Briefing Document

Endocrinologic and Metabolic Drugs
Advisory Committee
June 10, 2003

Humatrope®
(somatropin [rDNA origin] for injection)
for Non-Growth Hormone Deficient Short Stature

Volume 1

Lilly Research Laboratories
Eli Lilly and Company
Lilly Corporate Center
Indianapolis, Indiana  46285

AVAILABLE FOR PUBLIC DISCLOSURE WITHOUT REDACTION

Table of Contents

Section                                                                                                     Page

List of Abbreviations....................................................................................... 10

List of Clinical Studies..................................................................................... 15

Executive Summary......................................................................................... 16

1.      Introduction............................................................................................ 22

1.1.     Regulatory History for the Study of Non-Growth Hormone-Deficient Short Stature  22

1.2.     The Rationale for GH Treatment of Non-GHD Short Stature.............. 24

2.      Overview of Clinical Studies................................................................... 31

2.1.     Pivotal Study:  B9R-MC-GDCH....................................................... 31

2.2.     Supportive Study:  B9R-EW-E001.................................................... 31

2.3.     Supportive Peer-Reviewed Literature Studies..................................... 32

3.      Effectiveness of Humatrope..................................................................... 33

3.1.     Pivotal Clinical Study:  GDCH........................................................... 33

3.1.1.       Primary Objective..................................................................... 33

3.1.2.       Study Design............................................................................. 33

3.1.3.       Inclusion/Exclusion Criteria........................................................ 34

3.1.4.       Summary of Key Protocol Changes........................................... 35

3.1.4.1.       Entry Height Criterion.......................................................... 35

3.1.4.2.       Final Height Criterion........................................................... 36

3.1.4.3.       Poststudy Summary Visit..................................................... 36

3.1.4.4.       Termination of Blinded Treatment Period.............................. 36

3.1.5.       Population Definitions................................................................ 36

3.1.6.       Patient Disposition..................................................................... 38

3.1.7.       Baseline Patient Characteristics.................................................. 40

3.1.8.       Efficacy Data............................................................................. 43

3.1.8.1.       Primary Efficacy Analysis (Final Height SDS)....................... 43

3.1.8.2.       Sensitivity Analyses.............................................................. 44

3.1.8.3.       Additional Analyses of Interest............................................. 51

3.1.9.       Efficacy Summary...................................................................... 55

3.2.     Supportive Study:  B9R-EW-E001.................................................... 55

3.2.1.       Objectives................................................................................. 55

3.2.2.       Study Design............................................................................. 56

3.2.3.       Inclusion/Exclusion Criteria........................................................ 56

3.2.4.       Population Definitions................................................................ 57

3.2.5.       Patient Disposition..................................................................... 57

3.2.6.       Baseline Patient Characteristics.................................................. 59

3.2.7.       Efficacy Data............................................................................. 61

3.2.7.1.       Dose-Response Effect on Height Velocity............................ 61

3.2.7.2.       Dose-Response Effect on Height SDS................................. 62

3.2.7.3.       Significant Treatment Effect on Final Height.......................... 65

3.2.8.       E001 Efficacy Summary............................................................ 67

3.2.9.       Comparative Efficacy Summary................................................. 68

3.3.     Supportive Data:  Meta-Analysis of Effect of Growth Hormone Therapy on Height in Children with Idiopathic Short Stature................................................ 71

3.3.1.       Growth Hormone Effect on Final Height..................................... 71

3.3.1.1.       Controlled Studies............................................................... 71

3.3.1.2.       Uncontrolled Studies............................................................ 73

3.3.1.3.       Summary............................................................................. 75

3.4.     Height SDS Gain Similar to Height SDS Gain in Turner Syndrome...... 75

3.5.     Humatrope Dosage and Frequency of Administration......................... 77

3.5.1.       Humatrope Dosage................................................................... 78

3.5.2.       Frequency of Administration...................................................... 78

3.5.3.       Clinical Relevance of Height Gain in Non-GHD Short Stature..... 79

3.6.     Overall Efficacy Conclusions.............................................................. 80

4.      Safety..................................................................................................... 81

4.1.     Introduction....................................................................................... 81

4.2.     Overview of Clinical Studies Included in Safety Comparison............... 81

4.2.1.       Study GDAB............................................................................ 82

4.2.2.       Study GDCT............................................................................. 83

4.2.3.       Study GDCI.............................................................................. 83

4.2.4.       Study GDCH............................................................................ 84

4.2.5.       Study E001............................................................................... 84

4.3.     Exposure........................................................................................... 84

4.4.     Deaths............................................................................................... 87

4.4.1.       Summary Comparison............................................................... 87

4.4.2.       Growth Hormone Deficiency..................................................... 87

4.4.3.       Turner Syndrome...................................................................... 88

4.4.4.       Non‑Growth Hormone-Deficient Short Stature.......................... 88

4.5.     Discontinuations Due to Adverse Events............................................. 88

4.5.1.       Summary Comparison............................................................... 90

4.5.2.       Growth Hormone Deficiency..................................................... 90

4.5.3.       Turner Syndrome...................................................................... 90

4.5.4.       Non-Growth Hormone-Deficient Short Stature.......................... 90

4.6.     Serious Adverse Events..................................................................... 91

4.6.1.       Summary Comparison............................................................... 93

4.6.2.       Growth Hormone Deficiency..................................................... 93

4.6.3.       Turner Syndrome...................................................................... 94

4.6.4.       Non-Growth Hormone-Deficient Short Stature.......................... 95

4.7.     Treatment-Emergent Adverse Events................................................. 99

4.7.1.       Summary Comparison............................................................. 103

4.7.2.       Growth Hormone Deficiency................................................... 103

4.7.3.       Turner Syndrome.................................................................... 103

4.7.4.       Non-Growth Hormone-Deficient Short Stature........................ 103

4.8.     Adverse Events Referenced in the Current Humatrope Label............ 106

4.8.1.       Summary Comparison............................................................. 107

4.8.2.       Otitis Media............................................................................ 108

4.8.3.       Scoliosis.................................................................................. 108

4.8.4.       Hypothyroidism....................................................................... 109

4.8.5.       Alterations in Carbohydrate Metabolism.................................. 109

4.8.6.       Hypertension........................................................................... 109

4.8.7.       Slipped Capital Femoral Epiphysis........................................... 109

4.9.     Safety Information from the Literature on GH Treatment in Non-GHD Short Stature 110

4.10.   Clinical Laboratory Evaluation.......................................................... 111

4.10.1.     Carbohydrate Metabolism....................................................... 112

4.10.1.1.     Summary Comparison....................................................... 115

4.10.1.2.     Growth Hormone Deficiency.............................................. 115

4.10.1.3.     Turner Syndrome............................................................... 115

4.10.1.4.     Non-Growth Hormone Deficient Short Stature................... 115

4.10.2.     Insulin-Like Growth Factor-I................................................... 120

4.10.2.1.     Turner Syndrome............................................................... 122

4.10.2.2.     Non‑Growth Hormone Deficient Short Stature................... 122

4.10.3.     Thyroid Function..................................................................... 123

4.10.3.1.     Growth Hormone Deficiency.............................................. 126

4.10.3.2.     Turner Syndrome............................................................... 126

4.10.3.3.     Non‑Growth Hormone Deficient Short Stature................... 126

4.11.   Overall Safety Conclusions.............................................................. 127

5.      Benefit/Risk Assessment....................................................................... 128

6.      Risk Management................................................................................. 133

6.1.     Introduction..................................................................................... 133

6.2.     Specific Concerns............................................................................ 133

6.2.1.       Inappropriate Prescribing......................................................... 133

6.2.2.       Lack of Adequate Diagnostic Evaluation.................................. 134

6.2.3.       Emergence of New Adverse Events......................................... 134

6.3.     Elements of the Lilly Risk Management Process................................ 134

6.3.1.       Restrictive Humatrope Labeling for Non-GHD Short Stature.... 134

6.3.2.       Physician Education................................................................. 136

6.3.2.1.       Physician-to-Physician Programs (Lilly-Sponsored)............ 137

6.3.2.2.       Continuing Medical Education............................................ 137

6.3.3.       Limited Marketing................................................................... 137

6.3.3.1.       Limited Sales Force........................................................... 137

6.3.3.2.       Sales Force Training on the Humatrope Benefit/Risk Profile and Appropriate Use                                                                                          138

6.3.3.3.       No Direct-to-Consumer Advertising.................................. 138

6.3.4.       Controlled Distribution Process................................................ 138

6.3.5.       Safety Monitoring and Analysis................................................ 139

6.3.5.1.       Pharmacovigilance............................................................. 139

6.3.5.2.       Postmarketing Surveillance Research.................................. 139

6.4.     External Factors.............................................................................. 141

6.4.1.       The Pediatric Endocrine Community........................................ 141

6.4.2.       Professional Physician Societies............................................... 141

6.4.3.       Insurance Companies.............................................................. 142

6.5.     Conclusions..................................................................................... 142

7.      Summary and Conclusions.................................................................... 145

7.1.     Height Gain..................................................................................... 145

7.2.     Dosage............................................................................................ 145

7.3.     Dose Frequency.............................................................................. 145

7.4.     Safety.............................................................................................. 146

7.5.     Conclusion...................................................................................... 146

8.      References........................................................................................... 147

Table of Contents

Table                                                                                                         Page

Table 1.         Mean Height SDS of Patients with Growth Disorders at Initiation of Growth Hormone Treatment 28

Table 2.         Demographics and Other Baseline Characteristics a Study GDCH.. 41

Table 3.         Final Height Standard Deviation Score Analysis of Covariance Final Height Population Study GDCH.. 43

Table 4.         Modified Intent-to-Treat Analysis Efficacy Evaluable Population Study GDCH   45

Table 5.         Intent-to-Treat Analyses of Last Observed Height SDS All Randomized Population Study GDCH.. 46

Table 6.         Analyses of Adult Height Protocol Complete and Final Height Populations Study GDCH   47

Table 7.         Additional Endpoint Height Analyses Efficacy Evaluable Population Study GDCH   52

Table 8.         Additional Final Height Analyses Final Height Population Study GDCH   53

Table 9.         Additional Final Height Analyses Protocol Complete Population Study GDCH   54

Table 10.       Demographics and Other Baseline Characteristics Study E001. 60

Table 11.       Height Velocity Changes from Pretreatment to 2-Year Endpoint Two-Year Height Velocity Population Study E001. 62

Table 12.       Secondary Efficacy Analyses Two‑Year Height Velocity Population Study E001  63

Table 13.       Final Height Standard Deviation Score Analysis of Covariance Final Height Population Study E001. 64

Table 14.       Final Height Characteristics Final Height Population Study E001. 66

Table 15.       Final Height Results:  Meta-Analysis of Controlled Trials a. 72

Table 16.       Results of Meta‑Analysis of Uncontrolled Studies from Peer‑Reviewed Literature a 74

Table 17.       Clinical Studies Included in Safety Comparison. 82

Table 18.       Time on Study. 86

Table 19.       Patient Deaths During and After Study. 87

Table 20.       Discontinuations Due to Adverse Events. 89

Table 21.       Serious Adverse Events. 92

Table 22.       Patient Diagnosed with Hodgkin Disease:  Timecourse of Events. 97

Table 23.       Treatment-Emergent Adverse Events. 100

Table 24.       Summary of Treatment-Emergent Adverse Events by Clinically Relevant Categories Safety Population Study GDCH..... 105

Table 25.       Adverse Events Referenced in Humatrope Label 107

Table 26.       Somatropin Safety in Non-GHD Conditions Kabi International Growth Study (KIGS) 110

Table 27.       Somatropin Safety in the National Cooperative Growth Study (NCGS) 111

Table 28.       Carbohydrate Metabolism Changes from Baseline to Endpoint 113

Table 29.       Insulin-Like Growth Factor-I Changes from Baseline to Endpoint 121

Table 30.       Thyroid Function Changes from Baseline to Endpoint 124

Table 31.       Risk Management Elements and External Factors Related to Approval of Non-GHD Short Stature. 144

 

 

Table of Contents

Figure                                                                                                       Page

Figure 1.        A diagnostic algorithm for investigation of short stature.. 26

Figure 2.        Design of Study GDCH. 34

Figure 3.        Patient disposition for Study GDCH. 39

Figure 4.        Bone age versus year on study for Study GDCH. 48

Figure 5.        Increase in height SDS over baseline versus year on study relative to last observed height (year=0) in Study GDCH.. 49

Figure 6.        Height standard deviation score by year on study for the Efficacy Evaluable Population (Study GDCH). 51

Figure 7.        Design of Study E001. 56

Figure 8.        Patient disposition for Study E001. 58

Figure 9.        Bone age versus year on study in Study E001. 65

Figure 10.      Final height minus baseline predicted height (cm) in the Final Height Populations of Studies GDCH and E001. 67

Figure 11.      Comparative summary of Studies GDCH and E001:  Final height SDS. 69

Figure 12.      Significant number of GH treated patients achieved normal height in Studies GDCH and E001. 70

Figure 13.      Mean difference in height standard deviation scores between treatment and control groups for predicted adult height (at baseline) and achieved adult height for controlled studies.  73

Figure 14.      Mean adult height standard deviation scores predicted at baseline and achieved for uncontrolled studies. 75

Figure 15.      Distribution of height standard deviation score change (baseline to final height) in patients treated with Humatrope (Study GDCH and Study GDCI). 77

Figure 16.      Mean fasting glucose by year on study for Study GDCH.. 116

Figure 17.      Mean fasting insulin by year on study for Study GDCH.. 117

Figure 18.      Qualitative Insulin Sensitivity Check Index (QUICKI) baseline to endpoint in Study GDCH. 118

Figure 19.      Analysis of covariance of last on study QUICKI using baseline QUICKI as the covariate. 119

Figure 20.      No Humatrope dose effect on fasting glucose. 120

Figure 21.      IGF-I increased modestly in Humatrope-treated patients in Study GDCH. 122

Figure 22.      No significant GH-related change in free thyroxine in Study GDCH. 126

 

List of Abbreviations

Symbol

Definition

AACE

American Association of Clinical Endocrinologists

AAP

American Academy of Pediatrics

AE

adverse event

ANCOVA

analysis of covariance

ANOVA

analysis of variance

BA

bone age

BPH

baseline predicted final height

BSA

body surface area

CA

chronological age

CFR

Code of Federal Regulation

CIB

clinical investigator’s brochure

CME

continuing medical education

CRF

clinical report form or case report form

CRI

chronic renal insufficiency

CSR

clinical study report

CT

clinical trial

DHEAS

dehydroepiandrosterone sulfate

DSMB

Data and Safety Monitoring Board

ERB

Ethical Review Board

FDA

Food and Drug Administration

ELECT

Eli Lilly Event Classification Terms

FSH

follicle-stimulating hormone

GCP

good clinical practice

GeNeSIS

Genetics and Neuroendocrinology of Short Stature International Study

GH

growth hormone

GHD

growth hormone deficiency, or growth hormone deficient

GMP

good manufacturing practice

GRS

Growth Hormone Research Society

HbA1c

hemoglobin A1c (glycosylated hemoglobin)

HGHPRC

Human Growth Hormone Protocol Review Committee

HHS

Department of Health and Human Services

ICD

informed consent document

ICH

International Conference on Harmonisation

IGF-I

insulin-like growth factor-I, also known as somatomedin-C

IND

investigational new drug

IRB

Institutional Review Board

ISS

idiopathic short stature

IU

International Units

IUGR

intrauterine growth retardation

KIGS

Kabi International Growth Study

LH

luteinizing hormone

LOCF

last observation carried forward

LSM

least squares mean

MQA

Medical Quality Assurance

                                                                                                                                             Continued

 

Symbol

Definition

NCGS

National Cooperative Growth Study

NCHS

National Center for Health Statistics

NGHDSS

non-growth hormone deficient short stature

NIH

National Institutes of Health

NICHD

National Institute of Child Health and Human Development

NOS

not otherwise specified

NVSS

normal variant short stature

OGTT

oral glucose tolerance test

pre-sNDA

pre-supplemental New Drug Application

SAE

serious adverse event

SAP

statistical analysis plan

SCFE

slipped capital femoral epiphysis

SD

standard deviation

SDS

standard deviation score

SE

standard error of mean

SGA

small for gestational age

SHOX

short stature homeobox-containing gene on the X-chromosome

TEAE

treatment-emergent adverse event

TIW

three times per week

TSH

thyroid-stimulating hormone

WWPE

World-Wide Pharmacovigilance and Epidemiology


Definitions of Terms

Adult height

See final height.

