Introduction. Prader-Willi syndrome (PWS) is characterized by obesity, hypotonia, hypogonadism, hyperphagia, short stature, and a neurobehavioral profile that includes cognitive deficits, learning problems, and behavioral difficulties that increase in both quantity and severity over time. PWS results from an alteration in the molecular composition of a critical region of C#15q. Morbid obesity resulting from hyperphagia is amplified by decreased energy expenditure and reduced physical activity. The hyperphagia has proven refractory to all psychopharmocologic intervention; the behavioral components are equally resistant to psychotropic intervention. PWS patients’ body composition resembles that of individuals with growth hormone (GH) deficiency, including short stature and reduced lean body mass with concomitant increased fat mass. We hypothesized that GH administration to children with PWS, in addition to stimulating linear growth, would improve body composition, increase energy expenditure and fat utilization, and improve muscle strength, physical agility, and pulmonary function. Two recent reports from this study document significant positive effects of GH treatment on these children’s physical parameters measured in a 2-year, controlled study. However, the behavioral impact of GH treatment in this population remains incompletely described.
A psychosocial burden, including emotional, behavioral, and cognitive disturbances associated with short stature, has been previously described in a non-PWS population with GH deficiency and idiopathic short stature. An impaired quality of life and psychosocial status is also documented in otherwise normal adults with GH deficiency. In both populations, growth hormone replacement therapy (GHRT) is reported to improve alertness, activity level, endurance, irritability, tendency to worry, and extroversion resulting in better personal relationships with fewer conflicts. This report focuses on that portion of the study investigating the behavioral and psychosocial outcomes accompanying increased stature and improved physical status for persons with PWS treated with GHRT. We hypothesized that, as in other populations, GHRT for persons with PWS would have a significant positive effect on their psychosocial status as well as an improvement in their growth parameters.
Methods. A 2-year, controlled study with control group crossover in the second year was used. Fifty-four consecutive children with genetically confirmed PWS were enrolled. Patients were 4 to 16 years of age at time of enrollment, had skeletal maturation <13 for girls and <15 for boys; all but 3 participants remained prepubertal (Tanner stage 1) throughout the study. Children who had previous therapy with GH were excluded, as were children with a scoliosis >20°. After a 6-month growth assessment were randomized into a 60:40 treatment:control ratio. Treatment consisted of Nutropin (Genentech), 1 mg/m2/day.
A modified Offord Survey Diagnostic Instrument (SDI) was used to monitor behavior at 6-month intervals. The SDI is a 165-item behavioral checklist with items rated on a scale of 0 = Never or Not True, 1 = Sometimes or Somewhat True, and 2 = Often or Very True. The items are balanced between positively and negatively scored items. The present instrument was designed to derive diagnoses for the following Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition categories: Depression, Obsessive-Compulsive Disorder, Anxiety Disorder, Somatization Disorder, Conduct Disorder, and Attention-Deficit/Hyperactivity Disorder. The SDI was modified to include 10 items specifically inquiring about PWS (eg, denies having PWS, picks excessively at skin, nose, or other body parts). Because diagnoses are not mutually exclusive, an individual can meet criteria for 1 or more diagnostic categories.
The SDI contains a second section measuring behavior functioning in the school environment, in the family, and in personal and social relationships. A wider scoring range is used and is question-specific. Parallel forms of this measure are available for parents, teachers, and the child him/herself. We gathered data from both parents and teachers at 6-month intervals. No questionnaire was scored until the completion of the entire study to avoid any possibility of an inadvertent “feedback” or “self-fulfilling prophecy” effect. All questionnaires were scored by a Bachelor’s level research assistant blind to study assignment.
Family stress was monitored with the Family Inventory of Life Events. At study completion, the impact of GH was measured with a 13-item summary interview adapted from Wiren et al. After completion of all final study visits, a single research assistant blind to treatment assignment interviewed all families by phone. This method was chosen to minimize any positively biased demand characteristics.
