| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles |
State University of New York at Buffalo and Children’s Hospital of Buffalo, Buffalo, New York 14222
Address all correspondence and requests for reprints to: David E. Sandberg, Ph.D., Pediatric Psychiatry, Children’s Hospital of Buffalo, 219 Bryant Street, Buffalo, New York 14222. E-mail: dsandber{at}acsu.buffalo.edu
Abstract
Several studies have investigated the quality of life (QOL) of GH-deficient (GHD) adults who, as children, had been treated with GH. Variable findings are probably related to sample heterogeneity and disparate research methodologies and designs, particularly the choice of control or comparison groups. In addition to comparing a relatively large sample to questionnaire norms, the present study is the first to compare the QOL adjustment of GHD patients to that of same sex siblings. A total of 140 former patients (76% of those eligible; mean age, 26 yr; n = 95 isolated GHD, n = 45 multiple pituitary hormone deficiencies; 117 males and 23 females) and 53 same sex siblings (84% participation), 18 yr and older, participated in the telephone questionnaire survey. The majority of interviews with GHD patients (78%) and siblings (87%) were conducted blind to the subject’s clinical status. Comparisons between GHD patients and norms for standardized questionnaires indicated both better and worse functioning in several domains. In contrast, very limited differences were detected between GHD cases and same sex siblings. Isolated GHD patients were functioning better than those with multiple pituitary hormone deficiencies, but the effect sizes of these differences in most areas were relatively small. Adult height and degree of growth over the course of GH therapy were generally unrelated to QOL outcomes. Findings from the present study underscore the importance of selecting unbiased control/comparison groups in evaluating psychological outcomes among GHD adults.
THE ENDOCRINE management of growth failure and short stature has changed dramatically since the introduction of biosynthetic GH. To date, however, there are no published studies concerning the behavioral outcomes of individuals who had been treated entirely within the new era of biosynthetic GH therapy. Insofar as improvements in the treatment regimen have resulted in better growth and adult height outcomes (1), there is good reason to speculate that the psychosocial adjustment of individuals treated entirely in this new era will also be superior to that of patients who were treated when supplies of GH were limited. This belief is so pervasive that it has contributed to the expanded prescribing of GH to non-GH-deficient patient groups with short stature (2, 3). Until behavioral studies of these newer cohorts are completed, there remains much to be learned from the experiences and quality of life (QOL) outcomes of patients, now young adults or older, who have completed treatment regardless of the era in which it was received.
Nine studies, conducted between 1985 and 1995, focused on the psychosocial adaptation of formerly treated patients with child-onset GH deficiency (GHD) (4, 5, 6, 7, 8, 9, 10, 11, 12). Studies of the benefits of GH replacement in adult GHD patients were not included. The nine studies vary markedly in terms of sample sizes, final adult height achieved with treatment, and research design and statistical analyses used to assess QOL outcomes.
Educational attainment among GHD patients has typically been shown [but not without exception (4)] to be comparable to that of unaffected siblings and general population norms. Although not consistently addressed in the data analyses of these studies, most have also not detected a difference between the level of attainment of individuals with isolated GHD (IGHD) and those with multiple pituitary hormone deficiencies (MPHD) (5, 6, 7) despite the fact that the intellectual potential of MPHD patients appears to be lower than that in IGHD patients (8, 13). In a different, but probably overlapping partitioning of GHD cases, Dean and colleagues (9) did not detect a difference in educational attainment between those with idiopathic vs. organically based GHD. In an uncontrolled study, Frisch and colleagues (10) reported that 47% of their sample (16–26 yr) had repeated a grade in school. The median intelligence quotient (IQ) of the GHD sample was 105 (range, 75–122) if those cases (n = 8) with physical disabilities (related to brain tumor and visual deficits) were excluded from analyses. No correlation was detected between IQ and duration of GH therapy or standardized height at the time of assessment.
A review of employment patterns revealed a somewhat different, and less optimistic, picture. Several studies found higher rates of unemployment or underemployment among GHD patients compared to norms or comparison groups (6, 9, 11), although there were exceptions (4, 5, 7). Employment status (9) and income (4) were unrelated to final adult height of former patients, but it was suggested that GHD patients may occupy lower status occupations (6). This observation might be related to the experience among those GHD patients who had achieved a relatively poor adult height outcome and who reported that they experienced height-related discrimination when applying for jobs (4, 10). Finally, a consistent difference across studies in employment status of IGHD vs. MPHD patients did not emerge; two studies (7, 9) did not detect differences in employment rates of the patient subgroups, whereas a third (8) found that IGHD occupied higher status occupations than MPHD cases. Patients with idiopathic GHD were employed in the same proportion as those with organic GHD (9).
