paediatrics Brussels 17

297

Volume 88 Number 2 2014

Quality of survival in the PNET4 study

Table 2 Outcome z -scores by treatment group in all participants for the 4 principal domains of quality of survival and for decrement since diagnosis in height and weight z -score

Group mean (SD) z -scores Mean intergroup difference (95% CI) P HFRT (a) STRT (b) (b) e (a)

Outcome

n1, n2

Quality-of-survival measures Executive Function BRIEF-GEC z -score *

68, 71 0.25 (0.87) 55, 59 0.12 (0.86) 50, 50 0.11 (0.89) 62, 66 0.07 (1.02)

0.24 (1.06) 0.14 (1.18) 0.11 (1.10) 0.07 (0.98)

0.48 (0.16 to 0.81) 0.26 ( 0.65 to 0.12) 0.23 ( 0.17 to 0.63) 0.14 ( 0.49 to 0.21)

.004

Health Status HUI3 z -score *

.40 .25 .42

SDQ behavioral difficulties z -score y

Quality of Life z -score z

Height and weight x Height decrement from diagnosis Weight decrement from diagnosis

59, 56 1.27 (0.90) 59, 60 0.42 (1.02)

0.84 (0.87) 0.21 (0.91)

0.43 (0.10 to 0.76) 0.20 (0.15 to 0.55)

.011

.27 Abbreviations: BRIEF-GEC Z Behavior Rating Inventory of Executive Function Global Executive Composite; CI Z confidence interval; HUI Z Health Utilities Index; SDQ Z Strengths and Difficulties Questionnaire. Other abbreviations as in Table 1 . * By proxy-report if aged < 18 years; by self-report if aged 18 years. Higher BRIEF-GEC scores indicates worse executive function. Higher HUI3 scores indicate better health status. y By proxy-report if aged < 18 years; not available if aged 18 years. Higher SDQ scores indicate worse behavior. z By self-report Quality of Life Inventory if aged < 18 years; by self-report Core 30-item version of Quality of Life questionnaire, if aged 18 years. Higher scores indicate better quality of life. x Expressed as a z -score where mean Z 0, SD Z 1 for the healthy United Kingdom population. More negative scores indicate greater decrement in height and weight z -scores between dates of diagnosis and follow-up.

Childhood Cancer Survivor Study and contrast with estimates of 64-85% rates of impaired executive function obtained by direct assessment in adult survivors of medulloblastoma (7, 8, 24) . This discrepancy does not bias or explain the observed intergroup differences. The better BRIEF subscale scores for behavioral regulation and metacognition underlying the differences in global scores for executive function, after HFRT relative to STRT, could indicate a decrease of deficits in working memory, attention, and processing speed that have been previously reported after STRT (6-8, 10-12) . The 3-fold greater differences in executive function and growth decrement z -scores between treatment arms in study participants aged 3-8 years at diagnosis is consistent with previous observation of the greater effects of radiation, especially in combination with chemotherapy, on the CNS (9, 13) and on bony growth in this age group but must be treated with caution because stratification by age was an unplanned, exploratory post hoc analysis in the present study. The greater incidence of reported use of hearing aids after HFRT than after STRT is a concern but of uncertain significance because previously reported audiogram data from this study (30) were similar in the 2 treatment arms. Alteration of fields to spare the cochlea has become standard of care subsequent to the radiation therapy used in this study. Time from diagnosis, both to height measurement and also to GH replacement therapy, patient ages and genetic height potential, and numbers receiving GH and thyroid hormone were similar between the 2 treatment arms. Evidence of an additional central GH or thyroid hormonal secretory deficit associated with HFRT was lacking and, unlike spinal damage (47, 48) , is reversible with hormone therapies (15, 16) . The greater height decrement from diagnosis observed after HFRT is therefore likely to be due to relatively greater spinal shortening from radiation damage to both bony matrix and growth plate (48-50) and unlikely to result from differences in skeletal maturity or, thence, time to final adult height. This could be attributable to the higher biologically effective craniospinal dose of HFRT on bone, as predicted for a “late reacting” tissue (see next paragraph), a greater than predicted

good psychometric properties in this population and provides in- formation that is complementary to but different from perfor- mance measures (43-46) , but its use as the only measure of executive function is a limitation of the present study. The rela- tively low (10-23%) rates of BRIEF scores in the clinical range for executive dysfunction are similar to those observed using a 25-item neurocognitive questionnaire, based on the BRIEF, in the

HFRT STRT

0 .0 0 .2 0 .4 0 .6

=0.122

=0.011

=0.016

=0.005

P

P

P

P

-1 .6 -1 .4 -1 .2 -1 .0 -0 .8 -0 .6 -0 .4 -0 .2

=0.405

P

H e ig h t z -s c o re s

n = 67 62

64 64 59 56

59 56 59 55

X S (b )

B a s e lin e (a )

D e c re m e n t (b -a )

D e c re m e n t p e r y r

D e c re m e n t fro m M P H

Fig. 2. Group mean height z -scores by treatment allocation. Height z -scores (see Patients and Methods) (a) at baseline, (b) at cross-sectional (XS) follow-up, (c) decrement Z (b) minus (a), (d) decrement per year, (e) decrement from mid-parental height z -score (MPH). Error bars indicate 95% confidence intervals.