Adverse event

 

Clinical trial adverse event

AE

Any undesirable experience, unanticipated benefit, or pregnancy that occurs after informed consent for the study has been obtained, without regard to the possibility of a causal relationship and without regard to treatment group assignment, even if no study drug has been taken.

Clinical trial serious adverse event

SAE

Any adverse event in a clinical study patient that results in one of the following criteria:

·         Death;

·         Initial or prolonged inpatient hospitalization;

·         Life-threatening consequences;

·         Severe or permanent disability;

·         Cancer* (other than cancers diagnosed prior to enrollment in studies involving patients with cancer);

·         Congenital anomaly in the offspring of the patient;

·         Other significant consequence.

 

*As of 10 January 2001, cancer was removed from the SAE list based on International Conference on Harmonisation (ICH) guidelines.

Bone age

BA

Apparent developmental age of skeleton based on hand and wrist radiograph compared to normal standards (for example, normal bone age for a 12-year old child would be approximately 11-13 years).

Declaration of Helsinki

A document that defines an international standard for the conduct of clinical trials and has been adopted as legally enforceable by many countries and jurisdictions.

Eli Lilly Event Classification Terms

ELECT

A dictionary developed by Eli Lilly and Company that was used to describe, catalog, analyze, and report all adverse events (AEs).

Enrollment Process

 

Screen

The act of determining if an individual meets minimum requirements to become part of a pool of potential candidates for participation in a clinical study.

Enter

The act of obtaining informed consent for participation in a clinical study from individuals deemed potentially eligible to participate in the clinical study. Individuals entered into a study are those for whom informed consent documents (ICDs) for the study have been signed by the potential study participants or their legal representatives.

Randomization

In clinical trials, the assignment of a study participant to a treatment group in such a way that all possible treatment group assignments are equally probable, serving to avoid the introduction of known or unknown bias.

Enroll

For this study, enrollment was the act of assigning an individual to a treatment group.

A person who was entered into the study was potentially eligible to be enrolled in the study, but was required to meet all inclusion/exclusion specified in the protocol before being enrolled (assigned to a treatment group).  Individuals who entered into the study, but failed to meet inclusion/exclusion criteria were not eligible to participate in the study, and did not initiate therapy.

Final height

Generally a term used in clinical trials that refers to near-adult height, that is, the height at near-completion of growth.  The definition may vary between trials and is often defined as advanced bone age (>16 years in boys and >14 years in girls) and/or slowing of growth rate (0.5 –2.0 cm/y).

Final Height Population

 

Patients on whom a final height measurement was obtained.

Height standard deviation score (SDS)

 

The number of standard deviations from the mean for age and gender (normal range is –2 to +2 SDS).

Height velocity (cm/y)

 

Gain in height per time (normal:  5-7 cm/y before puberty, and 6-12 cm/y during puberty).

Two-Year Height Velocity
Population

 

Patients who had a height measurement at Visit 10 in Study E001.

Incidence

The incidence of adverse events is defined as the percent of patients reporting at least one adverse event at any time after baseline.

Intent-to-treat analysis

An analysis of study participants by the groups to which they were assigned by random allocation, even if the study participant did not take the assigned treatment, did not receive the correct treatment, or otherwise did not follow the protocol.  Such an analysis is sometimes referred to as “analyze as randomized” or “intention‑to‑treat.”

Patient-years

The sum of the days of exposure for all treated patients divided by 365.25.

Predicted Height

The predicted adult height is calculated on the basis of gender, current height, age, and bone age.

Pretreatment growth rate

The value obtained by computing the rate of growth between the height measurement taken approximately 12 months prior to Visit 2, and the height measurement taken at Visit 2.

Standard deviation score

SDS

The standard deviation score corresponding to a particular observation is a number that indicates how many standard deviations (SD) the observation is from the reference population mean.  It is positive or negative according to whether the observation lies above or below the mean.

Study drug

Refers to Humatrope or placebo.

Target height

The sex-adjusted average of parent’s heights (this is the patient’s genetic target).

Treatment-emergent adverse event

TEAE

Any adverse event that was not present at baseline or any pre-existing condition or event present at baseline that increased in severity during the study.

 

 

List of Clinical Studies

STUDY

TITLE

Pivotal:

 

 

B9R-MC-GDCH
Clinical Phase 3

 

Humatrope in Non-Growth Hormone Deficient Children with Short Stature.

 

Supportive:

 

 

B9R-EW-E001
Clinical Phase 3

 

The Efficacy and Safety of Biosynthetic Authentic Human Growth Hormone in Short Prepubertal Children with Normal Growth Hormone Response to Standard Provocation Tests.

 

 

Meta-analysis
Peer-Reviewed Literature

 

Finkelstein BS, Imperiale TF, Speroff T, Marrero U, Radcliffe DJ, Cuttler L. 2002. Effect of growth hormone therapy on height in children with idiopathic short stature.  A meta-analysis. Arch Pediatr Adolesc Med 156:230-240.

 

 

Executive Summary

This briefing document has been developed to aid the FDA Advisory Committee in evaluating Humatrope® (somatropin [rDNA origin] for injection) as a treatment for pediatric patients with non-growth hormone-deficient short stature (non-GHD short stature); meeting scheduled for 10 June 2003.  Throughout this document, the term somatropin refers to all brands of recombinant growth hormone (GH).  Humatrope refers specifically to the Lilly brand of somatropin.  Humatrope is a recombinant DNA-derived human growth hormone, identical in amino acid sequence to the 22-kd native human growth hormone.  It was approved on 08 March 1987 as replacement therapy “for the long-term treatment of children who have growth failure due to an inadequate secretion of normal endogenous growth hormone.”  On 11 March 1997, Humatrope was also approved for the treatment of short stature associated with Turner syndrome.  Currently these are the only two pediatric indications for which Humatrope is approved.  Humatrope has been approved at dosages up to 0.375 mg/kg/wk.  This document summarizes the clinical efficacy and safety data for Humatrope in pediatric patients with non-GHD short stature and the benefits and risks of such treatment.

INTRODUCTION

The 1983 International Conference on Uses and Abuses of Growth Hormone recognized a need for studies in “short children who do not have growth hormone deficiency” (Underwood 1984).  In 1987, the FDA Endocrinologic and Metabolic Drugs Advisory Committee further defined this need by recommending that a study to evaluate GH treatment in this population be a randomized, placebo-controlled trial to final height.

A pivotal trial, Study B9R-MC-GDCH, was designed and conducted in the US by Lilly and the National Institutes of Health between 1988 and 2001.  Study GDCH (n=71) was, as recommended by the FDA Advisory Committee, a double-blind, randomized, placebo-controlled study to final height.

Somatropin treatment is currently approved for 5 pediatric indications (growth hormone deficiency [GHD], chronic renal insufficiency, Turner syndrome, Prader-Willi syndrome, and children born small for gestational age), the latter 4 being non-GHD conditions.  The average height of patients with non-GHD short stature is very similar to that of other pediatric growth disorders.  Patients who do not pass the growth hormone treatment eligibility test (growth hormone response to stimulation falls above a defined threshold) and do not have one of the approved non-GHD indications have no approved treatment, despite an equivalent degree of short stature.

Over the past four decades the inequity of treatment availability for patients with non-GHD short stature led to a large volume of research (Finkelstein et al. 2002) on somatropin treatment in this patient population, culminating in Lilly’s pivotal and supporting studies, which were conducted between 1988 and 2001.

EFFICACY

Evidence for the efficacy of Humatrope treatment in pediatric patients with non-GHD short stature is presented.  Data sources include:  one pivotal trial - US, double-blind, randomized placebo-controlled study to final height; one supportive trial - European multicenter, 2‑year, three-arm, open-label, dose-response study with extension to final height; and a published meta-analysis on the effect of growth hormone treatment on height velocity and final height in patients with non-GHD short stature (Finkelstein et al. 2002).

Humatrope was effective in increasing final height as shown by the results of the pivotal study and the supportive dose-response study.  Study GDCH (0.22 mg/kg/wk, given in divided doses 3 times per week) involved patients with a baseline mean height well below the normal range (-2.8 standard deviation score [SDS]).  After a mean treatment duration of 4.4 years, and at a mean age of 18.8 years, the mean final height of the Humatrope-treated group was within the normal range, at -1.8 SDS, and was significantly greater than that of the placebo-treated group, which remained below the normal range, at -2.3 SDS.  The primary analysis, prespecified in the protocol, was an analysis of covariance (ANCOVA) of final height SDS, with baseline predicted height SDS as the covariate.  The mean treatment effect by this analysis was 0.51 SDS (95% CI:  0.10 to 0.92 SDS), corresponding to 3.7 cm (p=0.017).  Sensitivity analyses indicated a mean treatment effect of 2.8 to 5.0 cm.  These included intent-to-treat analyses, by both non-parametric and parametric methods that confirmed the significantly greater height SDS of the Humatrope-treated patients.  These gains in height SDS were achieved without any untoward effect on skeletal maturation or pubertal development.

Study B9R-EW-E001 (n=239) was a multicenter, 2-year, three-arm, open-label, dose-response study with extension to final height.  Patients were randomized to one of three treatment regimens:  0.24 mg/kg/wk; 0.24 mg/kg/wk the first year and 0.37 mg/kg/wk thereafter or 0.37 mg/kg/wk; all dosages were given in divided doses 6 times per week.  A dose-response effect for Humatrope was demonstrated by a greater increment in height velocity over the first 2 years of treatment for the patient group that received 0.37 mg/kg/wk compared with the group that received 0.24 mg/kg/wk (between-dose effect:  0.8 cm/y, 95% CI:  0.3 to 1.3 cm/y, p = 0.003).  Furthermore, a greater overall height gain, by approximately 3 cm, was observed at the higher dosages (incremental effect of 0.37 mg/kg/wk versus 0.24 mg/kg/wk).

In addition to the above evidence for dose-response, within-group analyses of final height minus baseline predicted height provided an estimate of treatment effect.  This is a conservative estimate of treatment effect because untreated patients with non-GHD short stature have been shown, on average, to reach an adult height below their baseline predicted height (Bramswig et al. 1990; Ranke et al. 1995; Buchlis et al. 1998; Rekers-Mombarg et al. 1999), as did the placebo-treated patients in Study GDCH.  The mean treatment effect sizes for this efficacy measure were 5.4 cm, 6.7 cm, and 7.2 cm for the dosages of 0.24 mg/kg/wk, 0.24→0.37 mg/kg/wk, and 0.37 mg/kg/wk, respectively.  Thus, the mean gain in adult height attributable to GH treatment with the 0.37 mg/kg/wk dosage was approximately 7 cm compared to the height that patients were predicted to achieve in the absence of treatment.

The supportive literature meta-analysis addressed the effect of GH therapy on height in children with non‑GHD short stature (referred to as idiopathic short stature in the paper).  The meta-analysis includes 38 studies, of which 4 studies that included a concurrent control group provide final height data.  For these 4 studies, the mean weighted GH dosage was 0.31 mg/kg/wk given in divided doses 6 times per week.  The mean duration of treatment was 5.3 years.  The between-group differences in achieved adult height for these 4 studies suggested a mean GH treatment effect of 5 to 6 cm (Finkelstein et al. 2002).  Thus, published studies support the efficacy of GH in non-GHD short stature, with the magnitude of benefit being similar to that observed in the Lilly pivotal and supportive dose-response studies.

The efficacy of Humatrope in increasing final height of patients with non-GHD short stature is similar to that seen in the approved indication for Turner syndrome.  Study B9R-CA-GDCT was a randomized, open-label study in patients with Turner syndrome, with an untreated control group as the comparator.  The Humatrope dosage was 0.30 mg/kg/wk, given in divided doses 6 times per week.  A planned interim analysis indicated a between-group difference in final height (t-test) of 3.9 cm (p=0.001).  A sensitivity analysis, an ANCOVA, with mid-parental height SDS as the covariate, indicated a treatment effect of 5.4 cm (p=0.001).  Thus, in the only other study to date with a long-term randomized control group to final height, the GH treatment effect was similar to that observed in the pivotal study of patients with non-GHD short stature.

Following the recommendation of the 1987 Endocrinologic and Metabolic Drugs Advisory Committee, Lilly conducted studies of patients with non-GHD short stature and focused on the treatment of their short stature.  Neither Study GDCH nor Study E001 provided evidence of potential benefits in quality of life or psychological well-being.  However, several lines of evidence suggest that the magnitude of GH-induced height gain in patients with non-GHD short stature was large enough to be clinically meaningful.  First, the GH-induced height gain in patients with non-GHD short stature was similar to that achieved in Turner syndrome.  Second, the mean heights of Humatrope-treated patients in Study GDCH, and of the 0.37 mg/kg/wk dosage group in Study E001, moved into the normal range during the course of treatment.  Third, whereas most final height SDS values of placebo-treated patients in Study GDCH were below normal, and all were below the 5th percentile, 94% of final height SDS values among the 0.37 mg/kg/wk dosage group of Study E001 were within the normal range.

SAFETY

Somatropin has a 16-year safety history and is currently approved for five pediatric indications and dosages up to 0.7 mg/kg/wk.  Worldwide it can be estimated that as many as 200,000 patients have been exposed to somatropin, representing over 500,000 patient-years of treatment.

In this document, the safety of Humatrope in patients with non-GHD short stature is evaluated by comparing the data collected in the non-GHD short stature clinical trials (Studies GDCH [n=68; Humatrope=37] and E001 [n=239]) with the safety data obtained in the clinical trials of Humatrope in patients with GHD (Study B9R-MC-GDAB [n=333]) and those with Turner syndrome (Study GDCT [n=136; Humatrope=74] and Study B9R-MC-GDCI [n=230]), the two pediatric populations for which Humatrope is currently approved.

Regarding deaths, discontinuations due to adverse events (AEs), or serious adverse events (SAEs) there were no meaningful differences identified between treatment groups or across studies or conditions.

Rates of SAEs were somewhat greater in the GHD and Turner syndrome studies than in the non-GHD short stature studies.  This probably relates to the higher baseline rates of serious illnesses in patients with GHD and Turner syndrome, particularly neurological disorders associated with GHD and ear and cardiac disorders associated with Turner syndrome, predisposing these patients to adverse events.  The rates of serious adverse events reported for the Humatrope-treated groups are as follows:  Study GDAB (GHD), n=90 (27%); Study GDCT (Turner syndrome), n=20 (27%); Study GDCI (Turner syndrome), n=41 (18%); Study GDCH (non-GHD short stature), n=5 (14%); Study E001 (non-GHD short stature), n=31 (13%).

In both the GHD and the non-GHD short stature studies, there were two cases of newly diagnosed neoplasia, described in detail in the Safety section (Section 4) of this document.  Neither case of neoplasia in the non-GHD short stature studies (Hodgkin lymphoma and desmoplastic small round cell tumor) was considered causally related to Humatrope exposure.

Patterns of treatment-emergent adverse events (TEAEs) differed somewhat between patient populations, mainly due to the presence of underlying disease in the GHD and Turner syndrome populations.  There were no statistically significant differences in TEAE rates between Humatrope and placebo groups in the pivotal study.  Except for scoliosis, all AEs currently referenced in the Humatrope label occurred at similar or lower rates in patients with non-GHD short stature.  Scoliosis was evaluated with added vigilance at the NIH and AE rates were found to be similar between the Humatrope and placebo treatment groups.  There was no evidence of a Humatrope effect on parameters of carbohydrate metabolism in either of the two non-GHD short stature studies, and IGF-I concentrations, measured only in Study GDCH, remained physiologic.

Overall, the safety profile of Humatrope treatment in this new patient population does not differ in a clinically meaningful way from that seen in the currently approved pediatric indications and no new safety language is required in the label.