The QOL outcome of marital status has been systematically assessed across several follow-up studies. For those studies in which comparison statistics were available, GHD patients were less frequently married (4, 7, 8, 9). This effect appeared to be more pronounced among MPHD than IGHD cases in some studies (4, 5), but not in others (6, 7). After controlling for treatment, demographic variables, and other individual characteristics, hypogonadotropism was not found to be predictive of lower marital rates (9). Examining a related aspect of QOL, several investigators have suggested that formerly treated GHD patients are more likely to be living with parents or relatives than living independently (5, 6, 7, 9, 10). Pituitary status (i.e. IGHD vs. MPHD) was not systematically related to this outcome across studies.
Beyond the QOL outcomes summarized above, individual studies have assessed additional domains of adaptation. Some of these studies have incorporated control/comparison groups into their designs, whereas others have relied upon established norms for the questionnaires selected. In general, GHD samples score more poorly relative to norms (4, 10, 12), although exceptions have also been reported (10, 11). Interestingly, final height was not significantly correlated with any of the adult outcome variables in which such a relationship was examined (4, 6, 9).
Several problems exist in this follow-up literature that make it difficult to interpret the findings. First, selection criteria for subject eligibility are not always explicitly stated and vary across studies. Several studies give only limited information concerning both inclusion and exclusion criteria (e.g. peak cut-off in provocative GH stimulation tests). Other studies do not permit the reader to know whether individuals with medical problems in addition to the pituitary deficiency remain within the sample frame. Obviously, the inclusion of such cases (e.g. individuals with blindness) could potentially confound the interpretation of various behavioral outcomes.
Previous studies have also commonly relied upon a postal survey methodology that provides little control over the circumstances under which (and by whom) forms are completed. An alternative would be telephone interviews, which are equivalent to in-person interviews in terms of reliability, validity, precision of estimates, and response rates (14, 15).
Given the amount of attention directed in the endocrine management of individuals with GHD toward improving growth velocity and optimizing adult height, it is surprising that only three of the nine cited QOL studies assessed the relationship between final height and psychosocial outcomes (4, 6, 9), and then only for selected dependent variables.
Finally, the aforementioned studies provide a broad mixture of control and comparison groups against which the psychosocial adaptation of the GHD samples are evaluated. In some studies and for some measures, no comparisons to physically healthy control subjects are provided; the results are thus of a totally descriptive nature (11). In other cases, population norms are invoked for an estimation of expected demographic parameters within the GHD sample without adjusting for differences in potentially influential background differences between the clinic-based sample and the general population (5). Along the same lines, questionnaire norms are adopted for comparison purposes with little consideration given to the suitability of such norms (12). Differences between the clinical and normative samples in terms of age, gender, socioeconomic status, etc. could easily result in spurious differences. Although an attractive alternative to norms, the potential methodological benefits of using a healthy sibling control group has not been maximized in previous studies because same and opposite sex siblings of varying ages are apparently lumped together with data derived from proxy reporting by patients or parents rather than from the siblings themselves (6, 9, 11).
In the case of statistical comparisons between subgroups of GHD patients (e.g. IGHD vs. MPHD), analyses do not adjust for potential background differences (e.g. age, gender, or socioeconomic status) between the groups that could influence the results.
The present study revisits the question of the adult psychosocial adjustment of formerly treated child-onset GHD individuals, but does so in a relatively large and endocrinologically well defined GHD sample. Telephone administration of well standardized questionnaires is introduced as a methodological innovation to enhance both the reliability and validity of the data collected. In addition to comparisons with published norms for the selected questionnaires, the present study is the first to directly contrast the adaptive functioning of GHD patients to unaffected, same sex siblings. Additional questions addressed include the relationship among adult height, patient diagnosis, and QOL.
Subjects and Methods
Target subjects
The potentially eligible sample (n = 212) comprised all
formerly treated GHD patients from 1 endocrinology program at a
regional pediatric hospital in a moderately sized city in the
northeastern United States who, at the start of this follow-up study
(November 30, 1993), were 18 yr or older. All patients had exhibited a
peak GH response to 2 standard provocative GH stimulation tests of less
than 10 ng/mL (polyclonal antibody) before therapy was
initiated.2 The presence of
mental retardation (IQ 
70 on standardized testing) or a chronic
physical disability/disease (e.g. blindness or
insulin-dependent diabetes mellitus) served as exclusion criteria
(n = 17) (12). An additional 10 cases were deceased at the time of
the follow-up. Of the remaining 185, 45 cases (24.3%) were lost to
follow-up, did not respond to repeated requests by mail to participate,
or refused. A total of 140 former patients (75.7%; 117 males and 23
females) participated. The mean age at follow-up was 26.1 (± 6.5)
yr (range, 18.8–46.9 yr). Sixty-five percent of the GHD families of
origin fell within social strata I (major business or professional) or
II (medium business, minor professional, technical) (16). Some 42.8%
of the GHD sample completed GHD therapy before 1985 (i.e.