Results. Both between-group and within-group contrasts were computed for baseline, 12 (time 1) and 24 month (time 2) measures. Because behavioral deterioration, as well as improvement, was a possibility, a 2-tailed hypothesis test was used for all comparisons. No differences were found between treatment and control groups, nor within groups across measurement points for attentional symptoms, anxiety, obsessive-compulsive complex, violence, or psychotic symptoms. Similarly, no differences were noted between groups on depressive symptoms; however, a significant positive effect (reduction of depressive symptoms) was noted for the treatment group from baseline to time 1, and was retained at time 2. The group was divided by age, with those 11.0 years and younger comprising one group and those older the second group. This analysis indicated that the major reduction in depressive symptoms occurred in those over 11 years old. When divided by age, a second unexpected finding emerged. There was a significant increase in attention-deficit/hyperactivity disorder symptoms from baseline to 24 months in those children 11 and under, independent of treatment status. The groups were subsequently further broken down by sex and by genetic status (deletion versus disomy) with no significant findings. At no time was the expected behavioral deterioration reported. We conclude that in addition to the previously detailed improvements in physical parameters for these children, behavioral improvement, including a lack of predictable behavioral deterioration during the treatment period, is a strong argument for the use of GHRT for this difficult syndrome.
Prader-Willi syndrome (PWS) is characterized by obesity, hypotonia, hypogonadism, hyperphagia, short stature, and a neurobehavioral profile that includes cognitive deficits, learning problems, and behavioral difficulties that increase in both quantity and severity over time.1–3 PWS results from an alteration in the molecular composition of a critical region of chromosome #15q.4 Morbid obesity resulting from hyperphagia is amplified by decreased energy expenditure and reduced physical activity. PWS patients’ body composition resembles that of individuals with growth hormone (GH) deficiency, including short stature and reduced lean body mass with concomitant increased fat mass.5–7 We hypothesized that GH administration to children with PWS, in addition to stimulating linear growth, would improve body composition, increase energy expenditure and fat utilization, and improve muscle strength, physical agility, and pulmonary function. Two recent reports detail the significant positive effects on these physical parameters that we documented in a 2-year controlled study of GH treatment for children with PWS.8,9 However, the behavioral impact of such treatment in this unique population remains incompletely described.10–11
A psychosocial burden associated with short stature has been previously described in a non-PWS population.12–13 This burden includes emotional, behavioral, and cognitive disturbances in both those with GH deficiency and those with idiopathic short stature. Furthermore, an impaired quality of life and psychosocial status is documented in otherwise normal adults with GH deficiency.14,15 In both populations, growth hormone replacement therapy (GHRT) is reported to improve alertness, activity level, endurance, irritability, tendency to worry, and extroversion resulting in better personal relationships with fewer conflicts.16 This report focuses on that portion of the study investigating the behavioral and psychosocial outcomes accompanying increased stature and improved physical status for GH-treated persons with PWS. We hypothesized that, as in other populations, GHRT for persons with PWS would have a significant positive effect on their psychosocial status, as well as an improvement in their growth parameters.
Behaviorally, the younger child with PWS is described as outgoing and affectionate despite the emerging hyperphagia that occurs between 3 to 6 years of age.17–18 By contrast, descriptions of adolescents and adults with PWS usually focus on the behavioral aspects.19,20 These persons are described as extraordinarily stubborn, clever, manipulative, irritable, mood labile, angry, perseverative, ego-centered, demanding, and prone to temper tantrums and rage attacks that may include aggressiveness toward themselves and others. Lying and stealing are noted as “not uncommon,” as are emotional difficulties in the form of depression, obsessions, and frank psychoses. For many, these behaviors remain resistant to most forms of behavioral and psychopharmacological management; indeed, frequently, these medications worsen behavior. Inability of caregivers to manage these behaviors can result in family turmoil, community placement failure, employment failure, referral for mental health consultation, and even institutional placement.
Thus, we hypothesized that in addition to predictable improvements in physical parameters resulting from GH treatment, the GH-mediated improvements in body composition, physical function, and a “normalizing of appearance” would have a favorable impact on the everyday life and mood/behavioral picture of children with PWS. Nonetheless, we were equally aware of the possibility that behavior could worsen as an idiosyncratic response of GH treatment. Our concern regarding behavioral deterioration was for both an increase in the quantity of negative behavior as well as in the severity of these episodes. At the most basic level, we worried that if a person with PWS was having a “typical tantrum” the improved strength and muscle tone resulting from GHRT might lead to the need for stronger external intervention and perhaps even the need for restraints, possibly causing an increased frequency of injury to the person with PWS and to those attempting external behavioral intervention.
