Intervention: The children were randomized to either a standard (43 g/kg d) or individualized ( g/kg d) GH dose.

Transcription

1 ORIGINAL ARTICLE Endocrine Care Growth Hormone (GH) Dosing during Catch-Up Growth Guided by Individual Responsiveness Decreases Growth Response Variability in Prepubertal Children with GH Deficiency or Idiopathic Short Stature Berit Kriström, A. Stefan Aronson, Jovanna Dahlgren, Jan Gustafsson, Maria Halldin, Sten A. Ivarsson, Nils-Östen Nilsson, Johan Svensson, Torsten Tuvemo, and Kerstin Albertsson-Wikland Göteborg Pediatric Growth Research Center (B.K., J.D., K.A.-W.), Department of Pediatrics, Institute of Clinical Science, The Sahlgrenska Academy at University of Gothenburg, SE Göteborg, Sweden; Department of Clinical Science, Pediatrics (B.K.), Umeå University, SE Umeå, Sweden; Department of Pediatrics (A.S.A., N.-O.N.), The Central County Hospital of Halmstad, SE Halmstad, Sweden; Department of Women s and Children s Health (J.G., M.H., T.T.), Uppsala University, SE Uppsala, Sweden; and Department of Clinical Sciences (S.A.I., J.S.), University Hospital Malmö, Lund University, SE Malmö, Sweden Context: Weight-based GH dosing results in a wide variation in growth response in children with GH deficiency (GHD) or idiopathic short stature (ISS). Objective: The hypothesis tested was whether individualized GH doses, based on variation in GH responsiveness estimated by a prediction model, reduced variability in growth response around a set height target compared with a standardized weight-based dose. Setting: A total of 153 short prepubertal children diagnosed with isolated GHD or ISS (n 43) and at least 1 SD score (SDS) below midparental height SDS (MPH SDS ) were included in this 2-yr multicenter study. Intervention: The children were randomized to either a standard (43 g/kg d) or individualized ( g/kg d) GH dose. Main Outcome Measure: We measured the deviation of height SDS from individual MPH SDS (diffmph SDS ). The primary endpoint was the difference in the range of diffmph SDS between the two groups. Results: The diffmph SDS range was reduced by 32% in the individualized-dose group relative to the standard-dose group (P 0.003), whereas the mean diffmph SDS was equal: and , respectively. Gain in height SDS 0 2 yr was equal for the GH-deficient and ISS groups: and , respectively, when ISS was classified on the basis of maximum GH peak on the arginine-insulin tolerance test or 24-h profile. Conclusion: Individualized GH doses during catch-up growth significantly reduce the proportion of unexpectedly good and poor responders around a predefined individual growth target and result in equal growth responses in children with GHD and ISS. (J Clin Endocrinol Metab 94: , 2009) ISSN Print X ISSN Online Printed in U.S.A. Copyright 2009 by The Endocrine Society doi: /jc Received July 14, Accepted October 30, First Published Online November 11, 2008 Abbreviations: AITT, Arginine-insulin tolerance test; diffmph SDS, deviation of height SDS from individual MPH SDS ; GHD, GH deficiency; GH max, GH maximum peak; HbA1c, glycosylated hemoglobin; ISS, idiopathic short stature; ITT, intention-to-treat; MPH SDS, midparental height SDS; NS, not significant; PP, per-protocol; SDS, SD score; SGA, small for gestational age. J Clin Endocrinol Metab, February 2009, 94(2): jcem.endojournals.org 483

2 484 Kriström et al. Individualized GH Dosing for Catch-Up Growth J Clin Endocrinol Metab, February 2009, 94(2): There is considerable variability in the growth response observed after GH treatment in short children, regardless of whether the cause of short stature is GH deficiency (GHD) or idiopathic short stature (ISS) (1 4). Although there is evidence of a dose-dependent effect of GH, a wide variation in growth response is seen even among children receiving the same dose of GH (5 8). Studies of the relationship between GH secretion and growth in children with normal or low growth rates have shown a positive correlation between the level of GH secretion and height (9). Despite this relationship, a broad range of serum GH levels was observed in children with similar growth rates, suggesting that there are differences in individual responsiveness to GH. Thus, GH secretion and GH responsiveness are both important for growth. It is believed that endogenous factors, such as underlying genetic polymorphisms, genetic abnormalities (10), or prenatal/early postnatal changes in metabolism may permanently influence the balance between GH secretion and responsiveness. Evidence-based models for predicting growth in response to GH treatment have been constructed using data from short children treated with GH (11 19). Models are available for GHdeficient and non-gh-deficient children, including those with ISS or born small for gestational age (SGA). Such models provide an indirect measure of individual responsiveness to GH and allow predictions to be made regarding the GH dose needed to attain a defined height target. This makes it possible to individualize GH treatment. The aim of this study was to