pituitary-derived GH era). The treatment for 22.9% spanned the
pituitary-derived GH and recombinant GH eras. Finally, 34.3% started
and completed their GH therapies during the recombinant GH era. Final
height was ascertained by one of two methods: 1) excerpts from the
patient’s endocrine chart at that visit when growth velocity was less
than 2 cm/yr, and bone age exceeded 16 yr for boys or 14 yr for girls;
or 2) measured height during the follow-up study either in the
endocrine clinic or in the office of the subject’s primary physician.
Selected diagnostic and treatment-related anthropometric variables are
summarized in Table 1
.
|
Of the 140 GHD patients eligible and agreeing to participate, a total of 63 subjects had an unaffected, same sex sibling 18 yr or older; 53 siblings (84%) participated in the study. Their mean age at follow-up was 29.7 ± 7.6 yr (18.4–52.9 yr). The GHD patients were significantly younger (27.9 ± 6.8 yr; 19.0–45.5 yr; by paired t test, P < 0.01). Forty-one (77%) of the GHD sibling pairs were male. Measuring adult heights of siblings was not feasible in the context of the present study.
Study protocol
The study was conducted in several steps. First, a letter describing the study was mailed to eligible GHD subjects. Those agreeing to participate were subsequently mailed response keys to facilitate the completion of questionnaires over the telephone. An interview appointment was then scheduled and performed by an interviewer who was blind to the medical history of the subject. After completion of the interview, the study coordinator requested permission of GHD subjects to contact their same sex siblings. Nowhere in the letter to siblings was it stated that they were included to serve as control subjects for their GHD brothers/sisters. This strategy was adopted to reduce the likelihood that siblings would bias their reports in the direction of positive psychological adjustment in concordance with them being labeled control subjects by the investigators (17). To reduce the likelihood that the interviewer’s knowledge of the subject’s status (GHD or sibling) might introduce a source of bias in data collection, contact letters to both target subjects and siblings requested that they not disclose to the interviewer any aspects of their medical background. This instruction was repeated at the outset of the telephone interview. These instructions were only partially successful, however: 77.9% and 86.8% of the interviews with GHD subjects and siblings, respectively, were conducted blindly. The study procedure was approved by the hospital’s institutional review board, and written informed consent was obtained from all subjects.
Four-factor index of social status (16). This measure was adopted in the present study to characterize the socioeconomic status of the families of origin of participating subjects as well to statistically control for demographic background differences in comparisons between subgroups of GHD patients.
Living arrangement. This variable was assessed at the outset of the telephone interview. Responses to the question "with whom are you currently living?" were classified into one of seven categories (alone; spouse; unmarried partner; roommate; parents or other relative; other adult; and children, including biological, adoptive, step, and foster children). Those subjects who were coded as having any living arrangement other than living with parents or other relative were coded as living independently.
SF-36 health survey (18) (Table 3). The SF-36 is a generic
QOL measure assessing eight health concepts relevant to functional
status and well-being. The psychometric properties of the SF-36 have
been extensively documented and demonstrate it to be both reliable and
valid in a range of medical and physically healthy populations. Age-
and gender-specific norms are available for a sample of the general
U.S. population.
|
) (19). This 26-item
scale was developed to assess both the structures and the processes of
social relationships. Five social support constructs are assessed.
Chronbach 
values for the five scales range from 0.78–0.89. General
population-based norms are not available.
|
) (20).
The SAS-SR measures either instrumental or expressive role performance
in six major areas of social functioning. To make SAS-SR items
applicable to the majority of subjects in the present study and also to
reduce the overlap in content between the SAS-SR and the SRS (see
above), only the items (13 in total) from the following role areas were
assessed: work as a student, and social and leisure. The latter scale
includes two items regarding frequency of and interest in romantic
dating that were broken out for data analysis. Because only a limited
set of SAS-SR items was incorporated in the present protocol, the use
of questionnaire norms for comparisons was not thought appropriate.
Brief symptom inventory (BSI; Table 5) (21). The BSI is a
53-item self-report symptom inventory designed to assess the
psychological symptom patterns of psychiatric and medical patients as
well as community nonpatient respondents. The scoring of the BSI
generates 9 primary symptom dimensions and 3 global indexes of
distress. Chronbach coefficients for all 9 primary symptom scales
are very good, ranging from 0.71–0.85. Test-retest coefficients over a
2-week interval were comparably high (0.68–0.91). Evidence of strong
discriminant and convergent validity for the BSI has also been
demonstrated.
|
GHD vs. same sex siblings (analysis set 1). 