Separate from the directionality of behavioral change issue, we reasoned that to the extent that behavioral change is a function of GH treatment rather than increased age, then the treatment versus control group comparisons at 1 year should show significant differences. If these changes are present, and are related to the length of treatment, then the differences should remain at 2 years. If, however, behavioral changes noted are primarily a function of increased age and/or environmental circumstances, then no differences should be apparent. Finally, if GH serves to improve (or worsen) behavior independent of the length of treatment, then the control group should look worse (or better) than the treatment group at 1 year, with those differences narrowing at 2 years after the control group has been treated for 1 year.
Fifty-four consecutive children with genetically confirmed PWS (GH secretory status unknown) were enrolled after informed consent was obtained from parents or guardians. All 54 patients had high-resolution cytogenetic, fluorescent in situ hybridization and/or methylation studies performed with supervision by a pediatric geneticist. Thirty-seven children had documented deletion of chromosome 15q11–13, 16 children demonstrated uniparental disomy, and 1 patient had an imprinting mutation. Patients were 4 to 16 years of age and had skeletal maturation of <13 years for girls and <15 years for boys. All but 3 subjects remained prepubertal throughout the study, although adrenarche was common. Children who had previous therapy with GH were excluded, as were children with a scoliosis >20.° This study was approved by the Human Subjects Committees of St Louis University and the University of Wisconsin, Madison.
After a 6-month growth assessment period, children were randomized in a 60:40 ratio (to provide an incentive for participation) to treatment (N = 35; Nutropin, Genentech, San Francisco, CA; 1 mg/m2/d) or nontreatment control (N = 19) groups. This dose is equivalent to 0.18 to 0.3 mg/kg/week, is within the range traditionally used to treat GH deficiency in childhood, and is similar to that used in previous studies of children with PWS. The dose calculation is based on body surface area to avoid excessive dosing of markedly obese subjects and to minimize the risk of side effects.
Stimulated GH levels (after administration of oral clonidine, 0.15 mg/m2) and fasting levels of insulin-like growth factor-1, insulin-like growth factor binding protien-3, osteocalcin, and type 1 procollagen were measured. Fasting and 2-hour postprandial glucose and insulin, total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, free fatty acid, triglyceride, free thyroxine, and thyroid-stimulating hormone levels were monitored at baseline and (with the exception of GH) at 12 and 24 months. These data are presented in Myers et al.21
Height was measured on a Harpenden stadiometer at −6, 0, 6, 12, 18, and 24 months. Bone age was determined annually by a single pediatric radiologist and spine films screening for scoliosis were obtained at baseline and, if indicated, at 6-month intervals. Percent body fat, percent lean body mass, and bone mineral density were measured by dual-energy radiograph absorptiometry at baseline, 12, and 24 months. Respiratory muscle strength and pulmonary function were assessed using spirometry and measurements of flow volume and respiratory muscle forces on the same time schedule. Physical strength and agility were assessed using a modified Bruinicks-Oseretsky test, including a timed run, a standing broad jump, and sit-ups. Arm curls with dumbbell weights were also included.
After a 1-year period, the control group began treatment using the same dosing.
Behavioral symptoms and symptom complexes were detailed with the Offord Survey Diagnostic Instrument (SDI, also called the Ontario Child Health Study Scales).22 The SDI is a 165-item behavioral checklist with items rated on a scale of 0 = Never or Not True, 1 = Sometimes or Somewhat True, and 2 = Often or Very True. The items are balanced between positively and negatively scored items. Originally designed, standardized, and normed against the Diagnostic and Statistical Manual of Mental Disorders, Third Edition (DSM-III) diagnostic criteria, the instrument has been revised and restandardized with each revision of DSM criteria. The present instrument was designed to derive diagnoses for the following Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) categories: Depression, Obsessive-Compulsive Disorder, Anxiety Disorder, Somatization Disorder, Conduct Disorder, and Attention-Deficit/Hyperactivity Disorder. The SDI was modified to include 10 items specifically inquiring about PWS (eg, denies having PWS, picks excessively at skin, nose, or other body parts). Because diagnoses are not mutually exclusive, an individual can meet criteria for 1 or more diagnostic categories.