BENEFIT-RISK ASSESSMENT

Benefits of Humatrope treatment in patients with non-GHD short stature are improved linear growth in childhood (allowing a degree of catch-up to peers) and increased height at completion of linear growth.  The magnitude of the benefit is similar to that seen in patients with Turner syndrome.  The risks to pediatric patients identified in the current Humatrope label are quite low, relatively mild, readily manageable, and in some cases, transient.  No new risks have been identified for patients with non-GHD short stature.  The benefit-risk profile of Humatrope treatment in patients with non-GHD short stature is similar to that seen in Turner syndrome.  Humatrope is safe and effective for the treatment of non-GHD short stature at a dosage of up to 0.37 mg/kg/wk.

In light of evidence for a positive benefit-risk profile, the following label indication is proposed:

Humatrope is indicated for the long-term treatment of non-growth hormone-deficient short stature, defined by height SDS £‑2.25, in pediatric patients whose epiphyses are not closed and in whom diagnostic evaluation excludes causes of short stature that should be treated by other means.

RISK MANAGEMENT PROGRAM

Lilly has identified areas of potential concern regarding approval of this new indication:  inappropriate prescribing, lack of adequate diagnostic evaluation prior to initiation of treatment, and emergence of new adverse events.

Potential concerns will be addressed by the following elements of the Lilly Risk Management Program:

[1]    Lilly is proposing a restrictive label to help establish appropriate use of Humatrope for this indication; the proposed indication excludes other causes of short stature and, unlike all previous pediatric indications, defines a maximum height threshold (height SDS £‑2.25) for initiating treatment.  This is a more conservative threshold than the definition of short stature of height SDS £‑2.0 (AAP 1997; AACE 2003).

[2]    Physicians will be trained (according to FDA guidelines) regarding the changes to the label and the restrictions for this patient population.

[3]    Marketing will be limited to endocrinologists only, with no direct-to-consumer advertising.

[4]    A proprietary controlled distribution process contributes to assuring appropriate prescribing and distribution of Humatrope to all patients including those with non-GHD short stature.

[5]    Pharmacovigilance and a post-marketing research program (Genetics and Neuroendocrinology of Short Stature International Study [GeNeSIS]) will continue to collect and analyze prospectively defined adverse events as well as spontaneously reported adverse events.  These data are analyzed annually and reported to investigators.

[6]    Additionally, there are a number of external factors that also mitigate these concerns independent of Lilly, including:  1) professional judgment of pediatric endocrinologists; 2) guidelines for growth hormone usage developed by professional endocrine societies; and 3) the requirement by insurance companies for demonstration of medical need.

CONCLUSION

Humatrope is safe and effective in pediatric patients with non-GHD short stature.  The approval of Humatrope for pediatric patients with non-GHD short stature will correct the current inequity in treatment availability for this population.

1. Introduction

Humatrope® (somatropin [rDNA origin] for injection) is a recombinant DNA-derived human growth hormone, identical in amino acid sequence to the 22-kd native human growth hormone.  Humatrope was approved on 08 March 1987 (NDA 19‑640) as replacement therapy “for the long-term treatment of children who have growth failure due to an inadequate secretion of normal endogenous growth hormone.”  On 11 March 1997, Humatrope was also approved for the treatment of short stature due to Turner syndrome.  Currently, these are the only two pediatric indications for which Humatrope is approved.

Eli Lilly and Company submits this briefing document to the Endocrinologic and Metabolic Drugs Advisory Committee of the Food and Drug Administration in support of the application for approval of Humatrope as treatment for non-growth hormone-deficient short stature (non-GHD short stature).  Lilly realizes that a number of issues and concerns exist regarding the implications of an approval for this indication.  Therefore, to support this application, this briefing document, in addition to presenting detailed analysis of safety and efficacy, will address the following questions:

·        Is it appropriate to treat patients whose short stature is not clearly associated with a defined “disease”?

·        Is GH effective in these patients, and is the magnitude of benefit clinically relevant?

·        Should psychological benefits be a required outcome of GH treatment?

·        Is this treatment safe in this patient population?

·        Why was the height cut-off of –2.25 SDS chosen for the label indication?

·        Will this new indication obviate the need for thorough diagnostic evaluation in children with growth disorders?

·        Will this new indication “open the floodgates” to inappropriate use?

These issues will be addressed within the appropriate sections of this document, and will be summarized in the Benefit-Risk (Section 5) and Risk-Management (Section 6) sections.

1.1.   Regulatory History for the Study of Non-Growth Hormone-Deficient Short Stature

This section summarizes key milestones, or interactions and agreements reached between the US Food and Drug Administration (FDA) Division of Metabolic and Endocrine Drug Products and Eli Lilly and Company (Lilly) regarding Lilly studies of efficacy and safety of Humatrope in pediatric patients with non‑GHD short stature.  Throughout this document, once each specific protocol has been identified by its full study code, all subsequent referrals will be by the final four letters alone (for example, Study B9R‑MC‑GDCH will be referred to as Study GDCH).

18 June 1986:  Lilly submitted an investigational new drug application (IND 28,574) to support studies of Humatrope for non-hypopituitary indications.

07 July 1987:  Lilly submitted to IND 28,574 the protocol for Study GDCH.  Study GDCH was a double‑blind, randomized, parallel, placebo‑controlled clinical study of Humatrope to final height in pediatric patients with non-GHD short stature.

28 September 1987:  The Endocrinologic and Metabolic Drugs Advisory Committee of the FDA met to provide guidance for GH manufacturers regarding studies of GH treatment in pediatric patients with non-GHD forms of short stature.  The committee unanimously agreed that the critical endpoint was final height and that such studies should include a control group.  Although there were concerns about the type and feasibility of the control, the committee recommended, “…the control group should be a placebo‑treated, parallel, randomized group of patients…” and  “…the subjects should be followed until their ultimate height is reached…” (FDA 1987, Dr Philip Troen).

22 January 1988:  Study GDCH was initiated by Lilly and the National Institute of Child Health and Human Development (NICHD) of the National Institutes of Health (NIH).

1992:  There was a challenge to the study by a third party who asserted that the study was being conducted contrary to the principles that should be followed when using children in medical research.  In response to this challenge, the NIH convened an external advisory panel, the Human Growth Hormone Protocol Review Committee (HGHPRC).  The HGHPRC concluded that the protocol addressed an important public health need and did not violate any of the applicable Department of Health and Human Services (HHS) regulations cited in the challenge (45 Code of Federal Regulations [CFR] Part 46).  It was recommended that a Data and Safety Monitoring Board (DSMB) be convened to conduct an independent review of the study on a regular basis.  The role of the DSMB was to provide an independent review of the accumulating data in the study, and to evaluate the appropriateness of continuing the study in the context of these data and any other relevant published data.  No formal statistical stopping rules were instituted.  The external DSMB was subsequently convened and reviewed interim data, unblinded at the treatment group level, at each of its meetings.

28 October 1993:  The external DSMB met for the first time to conduct an independent review of the study.  After a detailed examination of data, the DSMB recommended continuation of Study GDCH.  The DSMB convened again on 14 October 1994, 18 April 1996, 03 June 1997, 08 June 1998, and 24 June 1999, each meeting returning a recommendation for continuation of the study.

05 June 2000:  The DSMB met and recommended that the placebo‑controlled study be terminated, that active patients be offered the option to receive open‑label treatment, and that the results be disseminated as soon as possible.  At that time, patients were reaching final height at a rate of 2 patients per year and it would have required approximately 5 additional years before the remaining patients reached final height.  Therefore, the DSMB unanimously concluded, “…the study is not maturing sufficiently to justify the maintenance of a placebo injection control group.” (written communication, Data Monitoring Committee Report, 05 June 2000).

31 July 2001:  A pre‑supplemental New Drug Application (pre‑sNDA) meeting between Lilly and the FDA was held to discuss a plan by Lilly to submit a data package in support of an indication for non-GHD short stature.  The FDA indicated that the planned submission appeared to be acceptable for review.

1.2.   The Rationale for GH Treatment of Non-GHD Short Stature

Over the past 40 years there have been significant advances in the understanding and management of childhood growth disorders.  When human growth hormone (GH) was first introduced as a therapeutic agent in 1958, its use was restricted to children with the most severe forms of growth hormone deficiency (GHD) due to hypopituitarism.  Availability was limited by the supply of GH, due to both its human source and to the complex and time‑consuming extraction and purification procedures required.  Following the withdrawal of pituitary-derived human GH and the introduction of recombinant DNA-derived GH in the mid 1980s, treatment became available for children with less severe forms of GHD.  In addition, the potential value of GH treatment for impaired growth due to other conditions began to be investigated.  In 1985 recombinant methionyl GH (somatrem) was approved for treatment of pediatric patients with growth failure due to inadequate secretion of endogenous GH.  Natural sequence recombinant DNA-derived GH (somatropin) was first approved for the same indication in 1987 (initial dosage approved was 0.18 mg/kg/wk; current approved dosage range is 0.18 to 0.30 mg/kg/wk).  In the 16 years following its introduction, the safety and efficacy of somatropin has been established and treatment approved for four additional pediatric disorders in patients who are not GHD.  These approvals allowed physicians to provide safe and effective treatment for impaired growth in children with chronic renal insufficiency, Turner syndrome, Prader-Willi syndrome, and children born small for gestational age (SGA) who fail to catch up in height.  The goal of treatment in these patients is to treat the growth impairment and not the underlying condition.  Currently, the approved dosage range across these pediatric indications (all brands of somatropin combined) is 0.18 to 0.70 mg/kg/wk.

Growth failure and short stature are among the most common reasons for which children are evaluated by pediatric endocrinologists.  The term “growth failure” refers to the decline in the rate of linear growth (height velocity) that, if persistent, results in short stature.  Various definitions of short stature can be found in the literature; however, the most commonly used and generally accepted definition is provided by the American Association of Clinical Endocrinologists (AACE) and the American Academy of Pediatrics (AAP), and defines short stature as height more than 2.0 standard deviations below the population mean (AAP 1997; AACE 2003).  By US standards, this is equivalent to 5’3.6” for an adult male, and 4’11.1” for an adult female.  Because there are numerous endocrine and non-endocrine causes of growth failure and short stature, careful and comprehensive investigation is required to determine the cause.  The Growth Hormone Research Society (GRS) recommends investigation of children with short stature whose height falls below -2.0 standard deviation scores (SDS) (GRS 2000).

Investigation of growth failure or short stature includes detailed history, assessment of the patient’s and family’s patterns of growth over time, physical examination with specific attention to body proportions and phenotypic markers of certain syndromes, and a number of biochemical, radiological and sometimes genetic evaluations as outlined in the diagnostic algorithm presented in Figure 1.

* Patients with Turner syndrome have variable shortness of the forelimbs, which becomes more apparent with age.  These patients may also have other skeletal anomalies such as short fourth metacarpal, increased carrying angle, high-arched palate, and Madelung deformity.
Note:  This figure was compiled from the following references:  Parkin 1989; Schwartz and Bercu 1992; Van den Brande and Rappaport 1993; Blizzard and Johanson 1994; Reiter and Rosenfeld 1998; Hintz and Ritzen 1999.

Abbreviations:  GH = growth hormone; GHD = growth hormone deficiency; IGF‑I = insulin‑like growth factor‑I; IGFBP-3 = insulin‑like growth factor binding protein‑3; NGHDSS = non‑growth hormone deficient short stature.

Figure 1.                       A diagnostic algorithm for investigation of short stature.

The final step in the diagnostic process is a growth hormone stimulation test, in which pharmacological agents are administered to provoke release of pituitary GH stores.  Patients whose peak responses to stimulation fall below a defined threshold are classified as GHD and are eligible for GH treatment.  In contrast, patients whose GH responses exceed the specified threshold are deemed non-GHD and are ineligible for GH treatment, despite phenotypes and degrees of short stature essentially indistinguishable from those with GHD.  Patients, their families, and physicians find this inequity frustrating.  The inequity of this situation is further exacerbated by the fact that patients with chronic renal insufficiency, Turner syndrome, Prader-Willi syndrome, and those born small for gestational age are eligible for GH treatment irrespective of their GH secretion status or degree of short stature.

The lack of approved therapy for patients who have neither GHD, nor one of the 4 non-GHD conditions listed above, led to much discussion in the pediatric endocrine community over many years, regarding why such children should or should not be eligible for GH treatment.  In assessing this question, the following points should be considered:

  1. Children and adults with short stature may have disadvantages with respect to their peers, irrespective of the cause of the short stature.
  2. The growth failure in patients with non-GHD short stature is equivalent to that seen in other growth disorders.
  3. The majority of untreated patients with non-GHD short stature fail to achieve their adult height prediction.
  4. GH treatment in other pediatric conditions treats the growth failure or short stature, not the underlying condition or “disease”.
  5. Absence of a known etiology for the growth failure does not justify exclusion from treatment.
  6. The growth failure in patients with non-GHD short stature is responsive to GH treatment.

Each of the first five points above are discussed further in the following paragraphs, while the sixth point is discussed in the Efficacy Section (Section 3) of this document.

First, children and adults with short stature, irrespective of cause, may have a number of disadvantages in life relative to their normal-stature peers.  Short children may be subject to juvenilization, teasing, bullying, exclusion from activities and peer groups and impairment of the normal progression toward independence (Sandberg 1999; Voss and Mulligan 2000).  In adulthood there may be problems of social isolation, reduced marriage rates, perceptions of lower competence, and ineligibility for certain occupations that have specific minimum height requirements.  There are a number of potential employers, such as construction companies, the aviation and aerospace industries, and the military, that have minimum height standards for employees.  Furthermore, aspects of daily living such as driving a car, accessing cupboards, using kitchen or bathroom benches or sinks, and using standard height furniture in workplaces may provide additional challenges.

Second, there is marked concordance in the severity of growth failure and short stature among the various growth disorders for which somatropin treatment is typically prescribed (Table 1).  In two large postmarketing research programs (the National Cooperative Growth Study [NCGS] and the Kabi International Growth Study [KIGS]), height SDS at initiation of GH treatment was well below normal across the various conditions.  Table 1 shows a remarkable similarity between patients with non‑GHD short stature (referred to in these reports as idiopathic short stature) and patients with conditions for which somatropin is currently approved.  The values in the table represent mean height SDS ± standard deviation at the initiation of GH treatment.

Table 1.                         Mean Height SDS of Patients with Growth Disorders at Initiation of Growth Hormone Treatment

Condition

Mean Height SDS

Corresponding adult male height
 (feet, inches)

Corresponding adult female height
 (feet, inches)

Idiopathic GHD

-2.8 ± 1.1 a

5 ft, 1.3 in

4 ft, 9.1 in

Idiopathic Short Stature

-2.9 ± 0.9 a

5 ft, 1.0 in

4 ft, 8.8 in

Chronic Renal Insufficiency

-2.6 ± 0.8 b

5 ft, 1.9 in

4 ft, 9.6 in

Turner Syndrome

-2.8 ± 1.0 c

NA

4 ft, 9.1 in

Small for Gestational Age

-2.8 ± 0.9 d

5 ft, 1.3 in

4 ft, 9.1 in

Note:  Values represent mean ± SD.  Conversion of height SDS to feet and inches was based on Kuczmarski et al. 2000.

Abbreviations:  GHD = growth hormone deficiency; SDS = standard deviation score.

a     Root et al. 1998.

b    Fine et al. 1996.

c     Ranke et al. 2000.

d    Ranke et al. 2003.

 

These observational study data also demonstrate the relative severity of short stature in patients entering GH treatment compared to accepted definitions of short stature (AAP 1997; AACE 2003) and with the recommended height threshold for further investigation (GRS 2000), of –2.0 SDS.  This finding reflects the fact that pediatric endocrinologists carefully evaluate children with short stature and take a conservative approach, providing GH treatment to those with the greatest need.

Third, in addition to the equivalence of their short stature at baseline, as demonstrated in a number of studies, untreated patients with non-GHD short stature achieve adult heights that fall below the adult heights predicted for them during childhood (Bramswig et al. 1990; Ranke et al. 1995; Buchlis et al. 1998; Rekers-Mombarg et al. 1999).

