Supplementary Methods: Statistical Analysis

Health-related quality of life in high-risk low-grade glioma; results of a randomized controlled trial - Supplementary material web appendix

Supplementary methods: statistical analysis

All analyses were done using SAS v9.3 software. HRQOL was a secondary study end point evaluating the hypothesis that TMZ chemotherapy would have less negative impact on HRQOL than irradiation. The null hypothesis was that there was no difference in HRQOL between the treatment regimens, and the alternative hypothesis was that there was a difference, which may be in either direction. To reduce multiplicity, we preselected seven key scales for the primary analysis, namely global health/quality of life status (GH/Q), and role functioning, social functioning, communication deficit, visual disorder, motor dysfunction, and drowsiness, based on past clinical experience. Global health/quality of life status (GH/Q) was considered the primary scale of interest and the other six were secondary. The remaining HRQOL variables were examined on an exploratory basis only.

A linear mixed effects model was constructed with treatment, a time effect and time-treatment interactions as fixed effects and a patient specific random effect to account for the longitudinal nature of the data. The most suitable covariance structure was determined on the basis of Akaike’s Information Criterion by fitting the model to the data from all randomized patients according to the intention-to-treat principle 1. Score estimates, standard errors, the associated confidence intervals and resulting tests were obtained from the resulting model. The primary test consisted of a general overall post-baseline test for no differences between the two treatment arms at all post-baseline time points via an overall F-test statistic at 5% level of significance. As the sample size was powered to detect relevant differences in PFS rather than in HRQOL, statistical significance is used as a screening method rather than hypothesis testing. A change was considered as clinically relevant when >10 points 2.

In order to assess the robustness of the results with respect to missing data, a problem inherent to HRQOL outcomes, sensitivity analyses were pre-planned and performed, investigating the informative drop-out by graphical evaluation and assessing the relationship between selected variables (gender and the stratification factors) and questionnaire compliance via logistic regression. Factors found to be associated with the missing data process were included as additional factors in the mixed model. In addition, those factors were also used as predictors in a linear regression model to impute the value of missing HRQOL data and replicating the primary analysis 3. For the GH/Q scale, five summary statistics based on the change from the baseline score were evaluated: change to the last available score, to the lowest available score, to the highest available score and to the average of all post-baseline scores. The fifth summary statistic was the percentage of patients reporting a change from baseline of ≥10-points at any time on the GH/Q scale.

As both treatment arms differ in duration and character (irradiation on weekdays during six weeks versus 12 cycles of TMZ chemotherapy during 21 out of 28 days), comparison of HRQOL between both groups at pre-specified time points might not reveal all relevant information. As both treatment arms differ in duration and character (irradiation on weekdays during six weeks versus 12 cycles of TMZ chemotherapy during 21 out of 28 days), comparison of HRQOL between both groups at pre-specified time points might not reveal all relevant information. For example: assessment of HRQOL at month 9 (which is the 3rd follow-up assessment) means that LGG-patients allocated to irradiation have completed treatment more than six months ago, whereas LGG-patients allocated to TMZ chemotherapy still are on treatment. In order to determine whether the loss of HRQOL was different between both treatment arms during a certain interval, a post-hoc analysis compared the area under the curve (AUC) between the two arms. For each patient and for all primary HRQOL scales, two AUC values were calculated, one covering the period up to 15 months and one covering the period up to 24 months from inclusion, by calculating the product of the HRQOL scores over the time period by the duration of that time period through triangulation from randomization. For patients without baseline, the AUC is considered unknown. The score at 15 and 24 months itself was calculated as follows: if a HRQOL score is available both before and after the 15 and 24 months time point, an interpolation was made, while the AUC was censored at the last known time point if no further HRQOL score was available. For patients with incomplete follow-up (i.e. censored before 15 and 24 months), the AUC was censored at the last known time point. For patients who died before 15 or 24 months, the score at and after time of death was considered the worst score of the relevant scale (i.e. 0 for functioning scales and 100 for symptom scales). The resulting AUC was divided by the maximum possible AUC and multiplied by 100. The resulting standardized AUC score can therefore be interpreted as a percentage of the maximum AUC. The AUC are compared between the two groups by non-parametric logrank test (taking into account censoring for incomplete follow-up). Median AUCs were calculated using Kaplan-Meier method.

References

(1) Akaike H. Information measures and model selection. International Statistical Institute 1983;44:277-291.

(2) Osoba D, Rodrigues G, Myles J, Zee B, Pater J. Interpreting the significance of changes in health-related quality-of-life scores. J Clin Oncol 1998;16:139-144.

(3) Fairclough D. Design and Analysis of Quality of Life Studies in Clinical Trials. New York: 2002.