The SDI contains a second section measuring behavior functioning in the school environment, in the family, and in personal and social relationships. A wider, question-specific scoring range is used. Thus, for the question “Has your child’s behavior prevented you from taking the child out in public,” the scoring system is a 1 to 4 system, with 1 indicating never and 4 indicating always. For the question eliciting the number of days per week the child engages in activities with friends, the range is 1 to 5, with 1 indicating never and 5 indicating 6 to 7 days per week.
Parallel forms of this measure are available for parents, teachers, and the child him/herself. We gathered data from both parents and teachers at 6-month intervals. No questionnaire was scored until the completion of the entire study to avoid any possibility of an inadvertent “feedback” or “self-fulfilling prophecy” effect. All questionnaires were scored by a Bachelor’s level research assistant blind to study assignment.
Impact of GH was measured with a 13-item summary interview adapted from Wiren et al.15 After completion of all final study visits, a single research assistant blind to treatment assignment interviewed all families by phone. This method was chosen to minimize any positively biased demand characteristics. 0
Because persons with PWS are exquisitely sensitive to stress in their environment,23 family stress levels were monitored as a covariate. Current family stress was assessed by the Family Inventory of Life Events, a 71-item self-report instrument designed to record the normative and nonnormative life events and changes experienced by a family unit in the past year.24 Family Inventory of Life Events also records certain life events a family experienced before the past year, but are of such impact that their influence may still be present (eg, death of a child). Reliability and validity are adequate and normative data are readily available.
Data were analyzed by using a Student t test for paired samples or 2 related samples. The SPSS (SPSS Inc, Chicago, IL) statistical software package was used.
The randomization resulted in 35 children in the treatment group and 19 in the control group. Five of the children were from single-parent families (4 mothers, 1 father). All children were in special education, either full or part-time. Not all families returned all questionnaires. This report includes the 27 treatment and 14 control subjects who completed all questionnaires at each assessment point. No significant demographic differences were noted between the groups. We restrict this report to the parental behavior questionnaire data. A later report will address the teacher report and family measures.
Drawing on the previously cited studies, the SDI was used to examine the impact of GHRT on changes in alertness, activity level, endurance, irritability, tendency to worry, increased extroversion, and personal relationships. Symptom data were organized by DSM-IV25 diagnostic categories for Attention-Deficit/Hyperactivity Disorder, Anxiety, Obsessive-Compulsive Disorder, Depression, Conduct Disorder (violent type), and Psychoses (Table 1). The total symptom score for the previous time period was calculated for each diagnostic grouping, rather than computing an absolute diagnosis. Thus, for the category depression, the scores indicate a group average for the 14 symptoms assessed during the period since the previous assessment. This method was used as opposed to using normal threshold criteria (eg, depression requires a depressed mood for 2 or more weeks plus 5 or more symptoms representing a change from previous functioning), because the use of DSM-IV thresholds may not be appropriate in this population.26 In addition, to determine interparental reliability, both mother and father’s scores were tabulated and their scores tested for the degree to which they correlated, and for mean differences. Contrasts were computed for treated versus control at baseline and at 1 year. For the 24-month data, contrasts were made again between the continuous treatment group (2 years) and the control group who had now been treated for 1 year. Contrasts were also computed for a comparison of within group change using baseline, 1-year scores, and 2-year scores. A test of homogeneity of variance was conducted for each comparison, and when appropriate, recalculated based on a “no homogeneity assumption.” All contrasts were tested against a 2-tailed hypothesis.
