Increased height standard deviation scores in response to growth hormone therapy to near-adult height in older children with delayed skeletal maturation: results from the ANSWER Program

Treatment with recombinant human growth hormone (GH) is indicated for children with short stature or growth failure associated with a number of conditions in which there is a deficiency of, or decreased responsiveness to, endogenous growth hormone. The use of GH is approved by the US Food and Drug Administration for the treatment of growth failure associated with growth hormone deficiency (GHD), either isolated (IGHD) or part of multiple pituitary hormone deficiency (MPHD), and short stature associated with Turner syndrome (TS), Noonan syndrome (NS), children born small for gestational age (SGA), short stature homeobox (SHOX) gene haploinsufficiency, Prader-Willi syndrome (PWS), chronic kidney disease (CKD), or idiopathic short stature (ISS) [1]. Treatment with GH has been shown to increase short-term linear growth in children with various disorders associated with growth failure; however, some of the short-term clinical studies have shown varying treatment outcomes [2‐4]. A limited number of studies have investigated the effects of long-term GH therapy (GHT) on the adult height standard deviation score (HSDS) achieved and whether the attained adult height was within the normal range.

Attaining adult height within the normal range is an important treatment goal for GHT in children with short stature or growth failure disorders, especially with respect to the potential benefits in health-related quality of life (HRQoL) and physical and psychosocial well-being [5, 6]. The adult height of patients treated with GH is influenced by the height gained during the course of therapy, the rate of bone maturation, and the onset of puberty [7, 8]. Longitudinal analyses of large data sets of pediatric patients have consistently demonstrated that the early initiation and appropriate duration of GH treatment correlated significantly with the achievement of near-normal adult height and greater improvements in HSDS [2, 9, 10]. A multicenter, randomized, double-blind clinical study of children with short stature who were born SGA without signs of catch-up growth showed that GH treatment to adult height was associated with a significant improvement in HRQoL and the normalization of final height [5].

The American Norditropin Studies: Web-Enabled Research (ANSWER) Program (utilizing NovoNet®, the Novo Nordisk Web-based research platform) is an observational, noninterventional study evaluating the long-term effectiveness and safety of Norditropin (somatropin [recombinant DNA origin] injection, Novo Nordisk A/S, Bagsvaerd, Denmark), hereafter referred to as GH, therapy in pediatric and adult patients [11]. Enrollment in this patient registry is solely at the discretion of the participating physicians in patients for whom Norditropin® is prescribed for treatment of appropriate conditions of growth failure and short stature both within and outside the Norditropin® label. Data from the ANSWER Program have shown that children with IGHD, MPHD, ISS, TS, NS, or SGA demonstrated an increase in HSDS from baseline following 2 years of GH treatment [12, 13]. Height velocity (HV) at 4 months of GHT and baseline body mass index (BMI) SDS were significant predictive factors in patients with GHD (including patients with IGHD and MPHD) that positively correlated with a change in HSDS (ΔHSDS) [13]. Baseline age, baseline HSDS, and baseline serum insulin-like growth factor-1 (IGF-I) SDS were negatively correlated with ΔHSDS [12, 13]. No consistent effect of the sex of a patient on response to GH treatment was observed in these analyses from the ANSWER Program; however, a sex-related effect on ΔHSDS was reported by Savendahl et al. [14] in a combined analysis of short-term (2-year) GH treatment responses in children with GHD, MPHD, or SGA enrolled in the ANSWER Program and its European counterpart, the NordiNet® International Outcome Study. Greater growth responses were observed in younger, prepubertal children in all diagnostic groups studied in a report from the ANSWER Program, indicating that early initiation of GH treatment is important for optimizing linear growth response [12].

The principal aim of the current analysis of data from the ANSWER Program registry was to evaluate the effects of GH treatment in previously untreated pediatric patients with GHD, ISS, or TS with baseline HSDS ≤ -2 on near-adult height (NAH) outcomes. The effect of age at the start of treatment on attainment of NAH was also evaluated. NAH was evaluated only at the last clinic visit when GH was discontinued, whereas HSDS was determined at year 1, year 2, and at NAH (i.e., the last clinic visit).

The results of this analysis of GH treatment in pediatric patients with GHD, ISS, or TS demonstrated that GH-naïve children with short stature achieved a mean NAH within the normal reference range. Mean HSDS, corrected HSDS, and IGF-I SDS all improved over time with GH treatment, and the majority of patients were within the normal reference range (> - 2 HSDS) at the time of NAH. Overall, 83.5% of the study population had attained normal height (HSDS > -2) over the duration of GH treatment (mean duration, years: GHD 3.9; ISS 3.9; TS 4.4). It is possible that the results were biased towards good responders, since poor responders may have withdrawn from the study or stopped GH treatment before reaching adulthood. The proportion of TS patients who attained normal height (66%) in this study was smaller compared with those of the other indications studied (79-88%). The reasons for this finding are unclear; nevertheless, this result is consistent with findings from a previous analysis of patients with TS from the ANSWER program [21].