Supplementary results

Figure 1 and Table 1: Regression imputation with imputed values predicted from a regression model that included factors (treatment arm, assessment time, WHO performance status, age, gender, molecular testing, type of histology, and MRI contrast-enhancement):

Imputed Global Health/QoL
Assessment
Time / TMZ / RT / Difference (95% CI)
baseline / 69 (SD=1∙1) / 68∙4 (SD=1∙1) / 0∙7 (-2∙4,3∙7)
3 months / 68∙8 (SD=1∙12) / 64∙2 (SD=1∙1) / 4∙6 (1∙5,7∙7)
6 months / 68∙8 (SD=1∙13) / 70∙7 (SD=1∙11) / -1∙9 (-5,1∙2)
9 months / 71∙6 (SD=1∙15) / 71 (SD=1∙13) / 0∙6 (-2∙5,3∙8)
12 months / 70∙6 (SD=1∙15) / 72∙7 (SD=1∙14) / -2∙1 (-5∙3,1∙1)
15 months / 71∙2 (SD=1∙18) / 71∙7 (SD=1∙16) / -0∙5 (-3∙8,2∙7)
18 months / 72∙8 (SD=1∙23) / 71 (SD=1∙17) / 1∙8 (-1∙5,5∙1)
21 months / 73∙6 (SD=1∙24) / 71∙9 (SD=1∙19) / 1∙7 (-1∙6,5∙1)
24 months / 73 (SD=1∙28) / 73∙7 (SD=1∙23) / -0∙6 (-4∙1,2∙9)
27 months / 72∙1 (SD=1∙3) / 71∙7 (SD=1∙25) / 0∙4 (-3∙2,3∙9)
30 months / 71∙5 (SD=1∙34) / 73 (SD=1∙26) / -1∙5 (-5∙1,2∙1)
33 months / 72∙7 (SD=1∙39) / 73∙6 (SD=1∙31) / -0∙9 (-4∙7,2∙8)
36 months / 73∙6 (SD=1∙46) / 73∙3 (SD=1∙35) / 0∙4 (-3∙5,4∙3)

Table 2: Additional sensitivity analyses comparing summary statistics of the worst change from baseline for patients who have a valid baseline and at least one follow-up HRQOL assessment

Treatment / Total
(N=375)
RT
(N=181) / TMZ
(N=194)
Average global health/QoL
Median / -1∙4 / 0∙0 / 0∙0
Range / -52∙8 - 66∙7 / -54∙2 - 61∙1 / -54∙2 - 66∙7
Mean (SD) / -0∙07 (19∙14) / 0∙92 (16∙29) / 0∙44 (17∙71)
N obs / 181 / 194 / 375
Minimum global health/QoL
Median / -16∙7 / -16∙7 / -16∙7
Range / -75∙0 - 66∙7 / -83∙3 - 50∙0 / -83∙3 - 66∙7
Mean (SD) / -15∙56 (23∙95) / -15∙81 (22∙12) / -15∙69 (22∙99)
N obs / 181 / 194 / 375
Maximum global health/QoL
Median / 8∙3 / 8∙3 / 8∙3
Range / -41∙7 - 66∙7 / -50∙0 - 66∙7 / -50∙0 - 66∙7
Mean (SD) / 10∙64 (20∙13) / 12∙71 (17∙82) / 11∙71 (18∙97)
N obs / 181 / 194 / 375
Last global health/QoL
Median / 0∙0 / 0∙0 / 0∙0
Range / -66∙7 - 66∙7 / -66∙7 - 66∙7 / -66∙7 - 66∙7
Mean (SD) / -0∙37 (23∙73) / -0∙90 (22∙85) / -0∙64 (23∙25)
N obs / 181 / 194 / 375
10 pt decrease in global health/QoL
No / 81 (44∙8) / 91 (46∙9) / 172 (45∙9)
Yes / 100 (55∙2) / 103 (53∙1) / 203 (54∙1)

Table 3: Area under the curve analysis

Treatment / Nbr Patients / Median(95%CI) / P-Value
AUC-15m / RT / 232 / 71∙62 (65∙82, 73∙89) / 0∙606
TMZ / 231 / 71∙05 (65∙17, 74∙34)
AUC-24m / RT / 232 / 72∙21 (67∙07, 75∙34) / 0∙483
TMZ / 231 / 69∙27 (65∙27, 75∙40)

Table 4: Data from van de Poll et alcompared with our own baseline data for EORTC QLG-C30 scores.

General population (all age groups, all gender) / LGG-population (baseline)
Functioning scales
TMZ / RT
Physical / 90 (15) / 86∙8 (1∙24) / 86∙3 (1∙25)
Role / 89 (21) / 77∙4 (1∙89) / 73∙6 (1∙9)
Emotional / 89 (16) / 71∙8 (1∙67) / 71∙2 (1∙67)
Cognitive / 92 (15) / 76∙2 (1∙74) / 72∙7 (1∙75)
Social / 94 (16) / 75∙3 (1∙84) / 73∙6 (1∙84)
Global quality of life / 78 (17 / 69∙1 (1∙43) / 68∙3 (1∙43)
Symptom scales
Fatigue / 17 (20) / 30∙8 (1∙69) / 30∙7 (1∙69)
Nausea/Vomiting / 2,7 (10 / 5∙9 (0∙94) / 4∙4 (0∙95)
Pain / 15 (22) / 12∙6 (1∙54) / 15∙8 (1∙55)
Single items
Dyspnea / 7,1 (17) / 9∙5 (1∙38) / 10∙5 (1∙39)
Insomnia / 14 (23) / 22∙2 (1∙93) / 21∙7 (1∙94)
Appetite loss / 3,3 (12) / 9∙2 (1∙55) / 8 (1∙55)
Constipation / 4,8 (14) / 8∙4 (1∙46) / 7∙5 (1∙46)
Diarrhoea / 3,9 (14) / 5∙6 (1∙1) / 6∙2 (1∙1)
Financial difficulties / 3,1 (13) / 19∙8 (2∙02) / 19∙2 (2∙04)

