Original Paper
HOR MON E RE SE ARCH I N PÆDIATRIC S
Horm Res Paediatr 2012;77:241–249 DOI: 10.1159/000337975
Received: December 7, 2011 Accepted: March 13, 2012 Published online: April 19, 2012
When Do Short Children Realize They Are Short? Prepubertal Short Children’s Perception of Height during 24 Months of Catch-Up Growth Hormone Treatment John E. Chaplin a Berit Kriström b Björn Jonsson c Maria Halldin Stenlid c A. Stefan Aronson d Jovanna Dahlgren a Kerstin Albertsson-Wikland a a
Göteborg Pediatric Growth Research Centre, Institute of Clinical Science, Sahlgrenska Academy, University of Gothenburg, Gothenburg, b Department of Clinical Science, Umeå University, Umeå, c Department of Women’s and Children’s Health, Uppsala University Children’s Hospital, Uppsala, and d Department of Pediatrics, Central County Hospital of Halmstad, Halmstad, Sweden
Abstract Aim: To examine perceived height during the first 24 months of growth hormone (GH) treatment in short prepubertal children. Methods: Ninety-nine 3- to 11-year-old short prepubertal children with either isolated GH deficiency (n = 32) or idiopathic short stature (n = 67) participated in a 24-month randomized trial of individualized or fixeddose GH treatment. Children’s and parents’ responses to three perceived height measures: relative height (Silhouette Apperception Test), sense of height (VAS short/tall), and judgment of appropriate height (yes/no) were compared to measured height. Results: Children and parents overestimated height at start (72%, 54%) and at 24 months (52%, 30%). Short children described themselves as tall until 8.2 years (girls) and 9 years (boys). Prior to treatment, 38% of children described their height as appropriate and at 3 months, 63%. Mother’s height, parental sense of the child’s
© 2012 S. Karger AG, Basel 1663–2818/12/0774–0241$38.00/0 Fax +41 61 306 12 34 E-Mail
[email protected] www.karger.com
Accessible online at: www.karger.com/hrp
tallness and age explained more variance in children’s sense of tallness (34%) than measured height (0%). Conclusion: Short children and parents overestimate height; a pivotal age exists for comparative height judgments. Even a small gain in height may be enough for the child to feel an appropriate age-related height has been reached and to no longer feel short. Copyright © 2012 S. Karger AG, Basel
A motivating factor behind the demand by parents of short children for growth hormone (GH) treatment has been the argument that disparity in height could put the short-stature child at risk of psychosocial stress and lower quality of life (QoL) as a result of negative comparisons with peers, expectations of parents, or social labeling [1]. Secondary psychological consequences of short stature can also occur when negative stereotypes affect
The data in this paper has not been reported before. The methods of this study are based on a study first reported in Chaplin et al. [Horm Res Paediatr 2011;75:291–303].
John Eric Chaplin, PhD Göteborg Pediatric Growth Research Center Queen Silvia Children’s Hospital, Sahlgrenska Academy University of Gothenburg, SE–416 85 Gothenburg (Sweden) Tel. +46 31 343 5788, E-Mail john.chaplin @ vgregion.se
Downloaded by: Umea University Library 130.239.20.133 - 12/27/2016 7:34:20 PM
Key Words Children ⴢ Growth hormone treatment ⴢ Height determination ⴢ Height prediction ⴢ Puberty ⴢ Quality of life ⴢ Short stature ⴢ Height perception
242
Horm Res Paediatr 2012;77:241–249
their sense of tallness on a scale from short to tall. This is of importance in the interpretation of QoL measurements and in their use as an outcome measure in GH treatment.