2
analyses (with Yates’ correction for continuity) were used to assess
differences between GHD patients and siblings in the proportions of
cases who were attending higher education, employed vs.
unemployed, married, and living independently. Differences between GHD
subjects and siblings in the level of educational attainment and scores
on the SF-36 Health Survey, the SRS, the SAS-SR, and the BSI were
assessed by paired t tests. Because the same sex siblings
were significantly older than the GHD cases at the time of the study,
supplementary multiple regression (or logistic regression) analyses
were conducted in which subject’s age was statistically controlled.
Because these secondary analyses did not uncover results at variance
with those of the 2 or paired t tests, they
will not be discussed further.
GHD vs. questionnaire norms (analysis set 2). The total sample of GHD cases (those with and without a same sex sibling in the study) was compared by t test to population norms for the SF-36 Health Survey and BSI. To calculate relative effect sizes for the differences between the GHD and normative samples, we used the measure d (22), which estimates how far apart the means for samples are in terms of SD units. For instance, a d of 0.50 would mean that the means were 0.5 SD apart. Norms for the SF-36 are reported separately for each gender and by age group. Because the number of female GHD subjects falling into any one age range was too small for statistical analyses, comparisons to norms are restricted to males. In the case of the BSI, comparisons to gender-specific norms were performed for both males and females (BSI norms are not broken down by age group).
IGHD vs. MPHD (analysis set 3). The IGHD and MPHD subgroups of patients were unequal in the following, potentially confounding, variables: subject’s age [IGHD, 24.5 ± 5.2 yr; MPHD, 29.2 ± 7.9 yr; t(62) = 3.64; P = 0.001] and duration of GH therapy [IGHD, 3.6 ± 2.2 yr; MPHD, 6.3 ± 4.0; t(57) = 4.30; P < 0.001].3 Hierarchical multiple regression analyses were consequently performed to statistically control for the influence of these variables. The results of these analyses were essentially identical to those in which only the subject’s age and the socioeconomic status of family of origin were included. Presented here are results from the simpler hierarchical (or logistic in the case of dichotomous variables) regression models in which the subject’s age and socioeconomic status were entered in the first step, followed in step 2 by patient diagnosis (IGHD vs. MPHD).
Male vs. female GHD patients. The few nominally significant differences between males and females did not exceed the number expected by chance and, therefore, are not discussed further.
Relationship among final height, response to GH therapy, and
psychosocial outcomes (analysis set 4). To assess the influence of
patient’s adult height and growth response to GH therapy (as measured
by change in relative height (SD units) from the start of
treatment until adult height was achieved), these two dependent
variables were regressed on all psychosocial variables in separate
hierarchical multiple regression analyses. The potential confounding
influence of subject’s age, family of origin’s socioeconomic status,
subject’s diagnosis (IGHD vs. MPHD), and whether the
subject had a same sex sibling 
18 yr of
age,4 was statistically
controlled by entering these variables into the regression model
first, followed by either the variable final height or growth
response to GH therapy.
Results
Educational attainment, employment and marital status, and current
living arrangements (Table 2)
GHD vs. same sex siblings (analysis set 1; see Subjects and Methods). Restricting analyses to those GHD and sibling pairs not currently attending school (n = 30), GHD subjects did not differ in their level of educational attainment from sibling control subjects (P > 0.05). The mean level of education attained for both groups was approximately equivalent to partial completion of a college degree (or specialized training). There was also no difference in the proportions of cases (91.9%) and control subjects (95.5%) who were gainfully employed at the time of the study. Comparable proportions of GHD subjects (67.9%) and siblings (77.4%) were living independently from their parents and relatives (P > 0.10). In contrast, a higher proportion of siblings (50.9%) than GHD subjects (26.4%) were married at the time of the study (P < 0.05).
|
Health-related QOL (Table 3)
Transformed SF-36 scores were used in all data analyses. This transformation converts the lowest and highest possible scores on each scale to 0 and 100, respectively. Scores between these values represent the percentage of total possible score achieved (18).
GHD vs. same sex siblings (analysis set 1). The GHD sample differed significantly from the sibling control group on only one of eight QOL scales; GHD cases scored more negatively than control subjects on the general health scale (P < 0.05). The general health scale assesses perceptions of current health, resistance to illness, and health outlook.