Fourth, GH treatment in pediatric patients with growth disorders is intended to treat the growth failure or short stature, not the underlying condition or “disease”.  This is evidenced by the language of the label indication for each of the conditions for which somatropin is currently approved.  Somatropin is indicated for:  “the long‑term treatment of pediatric patients who have growth failure due to an inadequate secretion of normal endogenous growth hormone”; “the treatment of short stature associated with Turner syndrome in patients whose epiphyses are not closed”; “treatment of growth failure associated with chronic renal insufficiency up to the time of renal transplantation”; “long-term treatment of growth failure due to Prader-Willi syndrome”; “long-term treatment of short stature in children born small for gestational age who fail to manifest catch-up growth by age 2”.  Indeed, in each of the above conditions, while the etiology of the disorder itself may be known (for example, complete or partial loss of one X-chromosome in Turner syndrome), the cause of the growth disturbance is only partially understood, if at all.  The key distinction between patients with non-GHD short stature and those with conditions for which somatropin is currently approved is that most of the latter have additional problems beyond their growth disturbance (such as ovarian failure in Turner syndrome) that are not addressed by somatropin.  The exception is the child born small for gestational age, whose short stature is typically the only clinical abnormality.

Fifth, the fact that children with non-GHD short stature are regarded by some as having no “disease”, does not justify excluding them from effective treatment.  There are many such conditions in both pediatric and adult patients that deserve and receive treatment.  Examples include enuresis, hypertension, hypercholesterolemia, erectile dysfunction, alopecia, hirsutism, gynecomastia, anxiety disorder, and nicotine addiction.  Whether or not any of these conditions is formally considered a “disease” appears to have no bearing on the appropriateness of treating the condition.  Prevention of pregnancy is also an accepted therapeutic aim for a condition that is not considered a disease.

Recognizing the unmet medical need of patients with non-GHD short stature the 1983 International Conference on Uses and Abuses of Growth Hormone, was convened by the National Institute of Child Health and Human Development (NICHD) and issued this consensus statement:  “….there is an urgent need for therapeutic trials to determine the effect of growth hormone in short children who do not have growth hormone deficiency” (Underwood 1984).  Subsequently, the Endocrinologic and Metabolic Drugs Advisory Committee of the FDA provided guidance for GH manufacturers regarding studies of GH treatment in pediatric patients with non-GHD forms of short stature.  The committee unanimously agreed that the critical endpoint was final height and that such studies should include a control group.  Although there were concerns about the type and feasibility of the control, the committee recommended, “…the control group should be a placebo‑treated, parallel, randomized group of patients…” and  “…the subjects should be followed until their ultimate height is reached…” (FDA 1987, Dr Philip Troen).  In the 16 years following the introduction of recombinant GH, more than 40 studies have been undertaken in the non-GHD short stature patient population.  Efficacy has been demonstrated by improvements in height velocity, although few of these studies have followed patients to adult height and none have been placebo controlled (Finkelstein et al. 2002).

“Lessening the disability of severe short stature has been the goal of GH therapy for three decades” (Allen et al. 1994).  However, for patients whose growth failure is not associated with one of the five conditions for which GH is currently approved, this goal is currently unattainable.  To address this deficit, Lilly undertook two long-term studies of the safety and efficacy of GH treatment in patients with non‑GHD short stature.  The pivotal study, Study GDCH, specifically followed the FDA Advisory Committee’s 1987 recommendation and is the only randomized, double-blind, placebo-controlled study to final height in this patient population.  Study GDCH unequivocally demonstrates that GH treatment is effective in patients with non-GHD short stature.  The efficacy of GH in this condition is also supported by Study E001, a second large, long-term, randomized, dose-response study, and by data from a comprehensive meta-analysis of the peer-reviewed literature (Finkelstein et al. 2002).

The establishment of clear efficacy of GH treatment in patients with non-GHD short stature, in the absence of any new safety concerns, provides the scientific, medical, regulatory, and ethical justification for approval of Humatrope treatment for these patients.

 

2.     Overview of Clinical Studies

The efficacy of Humatrope for the treatment of non-GHD short stature is addressed by:  one randomized, placebo-controlled, pivotal clinical trial; one randomized, dose-response study; and a recent meta-analysis of the peer-reviewed literature.  Brief summaries of these studies are provided below, while detailed information regarding study design, patient demographics, and efficacy data are provided in the following section (Section 3).  The safety of Humatrope treatment in patients with non-GHD short stature is addressed (Section 4) by comparison of Lilly studies in this patient population with studies undertaken in patients with GHD and with Turner syndrome.

2.1.   Pivotal Study:  B9R-MC-GDCH

Study GDCH was a double‑blind, randomized, parallel, placebo‑controlled study to final height in pediatric patients with non-GHD short stature (n=71).  The primary objective of the study was to determine whether final height of patients with non-GHD short stature treated with Humatrope (0.22 mg/kg/wk, administered in divided doses 3 times per week [TIW]) would be greater than that of a placebo-treated group.  The primary efficacy variable was final height, expressed as a standard deviation score (SDS) relative to the general population of the same age and gender (final height SDS).

2.2.   Supportive Study:  B9R-EW-E001

Study E001 was an open-label, three-arm, randomized, parallel, dose-response study.  Pediatric patients with non-GHD short stature (n=239) were randomly assigned to receive one of the following three Humatrope regimens (administered in divided doses 6 times per week):

·        Dose 1:  0.24 mg/kg/wk;

·        Dose 2:  0.24 mg/kg/wk for 1 year, followed by 0.37 mg/kg/wk; or

·        Dose 3:  0.37 mg/kg/wk.

The primary objective of this study was to determine the efficacy of two different Humatrope dosages (0.24 versus 0.37 mg/kg/wk) in stimulating an increase in height velocity in pediatric patients with non-GHD short stature.  The increase in height velocity during the initial 2 years of treatment was the primary variable used to evaluate dose-response effect among the Humatrope dosage regimens.  Final height SDS was a secondary outcome measure, as were the following variables:  final height minus baseline height (cm and SDS); final height minus baseline predicted height (cm and SDS); and final height minus target height (cm and SDS).

2.3.   Supportive Peer-Reviewed Literature Studies

A recent meta‑analysis (Finkelstein et al. 2002) provides supportive evidence from the literature for the effectiveness of GH treatment in pediatric patients with non-GHD short stature.  The data reported in this meta‑analysis are derived from 10 controlled studies and 28 uncontrolled studies that used recombinant GH (somatropin) from several manufacturers.  Of these studies, 12 (4 controlled and 8 uncontrolled; total number of patients = 454) provide data on adult height, and will be used to support the use of Humatrope in pediatric patients with non-GHD short stature.

 

 

3.   Effectiveness of Humatrope

3.1.   Pivotal Clinical Study:  GDCH

Study GDCH was a double‑blind, randomized, parallel, placebo‑controlled study to final height in pediatric patients with non-GHD short stature.

3.1.1.  Primary Objective

The primary objective of the study was to determine whether final height, defined as the last height obtained after height velocity had fallen below 1.5 cm/y, would be greater in patients treated with Humatrope than in those who received placebo injections.

3.1.2.  Study Design

Study GDCH consisted of a screening period and a blinded treatment period.  Patients in the Humatrope group received 0.074 mg/kg Humatrope by subcutaneous injection, 3 times per week (total dose 0.22 mg/kg/wk:  standard of care at the time for GHD was 0.18 mg/kg/wk); patients in the placebo group received placebo injections 3 times per week.  Participation in the primary (blinded) treatment period ended either when the patient reached final height (protocol completion) or at the time of closure of the blinded study, by Lilly, at which time eligible patients were offered the opportunity to enter an optional, open‑label extension phase.  Figure 2 presents the study design.  The open‑label extension phase is omitted for clarity, since no efficacy data are being collected from the open-label, single arm (Humatrope only) extension.

 

+ Only the subset of patients on whom lower leg measurements were obtained attended study Visits 2 to 4 and 6.
* Poststudy summary visit (Visit 99):  1 year after protocol completion for patients completing the protocol; at final height for patients who discontinued the study before protocol completion.

Figure 2.                       Design of Study GDCH.

Patients who completed the study were asked to return for a final height measurement 1 year after protocol completion.  This poststudy follow‑up visit was referred to as Visit 99.  In addition, patients who discontinued the study prior to protocol completion were asked to return for a final height measurement after height velocity, measured locally, had fallen below 1.5 cm/y.  This poststudy follow‑up visit was also referred to as Visit 99.

Patient enrollment began in January 1988 and continued through July 1999, when it was ended at the request of the DSMB in view of the length of the treatment required.  The blinded treatment period ended in February 2001.

3.1.3.  Inclusion/Exclusion Criteria

Inclusion criteria were:  age 10-16 (boys) or 9-15 years (girls); bone age £13 (boys) or £11 years (girls); Tanner stage breast or genital development £2; proportionate short stature; and peak stimulated GH >7 mg/L.  Patients were included if their height standard deviation score (SDS) or predicted adult height SDS within 1 year prior to study entry was £‑2.5.  During the period from May 1988 to February 1993 a cutoff of -2.25 SDS was used.  Section 3.1.4.1 provides a detailed explanation of these changes.  Children with stimulated GH concentration >7 mg/L were considered GH-sufficient based on normative data generated with the same GH assay (Marin et al. 1994).  Thus, the term, “non-GHD short stature” refers to children who have normal stimulated GH concentrations as defined above.  It is not meant to imply that GH has no role in the etiology of the short stature.  Patients were excluded if they had a chronic illness, a known genetic syndrome, had ever received GH, estrogen, or androgen treatment or were currently receiving other drugs likely to affect growth, including methylphenidate and similar stimulants.  Patients with hypothyroidism were eligible to enroll after thyroid function tests had been normal for at least 3 months on replacement therapy.

When this study was designed in the 1980s, patients who were born small-for-gestational age (SGA) were not excluded by the protocol, and 6 of the 71 patients enrolled met criteria for SGA (birth weight SDS £‑2.0 according to Table III of Usher and McLean [1969]).  Because these patients met the inclusion criteria for the study, they are included in the study statistical analyses.  At the pre-supplemental New Drug Application meeting between Lilly and the FDA, on 31 July 2001, the FDA indicated that a formal statistical analysis of SGA and non-SGA groups would not be necessary.  Nonetheless, analyses showed no differential treatment effect between the effect in SGA patients and the effect in non-SGA patients.

3.1.4.  Summary of Key Protocol Changes

3.1.4.1.    Entry Height Criterion

In response to a suggestion by the FDA to improve the balance of the two treatment groups, a stratified randomization was added in May 1988 (a total of 3 patients had been enrolled in the study at that time), with patients grouped by gender and predicted adult height.  In addition, to increase the rate of patient enrollment, the entry height criterion was changed from height standard deviation score (SDS) or predicted height SDS £-2.5 to height SDS or predicted height SDS £-2.25 (based on stature data from the National Center for Health Statistics [NCHS] Growth Charts [1976] and measured within the 12 months prior to Visit 1).  Based upon current US height standards (Kuczmarski et al. 2000), a height SDS of -2.5 corresponds to an adult height of 5 feet, 2.2 inches (157.9 cm) in males and 4 feet, 9.8 inches (146.9 cm) in females.  A height SDS of -2.25 corresponds to 5 feet, 2.9 inches (159.8 cm) in males and 4 feet, 10.5 inches (148.5 cm) in females.  Patients with height SDS <-2.25 represent the shorter 54% of patients who meet the American Academy of Pediatrics definition of short stature (height SDS <-2.0 [AAP 1997]). 

The entry criteria for upper height limit and predicted adult height were changed in February 1993.  At the recommendation of the Human Growth Hormone Protocol Review Committee, an independent panel appointed by the NIH Director, the inclusion criterion of height SDS or predicted height SDS £-2.25 was changed back to that of the original protocol, stating that patients must have height SDS or predicted height SDS £‑2.5.  Thirty-seven patients of the final 68 who received study drug (Humatrope or placebo) were enrolled on the basis of height SDS or predicted height SDS £‑2.25, during the period from May 1988 to February 1993.  Of these 37 patients, six (2 Humatrope, 4 placebo) had a height SDS or predicted height SDS at eligibility assessment between ‑2.5 and -2.25.

3.1.4.2.    Final Height Criterion

The original protocol defined the criterion for protocol completion as the achievement of height velocity <0.5 cm/y, based on measurements made at 12‑month intervals.  Two patients (1 Humatrope, 1 placebo) completed the protocol with this criterion.  In January 1994 (a total of 45 patients had been enrolled in the study at that time), the criterion for protocol completion was changed from height velocity <0.5 cm/y to height velocity <1.5 cm/y.  This criterion was changed to address the issue of drop-outs that occur as the height velocity slows down upon the approach of final height.  As the slow‑down progresses, the patient is less likely to want to continue injections and more likely to drop out of the study.

3.1.4.3.    Poststudy Summary Visit

To gather height and safety data for an intent‑to‑treat analysis, a poststudy summary visit (Visit 99) was added in January 1994 (a total of 45 patients had been enrolled in the study at that time) for those patients who had completed the study or who had discontinued the study prior to protocol completion.

3.1.4.4.    Termination of Blinded Treatment Period

In response to the recommendation made by the DSMB on 05 June 2000, the blinded treatment period of the study was terminated in January 2001 (a total of 71 patients had been enrolled in the study at that time).  An open‑label extension phase was implemented to provide Humatrope‑treated patients the opportunity to continue on Humatrope treatment and to allow placebo‑treated patients the option to receive Humatrope treatment.

3.1.5.  Population Definitions

The following populations were defined in the protocol:

Randomized Patients (n=71):  Seventy‑one patients enrolled in the study and were randomized into treatment groups (38 Humatrope, 33 placebo).  Analysis of this population serves as an intent-to-treat analysis for this study.

Safety Population (n=68):  Of the 71 randomized patients, 3 patients discontinued the study prior to receiving any study drug (1 Humatrope [physician decision]; 2 placebo [protocol entry criteria not met]).  The remaining 68 patients were included in the Safety Population (37 Humatrope, 31 placebo).

Efficacy Evaluable Population (n=64):  Assessment of efficacy required at least 6 months study drug treatment.  Of the 68 patients in the Safety Population, 3 patients discontinued without a height measurement at 6 months (Visit 5:  2 Humatrope [adverse event (1), patient decision (1)]; 1 placebo [patient decision]).  One additional placebo patient (Patient 008-1201), described below, who received growth hormone (GH) outside the study, was excluded from the Efficacy Evaluable Population.  The remaining 64 patients were included in the Efficacy Evaluable Population (35 Humatrope, 29 placebo).  Analysis of this population serves as a modified intent‑to‑treat analysis for this study.

Patient 008-1201 was randomized to the placebo treatment group but was excluded from the Efficacy Evaluable and Final Height Populations.  This patient discontinued the study at Visit 5, 6 months after randomization, but returned for a final height visit, as requested.  Because the patient had a height measurement at Visit 5 and a final height measurement, she would have qualified for the Efficacy Evaluable and Final Height Populations.  However, it was learned that this patient had received growth hormone (GH) treatment for approximately 4 years (personal communication, Ellen Leschek, MD) after discontinuing the study.  Because of the documented receipt of GH treatment, this patient was excluded from the Efficacy Evaluable and Final Height Populations but was included in the Safety Population for the placebo group.  Patient 008-1201 did not have any adverse events after discontinuing the study. 

Protocol Complete Population (n=25):  Of the 64 patients from the Efficacy Evaluable Population, 39 patients (19 Humatrope, 20 placebo) discontinued the study prior to reaching final height (height velocity <1.5 cm/y).  Twenty-five patients completed the protocol (16 Humatrope, 9 placebo) and were included in the Protocol Complete Population.

Final Height Population (n=33):  The 25 patients in the Protocol Complete Population form the core of the Final Height Population.  In addition, 8 patients in the Efficacy Evaluable Population, who discontinued the study prior to protocol completion (after a treatment duration averaging 2.7 years) but returned for a final height measurement, while still blinded to treatment assignment (6 Humatrope, 2 placebo), were included in the Final Height Population.  Therefore, there were 33 patients in the Final Height Population (22 Humatrope, 11 placebo).  In addition, 4 patients returned for a final height measurement but were excluded from the Final Height Population because they were not included in the Efficacy Evaluable Population for the following reasons:  did not receive study drug (1 Humatrope, 1 placebo), discontinued at Visit 1 (1 placebo), received GH after discontinuing from the study (1 placebo).