Methods Ethics The protocol for this psychosocial clinical study was approved by the ethical boards of the University of Gothenburg (for Göteborg and Halmstad), Umeå and Uppsala and by the Medical Product Agency of Sweden. Written informed consent was obtained from all parents, and children if old enough. Study Design The study was carried out as an auxiliary independent study in four of five centers conducting a multicenter clinical trial of the effects of randomized, fixed versus individualized GH treatment on gain in height in GH-deficient and ISS prepubertal children. Our recent research has emphasized the importance of trying to individualize the GH dosing according to responsiveness [11] in order to achieve optimal and safe attainment of the target height [12, 13]. The study design and results of the clinical trial have been reported elsewhere [13–15]. Subjects and their parents in four of the centers completed psychosocial evaluations at study start, and after 3, 12 and 24 months of GH treatment. Children and their parents completed the questionnaire without reference to each other, sitting in separate rooms. Evaluations were carried out by chartered psychologists. Participants Four clinics participated in the psychological auxiliary study. In the multicenter trial there were 114 children from the four sites who fulfilled the clinical trial protocol [13], of those there were 99 who completed the psychosocial evaluations. Of the 15 who were excluded, 3 were small for gestational age (SGA) [16], 5 lacked baseline psychological data, and 7 had missing follow-up psychosocial data. The study population included girls 3–10 years of age and boys 3–11 years at study start with heightSDS no greater than –2 SDS compared to the population norms and 1 SDS or more below MPH (MPHSDS). Further details of the study sample can be found in our previous publications [13, 15]. At baseline, the 99 children between 3 and 11 years (70 boys: 3–11 years; 29 girls: 3–10 years) had a mean age of 7.3 years (82.08), and a mean heightSDS of –2.66 8 0.43 related to population reference data. The MPHSDS was –0.94 8 0.64 and the difference between the child’s heightSDS and the MPHSDS (diffMPHSDS) was –1.72 8 0.56 with no significant difference in subgroups. Of the 99 children, there were 32 who were defined as GH-deficient and 67 who had ISS. The criteria of GH deficiency was defined according to the maximum level of endogenous GH in the patient’s blood (GHmax) measured by both a spontaneous 24-hour profile and two stimulation tests (arginine-insulin tolerance test) of GH secretion. Questionnaires were also completed by the parent or parents attending the clinic with the participating child. In 98% of cases the mother of the child completed the questionnaire. Where the mothers’ ques-
Chaplin et al.
Downloaded by: Umea University Library 130.239.20.133 - 12/27/2016 7:34:20 PM
how the child is treated by others [2]. The way others respond to the child and the child’s self-perception of how he or she is treated will in turn influence the child’s behavior and self-esteem [3]. Therefore, the self-perception of height may be an important determinant of the child’s self-concept which helps to shape the child’s social interactions, life satisfaction and subjective QoL. The conceptualization of height in relation to QoL has been referred to as growth-related QoL (GrQoL) [4]. This view, however, does not go uncontested. It is also claimed that there is very little effect of height as an isolated characteristic on a child’s QoL at any age [2]. The differences in the literature may be explained by the fact that the age of the child is seldom considered. It has been suggested that the age of the child has a moderating effect on the expression of poor QoL in relation to height: some researchers have proposed that a reduction in QoL occurs only during adolescence [5] or not until adulthood [6, 7]. However, because GH treatment has a gradual and accumulative effect, if adult height is to be increased substantially, treatment must be started early in the child’s life in order to have a possibility of achieving the height target of mid-parental height (MPH). This suggests that any beneficial effect of increased height on QoL or self-esteem will only be seen after many years of treatment. A recent consensus document on GH treatment in children with idiopathic short stature (ISS) places emphasis on the measurement of QoL outcomes that can be achieved via increasing height [8]. Instruments for QoL assessment in children with short stature have been developed [9, 10] and a new instrument is currently being developed within the QOLISSY project [1]. However, measuring children’s individual QoL in relation to height implies that the child has an understanding of their height in relation to other children of the same age and places an emotional value on height as part of their self-concept. Therefore, the child’s perception and evaluation of their height become important elements in the interpretation of QoL results, and therefore there is a need to explore height from a psychological perspective in order to understand the basis upon which an evaluation of QoL is being made by both the child and the parents. The objective of this study within a population of both growth hormone deficiency (GHD) and ISS children was, therefore, to explore demographic, auxiological and perceptual variables related to measures of these children’s sense of relative height to others of the same age, their sense of being at the appropriate height for their age, and
