Is clinical expressiveness of Paget's disease of bone decreasing?

Recent data have suggested secular changes implying a current trend toward decreased clinical severity of Paget's disease of bone (PD). To test this hypothesis, we conducted a study comparing the characteristics of two groups of PD patients, as disclosed from a sample assessed systematically. The investigation was a hospital-based study of all cases followed up at our unit since 1980. Throughout the follow-up period, diagnosis was based on standard X-ray criteria and the same clinical assessment was applied. Group I (n = 124) represented patients born before 1926, whereas group II (n = 109) included those born after that year. A bone scan performed with 99mTc-EHDP was available for all patients. X-rays of the pelvis and spine, and views of any hot spot observed on the scintigraphy scans were reviewed. The skeletal extent of PD, based on bone scan uptake, was determined by using the index proposed by Coutris. Alkaline phosphatase and hydroxyproline excretion levels were determined in blood and urine, respectively. Baseline characteristics were recorded on a purpose-designed computerized database. The proportion of males (47% in group I vs. 65% in group II; p = 0.007) and the mean (+/-SD) age at diagnosis (69.0 +/- 8.15 vs. 54.3 +/-9.14; p < 0.001) differed significantly between groups. The year of birth showed a strong negative correlation with age at diagnosis (r = -0.83, p < 0.0001) and a weak, but significant, negative correlation with extent of bone lesion (r = -0.20; p = 0.002). Likewise, subjects born prior to 1926 showed a greater percentage of affected skeleton cases (9.6 plus minus 8.01 vs. 7.06 +/- 5.79; p = 0.001). Group I individuals who had pelvic and/or femoral bone lesions were more prone to suffer "pagetic coxopathy" (65% vs. 40%; p = 0.003) with "protrusio acetabuli" (32% vs. 17%; p = 0.01), and the percentage of patients showing radiographic Monckeberg-type vascular calcifications (36% vs. 14%; p = 0.0006) was higher than in those born after 1926. No other epidemiologically clinically, or biochemically relevant differences were seen in the crude analysis. Multivariate analysis identified extent of skeletal lesions (OR = 0.76; p = 0.01), age at diagnosis (OR = 0.79; p = 0.008), number of bones involved (OR = 1.53; p = 0.03), and occupation (p < 0.0001) as the predictive variables linked to year of birth. Our data are consistent with a temporal tendency toward a smaller number of bone lesions and a decreased percentage of instances of affected skeleton. An earlier age at recent diagnosis times and absence of any relevant clinical or biochemical differences seems more likely linked to recent changes in referral and sociological patterns.