Persistence With Growth Hormone Therapy in Pediatric Patients | Page 2
Published Online: February 19, 2014
Bradley S. Miller, MD, PhD; Deborah Rotenstein, MD; Larry C. Deeb, MD; John Germak, MD; and Tami Wisniewski, MPH
Demographics and baseline characteristics of patients with IGHD and MPHD, and of patients in each discontinuation category, were summarized using descriptive statistics. Significant differences between patient group baseline characteristics were determined using independent sample t tests. A x2 test evaluated differences in sex proportions among categories. The percentage of patients remaining on therapy over time in each category was determined from Kaplan-Meier survival curves. Adjusted for baseline age and sex, persistence was estimated with a computerized statistical model for survival data using proportional hazards without censoring and the Kolmogorov-Smirnov test for comparisons among categories. The same model was applied to analyze persistence in patients with IGHD or MPHD. The HSDS and corrected HSDS were determined using a last-observation-carried-forward approach and tested for differences between each category using independent sample t tests. All reported values are mean and standard deviation (SD). Differences between means were considered significant at P <.05.
Patient Demographics and Baseline Characteristics by Diagnosis
An initial total of 944 patients with GHD (IGHD, n = 876; MPHD, n = 68) met the inclusion criteria (Table 1). The mean age at baseline was 11.7 ± 3.48 years and 71.7% of the patients were male (IGHD: 72.9% male; MPHD: 55.9% male). Patients with IGHD were, on average, 2 years older (11.8 ± 3.27 years) than patients with MPHD (9.9 ± 5.22 years; P = .0038), and had a greater average delayed bone age versus chronologic age (IGHD, 1.6 years; MPHD: 0.2 years). The HSDS and target HSDS at baseline differed between patients with IGHD and MPHD (HSDS: −2.1 ± 0.85 and −1.8 ± 1.19, respectively, P = .0213; target HSDS: −0.4 ± 0.83 and 0.1 ± 0.75, respectively, P = .0005). Maximal stimulated serum GH levels (nanograms per milliliter) were significantly higher in patients with IGHD than MPHD (6.0 ± 2.59 and 4.4 ± 3.99, respectively, P = .0125). Baseline Characteristics by Discontinuation Category: Combined IGHD and MPHD Baseline characteristics were analyzed in a reduced subset of 826 patients for whom reason for discontinuation of GH treatment was available. Individuals with IGHD (n = 778) or MPHD (n = 48) were combined and categorized based on reason for discontinuation of GH treatment (Table 2). Baseline characteristics were similar among all discontinuation categories except for chronologic age and bone age. The mean baseline chronologic age (9.9 ± 3.8 years; P <.0001) and the mean baseline bone age (8.8 ± 3.3 years; P <.0001) were lowest in the insurance-issues patient category. No statistically significant differences (P = .4563) occurred in sex distributions among categories (ie, final height achieved, 70.8% male; insurance issues, 75.8% male; patient and caregiver decision, 70.1% male; and other, 69.4% male). The most common reason for discontinuation was final height achieved (34.9%), followed in order by insurance issues (28.0%), other (20.9%), and patient and caregiver decision (16.2%). Nonadherence (3.6%), healthcare provider recommendation (3.3%), adverse events (2.5%), change of physician or patient moved (1.5%), or lack of response (0.1%) were collectively included in the other category. The mean ± SD treatment duration was 46 ± 21 months for final height achieved, followed by patient and caregiver decision (35 ± 15 months), insurance issues (32 ± 21 months), and other (31 ± 18 months). Overall, no difference in treatment duration (mean ± SD) was observed between patients with IGHD (37 ± 19.9 months, n = 778) or MPHD (39 ± 26.7 months, n = 48). However, in patients who discontinued GH due to final height achieved, those with MPHD had a longer duration of treatment (57 ± 35.4 months, n = 13) than patients with IGHD (45 ± 19.6 months, n = 275); this was likely due to the younger age of MPHD patients at treatment start, but may also have been due to delayed entry into puberty of individuals with MPHD. The overall mean duration of GH treatment among all categories was 37 ± 20.3 months (n = 826).
Age- and Sex-Adjusted Persistence With GH Therapy
After adjusting for baseline age and sex, differences in persistence with GH therapy over time among the 4 discontinuation categories were estimated from analysis of survival curve functions (Figure 1A). Category comparisons indicated that patients who discontinued therapy due to final height being achieved were more persistent than patients in the other categories of insurance issues (P <.0001), other (P <.0001), and patient and caregiver decision (P = .0009). Patients who discontinued therapy due to patient and caregiver decision were also more persistent than patients who discontinued therapy due to insurance issues (P = .0086). Thus, patients who discontinued because of insurance issues or other reasons were the least persistent with GH treatment. Patients with either IGHD or MPHD were equally persistent with GH therapy (Figure 1B; P = .6953).
Height Outcome by Reason for Discontinuation
Overall, mean HSDS increased significantly in the combined IGHD and MPHD groups from baseline to the year 3 visit (Table 3) in patients in all discontinuation categories (Figure 2A, P <.0001). Pairwise comparisons showed that mean HSDS at last visit for patients who discontinued due to final height achieved (−0.6 ± 0.91, n = 288) was significantly higher than the mean HSDS for other (−1.3 ± 1.06, n = 173; P <.0001), insurance issues (−1.2 ± 0.90, n = 231; P <.0001), and patient and caregiver decision (−0.9 ± 1.19, n = 134; P = .0222). At year 3, prior to GH discontinuation and the last visit, the mean HSDS for final height achieved (−0.7 ± 0.82, n = 197) was significantly higher than that for other (−1.2 ± 1.06, n = 65; P = .0006) and insurance issues (−1.0 ± 0.83, n = 79; P = .0149), but it was not significantly higher than that for patient and caregiver decision (−1.0 ± 0.95, n = 51; P = .0679).
PDF is available on the last page.