Measures Attained Height. Age- and gender-related height SD scores (SDS) were used from the Swedish population-based longitudinal reference values from birth to 18 years [17] and, since all children were prepubertal, the calculated childhood component was used [18]. Measured parental heights were expressed in SDS according to the Swedish reference values [17]. Short Stature. Short stature is defined as a standing height 12 SDS below the mean for sex [8]. The following three measures of perceived height were included: (1) Sense of relative height was measured relative to the child’s age and gender peers. The Silhouette Apperception Test (SATSDSNOW) [19, 20] with five randomly ordered silhouettes representing the 3rd (–1.88 SDS), 25th (–0.67 SDS), 50th (0.00 SDS), 75th (0.67 SDS), and 97th (1.88 SDS) percentiles was completed by both the child and parent. The SAT result (SATSDSNOW) was also compared with attained heightSDS in order to calculate a measure of overestimation of height by both the child and parent. Overestimation was judged to be a difference of 11 silhouette or 0.67 SDS. (2) Sense of tallness was measured by both child and parent on a 100-mm visual analogue scale (VAS) between the bipolar adjectives ‘short’ and ‘tall’ (short = 0, tall = 100) [10]. When responding to the VAS item, respondents indicate their level of agreement marking a point on the line between tall and short. (3) Sense of appropriate height was measured by the use of a single question from the Swedish ‘I think I am’ scale (ITIA) [21], which asks the child: ‘Are you about the right height for your age?’ In Swedish, the word for ‘about right’ also implies contentment and satisfaction. Statistical Analysis Analyses were carried out using SPSS versions 15 and 17 (SPSS Inc., Chicago, Ill., USA). Results are expressed as mean 8 SD unless otherwise specified. Analyses of paired samples were performed using the Wilcoxon signed-ranks test. Independent samples were analyzed using the Mann-Whitney U test. In bivariate correlation analyses, Spearman’s was applied. When analyzing tables with nominal data, 2, Fisher’s exact and McNemar tests were used. A two-tailed p value !0.05 was considered significant. Linear and logistic regression analyses were conducted with demographic, auxiological, and perceptual variables from the child and parent at baseline and at the 24-month visit as independent variables to predict the three outcome variables. Linear regression models were constructed for (1) sense of relative height to others and (2) sense of tallness. A logistic regression model was constructed for (3) sense of appropriate height for age. Within each regression model, the two measures of perceived height that were not being used as the dependent variable were used to predict the third. The independent variables entered in all regression models were: heightSDS (i.e. relative height for age and gender) at each time point, age, gender, ISS/GHD, maternal and paternal heightSDS, the extent of the child’s height overestimation, the child’s sense of tallness, the extent of parental overestimation of child’s height, and parents’ sense of child’s tallness.
When Do Short Children Realize They Are Short?
Results
Gain in Height (Delta HeightSDS) There was a substantial increment in heightSDS at 24 months of GH treatment (mean, SD 1.33 8 0.48) for the group as a whole. No differences were noted in average gain in heightSDS between the GHD/ISS or between the randomized and fixed-dose groups (table 1). (1) Sense of relative height related to same-age peers (SATSDSNOW) (table 2): At baseline, 28% of the children had a height perception that corresponded with their attained heightSDS (i.e. 72% overestimated) and 46% of the parents had a height perception that corresponded with their child’s attained heightSDS (54% overestimated). At 3 months, the proportion of children who gave an accurate estimation of their heightSDS reduced to 18%. It was not until 24 months that a significant improvement in accuracy compared to baseline was found; at 24 months, 48% of children perceived their height to be within 1 SDS of their measured height, which was a significant improvement from baseline (p ! 0.002, McNemar test). At the same time point, 70% of parents’ judgments were within 1 SDS of the child’s measured height, a significant improvement from baseline (p ! 0.001, McNemar test). Therefore, after 24 months, 52% of children and 30% of parents overestimated the child’s height. (2) The child’s sense of tallness (VAS): At baseline, the average perceived tallness of the children was in the second quartile of the scale, 27 8 30.8, and rose to 47 8 27.1 at 24 months, i.e. the midpoint of the scale, neither short nor tall (p ! 0.001). At baseline, the majority of children (71%) indicated that they were short (VAS 0–33). At 24 months, the majority of children (48%) indicated they were of average height (VAS 34–66). A negative correlation was found between sense of tallness (VAS) and age (baseline: = –0.405, p ! 0.001; 24 months: = –0.326, p ! 0.002), indicating that younger children were more likely to describe themselves as tall than were older children. This can be seen in figure 1, where the pivotal point between the children describing themselves as short or tall was 8.2 years for girls and 9.0 years for boys. Figure 1 shows that the scores of the children varied between the extremes on the scale. For the parents, the short-tall estimation rose from 25 8 18.0 at baseline to 46 8 16.7 at 24 months (p ! 0.001), independent of the gender of the child (boys: 24.5 8 18.5 increasing to 46 8 16.6; girls: 25.5 8 17 increasing to 47.1 8 17.2). (3) The child’s sense of appropriate height: At baseline, 38% of the children expressed that they had an appropriate height (fig. 2). At 3 months, the proportion of children Horm Res Paediatr 2012;77:241–249
243
Downloaded by: Umea University Library 130.239.20.133 - 12/27/2016 7:34:20 PM
tionnaire was missing (3 cases), the fathers’ questionnaire was online supplemented (online suppl. table 1; for all online suppl. material, see www.karger.com/doi/10.1159/000337975).