determine whether catch-up growth in individual children with isolated GHD or ISS could be targeted by using individualized GH doses rather than a standard dose. The height target selected was midparental height SD score (MPH SDS ) for the individual. Catch-up growth was assessed by evaluating the difference between current height SDS and MPH SDS (diffmph SDS ) after 2 yr of treatment. Prediction modelbased estimates of individual responsiveness to GH were used to determine the individual GH dose within a set dose range ( g/kg d). Subjects and Methods Ethical consideration The protocol was approved by the ethical boards of the Universities of Göteborg (for Göteborg and Halmstad), Umeå, Uppsala, and Lund and the Medical Product Agency of Sweden. Written informed consent was obtained from all parents and from children if possible. The randomized, prospective, controlled, open, multicenter clinical trial was performed in accordance with the Declaration of Helsinki and Good Clinical Practice (GCP) in 153 children from five pediatric endocrinology units in Sweden Göteborg (n 79), Umeå (n 34), Uppsala (n 17), Malmö (n 13) and Halmstad (n 10)]. Subjects Comparison between standard-dose and individualizeddose groups according to the protocol The intention-to-treat (ITT) population. This population included 153 children with the following inclusion criteria: prepubertal; girls 3 10 yr of age and boys 3 11 yr at study start; height SDS no greater than 2 (all ISS children and the majority of GH-deficient children fulfilled this criterion) or growth velocity no greater than 1 SDS (a few children with severe GHD were included on this criterion); 1 SDS (20) or more below MPH SDS (21); and born at a gestational age above 30 wk. In addition, the study required access to length/height and weight measurements from birth, at 1 ( 3 months) and 2 yr ( 6 months) of age, and 1 yr ( 3 months) before the start of treatment, with at least two additional stadiometer measurements during the pretreatment year. Exclusion criteria were chronic illness or a clinical syndrome, disproportionate body measurements, catch-up growth during the pretreatment year, or an MPH SDS greater than 1.5. TABLE 1. Characteristics of the PP population at the start and end of the study Individualized dose n 87 (27 girls, 60 boys) Standard dose n 41 (11 girls, 30 boys) Variables Mean SD Median Min: Max Mean SD Median Min: Max At birth Gestational age (wk) : : 42.0 Length SDS : : 1.63 Weight SDS : : 1.28 At study start Age (yr) : : 10.9 Height SDS : : 1.77 BMI SDS : : 2.11 MPH SDS : : 0.14 diffmph SDS : : 0.86 GH max AITT (mu/liter) : : 42.6 GH max 24-h profile (mu/liter) : : 50.7 IGF-I SDS : : 0.76 Bone age delay TW II (yr) : : 0.90 After 2-yr GH diffmph SDS a : a : 0.92 Height SDS : : 0.20 Bone age delay TW II (yr) : : 1.70 There were no significant differences in any of the pretreatment variables between the groups receiving the individualized or standard doses. Fisher s exact test was used to assess significance for gender. t-tests were used for all other variables. At the end of the study, there was a significant difference ( a ) between the randomization groups in the distribution of diffmph SDS. BMI, Body mass index; TW II, Tanner Whitehouse II. a P

3 J Clin Endocrinol Metab, February 2009, 94(2): jcem.endojournals.org 485 Diagnostic classification. Diagnosis was made according to the arginineinsulin tolerance test (AITT) result, which was the gold standard at the time of this study. Of the 153 children, 110 were GH-deficient and 43 had ISS according to the GH maximum peak (GH max ) AITT. Inclusion of children with ISS was restricted to 25% of the total study group, of whom 30% had a height SDS between 2.0 and 2.5. All children had a birth length/weight above 2.5 SDS, but 13 of the children were born SGA (22) with a birth length/weight SDS between 2.0 and 2.5 SDS (n 11 in the individualized-dose group). Randomization. The children were randomly assigned to either a standard or individualized GH dose in a 1:2 ratio. A minimization procedure was applied to the following variables: age, gender, and auxological measurements. Variables included: weight SDS at birth; height SDS 1yr before GH start; height SDS at start; channel-parallel growth the year before start (yes/no); diffmph SDS ; the predicted change in height SDS during the first year of GH 33 g/kg d; the GH max AITT; and GH max 24 h. The per-protocol (PP) population. After 2 yr of GH treatment, 128 children (38 girls) had completed the protocol and remained prepubertal, constituting the PP population (90 GHD, 38 ISS; 11 were born SGA). During the study, 16 children entered puberty (15 during the second treatment year), five were withdrawn due to inclusion error, and four were excluded due to low compliance. Excluded children were equally distributed between the randomization groups according to gender and reason for exclusion. Characteristics at study start are given in Table 1 and Fig. 1. Comparison between children with GHD and ISS The GH status of the 117 non-sga children who