Reijneveld 2016

Methods	Design: Results of follow‐up of RCT Country: 19 countries (Australia, Austria, Belgium, France, Germany, Hungary, Italy, the Netherlands, Portugal, Spain, Sweden, Switzerland, Australia, New Zealand, Singapore, Canada, Egypt, Israel, UK) Accrual dates: December 2005 to December 2012 Trial reg: EudraCT. Number 2004‐002714‐11 and Clinical/Trials.gov, number NCT00182819 Funding: Merck Sharp & Dohme, Merck & Co (study chemotherapy drugs and grant), National Cancer Institute, Swiss Cancer League, National Centre for Health Research, Cancer Research UK, Canadian Cancer Research Institute, National Health and Medical Research Council, European Organisation for Research and Treatment of Cancer, Cancer Research Fund. It was stated that the funders of the research had no role in study design, data collection, data analysis, data interpretation or writing the report. Conflict of interest: two investigators report personal fees from Hoffmann La Roche outside the submitted work. One author reports grants and non financial support from Roche, Ipsen, and Astra‐Zeneca outside the submitted work. One author reports grants from Celgene, Novartis and Pharmamar and personal fees from Celgene, Boehringer, Genentec, Lilly and Merck‐Serono outside the submitted work. Grants from funders as above.
Participants	No. randomised: 477 assigned No. analysed: reported that 477 in ITT analyses (but considerable amounts of missing data for outcomes relevant to this review) Inclusion/exclusion criteria: Age: adults aged 18 years or more (median, 43 (36 to 52 interquartile range) in the radiotherapy group and 45 (37 to 53) in the chemotherapy group; 38% (92/240) less than 40 in the radiotherapy group and 36% (85/237) in the chemotherapy group) Gender: 58% men and 42% women in both groups. Glioma type: Astrocytoma WHO grade II 37% in the radiotherapy group and 33% in the chemotherapy group. Oligoastrocytoma WHO grade II 24% and 25%, Oligodendroglioma WHO grade II 39% and 41%. Glioma grade: low‐grade glioma confirmed 95% and 89%. Resection/biopsy: radiotherapy: biopsy 40%, partial removal 44%, total removal 15%; chemotherapy: biopsy 39%, partial removal 42%, total removal 19% Anti‐epileptics/SSRIs: reason for treatment: refractory seizures 12% and 14%. Duration of FU: time between biopsy or surgery to study treatment median 4.8 months in both groups but considerable variation (2.9 to 18.3 IQR months in the radiotherapy group and 2.6 and 26.4 months in the chemotherapy group). Time from initial diagnosis and study treatment medians 5.1 and 6.0 months.
Interventions	Arm 1: (n = 240) radiotherapy. Total dose of 50.4 Gy in 28 fractions of 1.8 Gy once daily for 5 days per week up to a maximum treatment period of 6.5 weeks. Reasons for treatment discontinuation included major worsening of neurological or mental status or other medical condition that would preclude continuation. Dose adjustments were not recommended. Arm 2: (n = 237) chemotherapy. 75 mg/m² oral temozolomide daily for 21 of 28 days (1 cycle) repeated for a maximum of 12 cycles until disease progression or unacceptable toxicity. The treatment was withheld if low neutrophil and platelet counts and resumed on recovery. Patients with severe recurrent toxicity despite dose reduction discontinued treatment.
Outcomes	Reported review outcomes: adverse events, health‐related quality of life (HRQoL scales including global health or quality of life status, role and functioning, social functioning, communication deficit, visual disorder, motor dysfunction, drowsiness) and cognitive functioning (MMSE). Outcomes reported up to 36 months. Other reported study outcomes: Primary outcome of trial was progression‐free survival.
Notes	Authors emailed 6 February 2019.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Centralised randomisation using a minimisation technique (stratified for WHO performance status, age, presence or absence of contrast enhancement on MRI, 1p status, and treatment centre
Allocation concealment (selection bias)	Low risk	Probably low risk as randomisation was centralised
Blinding of participants and personnel (performance bias) All outcomes	High risk	There was no attempt to blind participants and staff as treatments were different. The different types of treatment may have affected patient compliance
Blinding of outcome assessment (detection bias) All outcomes	High risk	For self‐assessed outcomes relevant to this review the different treatments may have affected response rates and those experiencing worse outcomes may have been less likely to respond. It was reported that response was lowest in patients with poor performance status
Incomplete outcome data (attrition bias) All outcomes	High risk	For long‐term outcomes relevant to this review response rates were less than 70% at 2 years; for early assessments response rates were greater in the chemotherapy group although this disparity between groups decreased over time. Denominators for some outcomes were not clear. There was variation between treatment centres in response rates.
Selective reporting (reporting bias)	Unclear risk	This was a registered trial with specified outcomes using standard measures. The frequency of testing may have introduced risk of multiple testing.
Other bias	Unclear risk	Treatment modalities differed significantly in duration and intensity and early differences detected in QoL may have reflected these differences in treatment modalities.