Table 1. Sample characteristics, dependent and independent variables (p < compared to baseline)
Child evaluations Child’s age Child’s attained heightSDS Child within 2 SDS of population mean Gain in heightSDS Child’s perception of height relative to same age children (SATNOWC) Child’s evaluation of tallness (VAS) Child’s judgment of appropriate heightSDS Child’s overestimation of heightSDS Parental evaluations
Baseline (n = 83)
3 months SD (n = 80)
12 months (n = 91)
24 months (n = 92)
7.382.1 –2.6680.43 4%
7.582.1 –2.3780.46b 22% 0.2980.17
8.382.1 –1.8080.54b 61% 0.8680.30
9.382.1 –1.3380.65b 82% 1.3380.48
–1.0480.82 SDS –0.6380.97 SDSa –0.5280.76 SDSb –0.2580.68 SDSb 27831 37830a 47831a 47827b b b 38% 63% 66% 64%b 72% 82% 65% 52%a (n = 92)
(n = 95)
(n = 96)
(n = 93)
Mother’s heightSDS –1.1180.09 Father’s heightSDS –0.0880.09 Parents’ perception of child’s relative height (SATNOWP) –1.4180.75 SDS –1.2980.67 SDS –0.84 8 0.57 SDSb –0.5680.63 SDSb Parents’ evaluation of child’s tallness (VAS) 38817b 46817b 25818 32817b Parents’ overestimation of child’s heightSDS 54% 45% 46% 30%b SDS = SD score; VAS = visual analogue scale; SATNOW = silhouette apperception test of the child’s current height – the judgment made by the child him-/herself (C) or by his/her parents (P). Mean 8 SD values are shown. Significant difference from baseline: a p < 0.01; b p < 0.001.
Table 2. Measured height, height perception and accuracy of perceived height using the SAT
n
Child Baseline 3 months 12 months 24 months
83 80 91 92
Parent Baseline 3 months 12 months 24 months
92 95 96 93
Measured heightSDS (heightSDS)
Perceived relative heightSDS (SATSDSNOW)1
Accuracy of perceived heightSDS (SATSDSNOW)1 % underestimation2
% accurate estimation3
% over-estimation4
–2.6680.43 –2.3780.46 –1.8080.54 –1.3380.65
–1.0480.819 –0.6380.109a –0.5280.763b –0.2580.684b
0 0 2 1
28 18 31 48a
72 82 67 51
–1.4180.713 –1.2980.083 –0.8480.827b –0.5680.784b
0 0 1 0
46 55 52 70b
54 45 47 30
244
Horm Res Paediatr 2012;77:241–249
Chaplin et al.
Downloaded by: Umea University Library 130.239.20.133 - 12/27/2016 7:34:20 PM
Mean 8 SD values are shown. From baseline: a p < 0.01; b p < 0.001. 1 SATSDSNOW = SAT current relative height transformed to SDS. 2 Under-estimation = greater than 1 SD below the attained heightSDS. 3 Accurate estimation = within 1 SD below and above attained heightSDS. 4 Over-estimation = greater than 1 SD above the attained heightSDS.