Although there was a tendency for fathers to rate their child with somewhat less severity than did mothers, no significant differences emerged between the 2 reports. For each symptom complex, no differences were found between the treatment and control groups or within groups across measurement points for attentional symptoms, anxiety, obsessive-compulsive complex, violence or psychotic symptoms (Table 1). Similarly, no differences were noted between groups on depressive symptoms; however, a significant positive effect (reduction of depressive symptoms) was noted for the treatment group from baseline to time 1, and was retained at time 2. To determine whether there were differential age-related effects, the group was divided by age, with those 11 years and younger comprising one group and those older the second group. This analysis indicated that the major reduction in depressive symptoms occurred in those over 11 years old. When divided by age, a second unexpected finding emerged. There was a significant increase in attention-deficit/hyperactivity disorder symptoms from baseline to 24 months in those children 11 and under, independent of treatment status. The groups were subsequently further broken down by sex and by genetic status (deletion versus disomy) with no significant findings.
To look at the impact of treatment on school, social, and family/personal behavior, a number of specific questions were examined (see Appendix A for specific questions and scoring direction). Because no significant differences were noted between parental reports, the mother’s report was used for this analysis. Table 2 indicates no significant differences between the treatment and control groups at any point. When comparing within groups across time, however, the treatment group at 12 months showed a slight but significant positive change from baseline in sports performance, but at the same time, a slight but significant negative change in school performance. The control group showed a significant improvement in school performance from baseline to 12 months, when no treatment was being administered, an effect that was not sustained after 1 year of treatment. Control children, however, tended to be younger and just starting school, suggesting that the academic decrements noted were more likely a function of age and increased academic demands rather than a treatment effect.
Next, those behaviors specific to PWS were monitored. These included arguing with others, a single symptom estimate of obsessional thoughts, destroying own or others property, stealing food at home, stealing food outside of home, storing up things unnecessarily, anger with the syndrome, annoying others, picking at nose or body, and tantrums. Table 3 indicates no differences between the treatment and control groups on any symptom at any measurement point. Comparing within groups across time, the treatment group showed significant reduction of obsessional thoughts and skin picking from baseline to 12 months, but both gains were reversed with a significant increase in these same behaviors from 12 months to 24 months. The control group showed a significant decrease in tantrums from the initiation of treatment at 12 months to the second measurement point at 24 months.
Finally, data from the final interview regarding the impact of GH was examined (Appendix B). Virtually all families indicated a positive effect of the GH on both growth and behavior (Table 4). Eleven families (30%) reported this as a sudden change, whereas 70% saw the change as more gradual. Six families reported change in the first month of treatment, whereas an additional 13 families reported change within the first 1 to 3 months. All but 2 families reported change within the first 6 months. All but 1 family reported that their child had more energy as a result of the GH therapy. All reported that their child’s trunk, hands, and feet were more normal in appearance. Eighty-one percent reported a normalization of the appearance of the head. All families reported that their child was more physically fit and that the improvement in fitness had been noticed by those outside the family. In addition, they reported that the increased energy, muscle mass, and muscle tone had now allowed their child with PWS to join “normal” sports activities such as T-ball and soccer, rather than either not playing at all or playing on a “modified” team. All but 4 families felt their child was easier to manage. Eighty-five percent reported improved memory, and 80% reported that their child was less easily annoyed and perseverated less. Eighty percent reported that their child slept less during nonsleep hours. In the area of social relationships, 89% report their child is now more social, and 70% report improved social relationships. All report that their child is now more satisfied with him/herself.
When asked to detail other effects of the GH treatment about which we had failed to inquire, several families noted that their other children were no longer embarrassed to be seen with the sibling with PWS, because of the normalization of the PWS child’s physical appearance and a perceived improvement in behavior. Parents reported being aware of “fewer snide remarks in public.” Several families reported that their child with PWS seems to take more responsibility for his/her behavior and in initiating and completing tasks in the home environment.