At the conclusion of the blinded phase of the study, there were 39 patients in the Efficacy Evaluable Population who had discontinued the protocol prior to protocol completion (19 Humatrope, 20 placebo).  Additional information was obtained regarding these patients from the NIH investigators regarding the efforts to obtain final height data for these 39 patients.  Of the 39 patients, 21 patients were still growing and were not eligible to be recalled for a final height measurement (11 Humatrope, 10 placebo).  Eighteen patients were known or considered likely to have a height velocity <1.5 cm/y and to be eligible to return for final height measurement (8 Humatrope, 10 placebo).  Six of the 18 patients (1 Humatrope, 5 placebo) were lost to follow-up and could not be contacted despite multiple attempts (phone calls, certified letters, and assistance of referring physicians).  Of the remaining 12 patients, 4 (1 Humatrope, 3 placebo) declined to return, and 8 patients (6 Humatrope, 2 placebo) returned and were included in the Final Height Population as described above.

The role of the Final Height Population and the Protocol Complete Population is to obtain a clinically interpretable assessment and estimate of treatment effect for patients with reasonably complete data.  The role of the modified intent-to-treat and intent-to-treat analyses are to demonstrate statistical existence of treatment effect and to verify that the estimates of treatment effect are comparable to estimates from the Final Height and the Protocol Complete Populations.  This is the paradigm by which Final Height, Protocol Complete, Efficacy Evaluable Populations and modified intent-to-treat and intent-to-treat analyses are complementary in confirming efficacy.

3.1.6.  Patient Disposition

Figure 3 illustrates patient disposition.

 

Figure 3.                       Patient disposition for Study GDCH.

The most common primary reason for early discontinuation in both treatment groups was patient decision.  Discontinuation due to Sponsor Decision (n=8) refers to the termination of the blinded treatment period in response to the DSMB recommendation of 05 June 2000.

3.1.7.  Baseline Patient Characteristics

Table 2 provides patient demographics at baseline.  The Humatrope and placebo treatment groups were well balanced and comparable at baseline in the five populations.  There were no significant differences between the Humatrope and placebo treatment groups for any of the variables in any of the patient populations.


Table 2.                         Demographics and Other Baseline Characteristics a
Study GDCH

 

All Randomized

Patients

 

Safety
Population

Variable

Humatrope

Placebo

 

Humatrope

Placebo

 

Number of patients

 

38

 

33

 

 

37

 

31

Male

29

26

 

29

24

Female

9

7

 

8

7

Ethnic origin

 

 

 

 

 

African descent

0

1

 

0

1

Asian

0

1

 

0

1

Caucasian

30

25

 

30

23

Hispanic

7

4

 

7

4

Other

1

2

 

0

2

Peak growth hormone concentration (μg/L)

16.2 ± 7.5

17.4 ± 9.7

 

16.3 ± 7.6

17.2 ± 9.8

IGF-I concentration SDS

-2.0 ± 1.1

-1.5 ± 1.5

 

-1.9 ± 1.1

-1.4 ± 1.5

Chronological age (y)

12.5 ± 1.6

12.3 ± 1.4

 

12.5 ± 1.6

12.2 ± 1.4

Bone age (y)

10.4 ± 1.9

10.4 ± 1.7

 

10.4 ± 1.9

10.3 ± 1.7

Height SDS

-2.7 ± 0.5

-2.8 ± 0.5

 

-2.8 ± 0.5

-2.8 ± 0.5

Pre-treatment height velocity (cm/y)

4.8 ± 1.8

4.8 ± 2.1

 

4.8 ± 1.8

4.8 ± 2.1

Predicted height SDS

-2.0 ± 0.8

-2.3 ± 0.8

 

-2.0 ± 0.8

-2.3 ± 0.8

Target height SDS b

-1.0 ± 1.0

-1.2 ± 0.7

 

-1.0 ± 1.0

-1.2 ± 0.8

                                                                                                                                                                                                    (continued)

Note:  Values represent mean ± standard deviation (SD).

Abbreviation:  IGF-I = insulin-like growth factor-I; SDS = standard deviation score.

a     There were no significant differences between the Humatrope and placebo treatment groups for any of the patient populations for any of the variables.

b    Target height represents the gender‑adjusted midparental height.

 


Table 2.                         Demographics and Other Baseline Characteristics a
Study GDCH (concluded)

 

 

Efficacy Evaluable
Population

 

Final Height
Population

 

Protocol Complete
Population

Variable

Humatrope

Placebo

 

Humatrope

Placebo

 

Humatrope

Placebo

 

Number of patients

 

35

 

29

 

 

22

 

11

 

 

16

 

9

Male

27

23

 

18

9

 

13

7

Female

8

6

 

4

2

 

3

2

Ethnic origin

 

 

 

 

 

 

 

 

African descent

0

1

 

0

1

 

0

1

Asian

0

0

 

0

0

 

0

0

Caucasian

29

22

 

18

7

 

14

5

Hispanic

6

4

 

4

1

 

2

1

Other

0

2

 

0

2

 

0

2

Peak growth hormone concentration (μg/L)

16.5 ± 7.7

17.3 ± 10.0

 

17.0 ± 7.7

17.6 ± 13.8

 

17.7 ± 7.6

18.6 ± 15.2

IGF-I concentration SDS

-1.9 ± 1.1

-1.5 ± 1.5

 

-1.8 ± 1.2

-1.7 ± 1.1

 

-1.9 ± 1.1

-1.8 ± 1.1

Chronological age (y)

12.5 ± 1.6

12.3 ± 1.3

 

12.5 ± 1.6

12.9 ± 1.1

 

12.4 ± 1.5

12.9 ± 1.2

Bone age (y)

10.4 ± 1.9

10.4 ± 1.6

 

10.4 ± 1.9

10.7 ± 1.1

 

10.2 ± 1.6

10.8 ± 1.2

Height SDS

-2.7 ± 0.5

-2.8 ± 0.5

 

-2.7 ± 0.6

-2.8 ± 0.6

 

-2.7 ± 0.6

-2.9 ± 0.6

Pre-treatment height velocity (cm/y)

4.9 ± 1.8

4.9 ± 2.1

 

5.2 ± 1.8

5.6 ± 2.4

 

5.2 ± 2.0

5.3 ± 2.2

Predicted height SDS

-2.0 ± 0.8

-2.3 ± 0.8

 

-2.1 ± 0.7

-2.3 ± 0.8

 

-2.2 ± 0.7

-2.4 ± 0.8

Target height SDS b

-0.9 ± 0.9

-1.2 ± 0.8

 

-1.1 ± 1.0

-1.3 ± 0.7

 

-1.1 ± 1.1

-1.4 ± 0.7

Note:  Values represent mean ± standard deviation (SD).

Abbreviation:  IGF-I = insulin-like growth factor-I; SDS = standard deviation score.

a     There were no significant differences between the Humatrope and placebo treatment groups for any of the patient populations for any of the variables.

b    Target height represents the gender‑adjusted midparental height.

 

 


3.1.8.  Efficacy Data

The protocol stated that the primary efficacy analysis would be an analysis of covariance (ANCOVA) of final height SDS, with baseline predicted height as the covariate, in the Final Height Population (n=33).

Protocol-specified sensitivity analyses included analysis of last observed height SDS by ANCOVA in the Efficacy Evaluable Population (n=64), which serves as a modified intent-to-treat analysis for the study, final height SDS by ANCOVA in the Protocol Complete Population (n=25); and final height minus baseline predicted height (cm) by t‑test in the Final Height Population.  In addition, a non-protocol specified, repeated measures analysis of height SDS at 18 years was included as an additional modified intent‑to‑treat analysis.  The repeated measures analysis uses repeated height SDS measurements over time, rather than just the last observed height SDS.  Lastly, intent-to-treat analyses, by both nonparametric and parametric methods, were performed in the Randomized Population (n=71).

3.1.8.1.    Primary Efficacy Analysis (Final Height SDS)

The primary efficacy variable was final height, expressed as a SDS relative to the general population of the same age and gender (final height SDS).  The primary efficacy analysis was of final height SDS for the Final Height Population.  Between‑group comparisons were performed using analysis of covariance (ANCOVA), with baseline predicted height SDS as the covariate (Table 3).  The two‑sided significance level for this analysis was set at a=0.05.

Table 3.                         Final Height Standard Deviation Score
Analysis of Covariance
Final Height Population
Study GDCH

 

 

Variable

 

Humatrope
(n=22)

 

Placebo
(n=10) a

Treatment

Effect b
(95% CI)

 

 

p-value

 

 

 

 

 

Final height SDS

     (ANCOVA using BPH     SDS as a covariate)

-1.81 ± 0.11

-2.32 ± 0.17

0.51 ± 0.20

(0.10-0.92)

0.017

Note:  Values represent least squares mean (LSM) ± standard error (SE).

Abbreviations:  ANCOVA = analysis of covariance; BPH = baseline predicted height; CI = confidence interval; n = number of patients; SDS = standard deviation score.

a       Only 10 patients were included in this analysis, as baseline predicted height was missing for 1 patient due to a missing baseline bone age x-ray.

b    Value represents the difference in the final height SDS between the Humatrope‑treated group and the placebo‑treated group.

 

The mean age at assessment of final height for the Final Height Population was 18.6 years for Humatrope‑treated patients and 19.1 years for placebo‑treated patients.

By ANCOVA, the patients who received Humatrope for 4.6 ± 1.6 (mean ± SD) years achieved a final height SDS of ‑1.81 ± 0.11 (least squares mean [LSM] ± standard error [SE]), while those who received placebo injections for 4.1 ± 1.7 years achieved a final height SDS of ‑2.32 ± 0.17, resulting in a mean Humatrope effect on final height SDS of 0.51 ± 0.20, 95% confidence interval (CI):  0.10 – 0.92 SDS (p=0.017).  The Humatrope effect of 0.51 SDS corresponds to a mean difference between groups of 3.7 cm.

Because one placebo patient in the Final Height Population was not included in the primary efficacy analysis due to missing data for the covariate (baseline predicted height SDS), the analysis was re-run for the full Final Height Population (n=33) by using a linear regression estimate for the missing baseline predicted height SDS value.  This analysis gave a similar mean Humatrope effect on final height SDS of 0.48 ± 0.19 SDS, 95% CI:  0.09 – 0.88 SDS (p=0.017).

The primary efficacy analysis was completed for all patients for whom final height data were available, including patients who discontinued the study before protocol completion (Final Height Population).  A number of patients in the Efficacy Evaluable Population either discontinued early and did not return for a final height measurement (n = 23) or remained in the study and were still growing at termination of the blinded treatment period (n = 8).  These patients are referred to as the Non‑Final Height subgroup of the Efficacy Evaluable Population (n = 31).  The Final Height Population (n = 33) and Non‑Final Height subgroup (n = 31) comprise the total Efficacy Evaluable Population (n = 64).  Since 31 of 64 patients from the Efficacy Evaluable Population were not available for final height measurement, the issue of potential dropout bias must be considered.

3.1.8.2.    Sensitivity Analyses

To address the potential dropout bias described above, two modified intent-to-treat analyses (Efficacy Evaluable Population) and four intent-to-treat analyses (Randomized Population) were performed to assess the robustness of the results of the primary analysis.

The first modified intent-to-treat analysis was an ANCOVA of last observed height SDS (using baseline predicted height SDS as the covariate) for the Efficacy Evaluable Population (Table 4).  The Humatrope effect for this analysis (0.52 ± 0.15 SDS, 95% CI:  0.22 – 0.82 SDS, p=0.001) was similar to that observed in the primary analysis (0.51 ± 0.20 SDS, p=0.017).

Table 4.                         Modified Intent-to-Treat Analysis
Efficacy Evaluable Population
Study GDCH

 

Analysis

Humatrope

n=35

Placebo

n=27 a

Treatment

Effect

 

p-value

 

 

 

 

 

Last observed height SDS (ANCOVA using BPH SDS as a covariate)

-1.89 ± 0.10

-2.40 ± 0.11

0.52 ± 0.15 b

0.001

 

 

 

 

 

Height SDS at age 18
(Repeated measures linear model)

-1.52 ± 0.11

-2.20 ± 0.12

0.69 ± 0.13 c

<0.0001

Note:  Values represent least squares mean (LSM) ± standard error (SE).

Abbreviations:  ANCOVA = analysis of covariance; BPH = baseline predicted height; n = number of patients; SDS = standard deviation score.

a     Two of the 29 patients in the placebo group did not have a baseline predicted height due to missing bone age x-rays and were not included in this analysis.

b    Value represents the difference in the last observed height SDS between the Humatrope-treated group and the placebo-treated group.

c     Value represents the difference in the height SDS at age 18 years between the Humatrope‑treated group and the placebo‑treated group.

 

To further address the issue of potential bias due to missing final height data, a repeated measures analysis of efficacy for the combined Final Height Population and Non‑Final Height subgroup of the Efficacy Evaluable Population was performed.  Repeated measures models are useful when repeated measurements are taken on the same patient and these measurements are correlated with each other.  This methodology is robust to the biases resulting from missing data (Verbeke and Molenberghs 2000).  Data were incorporated from 62 patients for whom baseline predicted height and all other necessary data for the statistical model were available.  A standard linear model would have used the endpoint height SDS values for each patient as the response variable, whereas this model used height SDS values throughout the course of the study (at ages 10-18).  Using these measured heights, the model estimated least squares mean height SDS at each age.  The comparison of interest was height SDS for Humatrope‑treated patients versus placebo‑treated patients at age 18 years.

Table 4 summarizes the results of the repeated measures analysis.  The mean effect of Humatrope on height SDS at age 18 years was 0.69 ± 0.13 SDS, 95% CI:  0.43 – 0.94 SDS (p<0.0001), corresponding to a mean between‑group height difference of 5.0 cm.  Thus, the two modified intent-to-treat analyses gave similar Humatrope treatment effects as the primary efficacy analysis.

To provide further evidence against dropout bias two nonparametric and two parametric intent-to-treat analyses of last observed height SDS for the entire Randomized Population (n=71) were performed.  The two non-parametric analyses were a rank analysis of covariance (ANCOVA) and a generalized Wilcoxon-Mann-Whitney test of last observed height SDS (Stokes et al. 2000).  The results of these analyses demonstrated that Humatrope was superior to placebo with p=0.0024 (rank ANCOVA) and p=0.0015 (generalized Wilcoxon-Mann-Whitney test).

As a second intent-to-treat approach, an ANCOVA, with baseline predicted height SDS as covariate, of last observed height SDS was performed.  This analysis yielded a Humatrope treatment effect of 0.40 ± 0.15 SDS (p=0.011) (Table 5).  For this analysis, the 5 missing baseline predicted height SDS values for the covariance analysis were imputed by linear regression using baseline height SDS and age as independent variables.  For patients missing postbaseline height data (n=2), their baseline height SDS was used as endpoint.  Without incorporating the effect of the covariate, ANOVA indicated a Humatrope treatment effect of 0.52 ± 0.17 SDS (p=0.003).  These modified intent-to-treat analyses (Efficacy Evaluable Population) and intent-to-treat analyses (Randomized Population), by their similarity to the primary efficacy analysis, provide strong evidence against dropout bias in the primary efficacy analysis.

Table 5.                         Intent-to-Treat Analyses of Last Observed Height SDS
All Randomized Population
Study GDCH

Analysis

Humatrope

n=38

Placebo

n=33

Treatment

Effect

p-value

ANCOVA (using BPH SDS as a covariate)

 

-1.96 ± 0.10

 

-2.36 ± 0.11

 

0.40 ± 0.15

 

0.011

 

ANOVA

 

-1.90 ± 0.11

 

-2.42 ± 0.12

 

0.52 ± 0.17

 

0.003

Abbreviations:  ANCOVA = analysis of covariance; ANOVA = analysis of variance; BPH = baseline predicted height; n = number of patients in treatment group; SDS = standard deviation score.

