Mortality Is Not Increased in Recombinant Human Growth Hormone-treated Patients When Adjusting for Birth Characteristics

Objective: This study aimed to investigate whether reported high mortality in childhood recombinant human GH (rhGH)-treated patients was related to birth-characteristics and/or rhGH treatment. Design and Setting: We sought to develop a mortality model of the Swedish general population born between 1973 and 2010, using continuous-hazard functions adjusting for birth characteristics, sex, age intervals, and calendar year to estimate standardized mortality ratio (SMR) and to apply this model to assess expected deaths in Swedish rhGH-treated patients with idiopathic isolated GH deficiency (IGHD), idiopathic short stature (155) or born small for gestational age (SGA). Participants:The general population: Swedish Medical Birth Register (1973-2010: 1 880 668 males; 1 781 131 females) and Cause of Death Register (1985-2010). Intervention Population: Three thousand eight hundred forty-seven patients starting rhGH treatment between 1985 and 2010 and followed in the National GH Register and/or in rhGH trials diagnosed with IGHD (n = 1890), ISS (n = 975), or SGA (n=982). Main Outcome Measures: Death. Results: Using conventional models adjusting for age, sex, and calendar-year, the SMR was 1.43 (95% confidence interval, 0.89-2.19), P = .14, observed/expected deaths 21/14.68. The rhGH population differed (P < .001) from the general population regarding birth weight, birth length, and congenital malformations. Application of an Advanced Model: When applying the developed mortality model of the general population, the ratio of observed/expected deaths in rhGH-treated patients was 21/21.99; SMR = 0.955 (0.591-1.456)P = .95. Model Comparison: Expected number of deaths were 14.68 (14.35-14.96) using the conventional model, and 21.99 (21.24-22.81) using the advanced model, P < .001, which had at all ages a higher gradient of risk per SD of the model, 24% (range, 18-42%; P < .001). Conclusions: Compared with the general Swedish population, the ratio of observed/expected deaths (21/21.99) was not increased in childhood rhGH-treated IGHD, ISS, and SGA patients when applying an advanced sex-specific mortality model adjusting for birth characteristics.