As we do not have the raw data of van de Poll et al, we did not perform statistical analysis. All scores are worse in the LGG population.

Table 5: Mean (SD) MMSE scores for both treatment arms at all time points. Exploratory 95% Confidence Intervals were computed for the difference of MMSE between groups at difference time point.

Assessment
Time / RT / TMZ / Difference (95% CI)
Baseline / 28∙5 (0∙14) / 28∙5 (0∙15) / 0 (-0∙4,0∙4)
3 months / 28∙5 (0∙16) / 28∙7 (0∙15) / 0 (-0∙5,0∙6)
6 months / 28∙6 (0∙19) / 28∙6 (0∙18) / 0∙1 (-0∙4,0∙6)
9 months / 28∙9 (0∙21) / 28∙8 (0∙2) / 0∙2 (-0∙2,0∙6)
12 months / 28∙7 (0∙16) / 29∙0 (0∙15) / -0∙1 (-0∙7,0∙4)
15 months / 28∙8 (0∙18) / 28∙9 (0∙17) / 0∙3 (-0∙1,0∙7)
18 months / 28∙9 (0∙15) / 29∙0 (0∙15) / 0∙1 (-0∙3,0∙5)
21 months / 28∙9 (0∙16) / 28∙9 (0∙16) / 0 (-0∙5,0∙4)
24 months / 28∙9 (0∙14) / 29∙1 (0∙14) / 0∙2 (-0∙2,0∙6)
27 months / 28∙7 (0∙19) / 28∙9(0∙19) / 0∙2 (-0∙2,0∙7)
30 months / 28∙7 (0∙16) / 29∙0 (0∙16) / 0∙2 (-0∙3,0∙7)
33 months / 28∙7 (0∙16) / 29∙0 (0∙16) / 0∙3(-0∙1,0∙7)
36 months / 28∙9 (0∙20) / 29∙0 (0∙21) / 0∙1(-0∙4,0∙6)

The difference was tested during 36 months of follow-up using the general mixed model with repeated measurement. The general conclusion of the test was that there was no difference between treatment arms overall.

Table 6: HRQoL scores for secondary scale communication deficit per timepoint in the intention-to-treat population.

Comm. deficit / TMZ / RT / Difference (95% CI) / P-value
baseline / 19.9 (SD=1.56) / 21 (SD=1.56) / -1.2 (-5.5,3.2) / 0.6005
3 months / 18.1 (SD=1.61) / 22.5 (SD=1.66) / -4.5 (-9,0.1) / 0.0542
6 months / 16.9 (SD=1.64) / 21.4 (SD=1.73) / -4.5 (-9.2,0.1) / 0.0571
9 months / 16.5 (SD=1.69) / 22 (SD=1.75) / -5.5 (-10.2,-0.7) / 0.0242
12 months / 15.8 (SD=1.72) / 21.1 (SD=1.8) / -5.3 (-10.2,-0.4) / 0.0339
15 months / 15 (SD=1.77) / 20.6 (SD=1.82) / -5.6 (-10.6,-0.6) / 0.0280
18 months / 16.2 (SD=1.83) / 21.9 (SD=1.85) / -5.7 (-10.8,-0.6) / 0.0274
21 months / 17.3 (SD=1.9) / 20.1 (SD=1.89) / -2.8 (-8.1,2.4) / 0.2902
24 months / 17.8 (SD=1.92) / 21.4 (SD=1.95) / -3.6 (-9,1.7) / 0.1862
27 months / 18.3 (SD=1.99) / 22.8 (SD=2.05) / -4.5 (-10.1,1.1) / 0.1131
30 months / 16.1 (SD=2.01) / 19.3 (SD=2.01) / -3.1 (-8.7,2.4) / 0.2683
33 months / 17.8 (SD=2.1) / 21.1 (SD=2.12) / -3.4 (-9.2,2.5) / 0.2615
36 months / 19.9 (SD=2.25) / 20.6 (SD=2.19) / -0.7 (-6.8,5.5) / 0.8283

Table 7: HRQoL scores for exploratory scale emotional functioning per timepoint in the intention-to-treat population.

