to represent a real change in disease incidence. There was a strong association between socioeconomic deprivation and disease with rates amongst the most deprived quintile more than twice those of the most affluent (RR 2.1 (1.5 – 2.9)). Similar incidence gradients for deprivation were seen in both urban and rural environments. Conclusions. These studies provide strong evidence to suggest a declining incidence of Perthes' disease within the UK and a strong relationship to socioeconomic deprivation. Although Perthes' disease incidence is falling it remains an important cause of child morbidity and exemplifies socioeconomic inequalities in health. The striking UK North–South divide is similar to that seen in adult diseases such as cardiovascular disease and osteoporosis. Latitude has an independent association with disease which may be through the action of sunlight and Vitamin D. A deprivation-related exposure (probably acting prenatally) appears critical but the aetiological determinants remain elusive