80 60 50 40 Boys 20 Girls 0 4
6 8 10 12 Age (years) at 24 months of GH treatment
14
Fig. 1. Child’s sense of tallness related to age and gender at 24 months. Regression lines are shown for girls (filled circles, R 2 = 0.115) and for boys (open circles, R 2 = 0.112). These lines cross the midpoint of the scale at 8.2 years for girls and at 9 years for boys. At this point, there is a change from the children perceiving themselves as tall (younger) to perceiving themselves as short (older).
% 100
% children within ±2 SDS of population mean
80
% children indicating they are of appropriate height
60 40 20 0 Baseline
3
12
24
Months
Fig. 2. HeightSDS compared with sense of appropriate height. The
percentage of children with heights within 82 SDS of the population norm are shown in columns. The line graph indicates the percentage of children at each time point who considered themselves to be of appropriate height. Appropriate height was measured at all time points using a single dichotomous question. At baseline, 38% of the children thought they were of appropriate height, whereas only 4% met the standard population measure of appropriate height (1 –2 SDS). At 24 months, 64% of the children thought they were of appropriate height, whereas 82% met the standard population measure of appropriate height (1 –2 SDS). No gender differences were found.
Horm Res Paediatr 2012;77:241–249
245
Downloaded by: Umea University Library 130.239.20.133 - 12/27/2016 7:34:20 PM
When Do Short Children Realize They Are Short?
100 Child’s sense of tallness (VAS) Short Tall
who thought their height was appropriate rose significantly to 63% (p ! 0.001, McNemar test) without any further significant incremental increase at either 12 months (66%) or at 24 months (64%). If normal height can be judged to be within 2 SDS of the population mean, then the proportion of children with normal height in this population was: at baseline 3%, at 3 months 22%, at 12 months 62%, and at 24 months 82%. At baseline, the children who found their height appropriate were more likely to be younger, 6.9 8 1.67 years, compared to those who found their height inappropriate, 8.0 8 1.83 years (p ! 0.01). At 24 months, those who judged their height as appropriate were 9.1 8 2.0 years compared to 9.9 8 1.9 years (not significant). The proportion of GHD children who estimated their height to be appropriate at baseline (42%) did not significantly increase at 24 months (54%: not significant). However, the proportion of ISS children who judged their height to be appropriate at baseline (37%) did significantly increase at 24 months (67%: p = 0.023). Despite this, both subgroups had grown in heightSDS equally over this period (heightSDS gain: GHD 1.35; ISS 1.32). If the population is divided between those children who thought they were of appropriate height and those who thought they were not of appropriate height, it is found that there is no correlation between tallness and SAT or tallness and measured height for those children who thought they were of appropriate height. In contrast, sense of tallness is significantly correlated (p ! 0.01) to measured heightSDS for those children whose height is not appropriate at the 12-month (r = 0.542) and 24-month (r = 0.447) time points (see table presented online). (4) Regression analyses (table 3): Separate regression analyses were conducted for the three following measures of perceived height: (i) Sense of relative height (SAT) (linear regression): No relationships were found for the sense of relative height at baseline. However, at 24 months a significant gender effect was found, namely that girls (mean age 9 years) tended to have a more accurate evaluation of their heights than boys (mean age 9.5 years) (p = 0.019). (ii) The child’s sense of tallness (VAS) (linear regression): At baseline, only the age of the child entered as a significant factor in the model and explained 16% of the variance of VAS. At 24 months, the parents’ sense of the child’s tallness, the mother’s height and the age of the child contributed to the regression analysis, explaining 34% of the variance. A positive correlation of 0.35 between the child’s sense of tallness and the mother’s height contributed to the model. Thus, the taller the mother, the more likely the child was to feel tall.
Table 3. Regression analyses evaluating contributions to the variance in the child’s sense of relative height, appropriate height, and
sense of tallness at baseline and 24 months Model Linear regression Child’s sense of relative height (SATNOWC) 24 months, R2 = 0.120 Child’s sense of tallness (VASc/p) Baseline, R2 = 0.164 24 months, R2 = 0.344
Logistic regression Child’s appropriate height Baseline, R2 = 0.114 24 months, R2 = 0.137
Variables
B
SE

gender (boys = 0; girls = 1)
0.379
0.158
0.257