At baseline, the mean ± SD overall age was 14.0 ± 2.1 years (12.8 years in females and 14.7 years in males), which was nearly 4 years older than the mean age of pediatric patients typically enrolled in the ANSWER Program (10.5 ± 4.0 years). Consequently, this analysis of patients who attained NAH was skewed toward children who were older at the time of GHT initiation, and thus may not be representative of the overall population in the ANSWER Program. The short duration of the Program (initiated in 2002) could have contributed to a selection bias towards older patients since a substantial proportion of enrolled patients would not have reached 18 years as of the cut-off date for this current analysis. The older age at baseline in this subgroup may have resulted in an underestimation of the overall adult height achievable by all patients treated with GH. Previous studies from the ANSWER Program have shown that earlier age at GH treatment initiation is associated with greater height gain [12, 13], reinforcing the importance of early referral for GHT in appropriate patients when growth failure or short stature is recognized. From the results of this analysis of the changes in mean HSDS and corrected HSDS from baseline to NAH, it would appear that these patients did experience a beneficial response to GH treatment. Sex was confirmed to be a statistically significant factor for NAH outcome, possibly due to the difference in referral age between sexes (females were younger than males). The significant negative correlation observed between ATS and near-adult HSDS for both sex subgroups indicates that, although a beneficial effect may be observed when GH therapy is initiated at an older childhood age, greater near-adult HSDS can be expected if these patients had started GHT at a younger age. Healthcare providers should therefore be aware of the potential limitations of this treatment among patients of a relatively advanced age. The correlation results of this study in relatively older children are consistent with recently published analyses of GH treatment on increasing linear growth to NAH or final height in younger children with IGHD or MPHD from other large patient databases [22‐26]. This has also been shown in patients from other diagnostic groups, such as Noonan syndrome. Recently published ANSWER Program data on patients with Noonan syndrome with a mean referral age of 9 years showed that baseline age was negatively correlated with ΔHSDS after 1 or 2 years of GH treatment, indicating that older baseline ATS resulted in lower ΔHSDS [27‐29]. In the present study, analysis of the relationship between age at initiation of GHT and near-adult HSDS in patients with GHD, ISS, and TS showed significant negative correlations, which is consistent with previous reports [13, 15, 28]. The results of the current analysis are among the first to describe the long-term effects of GH treatment, including an increase in HSDS at the time of NAH, in children who initiated GHT at an older age.

The effect of GH treatment on the BA/CA ratio was relatively small, and BA/CA remained <1 for patients across all indications over time. This suggests that, although GH treatment exerted positive effects on linear growth in this older pediatric patient population, there was a minimal effect on bone maturation relative to CA. The effect of pubertal stage in these patients was not analyzed and cannot be ruled out as a contributing factor. Also, the mean BA/CA ratio in the range of 0.82 to 0.88 indicates that, on average, a small amount of additional growth would have remained in some patients, consistent with near rather than final peak adult height.

Mean IGF-I SDS increased in patients with GHD, ISS, and TS at year 1 of GH treatment, remaining relatively stable thereafter. The difference in mean IGF-I SDS (GHD - ISS) was only -0.45 at baseline and 0.12 at NAH. From baseline to NAH, BMI SDS showed variable increases, but remained stable, during GHT among the diagnostic groups. This occurred in parallel with sustained increases in HSDS and corrected HSDS in each group during GH treatment. The negative mean BMI SDS at baseline observed in patients with ISS in the present study might suggest a degree of relative nutritional deficiency in these individuals that improved over time. Recent results published from the ANSWER Program registry indicated that for patients with GHD, a higher baseline BMI was positively correlated with the response to GH [13]. The association of BMI on GH response may reflect, in part, the importance of nutrition or nutritional status for optimizing outcomes in GH-treated patients [12, 13, 30, 31].

In conclusion, long-term GH therapy in GH-naïve pediatric patients with GHD, ISS, or TS resulted in increases in HSDS and corrected HSDS from baseline to NAH despite their older age at the start of GHT. Although the results of this study demonstrate good NAH outcomes in children with short stature or growth failure who initiated GH treatment at an older age, the negative correlations of ATS with near-adult HSDS when analyzed by sex indicate that earlier recognition and referral for growth failure or short stature, and efforts to start GH treatment at a younger age, would most likely have resulted in greater gains in height and an increase in adult HSDS.