 

The close similarity of the treatment effect results in the Final Height Population (n=33), Efficacy Evaluable Population (n=64), and Randomized Population (n=71) indicates that similar conclusions about efficacy are supported by the analyses in all 3 populations.

Two additional protocol-specified sensitivity analyses were performed (Table 6).

Table 6.                         Analyses of Adult Height
Protocol Complete and Final Height Populations
Study GDCH

 

Analysis

 

Humatrope

 

Placebo

Treatment

Effect

 

p-value

Final Height Analysis
(PC Population)

 

n=16

 

n=9

 

 

Final height SDS a
(ANCOVA using BPH SDS as a covariate)

 

-1.86 ± 0.14

 

 

-2.32 ± 0.18

 

 

0.46 ± 0.23

 

0.061

 

 

 

 

 

Final Height Analysis

(FH Population)

 

n=22

 

n=10

 

 

Final height minus BPH (cm) b
(t-test)

 

2.15 ± 0.84

 

-0.67 ± 1.31

 

2.83 ± 1.53

 

0.075

Abbreviations:  ANCOVA = analysis of covariance; BPH = baseline predicted height; FH = Final Height; n = number of patients; PC = Protocol Complete; SDS = standard deviation score.

a     Values represent least squares mean (LSM) ± standard error (SE).

b    Values represent mean ± standard error (SE).

 

First, an ANCOVA of final height SDS in the Protocol Complete Population, with baseline predicted height SDS as the covariate, indicated a mean ± SE treatment effect of 0.46 ± 0.23 SDS, corresponding to 3.3 cm, p=0.061.  Second, a between-group t-test of final height minus baseline predicted height yielded a mean ± SE treatment effect of 2.83 ± 1.53 cm, p=0.075.  Both of these sensitivity analyses yielded treatment effects that were similar to those of the protocol-specified primary and modified intent-to-treat analyses.  Although both treatment effects failed to reach statistical significance, this was not surprising because of the reduction in statistical power due to reduced sample size for these subgroup analyses.

Given that intent-to-treat analyses are preferred in clinical trials, the question arises  ‘Why the primary analysis was restricted to the final height data, which were available from only a subset of randomized patients?’  The answer relates to uncertainty about how the between-group difference in height SDS (GH-treated versus control) would change over time.  Specifically, there was concern that GH might accelerate not just height velocity, but also epiphyseal fusion, perhaps producing an earlier attainment of the same final height rather than an increase in adult height.  Such a result would be manifest as a transient increase in height, relative to the control group, that would not be sustained because of the earlier cessation of growth in the GH-treated patients.  Thus, the maximum GH treatment effect on height SDS would occur during treatment, and the inclusion of non-final height SDS data in an intent-to-treat analysis could lead to an overestimate of what the GH treatment effect would be if measured only at final height.  Concern about the possibility of overestimating the GH treatment effect led to the decision to limit the primary analysis to patients with final height measurements.

To evaluate the possibility of an earlier epiphyseal fusion during GH treatment, we examined whether or not bone maturation was accelerated by Humatrope treatment.  Figure 4 illustrates bone age versus year on study for the Humatrope and placebo arms of the Final Height Population, the non-final height subgroup of the Efficacy Evaluable Population, and the entire Efficacy Evaluable Population.  In each of these groups, there were no significant differences in bone age between the Humatrope and placebo arms.  Thus, there is no evidence that this GH treatment regimen advanced the tempo of skeletal maturation.  Consistent with these observations, there was no between-group difference in the mean age at which final height was attained (Humatrope, 18.6 ± 0.4 years; placebo, 19.1 ± 0.4 years, p=0.43).

Abbreviations:  EE = Efficacy Evaluable; FH = Final Height; H = Humatrope; n = number of patients; P = placebo; SE = standard error.

Figure 4.                       Bone age versus year on study for Study GDCH.

From these data, one would not predict that the GH-induced gains in height SDS would transiently increase and then decline, since bone age was not accelerated.  To examine the actual time-course of height SDS gains, Figure 5 shows the increase in height SDS for patients in the two treatment arms, plotted against the year before the last observed height SDS for each patient.  The time at which final height or last observed height was measured for each patient was set equal to zero to synchronize the observations in relation to the last height observation.  This allows one to focus on the between-group differences in height SDS gain during the several years before measurement of final height or last observed height.

Abbreviations:  EE = Efficacy Evaluable; FH = Final Height; n= number of patients; SE = standard error.

Figure 5.                       Increase in height SDS over baseline versus year on study relative to last observed height (year=0) in Study GDCH.

The temporal pattern of between-group differences in height SDS gain (a measure of GH treatment effect) was an increase during the early years of treatment, followed by stabilization of the GH treatment effect during the 3 years before measurement of final height or last observed height (Figure 5).  For the Final Height Population (Figure 5, left panel), the mean GH treatment effect on height SDS gain was 0.42 SDS, 3 years before final height measurement (at a mean age of 15.5 years), and 0.51 SDS at final height measurement (mean age 18.8 years).  For the non-final height subgroup of the Efficacy Evaluable Population (Figure 5, middle panel), the mean GH treatment effect at last observed height was 0.55 SDS (at a mean age of 15.1 years and mean treatment duration of 3.0 years).  Thus, the mean GH treatment effect for the non-final height subgroup at last observed height was similar to that for the Final Height Population at final height measurement.  After combining these two groups, which comprise the Efficacy Evaluable Population (Figure 5, right panel), the mean GH treatment effect on height SDS gain was 0.5495 SDS, 3 years before last observed height (at a mean age of 14.4 years), and 0.55 SDS at last observed height (mean age 17.0 years).  Thus, once the between-group difference in height SDS gain had stabilized, after approximately 3 years of treatment, the GH treatment effect remained stable during the several years until attainment of final height.

The evidence in Figure 5 against a transient GH treatment effect removes the principal objection to inclusion of non-final height data in the analysis of GH treatment effect.  Indeed, based on the data of Figures 4 and 5, the earlier concern that inclusion of such data would produce an overestimation of the GH treatment effect is not justified for the treatment regimen used in this study.  Moreover, the efficacy analyses described previously showed similar treatment effects for the primary efficacy analysis in the Final Height Population, for the modified intent-to-treat analyses in the Efficacy Evaluable Population, and for the intent-to-treat analyses in the Randomized Population.  From these observations we conclude that the Humatrope treatment effect is robust across the different study populations and that the primary efficacy analysis shows no indication of having been affected by dropout bias.

One potential misinterpretation of Figure 5 deserves comment.  The fact that the GH treatment effect, during continued GH administration, stabilizes in the 3 years before attaining final height does not imply that the treatment effect would remain stable if GH treatment were to be discontinued 3 years before attaining final height.  Studies have shown that such discontinuation is followed by a “catch-down” deceleration of height velocity to levels below those of the general population (Zadik et al. 1996; Lampit et al. 1998).  Up to 18 months may be required before height velocity returns to baseline levels.  For this reason GH administration is generally continued until a satisfactory adult height has been achieved or until the decline in height velocity indicates that near-final height has been attained.

To illustrate the mean height SDS for the two treatment arms in the initial years of treatment, Figure 6 displays height SDS by year on study for the Efficacy Evaluable Population.

Note:  This population includes all patients who received ³6 months study drug, whether or not they achieved final height.  Data are cross‑sectional.  The dashed line at –2 SDS represents the lower limit of the normal range for the general population (AAP 1997).

Abbreviation:  H = Humatrope; n = number of patients; P = placebo; SDS = standard deviation score.

Figure 6.                       Height standard deviation score by year on study for the Efficacy Evaluable Population (Study GDCH).

By 2 years of treatment, the mean height SDS for the Humatrope-treated patients was close to the lower limit of the normal range.  Discontinuation rates were similar during the early years of the study.  For example, at 3 years of treatment, the continuation rate was 70% (23 out of 33 patients) for the placebo group compared to 68% (26 out of 38 patients) for the Humatrope group.

3.1.8.3.    Additional Analyses of Interest

Tables 7-9 provide additional supportive evidence for efficacy of Humatrope in the Efficacy Evaluable, Final Height, and Protocol Complete Populations.  Statistically significant differences or trends in final height characteristics were observed between the treatment groups.  All of these analyses support the conclusion that Humatrope increases final height in pediatric patients with non-GHD short stature.


Table 7.                         Additional Endpoint Height Analyses
Efficacy Evaluable Population
Study GDCH

 

 

Analysis

Humatrope
(n=35)

Placebo
(n=29)

Treatment

Effect a (95% CI)

 

p-value

Endpoint Height (cm) - Baseline Predicted Height (cm) b

-2.20 ± 1.94

-6.02 ± 1.77

3.82  (-1.59 –  9.23)

0.163

Endpoint Height SDS - Baseline Predicted Height SDS b

0.13 ± 0.13

-0.21 ± 0.13

0.34  (-0.04 –  0.72)

0.079

Endpoint Height (cm)

157.14 ± 1.93

150.86 ± 1.90

6.27  (0.80 – 11.75)

0.025

Height Gain (cm) (Endpoint Height - Baseline Height) c

24.27 ± 1.56

19.55 ± 1.36

4.72  (0.49 –  8.95)

0.029

Endpoint Height SDS

-1.83 ± 0.12

-2.45 ± 0.10

0.62  (0.29 –  0.95)

0.000

Height SDS Gain (Endpoint Height SDS - Baseline Height SDS) c

0.91 ± 0.11

0.36 ± 0.07

0.55  (0.27 –  0.83)

0.000

Target Height (cm) - Endpoint Height (cm)

9.32 ± 1.89

13.04 ± 2.23 d

-3.73  (-9.59 –  2.14)

0.209

Target Height SDS - Endpoint Height SDS

0.88 ± 0.16

1.22 ± 0.17 d

-0.34  (-0.81 –  0.13)

0.156

Note:  Values represent mean ± standard error.  P-values are from t-tests.

Abbreviation:  CI = confidence interval; n = number of patients; SDS = standard deviation score.

a     Value represents the mean difference between the Humatrope-treated group and the placebo-treated group.

b    Two of the 29 patients in the placebo group did not have a baseline predicted height, due to missing bone age x-rays, and therefore could not
be included in this analysis.

c     Height gain is from start of treatment to endpoint height.

d    Four of the 29 patients in the placebo-treated group did not have a target height.


Table 8.                         Additional Final Height Analyses
Final Height Population
Study GDCH

 

Analysis

Humatrope
(n=22)

Placebo
(n=11)

Treatment

Effect a (95% CI)

 

p-value

Final Height SDS – Baseline Predicted Height SDS

0.32 ± 0.12

-0.14 ± 0.19 b

0.46 (0.02 –  0.89)

0.043

Final Height (cm)

161.12 ± 1.58

157.46 ± 1.77

3.66 (-1.58 –  8.90)

0.165

Height Gain (cm) (Final Height – Baseline Height) c

28.30 ± 1.57

22.58 ± 2.08

5.71 (0.27 – 11.15)

0.040

Final Height SDS

-1.77 ± 0.17

-2.34 ± 0.17

0.57 (0.03 –  1.10)

0.039

Height SDS Gain (Final Height SDS-Baseline Height SDS) c

0.93 ± 0.16

0.42 ± 0.07

0.51 (0.04 –  0.97)

0.034

Target Height (cm) – Final Height (cm)

4.71 ± 1.37

7.10 ± 1.81 d

-2.39 (-7.23 – 2.45)

0.321

Target Height SDS – Final Height SDS

0.66 ± 0.19

1.02 ± 0.25 d

-0.36 (-1.04 – 0.31)

0.280

Note:  Values represent mean ± standard errors.  P-values are from t-tests.

Abbreviation:  CI = confidence interval; n = number of patients; SDS = standard deviation score.

a     Value represents the mean difference between the Humatrope-treated group and the placebo-treated group.

b    n=10 for placebo, as one patient did not have a baseline predicted height due to missing bone age x-ray.

c     Height gain is from start of treatment to final height.

d    n=10 for placebo, as one patient did not have a target height value reported.


Table 9.                         Additional Final Height Analyses
Protocol Complete Population
Study GDCH

 

 

Analysis

Humatrope

(n=16)

Placebo
(n=9)

Treatment

Effect a (95% CI)

 

p-value

Final Height (cm) – Baseline Predicted Height (cm)

2.40 ± 1.05

-0.04 ± 1.29

2.44 (-1.08 – 5.96)

0.165

Final Height SDS – Baseline Predicted Height SDS

0.35 ± 0.15

-0.04 ± 0.18

0.40 (-0.10 – 0.89)

0.111

Final Height (cm)

160.69 ± 1.73

156.38 ± 1.96

4.31 (-1.37 – 9.98)

0.130

Height Gain (cm) (Final Height – Baseline Height) b

28.81 ± 1.69

22.15 ± 2.45

6.66 (0.65 – 12.68)

0.031

Final Height SDS

-1.81 ± 0.20

-2.41 ± 0.19

0.59 (-0.03 – 1.22)

0.062

Height SDS Gain (Final Height SDS – Baseline Height SDS) b

0.91 ± 0.16

0.45 ± 0.07

0.46 (-0.01 – 0.93)

0.055

Target Height (cm) – Final Height (cm)

5.44 ± 1.67

7.25 ± 2.02

-1.81 (-7.39 – 3.77)

0.509

Target Height SDS – Final Height SDS

0.75 ± 0.23

1.03 ± 0.28

-0.28 (-1.06 – 0.50)

0.459

Note:  Values represent mean ± standard error.  P-values are from t-tests.

Abbreviation:  n = number of patients; SDS = standard deviation score.

a     Value represents the mean difference between the Humatrope-treated group and the placebo-treated group.

b    Height gain is from start of treatment to final height.

 


3.1.9.  Efficacy Summary

The hypothesis of Study GDCH was that treatment with Humatrope would increase the adult height of patients with non-GHD short stature.  The primary analysis supports this hypothesis.  By ANCOVA, patients in the Final Height Population who received Humatrope had a significantly greater mean final height SDS than those who received placebo, indicating a Humatrope effect on final height SDS of 0.51 SDS, which corresponds to 3.7 cm.  Four additional analyses (3 protocol-specified and 1 nonprotocol‑specified; Section 3.1.8), which were performed to investigate the robustness of the results from the primary analysis, also support the efficacy of Humatrope in patients with non-GHD short stature.  These included two modified intent-to-treat analyses in the Efficacy Evaluable Population, which indicated a mean Humatrope effect of 3.8 to 5.0 cm, and two sensitivity analyses, in subpopulations of the Efficacy Evaluable Population, which indicated a mean treatment effect of 2.8 to 3.3 cm.  Lastly, intent-to-treat analyses, by both nonparametric and parametric methods, confirmed the significantly greater height SDS of the Humatrope treated patients.  These gains in height SDS were achieved without any untoward effect on skeletal maturation or pubertal development.  In conclusion, Humatrope increases the adult height of patients with non-GHD short stature.

3.2.   Supportive Study:  B9R-EW-E001

Study E001 was a multicenter, randomized, dose-response study conducted in 10 European countries.

3.2.1.  Objectives

The primary objective of this study was to determine the ability of two different Humatrope dosages (0.24 mg/kg/wk versus 0.37 mg/kg/wk) to increase height velocity during the first 2 years of treatment in patients with non-GHD short stature.

A secondary objective of this study was to determine whether increasing the dosage of Humatrope from 0.24 mg/kg/wk to 0.37 mg/kg/wk for the second year of treatment would sustain the first-year increase in height velocity during the second year of treatment.  Ordinarily, the improvement in height velocity over baseline is smaller in the second year of treatment than in the first year.  An additional secondary objective was to assess the long-term effect of different Humatrope dosages in patients who were followed to final height (defined as height measured after height velocity had fallen below 2 cm/y).  The less stringent final height criterion for Study E001 compared to Study GDCH (height velocity <2 cm/y versus <1.5 cm/y) reflected the independent design of the two studies and the lack of a generally accepted criterion for final height.