completed the 2-yr protocol was reassessed. A GH max of at least 32 mu/liter ( old 10 g/ liter ) on either GH max AITT or GH max 24-h profile resulted in a classification of ISS. Reclassification resulted in 37 children diagnosed with GHD and 80 with ISS; thus, in about one third of children the GH max AITT was falsely low compared with the GH max 24-h profile (23). Prediction model and dose selection GH responsiveness was assessed at the start of the study using a validated prediction model (11). The model was constructed on data from 269 short prepubertal children with GH secretion ranging from very low to normal. The model predicts the 2-yr growth response to a GH dose of 33 g/kg d, using data available from before the start of treatment (parental heights, auxological variables from the child including age, gender, weight, and height at GH start and at 1 yr before start, birth weight, gestational age, height and weight at 1 yr of age, height at 2 yr of age, and GH max 24 h) in an algorithm giving the predicted response with low error; the 2-yr prediction error was 0.28 SDS. The predicted growth response is an indirect measure of the degree of GH responsiveness and was used as a randomization variable. In the individualizeddose group, it was used in dose selection within the interval g/kg d ( IU/kg d) together with diffmph SDS. The lower the GH responsiveness and the greater the relative distance of current height from MPH SDS, the higher the dose for the individual child (Table 2). The dose for the standard-dose group was set to 43 g/kg d (0.13 IU/kg d), with the aim of achieving equal mean doses between the groups, based 200 Dose fixed individual 200 Dose fixed individual Height (cm) boys girls Age (years) FIG. 1. Height and age according to gender and randomization group in the PP population at the study start, plotted on the childhood (prepubertal) component of the Swedish reference growth chart (20). The curves represent the mean (continuous line) and 1, 2, and 3 SDS (dotted lines). Each curve starts with the prepubertal (childhood) component of growth and diverges as the total growth with addition of the pubertal growth component is visualized in the figure. Only prepubertal children are included in the study; thus SDS is calculated according to the prepubertal component of the growth curve at end of the study.

4 486 Kriström et al. Individualized GH Dosing for Catch-Up Growth J Clin Endocrinol Metab, February 2009, 94(2): TABLE 2. Treatment schedule for selection of dose in the group with individualized GH doses ( g/kg d) with use of the two variables predicted diffmph SDS and growth response after 2 yr of treatment Predicted height SDS after 2 yr of Predicted difference in MPH SDS after 2 yr of GH treatment with 33 g/kg d GH treatment with 33 g/kg d < to to to 0.5 > to on data from children who had started GH treatment in Sweden not long before the start of the study. However, because fewer children with severe GHD were enrolled than expected, the mean dose was higher for the individualized-dose group than for the standard-dose group: 50 g/ kg d (0.15 IU/kg d) vs. 43 g/kg d (0.13 IU/kg d) (P ). The distribution of doses in the individualized-dose group was 2% on 17 g, 21% on 33 g, 8% on 40 g, 20% on 50 g, 11% on 66 g, and 5% on 100 g. The remaining 33% was the control group receiving the standard dose, 43 g. Methods Growth references The childhood component of the Swedish population-based growth reference values (20) was used for the height-related inclusion criteria for the children and to assess changes in height, weight, and body mass index longitudinally (24). Reference standards from Niklasson et al. (22) were used for SDS at birth. Parental heights were measured with a Harpenden stadiometer, using the same technique as for the children and expressed in SDS (20, 21). Bone age was assessed by one person at study start and thereafter yearly according to Tanner-Whitehouse II. Hormone assays GH (25) and IGF-I (26) analyses were performed at the Goteborg Pediatric Growth Research Center laboratory (Swedac accredited no. 1899). The cutoff value for the diagnosis of GHD was 32 mu/liter using a polyclonal assay and World Health Organization International Reference Preparation 80/505, corresponding to the old 10 g/liter cutoff value (25). Fasting insulin, fasting glucose, and glycosylated hemoglobin (HbA1c) levels were measured in serum samples collected at baseline and after 1 and 2 yr of treatment. If the result required an intervention according to the protocol, a second sample was analyzed. Analyses were performed at accredited university hospital laboratories. Safety IGF-I, fasting glucose, fasting insulin, and HbA1c were monitored for safety reasons. If IGF-I SDS values were greater than 3 SDS in two consecutive samples in the individualized-dose group, the dose was stepwise reduced until the IGF-I SDS was below 3 SDS. Statistical methods Data were entered into a validated central database and checked against original case record forms according to GCP and were monitored