GHD vs. questionnaire norms (analysis set 2). Compared to the normative sample of males aged 18–44 yr (weighted means were calculated for the age groupings 18–24, 25–34, and 35–44 yr), the male GHD sample (n = 115) scored higher (indicating better functioning) on the bodily pain (d = 0.30) and emotional role (d = .31) scales (both P < 0.01). Bodily pain assesses the intensity of bodily pain experienced in the past 4 weeks and the extent to which pain interferes with normal work. Emotional role assesses problems with work or other daily activities as a result of emotional problems.
IGHD vs. MPHD (analysis set 3). The MPHD subgroup received significantly lower scores (i.e. indicating poorer functioning) on five of eight scales (physical functioning: r2 change = 0.078; P < 0.001; general health: r2 change = 0.053; P < 0.01; emotional role: r2 change = 0.051; P < 0.01; social functioning: r2 change = 0.038; P < 0.05; mental health, r2 change = 0.028; P < 0.05). Physical functioning assesses limitations in performing physical activities due to health problems; the influence of either physical health or emotional problems on the quantity and quality of social activities is rated on the social functioning scale, and mental health in- corporates items concerned with depression, anxiety, and psychological well-being.
Social adjustment (Table 4)
GHD vs. same sex siblings (analysis set 1). Statistically significant differences between the GHD sibling pairs were not detected on any of the five SRS scales or on the SAS-SR social and leisure scale, including the items assessing romantic dating. A comparison between the groups on the work as student scale of the SAS-SR was precluded because there were no pairs in which both the case and control subjects were simultaneously attending school.
IGHD vs. MPHD (analysis set 3). With the exception of the subjective integration scale of the SRS, there were no significant differences between the IGHD and MPHD subgroups. In contrast, the IGHD subgroup expressed feeling more a part of a group of friends than did those with MPHD (r2 change = 0.045; P < 0.01). Differences in the same direction (i.e. better functioning among IGHD subjects) were observed on the social and leisure scale (r2 change = 0.059; P < 0.01) and the two combined dating items (r2 change = 0.142; P < 0.001) of the SAS-SR.
Emotional distress (Table 5)
GHD vs. same sex siblings (analysis set 1). GHD subjects were not statistically distinguishable on any of the nine primary symptom dimensions or three global indexes of distress.
GHD vs. questionnaire norms (analysis set 2). In marked
contrast to the GHD-sibling comparisons, both the male and female GHD
patients (those with and without siblings) reported significantly more
distress than gender-specific nonclinical norms. The male GHD patients
scored higher than the norms on eight of nine primary symptom
dimensions and on two global indexes of distress. Effect sizes of the
significant differences ranged from 0.28 (phobic anxiety) to 0.78
SD units (paranoid ideation; mean d =
0.62). Using the recommended cut-off for the classification of
clinically significant emotional distress (i.e. a global
severity index t score 63 or any two primary dimension
t scores 63) (21), 23.9% of the male GHD sample would be
classified as psychiatric cases based upon the global severity index
score and 40.2% according to the method using primary dimension
scores.
The GHD females received significantly higher symptom scores on five of nine symptom dimensions and the same two of three global indexes of distress as the males. The effect sizes of the significant differences ranged from 0.56 (obsessive-compulsive) to 1.14 SD units (interpersonal sensitivity) (mean d = 0.85). These relatively large effect sizes are reflected in the numbers of subjects classified as cases: 30.4% and 52.2% according to the global severity index and elevated primary dimensions schemes, respectively.
IGHD vs. MPHD (analysis set 3). Patients with MPHD scored higher (i.e. were more symptomatic) than those with IGHD on only one primary symptom dimension scale, somatization (r2 change = 0.034; P < 0.05). The somatization dimension reflects distress arising from perceptions of bodily dysfunction.
Relationship between anthropometric/treatment and outcome variables (analysis set 4). Subjects’ final height and responsiveness to the growth-promoting action of GH therapy were associated with limited effects on the psychosocial outcome variables assessed in this study. Subjects’ change in relative height over the course of treatment was not a statistically significant predictor for any of the psychosocial outcome variables. The same was true for the relationship between adult height and scale scores on the SF-36 and BSI and whether a GHD patient was married, living independently, or was gainfully employed. A taller adult height was predictive, however, of higher (better) scores on the subjective integration scale of the SRS (r2 change = 0.028; P < 0.05). This scale assesses perceptions of feeling part of a group of friends. Similarly, taller patients scored higher on the social and leisure scale of the SAS-SR (r2 change = 0.032; P < 0.05), but there was no effect on the two dating items.
Discussion
The value of information regarding the psychosocial adjustment and QOL of formerly treated adult GHD patients rests in part in its potential to inform clinical management decisions for this specific patient group as well as for nontraditional groups of patients currently receiving GH therapy. The review of studies cited in the introduction suggests that GHD patients, as adults, remain at risk for significant problems of adjustment, although findings across studies are far from uniform. This variability in findings may in large part be due to sample heterogeneity and disparate research methodologies and designs, in particular the choice of control or comparison groups.