As was common in clinical studies in the 1980s, the original analytical plans were described at a high level rather than as detailed statistical analysis plans.  Furthermore, because the clinical study report for Study E001 was prepared after the investigators had published a substantial amount of data from this unblinded study, a prospective statistical analysis plan was not possible.  For consistency with the pivotal study, the analyses performed for Study E001 conform as closely as possible, given differences in study design, to those performed for Study GDCH.

3.2.2.  Study Design

The core study consisted of a screening phase of up to 12 months, during which patients were assessed for study eligibility, followed by a 2-year, three‑arm, randomized, open‑label, dose‑response phase (Figure 7).  Patients were randomized to one of three Humatrope treatment arms:  Dose 1 = 0.24 mg/kg/wk; Dose 2 = 0.24 mg/kg/wk for the first year, then 0.37 mg/kg/wk (abbreviated as 0.24→0.37); Dose 3 = 0.37 mg/kg/wk.  Humatrope was given in divided doses 6 times per week.  Participation in the core dose‑response phase of Study E001, for which the primary endpoint was the increase in height velocity measured from 0 to 2 years, ended when the patient completed 2 years on treatment (Visit 10).

 

Note:  Visit 11 occurred 12 months after Visit 10.  Visit 11 procedures were repeated every 12 months until final height was attained.

Figure 7.                       Design of Study E001.

3.2.3.  Inclusion/Exclusion Criteria

Patients were included who were prepubertal and had chronological age ³5 years, height SDS £‑2.0, plasma GH peak above 20 mU/L (10 mg/L) in response to a standard stimulation test, bone age <10 years (females) or <12 years (males), height velocity below the 25th percentile for age before age 10 years for girls and age 12 years for boys (or, if above these age limits, below the 25th percentile for bone age), and normal thyroid function.  The different cutoffs used to determine sufficient GH secretion in Study E001 versus Study GDCH (>10 μg/L versus >7 μg/L, respectively) reflect the lack of a generally accepted criterion due to differences in assay characteristics and in the particular stimulation protocol employed (DTCLWPES 1995; Sizonenko et al. 2001).  Patients were excluded if they had a chronic illness, a known genetic syndrome, had ever received GH, or were currently receiving other drugs likely to affect growth.

3.2.4.  Population Definitions

All Randomized Patients (n=239):  Of 261 patients who entered the study, 22 were ineligible or discontinued prior to randomization.  The remaining 239 patients comprise the All Randomized Patients population:  Dose 1, n=78; Dose 2, n=78; Dose 3, n=83.

Two-Year Height Velocity Population (n=209):  Of the 239 randomized patients, 30 (13%) discontinued prior to reaching the primary 2-year endpoint.  The remaining 209 patients (87% of randomized patients) comprise the Two-Year Height Velocity Population:  Dose 1, n=70; Dose 2, n=67; Dose 3, n=72.

Final Height Population (n=50):  Of the 209 patients who completed the core 2‑year height velocity phase of the study 173 entered the final height extension phase.  Fifty of these patients attained final height on study, or returned for final height measurement at a post-study follow-up:  Dose 1, n=17; Dose 2, n=16; Dose 3, n=17.  This includes one patient (Dose 1) who discontinued the study at Visit 3 (prior to 1 year of treatment) who was followed to final height post-treatment and was included in the Final Height Population.

3.2.5.  Patient Disposition

Figure 8 illustrates patient disposition.

 

a           One patient who discontinued at Visit 3 was followed to final height post-treatment and was therefore also included in the Final Height Population.

Figure 8.                       Patient disposition for Study E001.

 

3.2.6.  Baseline Patient Characteristics

Table 10 provides baseline demographics for All Randomized Patients, Two‑Year Height Velocity Population, and Final Height Population.

 


Table 10.                       Demographics and Other Baseline Characteristics
Study E001

Population

All Randomized Patients

Two-Year Height Velocity

Final Height

Humatrope Dosage
(mg/kg/wk)

Dose 1
0.24

Dose 2
0.24→0.37

Dose 3
0.37

Dose 1
0.24

Dose 2
0.24→0.37

Dose 3
0.37

Dose 1
0.24

Dose 2
0.24→0.37

Dose 3
0.37

 

Number of patients

 

78

 

78

 

83

 

70

 

67

 

72

 

17

 

16

 

17

Male

49

50

59

45

43

51

11

9

11

Female

29

28

24

25

24

21

6

7

6

Ethnic origin

 

 

 

 

 

 

 

 

 

Asian

0

2

0

0

2

0

0

0

0

Caucasian

78

76

83

70

65

72

17

16

17

Peak GH concentration (μg/L)

 

16.8 ± 7.5

 

17.6 ± 9.9

 

17.0 ± 6.2

 

16.6 ± 6.8

 

17.5 ± 10.5

 

16.9 ± 6.1

 

14.2 ± 4.7

 

16.8 ± 11.8

 

14.3 ± 3.4

IGF-I concentration (μg/L)

89.0 ± 44.4

100.2 ± 65.5

99.4 ± 48.6

88.9 ± 45.4

103.8 ± 67.4

102.6 ± 49.1

80.4 ± 32.7

113.2 ± 54.6

109.1 ± 61.3

Chronological age (y)

9.4 ± 2.4

9.9 ± 2.2

10.0 ± 2.2

9.4 ± 2.5

9.8 ± 2.1

10.0 ± 2.2

10.4 ± 2.3

10.4 ± 2.1

10.2 ± 2.1

Bone age (y)

7.4 ± 2.6

8.1 ± 2.3

8.0 ± 2.1

7.4 ± 2.6

8.0 ± 2.3

8.0 ± 2.0

8.5 ± 2.1

8.5 ± 2.1

8.9 ± 1.9

Height SDS a

-3.4 ± 0.8

  -3.2 ± 0.7

  -3.0 ± 0.5

  -3.4 ± 0.8

  -3.2 ± 0.7

  -3.0 ± 0.5

-3.3 ± 0.8

-3.1 ± 0.8

-2.9 ± 0.6

Predicted height SDS a

-2.7 ± 1.0

  -2.8 ± 1.1

  -2.4 ± 1.1

  -2.8 ± 1.0

  -2.9 ± 1.0

  -2.3 ± 1.1

-2.5 ± 1.1

-2.6 ± 0.9

-2.3 ± 0.9

Target height SDS b

-1.3 ± 0.9

-1.2 ± 1.0

-1.2 ± 0.9

-1.3 ± 0.9

-1.1 ± 0.9

-1.2 ± 0.9

-1.2 ± 1.1

-0.8 ± 1.1

-0.9 ± 0.9

Pretreatment height velocity (cm/y)

 

4.3 ± 1.1

 

4.4 ± 1.3

 

4.3 ± 1.1

 

4.2 ± 1.1

 

4.5 ± 1.3

 

4.4 ± 1.1

 

4.7 ± 1.4

 

5.1 ± 2.0

 

4.4 ± 1.5

Note:  Values represent mean ± standard deviation (SD).

Abbreviation:  GH = growth hormone; IGF-I = insulin-like growth factor-I; SDS = standard deviation score.

a     There were statistically significant differences (p<0.05) among the three dose groups for the All Randomized and the Two-Year Height Velocity populations.

b    Target height represents the gender‑adjusted midparental height.

 


There were significant differences (p<0.05) among the Humatrope dosage groups for height SDS and predicted height SDS for All Randomized Patients and Two‑Year Height Velocity Population.  To account for this, analyses were performed using baseline predicted height SDS as a covariate.  There were no other statistically significant differences among the dosage groups for baseline characteristics.

3.2.7.  Efficacy Data

3.2.7.1.    Dose-Response Effect on Height Velocity

The primary efficacy variable was increase in height velocity (cm/y) from baseline to 2‑year endpoint.  The protocol-specified primary efficacy analysis was of the difference in height velocity increase between the group that received the Humatrope dosage of 0.24 mg/kg/wk and the group that received 0.37 mg/kg/wk.  Between‑group comparisons were performed using analysis of variance (ANOVA) with a two‑sided significance level of 0.05.

Table 11 presents the effect of Humatrope dosage on height velocity from pretreatment to 2‑year endpoint.

Table 11.                       Height Velocity Changes from Pretreatment to 2-Year Endpoint
Two-Year Height Velocity Population
Study E001

Humatrope Dosage
(mg/kg/wk)

Dose 1

0.24

Dose 2

0.24®0.37

Dose 3

0.37

 

 

 

 

Number of Patients

68

66

71

 

 

 

 

Baseline

 

 

 

 

Height Velocity (cm/y)

 

      4.23 ± 0.14

 

         4.45 ± 0.14

 

4.35 ± 0.14

Difference (cm/y)

Dose 2 – Dose 1

Dose 3 – Dose 2

Dose 3 – Dose 1

 

0.23 ± 0.20

-0.10 ± 0.20

0.12 ± 0.20

p-value

0.264

0.608

0.534

 

 

 

 

Endpoint

 

 

 

Height Velocity (cm/y)

 

      7.49 ± 0.16

 

7.61 ± 0.16

 

8.39 ± 0.16

Effect (cm/y)

Dose 2 – Dose 1

Dose 3 – Dose 2

Dose 3 – Dose 1

 

0.11 ± 0.23

0.78 ± 0.23

0.90 ± 0.23

p-value

0.619

0.001

<0.001

 

 

 

 

Change

 

 

 

 

Height Velocity (cm/y)

 

      3.27 ± 0.18

 

3.16 ± 0.19

 

4.04 ± 0.18

Effect (cm/y)

Dose 2 – Dose 1

Dose 3 – Dose 2

Dose 3 – Dose 1

 

-0.11 ± 0.26

0.89 ± 0.26

0.78 ± 0.26

p-value

0.672

0.001

0.003

 

 

 

 

Note:  Values represent least squares mean ± standard error (SE).

 

By ANOVA, patients who received 0.37 mg/kg/wk Humatrope achieved a significantly greater pretreatment to 2‑year endpoint increase in height velocity than patients who received 0.24 mg/kg/wk (dose effect = 0.8 cm/y, 95% CI:  0.3 – 1.3 cm/y, p=0.003) or those who received 0.24 mg/kg/wk for the first year and then 0.37 mg/kg/wk thereafter (mean ± SE dose effect = 0.9 ± 0.3 cm/y, p=0.001).  There was no statistically significant difference in height velocity change between the 0.24 mg/kg/wk and the 0.24®0.37 mg/kg/wk groups (p=0.672).

3.2.7.2.    Dose-Response Effect on Height SDS

To evaluate long-term outcome in the broader population of study patients, ANCOVA and repeated measures analyses of height SDS were performed for the Two‑Year Height Velocity Population (Table 12).  These analyses are analogous to those performed in the Efficacy Evaluable Population of Study GDCH (Table 4).  The number of patients available for these analyses are less than for the primary height velocity endpoint because both the ANCOVA and repeated measures analysis used baseline predicted height SDS as a covariate.  Some children were too young to perform a height prediction because the Bayley-Pinneau method requires a minimum bone age of 6 to 7 years (depending on gender and relation of chronologic age to bone age).

Table 12.                       Secondary Efficacy Analyses
Two‑Year Height Velocity Population
Study E001

Humatrope Dosage
(mg/kg/wk)

 

Dose 1

0.24

 

Dose 2

0.24®0.37

 

Dose 3

0.37

 

Dose

Effect

 

p-value
(Dose 1 vs Dose 3)

 

 

 

 

 

 

Variable

 

 

 

 

 

ANCOVA

 

 

 

 

 

n

39

52

48

 

 

Last observed
height SDS a

 

-1.95 ± 0.13

 

-1.87 ± 0.12

 

-1.45 ± 0.12

 

0.51 ± 0.18 b

 

0.006

 

 

 

 

 

 

Repeated measures

 

 

 

 

n

39

52

47

 

 

Height SDS at age 18 years c

 

-1.26 ± 0.16

 

-1.56 ± 0.15

 

-0.82 ± 0.14

 

0.44 ± 0.17 d

 

0.012

Abbreviations:  ANCOVA = analysis of covariance; n = number of patients who had a baseline predicted height measurement, required for the ANCOVA; SDS = standard deviation score; vs = versus.

a     Data are expressed as least squares mean (LSM) ± standard error (SE) from ANCOVA, with baseline predicted height (BPH) SDS as the covariate.

b    Value represents the difference in the endpoint height SDS between the Dose 1 group and the Dose 3 group.

c     Data are expressed as least squares mean (LSM) ± standard error (SE) from repeated measures linear model for measured or estimated height SDS at age 18 years. 

d    Value represents the difference in the height SDS at age 18 years between the Dose 1 group and the Dose 3 group.

 

Patients in the Two-Year Height Velocity Population had a mean age of 15 years at last observed height SDS after a mean treatment period of 5.1 years.  By ANCOVA, patients in the Two-Year Height Velocity Population who received 0.37 mg/kg/wk Humatrope had a greater last observed height SDS than those who received 0.24 mg/kg/wk (p=0.006).  The mean between-dose effect size, that is, the incremental increase in last observed height SDS for the 0.37 mg/kg/wk dose group versus the 0.24 mg/kg/wk group, was 0.51 SDS (corresponding to 3.3 cm), 95% CI:  0.15 – 0.87 SDS (p=0.006).

By repeated measures analysis of height SDS at age 18 years, for the same patients, the mean between-dose effect size was 0.44 SDS (corresponding to 2.8 cm), 95% CI:  0.10 – 0.78 SDS (p=0.012).  These data indicate that the higher dose resulted in last observed height SDS and height SDS at age 18 years that were 2.8 to 3.3 cm greater for the 0.37 mg/kg/wk than for the 0.24 mg/kg/wk dose group.

The analyses of the between-dose effect size do provide insight into the overall height SDS gain in the 0.37 mg/kg/wk dosage group.  The overall treatment effect of the 0.37 mg/kg/wk dosage can be conceptualized as the between-dose effect of the 0.37 mg/kg/wk dosage versus the 0.24 mg/kg/wk dosage (2.8 to 3.3 cm in the above analyses) plus the treatment effect of the lower 0.24 mg/kg/wk dosage.  The treatment effect of the lower dose could be estimated roughly, based upon the results of the slightly lower (0.22 mg/kg/wk) dosage utilized in Study GDCH.  The treatment effect in Study GDCH was 3.7 cm, suggesting that the overall treatment effect of the 0.37 mg/kg/wk dosage would be approximately 6 to 7 cm, or 1 SDS.

Table 13 provides the between-group dose effect analysis for final height SDS in the Final Height Population.

Table 13.                       Final Height Standard Deviation Score
Analysis of Covariance
Final Height Population
Study E001

 

Humatrope Dosage (mg/kg/wk)

 

Dose 1
(0.24)

 

Dose 2
(0.24→0.37)

 

Dose 3

(0.37)

 

Dose

Effect a

p-value
(Dose 1 vs Dose 3)

Final Height Analysis

n=13

n=13

n=13

 

 

 

Final height SDS (ANCOVA using BPH SDS as a covariate)

 

-1.65 ± 0.18

 

-1.38 ± 0.18

 

-1.19 ± 0.18

 

0.45 ± 0.26

 

0.086

Note:  Values represent least squares mean ± standard error (SE).

Abbreviations:  ANCOVA = analysis of covariance; BPH = baseline predicted height; n = number of patients; SDS = standard deviation score.

a     Value represents the difference in final height SDS between Dose 3 and Dose 1.

 

After a mean treatment period of 6.5 years for the Final Height Population, and at a mean age of 18 years, the mean between-dose effect (0.37 mg/kg/wk versus 0.24 mg/kg/wk) on final height SDS, by ANCOVA with baseline predicted height SDS as the covariate, was 0.45 SDS, corresponding to 2.9 cm (p=0.09).  This between-dose effect size was similar to the dose effect on last observed height SDS and on height SDS at age 18 years for the patients who completed 2 years of treatment (0.51 and 0.44 SDS, respectively).  Although the effect did not reach statistical significance, this was not surprising because of the reduction of statistical power due to small sample size in this subgroup analysis.

As mentioned in relation to Study GDCH, the major rationale for continuing GH treatment studies to adult height has been the concern that GH might accelerate bone maturation, epiphyseal fusion, and cessation of linear growth, producing an earlier attainment of the same final height rather than an increase in adult height.  To address this concern, we examined whether the higher GH dosage of 0.37 mg/kg/wk accelerated bone maturation compared to the 0.24 mg/kg/wk dosage (Figure 9).