externally. Statistical analyses were performed using SAS version 8.2. All tests were two-tailed and were conducted at a 5% significance level. Equality of variance was analyzed using the F-test. Student s t test was used for comparison between groups. Fisher s exact test was used for comparisons of proportions between groups. Wilcoxon signed rank test was used to analyze changes within groups. Comparison between standard-dose and individualized-dose groups according to the protocol. The PP population analysis included those children (n 128) who had completed the protocol and remained prepubertal. For all ITT analyses, the last-observation-carried-forward method was applied. Safety assessment was carried out in the ITT population (n 153). The primary endpoint was the diffmph SDS from each of the two randomization groups. Because this variable was found to be normally distributed, the two variances were compared with an F-test, using the ratio of the variance estimates: F s 1 2 /s 2 2. The F-test P value was used for difference in the range of values (SD) except for absolute diffmph SDS to 0 at 2 yr, where differences between means were tested. The hypothesis tested was that individualizing the GH dose would reduce the range of the variable diffmph SDS by 30% relative to the control group. The limit of 30% was chosen before study start as representing a clinically relevant improvement. Height SDS and diffmph SDS were found to be normally distributed. Insulin was not normally distributed, and for comparison of change in insulin levels between the two groups Mann-Whitney U test was used. Comparisons between children with GHD and ISS. After the children included in the PP analysis were reclassified as having GHD or ISS on the basis of their GH max 24-h profile and GH max AITT results, there was no statistical relevance of the randomization. The ISS and GHD groups remained equivalent regarding baseline factors, and there were equal distributions of standard (33%) or individualized (67%) doses in the two groups. Results Comparison between standard-dose and individualized-dose groups Range in growth responses and comparison with parental heights The primary endpoint of the 2-yr study was the range of the distribution of the diffmph SDS because the goal was to avoid overtreating or undertreating any child. The range of the distribution was 32% narrower in the individualized-dose group than in the standard-dose group: the absolute distance between the maximum and minimum values of diffmph SDS was 2.25 SDS for the individualized-dose group and 3.36 SDS for the standarddose group. Mean values ( SD) were for the individualized group and for the standard-dose group, which is nonsignificant for the means but P for the difference in SD (Fig. 2). Comparison with population heights Mean height SDS after 2 yr was for the individualized-dose group and not significant (NS)] for the standard-dose group, compared with baseline height SDS of and (NS), respectively. Thus, the mean gain in height SDS was 1.32 in both groups.

5 J Clin Endocrinol Metab, February 2009, 94(2): jcem.endojournals.org 487 P<0.003 ns FIG. 2. The distributions of values at the end of study for the difference in current height SDS to midparental height SDS (diffmph SDS ) according to randomization group in the PP population. Mean value was for the individualized-dose group and for the standard-dose group (P for the difference in SD). The range in distribution around the treatment goal, the individual MPH SDS, is significantly narrower (32%) in the group receiving an individualized GH dose compared with the group receiving a standard dose. The boxes represent 50% of the study group, with the line indicating the median (the plus sign indicating the mean); the whiskers indicate the maximum and minimum. Influence of including data from children born SGA When the PP population analysis was rerun excluding data from SGA children, the range of the distribution of values for the diffmph SDS in the individualized-dose group was 34% narrower than for the standard-dose group (P 0.003). The ITT population The range of the distribution of the values for the diffmph SDS in the ITT and PP populations was not significantly different. Bone age There were no significant differences between the randomization groups in bone age delay at study start or in the change in bone age after 2 yr of treatment (Table 1). Age at treatment start, gender, and GH dose did not influence bone maturation. Comparison between children with GHD and ISS After the reclassification of the non-sga PP population, 80 children were found to have ISS, and 37 remained GH-deficient. After 2 yr of treatment, there was no difference in the gain in height SDS between children with GHD and ISS (Fig. 3). Mean gain in height SDS from baseline was for the GHdeficient group and for the ISS group. Evaluation of the GH doses given according to the protocol when individual GH responsiveness was accounted for in the majority of children (67%) revealed that the mean daily GH dose was FIG. 3. Gain in height SDS after 2 yr of GH treatment after reclassification of children in the study as having ISS (mean, 1.31 SDS) or GHD (mean, 1.36 SDS) according to GH max level above cutoff in the AITT or during a spontaneous 24-h GH profile. The groups include children treated with both standard and individualized doses but were equal regarding baseline factors and distributions of standard (33%) and individualized (67%) doses. 