The present research design includes a number of strengths. First, subject recruitment was restricted to an endocrinologically well defined sample from which only a limited number of cases (n = 17) were excluded for a priori reasons because of accompanying intellectual or physical disabilities. An additional strength of the study is its relatively large sample size, which allows for increased statistical power to detect both within- and between-group differences (including differences between cases and controls). Furthermore, because the effects of background characteristics such as subject’s age and the socioeconomic status of the family of origin were statistically controlled in data analyses, our results are likely to have wider generalizability than previous studies. The findings in the present study are unlikely to be influenced by a volunteer bias because the refusal rate for participation was quite low (between 10.3–15.1%, depending upon whether one classifies as refusers those who failed to respond to subject recruitment correspondence). Similarly, the majority of the interviews were completed blindly, reducing the likelihood that awareness by the interviewer of the subject’s clinical status influenced the subject’s responses. Telephone interviews enhanced the quality of the data collected by allowing greater control and uniformity over the circumstances under which the questionnaires were completed by respondents. Finally, in addition to using normative data for those questionnaires that have been extensively standardized, this study is the first to introduce a same sex sibling control group that completed all questionnaires in the same manner as the GHD cases. This latter procedure provides optimal control for family demographic and genetic factors that could influence responses to study questions.
Replicating earlier studies (5, 6, 7, 8, 9, 11), the educational attainment of GHD patients in the present study appears unaffected by clinical status. We detected no significant difference in this variable between GHD cases and healthy, same sex siblings. This observation is particularly noteworthy in view of findings from other studies indicating that subgroups of GHD patients exhibit lower IQ scores (8, 13) and/or cognitive deficits (23, 24). These latter effects might be related to the observation that GHD patients are at risk for experiencing academic underachievement and grade retention (for review, see Ref.25). The apparent discrepancy between the different sets of findings might be understood as indicating that GHD is a better predictor of academic achievement during the early years of schooling than it is of the ultimate level of educational attainment. Early academic challenges may be successfully ameliorated through specialized remediation. Because data concerning early educational and remedial experiences were not collected, this hypothesis remains untested. Recent findings from a separate study, however, support this idea (26).
A comparison between GHD cases and sibling controls also did not reveal differences in the proportions of each group who were gainfully employed at the time of the study or who were living independently from parents or other relatives. In the case of marriage, however, a significantly lower proportion of GHD cases than siblings were married. More will be said about this finding later.
When GHD cases were compared to general population norms for the SF-36 Health Survey, a health-related QOL measure, an unexpected finding emerged; former patients scored higher (indicating better functioning) on two of eight scales, without any significant differences in the opposite direction. In contrast, a comparison of GHD cases with norms for the BSI, a measure of psychological distress, revealed much higher symptom rates within the patient group. A more consistent and interpretable picture developed from comparisons between GHD cases and same sex siblings. In the case of both the SF-36 and BSI, there were very limited differences between the groups. The one statistically significant difference that did emerge was predictable and relatively easy to understand; the GHD cases reported more problems than siblings on the general health scale of the SF-36. These findings suggest that the choice of control or comparison groups in studies of GHD patients can have profound effects on the results and, thus, on the conclusions drawn from the study. We favor those comparisons using same sex siblings because of the greater ability to control for genetic and social-environmental factors within individual families, which could influence the psychosocial adaptation as assessed by the questionnaires used in this study. An alternative interpretation of the relative equality of the GHD and sibling groups posits that the siblings were themselves experiencing problems of adjustment (e.g. on the BSI) that masked the difficulties experienced by the GHD group. Increased problems in the control group would be conceptualized as stress-related and stemming from the presence of a sibling with a chronic medical problem. Weakening this argument, the empirical literature on this subject has not consistently demonstrated an association between having a sibling with a chronic illness and psychological maladjustment (27). Furthermore, speculation that the siblings in this study were maladjusted conflicts with self-reports of better adjustment compared to norms on the SF-36, a measure of health-related QOL.