Emotional Functioning / TMZ / RT / Difference (95% CI) / P-value
baseline / 71.7 (SD=1.65) / 71.2 (SD=1.65) / 0.5 (-4.1,5.1) / 0.8285
3 months / 75.6 (SD=1.7) / 69.2 (SD=1.77) / 6.4 (1.5,11.2) / 0.0097
6 months / 74.7 (SD=1.73) / 72.3 (SD=1.85) / 2.3 (-2.7,7.3) / 0.3600
9 months / 77.9 (SD=1.8) / 73.8 (SD=1.87) / 4.1 (-1,9.2) / 0.1173
12 months / 76.4 (SD=1.84) / 73.5 (SD=1.94) / 2.9 (-2.3,8.1) / 0.2783
15 months / 78 (SD=1.89) / 75.2 (SD=1.97) / 2.8 (-2.6,8.1) / 0.3103
18 months / 74.5 (SD=1.98) / 72.1 (SD=1.99) / 2.4 (-3.1,7.9) / 0.3931
21 months / 75.8 (SD=2.06) / 76.1 (SD=2.05) / -0.2 (-5.9,5.5) / 0.9346
24 months / 76.4 (SD=2.08) / 76.5 (SD=2.11) / -0.1 (-5.9,5.7) / 0.9797
27 months / 76.5 (SD=2.15) / 75.6 (SD=2.24) / 0.9 (-5.2,7) / 0.7664
30 months / 76.1 (SD=2.17) / 76 (SD=2.19) / 0.2 (-5.9,6.2) / 0.9593
33 months / 75.6 (SD=2.3) / 73.7 (SD=2.31) / 1.9 (-4.5,8.3) / 0.5587
36 months / 78.2 (SD=2.48) / 75.6 (SD=2.4) / 2.6 (-4.2,9.4) / 0.4550

Table 8: HRQoL scores for exploratory scale fatigue per timepoint in the intention-to-treat population.

Fatigue / TMZ / RT / Difference (95% CI) / P-value
baseline / 30.8 (SD=1.66) / 30.6 (SD=1.66) / 0.2 (-4.4,4.8) / 0.9420
3 months / 37 (SD=1.71) / 41.8 (SD=1.78) / -4.9 (-9.7,0) / 0.0493
6 months / 34 (SD=1.74) / 31.7 (SD=1.86) / 2.3 (-2.7,7.3) / 0.3678
9 months / 30.1 (SD=1.81) / 29.2 (SD=1.88) / 0.9 (-4.2,6) / 0.7355
12 months / 30.6 (SD=1.85) / 29.3 (SD=1.95) / 1.3 (-4,6.5) / 0.6397
15 months / 27.2 (SD=1.9) / 28.5 (SD=1.99) / -1.4 (-6.8,4) / 0.6193
18 months / 28.4 (SD=1.99) / 30.3 (SD=2.01) / -2 (-7.5,3.6) / 0.4838
21 months / 28.3 (SD=2.08) / 28 (SD=2.07) / 0.3 (-5.4,6) / 0.9193
24 months / 26.4 (SD=2.1) / 27.6 (SD=2.13) / -1.2 (-7.1,4.6) / 0.6793
27 months / 25.9 (SD=2.18) / 30.8 (SD=2.23) / -4.9 (-11,1.2) / 0.1166
30 months / 28.6 (SD=2.2) / 26.7 (SD=2.21) / 1.8 (-4.3,7.9) / 0.5608
33 months / 29.3 (SD=2.34) / 28.8 (SD=2.34) / 0.5 (-6,6.9) / 0.8911
36 months / 24.9 (SD=2.52) / 27.7 (SD=2.43) / -2.8 (-9.7,4) / 0.4159

Table 9: HRQoL scores for exploratory scale cognitive functioning per timepoint in the intention-to-treat population.

Cognitive Functioning / TMZ / RT / Difference (95% CI) / P-value
baseline / 76.2 (SD=1.73) / 72.9 (SD=1.72) / 3.3 (-1.5,8.1) / 0.1758
3 months / 75.8 (SD=1.77) / 70.8 (SD=1.83) / 5 (0.1,10) / 0.0472
6 months / 76 (SD=1.79) / 72 (SD=1.9) / 4 (-1.1,9.1) / 0.1278
9 months / 77 (SD=1.85) / 72.9 (SD=1.92) / 4.1 (-1.1,9.4) / 0.1208
12 months / 75.4 (SD=1.89) / 75.3 (SD=1.97) / 0.2 (-5.2,5.5) / 0.9547
15 months / 78.6 (SD=1.93) / 73.9 (SD=2) / 4.7 (-0.8,10.1) / 0.0918
18 months / 78.6 (SD=2.01) / 72.2 (SD=2.02) / 6.4 (0.8,12) / 0.0252
21 months / 76.3 (SD=2.08) / 73.2 (SD=2.07) / 3.1 (-2.7,8.9) / 0.2919
24 months / 75.9 (SD=2.1) / 71.3 (SD=2.12) / 4.5 (-1.3,10.4) / 0.1282
27 months / 79.7 (SD=2.16) / 71.8 (SD=2.23) / 7.9 (1.8,14) / 0.0108
30 months / 76.7 (SD=2.18) / 74.6 (SD=2.19) / 2.1 (-3.9,8.2) / 0.4941
33 months / 73.9 (SD=2.29) / 72.3 (SD=2.3) / 1.6 (-4.8,8) / 0.6260
36 months / 78 (SD=2.46) / 72.8 (SD=2.38) / 5.2 (-1.5,11.9) / 0.1306

Table 10: HRQoL scores forexploratory scalenausea/vomiting per timepoint in the intention-to-treat population.