As an external index of behavioral impact, the use of psychotropic medications for behavior management was monitored. Two participants entered the study on psychostimulant mediations, both in the treated group (Table 5). Three subjects, 2 in the treatment group and 1 in the control group entered the study on psychotropic medication, 1 each on Prozac (Eli Lilly and Company, Indianapolis, IN), Zoloft (Pfizer, New York, NY), and Risperdal (Janssen Pharmaceuticals, Titusville, NJ; this participant was also one of the participants on psychostimulants). Nobody initiated medication support during the first 12 months of the protocol. During the 12- to 24-month period when all were receiving treatment, 1 subject in the continuously treated group discontinued Prozac. Three initiated psychotropic medications (2 on Prozac, 1 on Zoloft). Among the control group who were receiving treatment for the first time, 1 initiated treatment with Luvox, and 1 was started on both Depakote (Abbott Laboratories, Chicago, IL) and Celexa (Forest Pharmaceuticals, Inc, St Louis, MO). With the exception of one 5-year-old girl in the control group who was started on Luvox during the 12- to 24-month period, all participants who were on medication at baseline and who initiated medication during the study were 11 years old and above. The use of the selective serotonin reuptake inhibitors in this population is primarily for the purpose of increasing cognitive flexibility and decreasing neurologic perseveration.27 At the conclusion of the 2-year period, 18% of the treated group and 14% of the control group were on psychotropic medication. This figure is in line with those previously obtained for persons with PWS aged 13 years and older, living either at home or in alternative care settings. Thus, it seems that there is no excess initiation of medication to suppress behaviors resulting from GH therapy.
Although 100% of the families report positive effects from GH, 11 families (31%) reported negative side effects as well. In all instances, the negative effects were physical rather than behavioral. Thus, these data will be reported in a separate study.
The use of GHRT has now been shown to have significant positive impact on persons with PWS across a number of physical parameters including linear height, body composition, activity level, and strength. However, major concern has revolved around a possible direct negative impact of the hormone on an already difficult neurobehavioral picture, and indirectly on the role of increased size and strength in light of the neurobehavioral issues. Our data are encouraging in suggesting no apparent behavioral deterioration associated with GH administration. Furthermore, some positive behavioral effects were noted.
We monitored the impact of GHRT on activity level, emotions and mood, school, family, and social relationships using 2 methods: a structured interview with repeated measures throughout the study and a single interview at the completion of the study. The structured interview provided quantitative data on an ongoing basis, whereas the “terminal” interview provided an overall subjective qualitative measure.
In the area of activity, we had previously documented significant improvement in muscle tone, strength, and endurance in these patients using objective physiologic measures. Our interest here was the impact of this improvement on daily activity levels as reported by parents. On the structured questionnaire, 2 questions directly addressed this issue: How well does he/she do in sports compared with other children his/her age, and is your child underactive. Although the differences between groups did not reach significance, parents did report a significant improvement in sports ability compared with their peers in the treatment group from baseline to time 1, an improvement sustained at time 2. At the same time, parents still report excessive underactivity when compared with other children. Although the data indicated no increased involvement in extracurricular structured (coached) sports, nor more time spent on hobbies, parents did indicate on the terminal interview that their children were more able to participate in “normal” sports activities through their physical education classes, rather than needing “modified and/or adaptive” physical education. Furthermore, throughout the project, parents subjectively reported a number of areas of improvement not addressed by the structured questionnaire. These included more readily participating in prescribed exercises such as physical therapy, occupational therapy, physical education, walking, and swimming. Thus, we were able to document that an improvement in physical ability translated to a positive impact in daily life activities including more willingness to participate and more normal participation in sports and physical education. We were not, however, able to show a significant increase in daily activities. This may result, in part, from too narrow a domain of activities sampled. In addition, the restricted range of the measurement scale used (0–2, 0–5) may mask significant findings by restricting both the range of scores and the range of the variance. This phenomenon has previously been reported by Fleiss et al.28 Third, a parental protection factor may further mask positive changes. Children with developmental disabilities are often excluded from mainstream sports activities, and children with PWS have an additional restriction factor based on the need to regulate the food environment. It may well be that the measured changes in tone, strength, and endurance did not translate to improvements in the measured activities because parents’ need to regulate the food environment did not automatically allow expansion of their child’s exposure to or participation in a greater range of activities.
In the area of mood, we were able to show a decrease in parent reported symptoms of depression in the children. The change in parents’ perception of their child as depressed was significant on the structured questionnaire at both 12 months and 24 months in the treated group. The control group also showed a reduction of depressive symptomatology after 1 year of treatment, but this reduction did not reach significance. This may be a statistical phenomenon resulting from the smaller size of the control group. These positive findings were amplified by the “terminal” interview on which parents reported their children were less easily annoyed, had improved relationships with others, were less difficult to manage, and were more satisfied with themselves.