On measures of social support and adjustment, there were no differences between GHD cases and sibling control subjects. How, then, is one to understand the significantly lower marital rates among GHD cases compared to siblings despite the overall equivalence in functioning across a wide range of psychosocial measures? Supplementary statistical analyses assessing the potential influence on marital status of within-group (GHD) variables such as final height, gonadotropin deficiency, compliance with the sex hormone replacement regimen when applicable, and etiology of GHD (i.e. idiopathic vs. organic) did not uncover an explanation. Increased change in relative height from the start of treatment until final height was significantly predictive of marriage for the subgroup of GHD patients with a same sex sibling (i.e. those patients who had grown more were more likely to be married at follow-up). This effect could not be replicated, however, in the remainder of the GHD sample without a same sex sibling. Adult height was not correlated with marital status within the GHD sample. Further research is thus needed to explain why, in this and several other studies, marriage rates for adult GHD patients are lower than expected despite apparent similarities between GHD and non-GHD cases on several other indexes of psychosocial adjustment. Future studies will ideally obtain measured adult heights of both cases and control subjects to statistically isolate the influence of this variable.
IGHD and MPHD patients could not be differentiated based upon educational attainment, percentage employed, or marital or living status. Similar findings have been reported in some studies (5, 6, 7, 9) but not in another (4). In contrast, the IGHD patients described themselves as higher functioning than patients with MPHD. The effect sizes of the group differences on five of eight SF-36 scales, although statistically significant, were relatively small (between 2.8% for mental health and 7.8% of the variance for physical functioning). The IGHD group also reported more positive self-perceptions regarding their integration within a network of peers and involvement in spare-time activities (SRS and SAS-SR questionnaires). These findings contrast with those of Björk and colleagues (12), who did not detect differences between these subgroups on similar measures, although their sample size was substantially smaller (n = 23) than that in the present study. In the area of emotional distress, as assessed by the BSI, the MPHD subgroup reported more distress on only one scale, somatization. This difference should be interpreted cautiously. The somatization scale is designed to assess distress arising from perceptions of bodily dysfunction in generally healthy individuals. The higher scores among the MPHD than IGHD patients may reflect true differences in somatic symptoms attributable to the patient’s clinical status rather than a tendency to somatize psychological difficulties. One potential explanation for the differences in functioning between the IGHD and MPHD subgroups involves MPHD patients’ compliance with recommended hormone replacement regimens. One of the items included in the telephone interview concerned self-ratings of compliance. Statistically significant correlations between these self-ratings of compliance and scores on the outcome variables were not detected. In view of the common observation that adherence to medical regimens is typically far from perfect (28), future studies ideally will include laboratory assessments to verify patients’ self reports.
One of the goals of GH therapy for GHD is to allow the child to achieve his/her full genetic growth potential. Furthermore, it is often assumed that increases in growth velocity and adult height will translate into enhanced QOL (3). It is thus surprising that only a minority of studies of formerly treated GHD patients have directly assessed the relationship between adult height or magnitude of catch-up growth over the course of therapy and QOL outcomes in adulthood (4, 6, 9). Findings from these analyses do not support the presence of a statistically significant relationship between adult height and psychosocial adjustment. The present study generally confirms this finding in a relatively large and complete sample of GHD patients. Although final height (but not change in relative height over the course of therapy) was related to a limited set of variables, the effect sizes were rather small (between 2–3% of the variance in outcome scores). This was the case despite the very wide range in adult heights (from -4.7 SD to +0.8 SD) and changes in relative height with treatment (-2.0 SD to +6.0 SD). It would thus appear that maximizing adult height does not automatically translate into improved QOL outcomes. This statement should not be interpreted as suggesting that GH treatment was ineffective in enhancing patients’ QOL as adults. To make such a statement, one would need to compare treated and untreated GHD patients after a similar follow-up, an option that is both unrealistic as well as unethical. On the other hand, the observation that GH-induced increases in relative height are not associated with improved psychosocial adjustment may be relevant to the on-going debate regarding the benefits of providing GH to short stature patient groups who do not exhibit clear evidence of a GHD state (1, 2). The present findings are discouraging of the expectation that improvements in psychosocial adjustment will track increases in the individual’s relative height.
In conclusion, the present study confirms and challenges earlier findings. When formerly treated GHD patients are compared to norms, they appear dysfunctional on one measure and function better than average on another. Other studies using questionnaire norms in lieu of an internal control group have yielded comparable (i.e. inconsistent) findings. In marked contrast, when GHD patients are compared to same sex siblings, virtually all of the differences previously observed disappear. Although the subgroup of patients with siblings is smaller and thereby affords less statistical power, the sample size (53 pairs) exceeds that employed in almost half of the previous studies on this topic. If limited statistical power because of smaller sample size is an issue here, then it serves to inform us of the magnitude of differences we are likely to detect between GHD patients and healthy individuals, i.e. that they are quite subtle. The more marked deficits in the adaptive functioning of MPHD compared to IGHD patients that have been commonly reported were also evident in the present study. These differences, however, were not enough to distinguish a combined group of patients from their siblings. It thus appears that adult GHD patients are functioning better than one would have expected based upon findings from other studies using different control or comparison groups. This observation is relevant to the controversy regarding the risks and benefits of maintaining GHD patients on GH therapy as adults. A cautionary note regarding the QOL of adults with GHD is in order; GHD in patients in this study was of childhood onset. A recent report indicates that individuals with GHD of adult onset are more severely impacted, psychologically and metabolically, by their clinical status. They, in turn, may benefit from replacement of GH in adulthood to a greater degree than individuals with childhood-onset GHD (29).