Nausea / Vomiting / TMZ / RT / Difference (95% CI) / P-value
baseline / 6 (SD=0.91) / 4.3 (SD=0.91) / 1.6 (-0.9,4.1) / 0.2152
3 months / 12.5 (SD=0.96) / 10.4 (SD=1.02) / 2 (-0.7,4.8) / 0.1444
6 months / 11.4 (SD=0.99) / 2.8 (SD=1.09) / 8.6 (5.7,11.5) / <.0001
9 months / 8.3 (SD=1.05) / 4.3 (SD=1.12) / 3.9 (0.9,6.9) / 0.0101
12 months / 7.5 (SD=1.08) / 3.7 (SD=1.17) / 3.8 (0.6,6.9) / 0.0183
15 months / 5.1 (SD=1.14) / 3.9 (SD=1.21) / 1.2 (-2.1,4.4) / 0.4869
18 months / 4.4 (SD=1.21) / 2.5 (SD=1.23) / 1.9 (-1.5,5.3) / 0.2635
21 months / 4.6 (SD=1.29) / 3 (SD=1.28) / 1.6 (-2,5.2) / 0.3805
24 months / 5.7 (SD=1.31) / 2.8 (SD=1.34) / 2.9 (-0.8,6.6) / 0.1216
27 months / 4.4 (SD=1.38) / 2.9 (SD=1.43) / 1.5 (-2.4,5.4) / 0.4484
30 months / 5.5 (SD=1.4) / 2.3 (SD=1.41) / 3.2 (-0.7,7.1) / 0.1052
33 months / 4.5 (SD=1.51) / 3 (SD=1.52) / 1.5 (-2.7,5.7) / 0.4748
36 months / 4.4 (SD=1.67) / 4.6 (SD=1.6) / -0.2 (-4.7,4.3) / 0.9262

Table 11: HRQoL scores for exploratory scale constipation per timepoint in the intention-to-treat population.

Constipation / TMZ / RT / Difference (95% CI) / P-value
baseline / 8.4 (SD=1.46) / 7.6 (SD=1.45) / 0.8 (-3.2,4.8) / 0.6950
3 months / 17.4 (SD=1.52) / 9.6 (SD=1.6) / 7.8 (3.4,12.1) / 0.0005
6 months / 15.5 (SD=1.55) / 8.4 (SD=1.69) / 7.1 (2.6,11.6) / 0.0021
9 months / 13.5 (SD=1.63) / 8.4 (SD=1.73) / 5.1 (0.4,9.7) / 0.0322
12 months / 12.4 (SD=1.68) / 7.7 (SD=1.8) / 4.7 (-0.1,9.6) / 0.0545
15 months / 8.8 (SD=1.75) / 9.7 (SD=1.84) / -1 (-6,4) / 0.7022
18 months / 11.3 (SD=1.85) / 8.1 (SD=1.87) / 3.2 (-1.9,8.4) / 0.2222
21 months / 9.9 (SD=1.95) / 7.9 (SD=1.94) / 2 (-3.4,7.4) / 0.4776
24 months / 9.3 (SD=1.98) / 12.3 (SD=2.01) / -3 (-8.5,2.6) / 0.2923
27 months / 10.7 (SD=2.07) / 12.7 (SD=2.16) / -2.1 (-7.9,3.8) / 0.4874
30 months / 8 (SD=2.09) / 11.2 (SD=2.11) / -3.3 (-9.1,2.5) / 0.2707
33 months / 8.7 (SD=2.24) / 13 (SD=2.25) / -4.3 (-10.5,1.9) / 0.1757
36 months / 8.5 (SD=2.45) / 7.8 (SD=2.36) / 0.7 (-6,7.4) / 0.8378

Table 12: HRQoL scores for exploratory scale appetite loss per timepoint in the intention-to-treat population.