40 g/kg for GH-deficient children and 50 g/kg for children with ISS (P 0.001). Safety Serious adverse events There were no reports of malignancy, slipped caput femoral epiphysis, intracranial hypertension, or any other serious adverse events related to GH. Adverse events Adverse event rates did not differ between the two groups randomized as part of the ITT analysis. Almost all patients reported at least one adverse event during the 2-yr study period. These were mostly minor and temporary and were not thought to be related to the study drug, such as infectious disease or minor trauma. Shortly after the start of treatment, many children experienced an increased appetite, and some parents reported transient changes in temper in their children. Glucose homeostasis Fasting blood glucose, fasting insulin, and HbA1c were normal at baseline. Although fasting blood glucose and HbA1c did not change significantly during treatment, fasting insulin levels increased significantly (P 0.001), but without significant difference between the two groups, and remained within normal limits. In one obese boy who received dietary information, fasting insulin level above 30 mu/liter was seen, whereas fasting glucose and HbA1c levels remained normal. IGF-I levels IGF-I levels above 3 SDS were found in nine children in the individualized-dose group after 1 yr in three girls (doses of 33, 100, and 100 g), and after 2 yr in four girls (doses of 33, 66, 100, and 100 g) and two boys (doses of 33 and 66 g)], and in five

6 488 Kriström et al. Individualized GH Dosing for Catch-Up Growth J Clin Endocrinol Metab, February 2009, 94(2): Within individual dose group 5 Randomization groups IGF-I SDS Individual dose Standard dose FIG. 4. IGF-I levels in SDS at the 2-yr visit. Left, Within the individualized-dose group, divided according to the GH-dose steps ( g/kg d). Right, According to the randomization groups. No significant differences or dose trends were found within the individualized-dose group or between the randomization groups. children in the standard-dose group after 1 yr in two girls and two boys and after 2 yr in one boy (NS)]. IGF-I levels were not significantly different comparing the two randomization groups (mean level, and SDS, respectively). When the group with individualized doses was divided according to dose, there was considerable overlap in IGF-I levels (Fig. 4). Discussion The proportion of unexpectedly good and poor responders to GH treatment was significantly reduced when the dose given was guided by an estimate of individual GH responsiveness obtained from a model for prediction of growth response to GH treatment. This is shown by the range of the variable diffmph SDS being significantly reduced by 32% in the group of children randomized to receive an individualized dose compared with the group that received a conventional weight-based dose. This means that fewer children are undertreated (more children reach a height SDS closer to their MPH SDS ) or overtreated (fewer children reach above MPH SDS ). This study also resulted in equal growth responses for children diagnosed with GHD and ISS. Not surprisingly, GH responsiveness was lower in the ISS group; the mean dose for the group was 50 g/kg d compared with 40 g/kg d for the GHdeficient group. Genetic abnormalities or polymorphisms have been studied because they are believed to be responsible for variations in GH secretion and GH sensitivity associated with different diagnoses of short stature, including GHD and ISS (10, 27 30). Validated prediction models that indirectly take account of many of these factors have been developed to select the appropriate dose of GH for individual children (11 19). The models, based on large groups of children, integrate individual auxological measurements and GH secretion and responsiveness. Validation of such models in different cohorts of children has demonstrated their usefulness in predicting GH response in clinical practice, although some of the variation remains unexplained. The prediction model used in the present study was the best available when the study was designed. Because children with GHD or ISS in Sweden have been treated with a standard daily dose of GH of 33 g/kg since 1986, a dose effect could not be included when the model was constructed. The GH dose range used in the present study ( g/ kg d) was selected based on previously published efficacy and safety studies. The highest dose had been used in randomized studies without adverse effects (6, 31, 32), a finding confirmed later (6). Dose-related adverse events were not seen in the present study. One reason for this may be that higher GH doses