We were unable to document positive change in anxiety or obsessive-compulsive symptoms. This may be, in part, attributable to the relatively low levels of these symptoms reported at baseline. Of a possible score of 32 for anxiety, the baseline levels for both the treated and control groups was an average score of 4, with no change over time. Similarly for the items comprising the Obsessive-Compulsive Diagnostic category, the maximum achievable score was 10, with both groups scoring an average of 4 at baseline, and 3 for the treated and 2 for the control group at 2 years. Although this does not indicate a statistically significant change in the positive direction for either group, it may well be that the impact of the treatment was the prevention of a serious escalation of these symptoms–a predictable escalation well described in this population.29 Similarly, no changes were noted in the symptom complexes of violence and psychoses. Again, this may reflect the relatively low levels of these symptoms reported at baseline.
In the area of school performance, the quantitative data suggested a worsening of school performance across time (Table 2). However, the terminal interview depicted a very different picture with 85% of the families reporting improved memory, and 70% reporting improved school performance. This difference may reflect a difference in the way the questions are phrased on the 2 measures. The structured interview inquiry asks “Which of these statements best describes how well your child has done in school during the past 6 months” with 1 = very well, excellent student and 5 = not well at all, poor student. The terminal interview, however, simply inquires about an improvement in school performance. Thus, parents may well see a noticeable improvement in school performance for this child, while still recognizing that overall, when compared with other children, this child is not a particularly good student. However, any improvement in school performance will hopefully translate to improved long-term outcomes for this population, so even a minimal improvement is a positive outcome.
Finally, in the area of family and social relationships we inquired about specific improvements in these areas for the affected individual, and also separately about an overall improvement in family functioning as a result of having this child treated with GH. Again the structured interview data did not indicate improvement in these areas, whereas the terminal interview indicated that most families perceived improvement in these areas. For instance, on the structured interview one question is posed “During the past 6 months, how well has he/she gotten along with his/her teacher(s) at school?” with the response range of 1 = very well, no problems, to 5 = not well at all, constant problems. The quantitative data indicated an answer of “pretty well, occasional problems at all measurement points,” which may for this population be a socially accurate picture. However, the terminal interview simply inquired about improved relationships, with 89% of the families indicating that their child was more social and 70% of the families that their child had improved relationships. Thus, the differences between these data may result from the different approach to the question.
When asked if family relationships were improved as a result of GH treatment of the child with PWS, most families indicated no significant change in family, marital, or parental relationships. This finding held up across measures. Thus, although parents report an improvement on a number of social dimensions for their child, this improvement was not perceived to translate to an easier time in ongoing family relationships. The exception was that siblings found the improvement in the physical appearance of their sibling with PWS allowed them to feel less embarrassed by this child, and the parents reported fewer “snide remarks” when in public. As an index of “family pain,” these improvements may be salutary.
Our data are similar to those reported by Lindgren et al,10 who reported no decrement in behavior as a result of GH treatment in a population of children with PWS, and by Eiholzer et al,7 who noted that their population of children with PWS was “more attentive and more lively, which made them more independent, more self-assured and less anxious.” In addition, these findings are similar to those reported in other populations treated with GHRT.30–32 The apparent (if small) improvement in behaviors is an unexpected benefit of GHRT in this population. The origins of these beneficial effects, however, must remain an open question. The apparent improvement may be a result of improved body composition and improved appearance leading to an improved self-esteem, or the improvement might be related to indirect effects of GH treatment on brain functioning in this population. However, the time span encompassed by this data are insufficient to unequivocally attribute the origins of these beneficial effects to GHRT. At this point, a placebo effect remains an equally plausible explanation. This may well explain the improvement in specific PWS related behaviors (eg, obsessional thoughts, skin-picking) that occurred in the first year of treatment in the continuously treated group, but which reverted to baseline levels at the end of the second year of treatment. An expectation for positive effects may have been inadvertently conveyed by parents and study staff, and may have the dual effect of a placebo based short term improvement in the treated child and a tendency on the part of observers to see a more positive outcome than is actually present.