Appetite loss / TMZ / RT / Difference (95% CI) / P-value
baseline / 9.3 (SD=1.5) / 8.1 (SD=1.5) / 1.2 (-3,5.4) / 0.5718
3 months / 17.4 (SD=1.56) / 21.7 (SD=1.65) / -4.3 (-8.7,0.2) / 0.0601
6 months / 14.7 (SD=1.6) / 14.2 (SD=1.74) / 0.5 (-4.2,5.1) / 0.8433
9 months / 13.6 (SD=1.68) / 8.5 (SD=1.77) / 5.1 (0.3,9.9) / 0.0361
12 months / 14.8 (SD=1.73) / 9.8 (SD=1.85) / 5 (0,9.9) / 0.0495
15 months / 8.4 (SD=1.8) / 8.9 (SD=1.89) / -0.5 (-5.6,4.6) / 0.8457
18 months / 10.9 (SD=1.9) / 8.1 (SD=1.93) / 2.8 (-2.5,8.2) / 0.2924
21 months / 8 (SD=2.01) / 7.9 (SD=1.99) / 0.1 (-5.5,5.6) / 0.9755
24 months / 11.2 (SD=2.03) / 9.3 (SD=2.07) / 1.9 (-3.8,7.6) / 0.5115
27 months / 9.4 (SD=2.12) / 6.2 (SD=2.19) / 3.1 (-2.8,9.1) / 0.3048
30 months / 11.5 (SD=2.16) / 8.1 (SD=2.17) / 3.3 (-2.7,9.3) / 0.2755
33 months / 5.7 (SD=2.3) / 8.2 (SD=2.31) / -2.5 (-8.9,3.9) / 0.4405
36 months / 7.9 (SD=2.52) / 6.8 (SD=2.42) / 1.1 (-5.7,8) / 0.7491

Table 13: HRQoL scores for exploratory scale hair loss per timepoint in the intention-to-treat population.

Hair loss / TMZ / RT / Difference (95% CI) / P-value
baseline / 6.8 (SD=1.6) / 8.7 (SD=1.62) / -1.9 (-6.3,2.6) / 0.4104
3 months / 6.5 (SD=1.67) / 36.1 (SD=1.74) / -29.6 (-34.4,-24.9) / <.0001
6 months / 3.7 (SD=1.71) / 18.8 (SD=1.85) / -15.1 (-20,-10.2) / <.0001
9 months / 5.6 (SD=1.79) / 14.9 (SD=1.88) / -9.2 (-14.3,-4.1) / 0.0004
12 months / 4.1 (SD=1.87) / 13.6 (SD=1.99) / -9.6 (-14.9,-4.2) / 0.0005
15 months / 4 (SD=1.93) / 10.7 (SD=2.02) / -6.7 (-12.1,-1.2) / 0.0172
18 months / 7.4 (SD=2.03) / 9.6 (SD=2.05) / -2.2 (-7.9,3.4) / 0.4431
21 months / 6.8 (SD=2.19) / 9 (SD=2.12) / -2.2 (-8.2,3.7) / 0.4624
24 months / 4.7 (SD=2.22) / 10.6 (SD=2.21) / -6 (-12.1,0.2) / 0.0565
27 months / 5.5 (SD=2.3) / 11.4 (SD=2.39) / -5.8 (-12.3,0.7) / 0.0791
30 months / 6.1 (SD=2.33) / 12.5 (SD=2.32) / -6.3 (-12.8,0.1) / 0.0541
33 months / 4.4 (SD=2.46) / 9.1 (SD=2.5) / -4.7 (-11.6,2.2) / 0.1797
36 months / 8.9 (SD=2.79) / 10.8 (SD=2.59) / -1.9 (-9.4,5.6) / 0.6201