were reserved for children who, through the use of the prediction model, were found to have low GH responsiveness (11). Safety in patients receiving GH replacement therapy has previously been proposed to be a matter of responsiveness (33). In the present study, the GH dose was reduced if IGF-I SDS exceeded 3 due to the theoretical concern of long-term negative effects. No such effects have been reported, but long-term follow-up is justified. The aim of the study was to show that by using an estimate of GH responsiveness for dosing, it was possible to reach a set target for growth. The target set was the MPH SDS, which has previously been found to be an important factor in explaining the variance in growth response in GH-treated children with GHD or ISS (1). MPH SDS probably reflects the genetic potential of the child and is therefore an appropriate measure against which to evaluate the effect of treatment, except in families with extreme tall or short

7 J Clin Endocrinol Metab, February 2009, 94(2): jcem.endojournals.org 489 stature. The 2-yr study duration was chosen based on the time taken to achieve catch-up growth in other studies using an optimal treatment regimen (34, 35). However, some children in our study group needed a third year for their height to reach MPH SDS (data not shown). The observed 2-yr gain in height SDS (mean, 1.30 SDS) for the standard-dose group was in accordance with results from a previous study in more than 300 children with GHD and ISS treated with a GH dose of 33 g/kg d (36). In the present study, 25% of the children in both groups reached a height SDS above their MPH SDS. The median difference to MPH SDS was approximately 0.4 SDS. Thus, to reach their MPH SDS, higher GH doses would have been needed for many children. However, when the study was designed, the maximum dose was restricted to 100 g/kg d, a decision that had an impact on the standard dose. In this study, we confirm that in children with GHD or ISS, the velocity and magnitude of catch-up growth is dependent on both GH dose and GH responsiveness. Similar results were shown by Cohen et al. (6) in GH-deficient children on GH doses of 25, 50, or 100 g/kg d, but without using individual growth targets or taking GH responsiveness into account. There was great variation in growth responses in all three randomization groups, showing the wide range of GH responsiveness. Similar differences in responsiveness have been shown for children with ISS (2). In an IGF-I titration study performed in short prepubertal children with low IGF-I levels, the range in individual GH sensitivity was demonstrated by the broad range of GH doses needed: g/kg d for an IGF-I level of SDS and g/kg d for an IGF-I level of SDS (7). The separation between the diagnoses GHD and ISS is a matter of definition, with an arbitrary cutoff between the groups (30, 37). Diagnosis can therefore change when an additional test is applied. In this study, one third of the children were transferred from the GH-deficient group to the ISS group when the result from a 24-h spontaneous GH profile was considered. Using the GH max from the profile has the advantage that possible refractoriness of the somatotrophs during provocation testing (23) does not confound the data. Importantly, the prediction model used in the study is validated for both children with GHD and children with ISS (11). Thus, these diagnostic issues do not interfere with study reliability. Although new biomarkers are being found (38), the GH max 24-h profile is so far the most informative GH-related variable to use in a prediction model (11) in addition to information on early growth patterns (39). Many children with short stature present late for medical advice. The response to GH treatment decreases with age, and the highest growth response is obtained during the first year of treatment. Therefore, it is important that outcome is evaluated on an individual basis and that treatment is started at a dose high enough to have a substantial effect on growth, but low enough not to jeopardize long-term health. Doses of GH need to be individualized, as is the case for other hormones. For some children, the dose clearly needs to be higher than the standard doses currently used, whereas for others the dose may be reduced while still achieving the desired height target. Conclusion Irrespective of the diagnosis of GHD or ISS, most children respond well to GH treatment. The challenge is to identify those individuals who will respond extremely well or badly to GH at an early age and to manage their treatment in an efficient, safe, and cost-effective manner. In many children it is possible to collect pretreatment information that can be used in a prediction model to aid medical decision-making. The present study shows that prediction models can be used not only to decide whether or not to treat, but also for individualized dose selection within the range of g/kg d. Individualized treatment was found to be equally effective in promoting catch-up growth in children with GHD and ISS because the dose