These effects aside, however, the failure of behavior to worsen can only be seen as a positive outcome for parents of children with PWS grappling with the decision to initiate GHRT. However, in addition to the question of sustainability of effect, is the question of a possible “rebound” effect or worsening of behavior with the cessation of GHRT. These questions cannot be answered for several years, so any family considering initiation of GHRT must be aware that although there is evidence that GHRT does not negatively impact behavior in the first 2 years of treatment, there remain many open questions to be considered in the decision.
We thank the Genentech Foundation for Growth and Development which provided partial funding, and Genentech, Inc, which provided GH for this project. We are grateful to the Columbian Charities of Missouri which provide ongoing support for the work of Dr Whitman.
- Received May 11, 2001.
- Accepted September 28, 2001.
- Address correspondence to Barbara Y. Whitman, PhD, St Louis University School of Medicine, Department of Pediatrics, 1465 S Grand Blvd, St Louis, MO 63104-1095
Dr Allen is a member of the advisory board for the Genentech Center for Research and Education.
- ↵Prader A, Labhart A, Willi H. Ein syndrom von adipositas, kleinwuchs, kryptorchismus, und oligophrnie nach myatonieartigem zustand im neugeborenenalter. Schweizerische Med Wochenschr.1956;8 :1260– 1261
- Zellweger H, Schneider HJ. Syndrome of hypotonia-hypomentia-hypogonadism-obesity (HHHO) or Prader-Willi syndrome. J Dis Child.1968;115 :588– 598
- ↵Zellweger H. Diagnosis and therapy in the first phase of Prader-Willi syndrome. In: Holm V, Sulzbacher SJ, Pipes PL, eds. The Prader Willi Syndrome.Baltimore, MD: University Park Press;1981:55– 68
- ↵Lee P, Hwu K, Henson et al. Body composition studies in Prader-Willi syndrome (PWS): effects of growth hormone (GH) therapy. In: Ellis KJ, Eastman JD, eds. Human Body Composition. In Vitro Methods, Models, and Applications. Newark, NJ: Plenum Press;1993:201– 206
- Brambilla P, Bosio L, Manzoni P, et al. Peculiar body composition in patients with Prader-Labhart-Willi syndrome. Am J Clin Nutr.1997;65 :1369– 1374
- ↵Myers SE, Carrel AL, Whitman BY, Allen DB. Physical effects of growth hormone treatment in children with Prader-Willi syndrome. Acta Paediatr Suppl.1999;433 :112– 114
- ↵Davies PSW. Body composition in Prader-Willi syndrome: assessment and effects of growth hormone administration. Acta Paediatr Suppl.1999;433 :105– 108
- ↵Sandberg D, Brook A, Campos S. Short stature: a psychosocial burden requiring growth hormone therapy? Pediatrics.1994;94 :832– 840
- ↵Sulzbacker S, Crnic KA, Snow J. Behavioral and cognitive disabilities in Prader-Willi syndrome. In: Holm VA, Sulzbacher SJ, Pipes PL, eds. The Prader-Willi Syndrome. Baltimore, MD: University Park Press;1981:147– 160
- ↵Curfs L, Verhulst F, Fryns JP. Behavioral and emotional problems in youngsters with Prader-Willi syndrome. Genet Counseling.1991;2 :33– 41
- ↵Whitman B, Greenswag L. Psychological and behavioral management. In: Greenswag L, Alexander R, eds. Management of Prader-Willi Syndrome. 2nd ed. New York, NY: Springer-Verlag;1995:125– 141
- ↵McCubbin H, Thompson A. Family Assessment Inventories for Research and Practice. Madison, WI: Family Stress Coping and Health Project;1987:81– 99
- ↵American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Press;1994:327
- ↵Whitman B, Greenswag L, Boyt M. The use and impact of psychotropic medications for managing and behavior in persons with Prader-Willi syndrome. Proceedings 13th Annual Prader-Willi Association Scientific Conference; July 22, 1998; Columbus, OH
- ↵Fleiss JL. Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: Wiley and Sons;1988
- Copyright © 2002 by the American Academy of Pediatrics