Luis A García Rodríguez1, Antonio González-Pérez2, Alexander Michel3, María E Sáez4. 1. Spanish Centre for Pharmacoepidemiologic Research (CEIFE), Almirante 28, Madrid 28004, Spain. Electronic address: lagarcia@ceife.es. 2. Spanish Centre for Pharmacoepidemiologic Research (CEIFE), Almirante 28, Madrid 28004, Spain; Andalusian Bioinformatics Research Centre (CAEBi), Seville, Spain; Pharmacoepidemiology Research Group, Institute for Health Research IRYCIS, Madrid, Spain. 3. Epidemiology, Bayer Consumer Care AG, Division Pharmaceuticals, Basel, Switzerland. 4. Spanish Centre for Pharmacoepidemiologic Research (CEIFE), Almirante 28, Madrid 28004, Spain; Andalusian Bioinformatics Research Centre (CAEBi), Seville, Spain.
Abstract
OBJECTIVE: To estimate the contemporary epidemiology of systemic sclerosis (SSc) in the UK and to explore the validity of using The Health Improvement Network (THIN) primary care database to study SSc. METHODS: 4,520,299 individuals (2000-2012) aged 1-99 years were followed to identify potential incident cases of SSc; potential prevalent cases were identified at start of follow-up. Patient profiles, including free-text comments, were manually reviewed to verify cases of SSc, and case validation was undertaken for a sample (n = 100) using questionnaires to primary care physicians. Incidence, prevalence and mortality rates were calculated. RESULTS: Following manual review, we identified 1321 cases of SSc (689 incident and 632 prevalent) over a mean follow-up of 7.6 years; mean age at diagnosis 59.1 years. Using information from 91/100 valid questionnaires returned, the positive predictive value of SSc diagnoses in THIN following manual review was 94%. Incidence rates of SSc per 100,000 person-years (95% CI) were 2.02 (1.87-2.17) overall, [3.36 (3.09-3.64) in females, 0.61 (0.49-0.74) in males], ranging from 1.53 (1.11-2.06) to 2.56 (1.78-3.56) across calendar years. Prevalence of SSc per 100,000 (95% CI) increased from 17.13 (14.97-19.51) to 25.38 (23.68-27.16) over the study period, and was higher in females. Using Poisson regression, the adjusted mortality rate ratio was 2.82 (95% CI: 2.55-3.13) among SSc cases versus the general population. CONCLUSIONS: THIN enables precise and valid estimates of SSc occurrence to be determined. The observed increase in SSc prevalence has not been driven by increasing incidence.
OBJECTIVE: To estimate the contemporary epidemiology of systemic sclerosis (SSc) in the UK and to explore the validity of using The Health Improvement Network (THIN) primary care database to study SSc. METHODS: 4,520,299 individuals (2000-2012) aged 1-99 years were followed to identify potential incident cases of SSc; potential prevalent cases were identified at start of follow-up. Patient profiles, including free-text comments, were manually reviewed to verify cases of SSc, and case validation was undertaken for a sample (n = 100) using questionnaires to primary care physicians. Incidence, prevalence and mortality rates were calculated. RESULTS: Following manual review, we identified 1321 cases of SSc (689 incident and 632 prevalent) over a mean follow-up of 7.6 years; mean age at diagnosis 59.1 years. Using information from 91/100 valid questionnaires returned, the positive predictive value of SSc diagnoses in THIN following manual review was 94%. Incidence rates of SSc per 100,000 person-years (95% CI) were 2.02 (1.87-2.17) overall, [3.36 (3.09-3.64) in females, 0.61 (0.49-0.74) in males], ranging from 1.53 (1.11-2.06) to 2.56 (1.78-3.56) across calendar years. Prevalence of SSc per 100,000 (95% CI) increased from 17.13 (14.97-19.51) to 25.38 (23.68-27.16) over the study period, and was higher in females. Using Poisson regression, the adjusted mortality rate ratio was 2.82 (95% CI: 2.55-3.13) among SSc cases versus the general population. CONCLUSIONS: THIN enables precise and valid estimates of SSc occurrence to be determined. The observed increase in SSc prevalence has not been driven by increasing incidence.