could be adjusted according to individual responsiveness. Thus, we propose that individualized GH-dosing should be used for GH treatment in short prepubertal children with either GHD or ISS. Acknowledgments The authors thank participating children and families, the national study team led by Carola Pfeiffer-Mosesson, regional study teams, and colleagues at University Pediatric Endocrinology centers and at county hospitals responsible for caring for the children close to home. We also thank the Goteborg Pediatric Growth Research Center laboratory led by Irène Leonardsson. The authors are grateful for the statistical support of Nils- Gunnar Pehrsson and his colleagues Gunnar Ekeroth, Fredrik Parknäs, and Aldina Pivodic; for statistical support and discussions with Andreas F. M. Nierop, Anders Lindberg, and Sten Rosberg; and for valuable discussion with Rolf Gunnarsson, Pharmacia, before the study started. Address all correspondence and requests for reprints to: Kerstin Albertsson Wikland, GP-GRC/Växthuset, The Queen Silvia Children s Hospital, SE Göteborg, Sweden. kerstin.albertsson-wikland@ pediat.gu.se. This investigator-initiated study (TRN ) was funded by Pharmacia/Pfizer and by grants from the Swedish Research Council (no. 7509), Sahlgrenska University Hospital (ALF), and West Sweden Region (VGR). Disclosure Information: Research grants were received by K.A.-W., T.T., and J.G. Consulting fee was obtained by B.K., J.D., J.G., and T.T. Stock options were held by A.S.A., J.G., and J.D.M.H., S.A.I., N.-Ö.N., and J.S. have nothing to declare. References 1. Kristrom B, Karlberg J, Albertsson-Wikland K 1995 Prediction of the growth response of short prepubertal children treated with growth hormone. Swedish Paediatric Study Group for GH treatment. Acta Paediatr 84: Leschek EW, Rose SR, Yanovski JA, Troendle JF, Quigley CA, Chipman JJ, Crowe BJ, Ross JL, Cassorla FG, Blum WF, Cutler Jr GB, Baron J 2004 Effect of growth hormone treatment on adult height in peripubertal children with idiopathic short stature: a randomized, double-blind, placebo-controlled trial. J Clin Endocrinol Metab 89: Moore KC, Donaldson DL, Ideus PL, Gifford RA, Moore WV 1993 Clinical diagnoses of children with extremely short stature and their response to growth hormone. J Pediatr 122: Tanner JM, Hughes PC, Whitehouse RH 1977 Comparative rapidity of response of height, limb muscle and limb fat to treatment with human growth hormone in patients with and without growth hormone deficiency. Acta Endocrinol (Copenh) 84: Albertsson-Wikland K, Aronson S, Gustafsson J, Hagenäs L, Häger A, Ivarsson S, Jonsson B, Kriström B, Marcus C, Nilsson K-O, Ritzén M, Tuvemo T, Westphal

8 490 Kriström et al. Individualized GH Dosing for Catch-Up Growth J Clin Endocrinol Metab, February 2009, 94(2): O, Åman J 2008 Dose-dependent effect of growth hormone (GH) on final height in non-ghd short children. J Clin Endocrinol Metab 93: Cohen P, Bright GM, Rogol AD, Kappelgaard AM, Rosenfeld RG 2002 Effects of dose and gender on the growth and growth factor response to GH in GHdeficient children: implications for efficacy and safety. J Clin Endocrinol Metab 87: Cohen P, Rogol AD, Howard CP, Bright GM, Kappelgaard AM, Rosenfeld RG 2007 Insulin growth factor-based dosing of growth hormone therapy in children: a randomized, controlled study. J Clin Endocrinol Metab 92: Wit JM, Rekers-Mombarg LT, Cutler GB, Crowe B, Beck TJ, Roberts K, Gill A, Chaussain JL, Frisch H, Yturriaga R, Attanasio AF 2005 Growth hormone (GH) treatment to final height in children with idiopathic short stature: evidence for a dose effect. J Pediatr 146: Albertsson-Wikland K, Rosberg S 1988 Analyses of 24-hour growth hormone profiles in children: relation to growth. J Clin Endocrinol Metab 67: Rosenfeld RG 2005 The molecular basis of idiopathic short stature. Growth Horm IGF Res 15 Suppl A:S3 S5 11. Albertsson-Wikland K, Kristrom B, Rosberg S, Svensson B, Nierop AF 2000 Validated multivariate models predicting the growth response to GH treatment in individual short children with a broad range in GH secretion capacities. Pediatr Res 48: Dahlgren J, Kriström B, Niklasson A, Nierop AF, Rosberg S, Albertsson- Wikland K 2007 Models predicting the growth response to growth hormone treatment in short children independent of GH status, birth size and gestational age. BMC Med Inform Decis Mak 7: de Ridder MA, Stijnen T, Hokken-Koelega AC 2003 Validation and calibration of the Kabi Pharmacia International Growth Study prediction model for children with idiopathic growth hormone deficiency. J Clin Endocrinol Metab 88: de Ridder MA, Stijnen T, Hokken-Koelega AC 2007 Prediction of adult height in growth-hormone-treated children with growth hormone deficiency. J Clin Endocrinol Metab 92: de Ridder MA, Stijnen T, Hokken-Koelega AC 2008 Prediction model for adult height of small for gestational age children at the start of growth hormone treatment. J Clin Endocrinol Metab 93: Kristrom B, Jansson C, Rosberg S, Albertsson-Wikland K 1997 Growth response to growth hormone (GH) treatment relates to serum