Randomisation per centre

Institution / Principle investigator / Randomised / % of total
Erasmus MC Cancer Institute - Location Daniel den Hoed, Rotterdam, The Netherlands / van den Bent-Bromberg / 39 / 8∙18
Hopitaux de Paris - La Pitié Salpétrière∙ Parios, France / Hoang-Xuan / 37 / 7∙76
Princess Margaret Hospital, Toronto. Canada / Mason / 24 / 5∙03
Ospedale Bellaria, Bologna, Italy / Brandes / 18 / 3∙77
Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland / Stupp / 17 / 3∙56
Institut Bergonie, Bordeaux, France / Kantor / 16 / 3∙35
Medisch Centrum Haaglanden – Westeinde, Den Haag, The Netherlands / Taphoorn / 16 / 3∙35
Centre Eugene Marquis, Rennes, France / Vauleon / 15 / 3∙14
Peter MacCallum Cancer Institute, Melbourne, Australia / Ryan / 14 / 2∙94
Universitaetsklinikum Heidelberg - Head Hospital, Heidelberg, Germany / Wick / 13 / 2∙73
University Medical Center, Groningen, The Netherlands / Enting / 12 / 2∙52
IRCCS Ospedale San Raffaele Scientific Institute, Milano, Italy / Reni / 12 / 2∙52
Institut Gustave Roussy, Villejuif, France / Dhermain / 11 / 2∙31
Hopitaux de Marseille - Hôpital de La Timone, Marseille, France / Chinot / 11 / 2∙31
Radboud University Medical Center, Nijmegen, The Netherlands / Gijtenbeek / 11 / 2∙31
VU University Medical Center, Amsterdam, The Netherlands / Reijneveld / 11 / 2∙31
BC Cancer Agency, Vancouver, Canada / Thiessen / 11 / 2∙31
U.Z. Leuven - Campus Gasthuisberg, Belgium / Clement / 10 / 2∙1
Institut Catala d’Oncologia (ICO). Hospital Germans Trias Pujol, Badalona, Barcelona, Spain / Balan / 10 / 2∙1
I.P.O. Francisco Gentil - Centro De Lisboa, Lisboa, Portugal / Bravo-Marques / 10 / 2∙1
General Hospital AKH, Medical University Vienna, Austria / Marosi / 9 / 1∙89
Rambam Health Care Campus, Oncology Institute, Haifa, Israel / Tzuk-Shina / 8 / 1∙68
Maastro Clinic (Maastricht Radiation Oncology) and MUMC Universiteit Maastricht, The Netherlands / Baumert / 8 / 1∙68
Tom Baker Cancer Centre, Calgary, Canada / Nordal / 7 / 1∙47
Hospital Clinic Universitari de Barcelona, Barcelona, Spain / Verger / 6 / 1∙26
Universitaetskliniken Regensburg, Regensburg, Germany / Bogdahn / 6 / 1∙26
Royal North Shore Hospital, Sydney, Australia / Back / 6 / 1∙26
Universitaetsklinikum Tubingen, Tubingen, Germany / Borchers / 6 / 1∙26
Universitair Ziekenhuis Brussel, Belgium / Neyns / 5 / 1∙05
Academic Medical Center, Amsterdam, The Netherlands / Reijneveld / 5 / 1∙05
Austin-Repatriation Medical Centre, Melbourne, Australia / Wada / 5 / 1∙05
National Cancer Institute, Cairo, Egypt / Samy El-Badawy / 5 / 1∙05
Hopitaux Universitaires Bordet-Erasme - Hopital Universitaire Erasme, Belgium / de Witte / 4 / 0∙84
Istituto Regina Elena, Roma, Italy / Carapella / 4 / 0∙84
London Regional Cancer Center, London, United Kingdom / MacDonald / 4 / 0∙84
Umea Universitet, Umea, Sweden / Henriksson / 4 / 0∙84
Royal Brisbane And Women's Hospital, Brisbane, Australia / Fay / 4 / 0∙84
Universitaetsklinikum Leipzig, Leipzig, Germany / Kortmann / 4 / 0∙84
Azienda Ospedaliera Città della Salute e della Scienza di Torino - Ospedale San Giovanni - Dipartimento Neuroscienze, Torino, Italy / Soffietti / 4 / 0∙84
Royal Free Hospital, London, United Kingdom / Wilson / 3 / 0∙63
University College Hospital London, United Kingdom / Short / 3 / 0∙63
Clatterbridge Cancer Centre NHS Foundation Trust, Bebington, Wirral, United Kingdom / Haylock / 3 / 0∙63
Centre Hospitalier de l'Université de Montreal - Hopital Notre-Dame, Montreal, Canada / Belanger / 3 / 0∙63
Linkoping University Hospital, Linkoping, Sweden / Hallbeck / 3 / 0∙63
Princess Alexandra Hospital, Brisbane, Australia / Lehman / 3 / 0∙63
Hospital General Vall D'Hebron, Barcelona, Spain / Maldonado Pijoan / 3 / 0∙63
Centre Paul Strauss, Strasbourg, France / Schott / 2 / 0∙42
Dr. Bernard Verbeeten Instituut, Tilburg, The Netherlands / Veninga / 2 / 0∙42
Universitaetsspital Zurich, Zurich, Switzerland / Hofer / 2 / 0∙42
Hamilton Health Sciences, Juravinski Cancer Centre, Hamilton, Canada / Whitton / 2 / 0∙42
Western General Hospital, Edinburgh, United Kingdom / Erridge / 2 / 0∙42
Alfred Hospital, Melbourne, Australia / Ruben / 2 / 0∙42
Sir Charles Gairdner Hospital, Perth, Australia / Joseph / 2 / 0∙42
Institut Catala d’Oncologia (ICO). Hospital Doctor Josep Trueta Girona, Spain / Fuentes Raspall / 2 / 0∙42
Canterbury District Health Board, Christchurch, New Zealand / James / 2 / 0∙42
Institut Claudius Regaud, Toulouse, France / Cohen-Moyal / 1 / 0∙21
CHU de Bordeaux - Groupe Hospitalier Saint-André, Bordeaux, France / Maire / 1 / 0∙21
Oncology Institute of Southern Switzerland - Ospedale Regionale Bellinzona e Valli, Bellinzona, Switzerland / Pesce / 1 / 0∙21
The Christie NHS Foundation Trust, Manchester, United Kingdom / Mcbain / 1 / 0∙21
Royal Marsden Hospital - Sutton, Surrey and Chelsea, London, United Kingdom / Brada / 1 / 0∙21
Cambridge University Hospital NHS - Addenbrookes Hospital, Cambridge, United Kingdom / Jefferies / 1 / 0∙21
Skane University Hospital, Lund, Sweden / Kinhult / 1 / 0∙21
National Institute of Neurosurgery, Budapest, Hungary / Sipos / 1 / 0∙21
Uppsala University Hospital - Akademiska Sjukhuset, Uppsala, Sweden / Smits / 1 / 0∙21
Nova Scotia Cancer Centre, Halifax, Nova Scotia, Canada / Rheaume / 1 / 0∙21
Liverpool Hospital, Sydney, Australia / Koh / 1 / 0∙21
Prince Of Wales Hospital, Sydney, Australia / Hovey / 1 / 0∙21
Calvary Mater Hospital, Newcastle, Australia / Ludbrook / 1 / 0∙21
Mcgill University Health Centre - Gerald Bronfman Centre - Dept Of Oncology, Montreal, Canada / Kavan / 1 / 0∙21
Lancashire Teaching Hospitals NHS Foundation Trust - Royal Preston Hospital, Preston, United Kingdom / Hindley / 1 / 0∙21
National University Hospital, Singapore, Singapore / Tang / 1 / 0∙21
Total / 477 / 100