Acknowledgments
We thank the patients and siblings who participated in the study.
Footnotes
1 This work was supported in part by grants from the Genentech
Foundation for Growth and Development, Inc. Portions of this work were
presented at the Fifth Joint Meeting of the European Society for
Pediatric Endocrinology and the Lawson Wilkins Pediatric Endocrine
Society, June 22–26, 1997, Stockholm, Sweden. 
2 Data analyses did not reveal a statistically
significant relationship between peak GH response to provocative
stimulation tests and scores on the dependent variables.
3 t tests based upon separate
variance estimates were employed because of inequality of variance in
the two samples.
4 GHD patients with same sex siblings 18 yr of
age exhibited poorer psychosocial adaptation on several outcome
variables.
Received September 17, 1997.
Revised December 5, 1997.
Accepted December 16, 1997.
References
This article has been cited by other articles:
 |
|
 |
|
  H. Urushihara, S. Fukuhara, S. Tai, S. Morita, and K. Chihara Heterogeneity in responsiveness of perceived quality of life to body composition changes between adult- and childhood-onset Japanese hypopituitary adults with GH deficiency during GH replacement Eur. J. Endocrinol., June 1, 2007; 156(6): 637 - 645. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. D Voss Is short stature a problem? The psychological view Eur. J. Endocrinol., November 1, 2006; 155(suppl_1): S39 - S45. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Koltowska-Haggstrom, A. F Mattsson, J. P Monson, P. Kind, X. Badia, F. F Casanueva, J. Busschbach, H. P F Koppeschaar, and G. Johannsson Does long-term GH replacement therapy in hypopituitary adults with GH deficiency normalise quality of life? Eur. J. Endocrinol., July 1, 2006; 155(1): 109 - 119. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Attanasio, E. P. Shavrikova, W. F. Blum, and S. M. Shalet Quality of Life in Childhood Onset Growth Hormone-Deficient Patients in the Transition Phase from Childhood to Adulthood J. Clin. Endocrinol. Metab., August 1, 2005; 90(8): 4525 - 4529. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Mauras, O. H. Pescovitz, V. Allada, M. Messig, M. P. Wajnrajch, B. Lippe, and on behalf of the Transition Study Group Limited Efficacy of Growth Hormone (GH) during Transition of GH-Deficient Patients from Adolescence to Adulthood: A Phase III Multicenter, Double-Blind, Randomized Two-Year Trial J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 3946 - 3955. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. K. van Pareren, H. J. Duivenvoorden, F. S. M. Slijper, H. M. Koot, and A. C. S. Hokken-Koelega Intelligence and Psychosocial Functioning during Long-Term Growth Hormone Therapy in Children Born Small for Gestational Age J. Clin. Endocrinol. Metab., November 1, 2004; 89(11): 5295 - 5302. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Wiren, G. Johannsson, and B.-A. Bengtsson A Prospective Investigation of Quality of Life and Psychological Well-Being after the Discontinuation of GH Treatment in Adolescent Patients Who Had GH Deficiency during Childhood J. Clin. Endocrinol. Metab., August 1, 2001; 86(8): 3494 - 3498. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. L. Barkan The "Quality of Life-Assessment of Growth Hormone Deficiency in Adults" Questionnaire: Can It Be Used to Assess Quality of Life in Hypopituitarism? J. Clin. Endocrinol. Metab., May 1, 2001; 86(5): 1905 - 1907. [Abstract] [Full Text]
|
|
|
|
|
|
H. Guyda Four Decades of Growth Hormone Therapy for Short Children: What Have We Achieved? J. Clin. Endocrinol. Metab., December 1, 1999; 84(12): 4307 - 4316. [Full Text]
|
|
|
|
|
|
A. Sartorio and S. Saraifoger Comments on Adults with Childhood-Onset GH Deficiency: the Psychosocial Adjustment Can Be Negatively Conditioned by a Final "Unsatisfactory" Heightg J. Clin. Endocrinol. Metab., December 1, 1998; 83(12): 4533 - 4534. [Full Text]
|
|
|
|
|
|
D. E. Sandberg and M. H. MacGillivray Adult Height and Psychological Adaptation--Authors' Responseh J. Clin. Endocrinol. Metab., December 1, 1998; 83(12): 4534 - 4534. [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