insulin-like growth factor I (IGF-I) and IGF-binding protein-3 in short children with various GH secretion capacities. Swedish Study Group for Growth Hormone Treatment. J Clin Endocrinol Metab 82: Land C, Blum WF, Shavrikova E, Kloeckner K, Stabrey A, Schoenau E 2007 Predicting the growth response to growth hormone (GH) treatment in prepubertal and pubertal children with isolated GH deficiency model validation in an observational setting (GeNeSIS). J Pediatr Endocrinol Metab 20: Ranke MB, Lindberg A, Chatelain P, Wilton P, Cutfield W, Albertsson- Wikland K, Price DA 1999 Derivation and validation of a mathematical model for predicting the response to exogenous recombinant human growth hormone (GH) in prepubertal children with idiopathic GH deficiency. KIGS International Board. Kabi Pharmacia International Growth Study. J Clin Endocrinol Metab 84: Ranke MB, Lindberg A, Cowell CT, Albertsson-Wikland K, Reiter EO, Wilton P, Price DA 2003 Prediction of response to growth hormone treatment in short children born small for gestational age: analysis of data from KIGS (Pharmacia International Growth Database). J Clin Endocrinol Metab 88: Albertsson-Wikland K, Luo ZC, Niklasson A, Karlberg J 2002 Swedish population-based longitudinal reference values from birth to 18 years of age for height, weight and head circumference. Acta Paediatr 91: Karlberg P, Taranger J 1976 The somatic development of children in a Swedish urban community. Acta Paediatr Scand Suppl 258: Niklasson A, Ericson A, Fryer JG, Karlberg J, Lawrence C, Karlberg P 1991 An update of the Swedish reference standards for weight, length and head circumference at birth for given gestational age ( ). Acta Paediatr Scand 80: Gelander L, Albertsson-Wikland K 1989 Growth hormone (GH) release after administration of GH-releasing hormone in relation to endogenous 24-h GH secretion in short children. J Endocrinol 122: Karlberg J, Luo ZC, Albertsson-Wikland K 2001 Body mass index reference values (mean and SD) for Swedish children. Acta Paediatr 90: Jansson C, Boguszewski C, Rosberg S, Carlsson L, Albertsson-Wikland K 1997 Growth hormone (GH) assays: influence of standard preparations, GH isoforms, assay characteristics, and GH-binding protein. Clin Chem 43: Löfqvist C, Andersson E, Gelander L, Rosberg S, Blum WF, Albertsson- Wikland K 2001 Reference values for IGF-I throughout childhood and adolescence: a model that accounts simultaneously for the effect of gender, age, and puberty. J Clin Endocrinol Metab 86: Domene HM, Bengolea SV, Martinez AS, Ropelato MG, Pennisi P, Scaglia P, Heinrich JJ, Jasper HG 2004 Deficiency of the circulating insulin-like growth factor system associated with inactivation of the acid-labile subunit gene. N Engl J Med 350: Dos Santos C, Essioux L, Teinturier C, Tauber M, Goffin V, Bougneres P 2004 A common polymorphism of the growth hormone receptor is associated with increased responsiveness to growth hormone. Nat Genet 36: Salerno M, Balestrieri B, Matrecano E, Officioso A, Rosenfeld RG, Di Maio S, Fimiani G, Ursini MV, Pignata C 2001 Abnormal GH receptor signaling in children with idiopathic short stature. J Clin Endocrinol Metab 86: Wit JM, Clayton PE, Rogol AD, Savage MO, Saenger PH, Cohen P 2008 Idiopathic short stature: definition, epidemiology, and diagnostic evaluation. Growth Horm IGF Res 18: de Zegher F, Du Caju MV, Heinrichs C, Maes M, De Schepper J, Craen M, Vanweser K, Malvaux P, Rosenfeld RG 1999 Early, discontinuous, high dose growth hormone treatment to normalize height and weight of short children born small for gestational age: results over 6 years. J Clin Endocrinol Metab 84: Kamp GA, Wit JM 1998 High-dose growth hormone therapy in idiopathic short stature. Horm Res 49(Suppl 2): Wilton P 2001 Safety in growth hormone replacement therapy: a matter of varied responsiveness? Horm Res 55(Suppl 2): Boersma B, Rikken B, Wit JM 1995 Catch-up growth in early treated patients with growth hormone deficiency. Dutch Growth Hormone Working Group. Arch Dis Child 72: Ranke MB, Schwarze CP, Mohnike K, von Muhlendahl KE, Keller E, Willgerodt H, Kiess W 1998 Catch-up growth after childhood-onset substitution in primary hypothyroidism: is it a guide towards optimal growth hormone treatment in idiopathic growth hormone deficiency? Horm Res 50: Kriström B 1999 Growth hormone treatment in children prediction of growth response. Department of Clinical Science, Pediatrics and Göteborg Pediatric Growth Research Centre, Department of Pediatrics. Umeå, Sweden: Umeå University and Göteborg University 37. Rosenfeld RG, Albertsson-Wikland K, Cassorla F, Frasier SD, Hasegawa Y, Hintz RL, Lafranchi S, Lippe B, Loriaux L, Melmed S 1995 Diagnostic controversy: the diagnosis of childhood growth hormone deficiency revisited. J Clin Endocrinol Metab 80: Andersson B, Hellgren G, Nierop A, Rosberg S, Dahlgren J, Albertsson-Wikland K 2007 Serum protein profiling for prediction of growth hormone (GH) treatment response. Horm Res 68(Suppl 1): Hochberg Z, Albertsson-Wikland K 2008 Evo-devo of infantile and childhood growth. Pediatr Res 64:2 7