Participating clinical investigators

Austria:

• Christine Marosi, General Hospital AKH, Medical University Vienna, Austria

Australia:

• Michael Fay, Royal Brisbane And Women's Hospital, Brisbane, Australia
• Elizabeth Hovey, Prince Of Wales Hospital, Sydney, Australia
• David Joseph, Sir Charles Gairdner Hospital, Perth, Australia
• Eng-Siew Koh, Liverpool Hospital, Sydney, Australia
• Joanna Ludbrook, Calvary Mater Hospital, Newcastle, Australia
• Michael Back, Gillian Lamoury, Royal North Shore Hospital, Sydney, Australia
• Margot Lehman, Princess Alexandra Hospital, Brisbane, Australia
• Andrew Pullar, Mater Adult Hospital, Brisbane, Australia
• Jeremy Ruben, Alfred Hospital, Melbourne, Australia
• Gail Ryan, Peter MacCallum Cancer Institute, Melbourne, Australia
• John Simes, Royal Prince Alfred Hospital, Sydney, Australia
• Mori Wada, Austin-Repatriation Medical Centre, Melbourne, Australia

Belgium:

• Paul Clement, U.Z. Leuven - Campus Gasthuisberg, Belgium
• Bart Neyns, Universitair Ziekenhuis Brussel, Belgium
• Olivier de Witte, Hopitaux Universitaires Bordet-Erasme - Hopital Universitaire Erasme, Belgium

Canada:

• Karl Belanger, Centre Hospitalier de l'Université de Montreal - Hopital Notre-Dame, Montreal, Canada
• Margot Burnell, Saint John Regional Hospital, Saint John, New Brunswick, Canada
• Petr Kavan, Mcgill University Health Centre - Gerald Bronfman Centre - Dept Of Oncology, Montreal, Canada
• Warren Mason, University Health Network - Oci / Princess Margaret Hospital, Toronto. Canada
• Robert Nordal, Tom Baker Cancer Centre, Calgary, Canada
• Marshall Pitz, Cancercare Manitoba, Winnipeg, Canada
• Dorianne Rheaume, Nova Scotia Cancer Centre, Halifax, Nova Scotia, Canada
• John P. Rossiter, Dept. of Pathology and Molecular Medicine, Queen’s University, Kingston, Ontario, Canada
• Muhammed Salim, Allan Blair Cancer Centre, Regina, Saskatchewan, Canada
• Brian Thiessen, BC Cancer Agency, Vancouver, Canada
• May Tsao, Odette Cancer Centre - Sunnybrook Health Sciences Centre, Toronto, Canada
• Chris Watling, London Regional Cancer Center, London, Canada
• Anthony Whitton, Hamilton Health Sciences, Juravinski Cancer Centre, Hamilton, Canada

Egypt:

• Samy El-Badawy, National Cancer Institute, Cairo, Egypt

France:

• Michel Bolla, CHU de Grenoble - La Tronche - Hôpital A. Michallon, France
• Mario Campone, Institut de Cancerologie de l’Ouest (ICO) - Centre Rene Gauducheau, Nantes, France
• Elizabeth Cohen-Jonathan Moyal, Institut Claudius Regaud, Toulouse, France
• Olivier Chinot, Assistance Publique - Hopitaux de Marseille - Hôpital de La Timone, Marseille, France
• Frederic Dhermain, Institut Gustave Roussy, Villejuif, France
• Khe Hoang-Xuan & Loic Feuvret, Assistance Publique - Hopitaux de Paris - La Pitié Salpétrière, Paris, France
• Guy Kantor, Institut Bergonie, Bordeaux, France
• Roland Schott, Centre Paul Strauss, Strasbourg, France
• Elodie Vauleon, Centre Eugene Marquis, Rennes, France

Germany:

• Ulrich Bogdahn, Universitaetskliniken Regensburg, Regensburg, Germany
• Christian Borchers, Michael Weller, Universitaetsklinikum Tubingen, Tubingen, Germany
• Rolf-Dieter Kortmann, Universitaetsklinikum Leipzig, Leipzig, Germany
• Gabriele Schackert, Universitaetsklinikum Carl Gustav Carus, Dresden, Germany
• Wolfgang Wick, Andreas von Deimling, Christian Hartmann & David Capper, Universitaetsklinikum Heidelberg - Head Hospital, Heidelberg, Germany

Hungary