 

Abbreviations:  Dose 1 = 0.24 mg/kg/wk; Dose 2 = 0.24 →0.37 mg/kg/wk; Dose 3 = 0.37 mg/kg/wk; FH = final height; HV = height velocity; n = number of patients; SE = standard error; Yr = year.

Figure 9.                       Bone age versus year on study in Study E001.

Figure 9 shows bone age versus year on study for the 0.24 mg/kg/wk, 0.24→0.37 mg/kg/wk, and 0.37 mg/kg/wk dosage groups of the Final Height Population (left panel), for patients who completed 2 years of treatment but did not have a final height measurement (middle panel), and for the entire Two-Year Height Velocity Population (right panel).  In each population or subgroup, there were no apparent between-dose effects on the rate of bone age progression.  Thus, the higher dosage of 0.37 mg/kg/wk had no discernible effect on bone age progression compared to the 0.24 mg/kg/wk dosage.

3.2.7.3.    Significant Treatment Effect on Final Height

The previous section focused on the between-dose effect size on height velocity and on height SDS.  This section will assess the overall treatment effect on final height for each of the three dosage groups.  Because there was no untreated control group in this study, this was done by comparing the final height of patients in each dose group with the height that they were predicted to achieve without treatment.

Table 14 provides a summary of final height characteristics for the Final Height Population.  Mean duration of treatment was 6.1 ± 2.3, 6.3 ± 2.2, and 7.0 ± 2.0 years for the 0.24 mg/kg/wk, 0.24®0.37 mg/kg/wk, and 0.37 mg/kg/wk groups, respectively.

Table 14.                       Final Height Characteristics
Final Height Population
Study E001

Humatrope Dosage
(mg/kg/wk)

Dose 1

0.24

Dose 2

0.24®0.37

Dose 3

0.37

 

 

 

 

Number of patients

13

13

13

FH - BPH (cm) a

5.36 ± 0.89

6.66 ± 1.14

7.21 ± 1.66

p-value b

<0.001

<0.001

0.001

 

 

 

 

Number of patients

17

16

17

FH SDS - BH SDS a

1.55 ± 0.14

1.52 ± 0.27

1.85 ± 0.20

p-value b

<0.001

<0.001

<0.001

 

 

 

 

Number of patients

17

16

17

TH - FH (cm) a

3.78 ± 1.78

5.31 ± 2.42

1.33 ± 1.21

p-value b

0.050

0.045

0.288

Abbreviations:  BH = baseline height; BPH = baseline predicted height; FH = final height; SDS = standard deviation score; TH = target height.

a     Data are expressed as mean ± standard error (SE).

b    p‑values refer to a within‑group t test of the null hypothesis that mean value equals zero.

 

All three treatment groups showed a significant treatment effect, as evidenced by mean final height minus baseline predicted height that ranged from 5.4 to 7.2 cm (lower dose to higher dose) and mean final height SDS minus baseline height SDS that ranged from 1.6 to 1.9 SDS (lower dose to higher dose).  Furthermore, patients who received the Humatrope dosage of 0.37 mg/kg/wk reached a final height that was not significantly below target height (gender‑adjusted midparental height), indicating that they came close to achieving their genetic potential for height.

The validity of final height minus baseline predicted height as an efficacy measure depends on the accuracy of the Bayley-Pinneau method in predicting the adult height that patients with non-GHD short stature would have achieved in the absence of treatment.  Since published studies in approximately 400 untreated patients show that the actual adult height of untreated patients falls short of the Bayley-Pinneau predicted height by up to 5 cm in males [Bramswig et al. 1990; Ranke et al. 1995; Buchlis et al. 1998; Rekers-Mombarg et al. 1999]), final height minus baseline predicted height should provide a conservative estimate of treatment effect.  In Study GDCH, placebo-treated patients failed to achieve their baseline predicted height, consistent with the published studies cited above, whereas patients treated with 0.37 mg/kg/wk, administered in divided doses 6 times per week, exceeded their baseline predicted height by 7.2 cm (Figure 10).

 

Note:  Values represent mean ± SE

Abbreviations:  n = number of patients; SE = standard error.

Figure 10.                     Final height minus baseline predicted height (cm) in the Final Height Populations of Studies GDCH and E001.

The above data support the validity of final height minus baseline predicted height as a conservative measure of GH treatment effect in patients with non-GHD short stature.  From these data we conclude that the mean gain in adult height attributable to GH treatment with the 0.37 mg/kg/wk dosage is at least 7.2 cm.

3.2.8.  E001 Efficacy Summary

The primary objective of this study was to determine the efficacy of two different Humatrope dosages (0.24 mg/kg/wk versus 0.37 mg/kg/wk) in stimulating an increase in height velocity during the first 2 years of treatment in patients with non-GHD short stature.  By ANOVA, patients who received 0.37 mg/kg/wk Humatrope had a greater increase in height velocity after 2 years of treatment than patients who received 0.24 mg/kg/wk (between-dose effect:  0.8 cm/y, 95% CI:  0.3 – 1.3 cm/y, p=0.003).  Furthermore, patients who received 0.37 mg/kg/wk from the start of treatment had a significantly greater increase in height velocity than those who received 0.24 mg/kg/wk for the first year, followed by 0.37 mg/kg/wk for the second year (between-dose effect: 0.9 ± 0.3 (SE) cm/y, p=0.001).  Thus, the primary efficacy analysis supported the hypothesis that the dosage of 0.37 mg/kg/wk is more effective in increasing two-year height velocity than either of the lower dosage regimens.

Secondary analyses indicated a mean between-dose effect (incremental effect of 0.37 mg/kg/wk versus 0.24 mg/kg/wk) on last observed height SDS, height SDS at 18 years, and final height SDS corresponding to 3.3 cm, 2.8 cm, and 2.9 cm, respectively.  Each of these analyses was statistically significant except for the last analysis (p=0.09), which had reduced statistical power because of the smaller size of the Final Height Population.  From the above dose-response analyses, we conclude that the dosage of 0.37 mg/kg/wk is more effective than the dosage of 0.24 mg/kg/wk, producing a significantly greater height velocity, by 0.8 cm/y, and a significantly greater last observed height SDS and height SDS at age 18 years, by 2.8 to 3.3 cm.

In addition to evidence for dose-response, within-group analyses of final height minus baseline predicted height, which provide a conservative estimate of treatment effect for this population (since untreated patients on average fail to achieve their baseline predicted height [Bramswig et al. 1990; Ranke et al. 1995; Buchlis et al. 1998; Rekers-Mombarg et al. 1999]), showed that GH treatment significantly increased final height above baseline predicted height for each of the three dosage groups.  The mean treatment effect size for this efficacy measure ranged from 5.4 cm at the 0.24 mg/kg/wk dosage to 7.2 cm at the 0.37 mg/kg/wk.  Thus, the mean gain in adult height attributable to GH treatment with the 0.37 mg/kg/wk dosage was approximately 7 cm compared to the height that the patients were predicted to achieve in the absence of treatment.

3.2.9.  Comparative Efficacy Summary

Figure 11 presents a comparative summary of final height SDS from the two studies, Study GDCH (0.22 mg/kg/wk, administered 3 times per week) and Study E001 (0.24 or 0.37 mg/kg/wk, administered 6 times per week).

 

 

*Analysis of covariance (ANCOVA) model incorporating effect for baseline predicted height standard deviation score (SDS); values are least squares mean ± standard error (SE).

Note:  The dashed line represents the lower limit of the normal range for the general population (AAP 1997).

Abbreviations:  n = number of patients.

Figure 11.                     Comparative summary of Studies GDCH and E001:  Final height SDS.

In Study GDCH the mean increase in height SDS between the Humatrope‑treated and placebo-treated groups was 0.51 SDS (corresponding to 3.7 cm).  In Study E001, the mean between-dose effect between the higher (0.37 mg/kg/wk) and lower (0.24 mg/kg/wk) dosage was 0.45 SDS (corresponding to 2.9 cm).  Although the latter effect did not achieve statistical significance (p=0.086), analyses of last observed height SDS and height SDS at age 18 years both gave statistically significant between-dose effects of a similar magnitude (0.51 SDS [p=0.006] and 0.44 SDS [p=0.012], respectively).  Thus, the overall GH treatment effect of the 0.37 mg/kg/wk dosage can be conceptualized as being comprised of 2 components:  the incremental effect of the dosage of 0.37 mg/kg/wk compared to 0.24 mg/kg/wk, and the effect of the 0.24 mg/kg/wk dosage compared to the height that an untreated group would have achieved.  From Figure 10, this overall treatment effect was 7.2 cm, or approximately 1 SD, obtained by comparing the final height of these patients to the height that was predicted to have occurred in the absence of treatment.  Figure 12 shows the individual final height data of these patients.

 

Note:  Values represent mean ± standard error (SE).  The dashed line at ‑2.0 SDS represents the normal range for the general population (AAP 1997).

Abbreviations:  SDS = standard deviation score; SE = standard error.

Figure 12.                     Significant number of GH treated patients achieved normal height in Studies GDCH and E001.

For the placebo patients, final height SDS for most patients remained below the normal range.  At the 0.22 mg/kg/wk dosage, given in divided doses 3 times per week, 55% of final height SDS values were within the normal range.  At the 0.24 mg/kg/wk dosage, given in divided doses 6 times per week, 71% of final height SDS values were within the normal range.  For the 0.37 mg/kg/wk dosage, all but one final height SDS, or 94%, were within the normal range.  The one patient with final height SDS below normal had a gain in height SDS of approximately 1 during treatment.  Thus, the 0.37 mg/kg/wk dosage resulted in nearly all adult height measurements falling within the normal range.

3.3.   Supportive Data:  Meta-Analysis of Effect of Growth Hormone Therapy on Height in Children with Idiopathic Short Stature

The unmet medical need of patients with non-GHD short stature has been recognized since modern GH testing enabled the differential diagnosis of GHD and non-GHD short stature in the 1960s.  To address this need, a large number of studies have been undertaken.  A recent meta‑analysis of 38 studies, which fulfilled specific inclusion criteria, provides a careful and comprehensive analysis of recombinant GH treatment in patients with non-GHD short stature (Finkelstein et al. 2002) and is summarized here as supportive evidence of efficacy.  The objective of this meta‑analysis was to evaluate short‑term and long‑term effects of treatment with recombinant GH in patients with non‑GHD short stature (referred to as idiopathic short stature [ISS] in this paper) by a review of the literature from 1985 to 2000.  Thirty-eight studies (10 controlled, 28 uncontrolled) met the following principal inclusion criteria:  patients’ initial height below the 10th percentile (most of the long-term studies, however, had entry criterion of height SDS£-2.0 [2.3 percentile]); no previous treatment with GH, sex steroids or anabolic agents; normal stimulated GH concentrations (³10 μg/L); absence of comorbid conditions; on-study treatment with recombinant GH; and inclusion of major outcome measures of height velocity or height SDS.  Two types of analyses were performed:  aggregate and paired.  Aggregate analyses provided pooled estimates across all studies reporting each growth variable.  Paired analyses provided pooled estimates across those studies reporting the given variable both at baseline and as an outcome.

Controlled studies were defined as those having a concurrent control group, either randomized or nonrandomized.  For brevity, only the data reported for the controlled studies that included adult or final height as an outcome will be discussed.  Similar GH treatment effects, however, were reported in the uncontrolled final height studies.

3.3.1.  Growth Hormone Effect on Final Height

3.3.1.1.    Controlled Studies

Among the 10 controlled studies included in the meta-analysis, adult height was measured only in the following four controlled studies, representing data from 188 children:  Zadik et al. 1992, Hindmarsh and Brook 1996, Buchlis et al. 1998, and McCaughey et al. 1998.  Across these four studies, the weighted mean age at study start was 10.8 years and mean duration of treatment was 5.3 years.  The weighted average GH dosage for the children in these studies was 0.31 mg/kg/wk, and in each study the dosage was given in divided doses 6 times per week.  Table 15 presents the final height results for the meta‑analysis of the controlled studies.

Table 15.                       Final Height Results:  Meta-Analysis of Controlled Trials a

 

 

Growth Variable

 

Patients (N)
(Studies [n])

Difference Between Treatment and Control Groups:
Pooled Estimate,
Mean
± SD (95% CI)

Childhood height SDS

 

 

Baseline

 

 

Aggregate

408 (9)

0.02 ± 0.05 (-0.08 to 0.12)

Paired b

36 (2)

0.12 ± 0.11 (-0.09 to 0.33)

1 year

36 (2)

0.60 ± 0.18  (0.26 to 0.95)

 

 

 

Adult height SDS

 

 

Predicted

 

 

Aggregate

118 (4)

0.30 ± 0.12  (0.07 to 0.53)

Paired b

106 (3)

0.13 ± 0.16 (-0.18 to 0.44)

Achieved

 

 

Aggregate

125 (4)

0.84 ± 0.19 (0.46 to 1.22)

Paired b

112 (3)

0.78 ± 0.22 (0.35 to 1.21)

Abbreviations:  CI = confidence interval; SD = standard deviation; SDS = standard deviation score.

a    Table modified from Finkelstein et al. (2002) Table 2.

b    “Paired” indicates analysis of only those studies reporting this variable at baseline and follow-up.

 

While no significant differences between treatment and control groups were noted at baseline, the mean difference in adult height SDS between the treatment group and the control group ranged from 0.78 (paired analysis) to 0.84 (aggregate analysis), which corresponds to 5 to 6 cm.

The between‑group difference (GH versus control) for achieved adult height SDS was also compared with the between‑group difference for baseline predicted height SDS.  In this analysis, the adult height SDS achieved by the GH‑treated patients exceeded baseline predicted height SDS by 0.54 (aggregate analysis) to 0.65 (paired analysis), which corresponds to 3.6 to 4.6 cm.  Figure 13 presents this comparison of the mean difference in height SDS between treatment and control groups for baseline predicted adult height and achieved adult height.

Note:  Recreated from Finkelstein et al. (2002).  For Government Use Only - No Further Reproduction Permitted.

*  No measure of variation was provided for predicted adult height in this study; therefore, it was not included in the analysis of differences between treatment and control groups.

Figure 13.                     Mean difference in height standard deviation scores between treatment and control groups for predicted adult height (at baseline) and achieved adult height for controlled studies.

The GH effect on final height is reflected by the significantly greater difference between treatment and control groups for achieved adult height SDS and by the difference between groups for the gain in height SDS over baseline predicted height SDS.

3.3.1.2.    Uncontrolled Studies

Adult height was measured in the following 8 uncontrolled studies:  Loche et al. 1994, Lopez‑Siguero et al. 1996, Zadik et al. 1996, Bernasconi et al. 1997, Schmitt et al. 1997, Zadik and Zung 1997, Hintz et al. 1999, and Pasquino et al. 2000.  Across these 8 studies, the mean duration of treatment was 4.7 years, and the weighted average GH dosage was 0.27 mg/kg/wk, given in divided doses 6 times per week.  Table 16 presents the results of the meta‑analysis of the uncontrolled studies.

Table 16.                       Results of Meta‑Analysis of Uncontrolled Studies from Peer‑Reviewed Literature a

 

 

Growth Variable

 

Patients (N)
(Studies [n])

Outcome:
Pooled Estimate,
Mean
± SE (95% CI)

Childhood height SDS

 

 

Baseline

 

 

Aggregate

550 (25)

-2.72 ± 0.05 (-2.82 to -2.63)

Paired b

209 (10)

-2.62 ± 0.09 (-2.79 to -2.44)

1 year

209 (10)

-2.19 ± 0.10 (-2.39 to -1.99)

 

 

 

Adult height SDS

 

 

Predicted

 

 

Aggregate

311 (9)

-2.18 ± 0.17 (-2.52 to -1.85)

Paired b

212 (6)

-2.25 ± 0.23 (-2.74 to -1.77)

Achieved

 

 

Aggregate

246 (8)

-1.62 ± 0.07 (-1.77 to -1.47)

Paired b

208 (6)

-1.62 <