Hoff 2011.

Study characteristics
Methods	Design: 2 RCTs, parallel 2‐arm, multicentre trial (DAHANCA 7 is a follow‐on trial from DAHANCA 5, undertaken to extend its usefulness) Setting: multiple oncology departments in Denmark Recruitment: January 1986 to September 1990 (DAHANCA 5) and January 1992 and October 1997 (DAHANCA 7) Maximum follow‐up: 5 years or until death
Participants	People with diagnosed invasive squamous cell carcinoma of the supraglottic larynx or pharynx and oral cavity, with no evidence of distant metastases Participants were stratified before randomisation according to sex, institution, site of tumour, tumour status, and haemoglobin status DAHANCA 5 assessed effects of the hypoxic radiosensitiser nimorazole plus radiotherapy vs placebo plus radiotherapy, and contained a subrandomised group of trial participants presenting with condition‐specific low Hb to either transfusion or no transfusion DAHANCA 7 assessed effects of the same dose of nimorazole given with 5 fractions of radiotherapy/week vs 6 fractions of radiotherapy/week. In this trial, again, all participants presenting with a condition‐specific low Hb were subrandomised to transfusion or no transfusion Trial flow: within the original DAHANCA 5 trial (n = 414), 171 participants with low pre‐irradiation Hb (women < 13 g/dL; men < 14.5 g/dL) were randomised to receive or not receive transfusion before final randomisation to nimazorole vs placebo. Within the DAHANCA 7 trial (n = 786), 321 participants with low haemoglobin (as above) were randomised to different regimens of administration of nimazorole (5 vs 6 radiotherapy fractions per week) Of the 492 participants from both trials, 26 were not randomised (no reason given); 1 participant in the no‐transfusion group was 'in the wrong strata' and was not considered evaluable Analysis included: Liberal group: n = 235; M/F: 195/35; 50% were ≥ 60 years old and 50% < 60 years old Restrictive group: n = 231; M/F: 195/35; 50% were ≥ 60 years old and 50% < 60 years old
Interventions	Liberal group: received transfusion Restrictive group: received no transfusion For those who met the criteria for low Hb, transfusions were given with packed RBCs to achieve Hb in the 'high' value range (1 unit at a time; without leucocyte depletion as standard). If during treatment, Hb fell below the values indicated above, transfusion was repeated. Haemoglobin level was measured every fortnight
Outcomes	Primary outcome (both trials): locoregional control after radiotherapy (defined as complete and persistent disappearance of disease in the primary tumour site and also the regional lymph nodes)  Secondary outcomes: local and regional control (with and without salvage surgery), overall survival, early and late treatment‐related morbidity The influence of haemoglobin on tumour response to irradiation was also a stated goal (Overgaard 1998 p 136)
Notes	Trial registration: none confirmed Trials (original trials and analyses) funding: sources of support included the Danish Cancer Society; Legatstiftelsen, Pedersholm and the Danish Cancer Society; Clinical Resarch Unit at Aarhus Oncological Centre; the Lundbeck Foundation Center for Interventional Reseach in Radiation Oncology; and the Danish Council for Strategic Research COI statement by investigators: no conflicts of interest reported by trial authors
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"In patients where all eligibility criteria were fulfilled, patient data was entered into a local computer which generated the correct strata and randomization number and at the same time printed a confirmation letter which was sent to the data center ... " (Overgaard 1998 p 136) Procedures for the later trial (DAHANCA 7) are described as similar, although here, it is noted that randomisation confirmation was made by telephone, not by letter (Overgaard 2003)
Allocation concealment (selection bias)	Low risk	"Each center was supplied with sealed envelopes indicating the randomization code. These envelopes were kept outside the radiotherapy department (at the hospital pharmacy) and could only be reached in the case of a clinical situation where knowledge of the presence of active drug was crucial for the further treatment of the patient. This did not happen in the present study and all envelopes were returned intact to the data center after completion of the trial. The trial has been maintained blinded during followup and the involved institutions are still unaware of which drug treatment the individual patients received" (Overgaard 1998 p 136) No details of allocation concealment were given for the later trial, but as was already mentioned, procedures were described as similar
Blinding of participants and personnel (performance bias) Objective outcomes	Low risk	Blinding of personnel for the pharmacological component of DAHANCA 5 was manifestly met (see above) and could not be met for DAHANCA 7 (where the intervention involved number of doses). Blinding of the transfusion component of both trials was, as ever, not feasible, but in our view, for objective outcomes such as mortality (the primary outcome used within this review), risk of bias remains 'low'
Blinding of outcome assessment (detection bias) Objective measures	Low risk	Blinding of mortality (the primary outcome used within this review) is not relevant. In addition, we note that "all diagnostic, therapeutic and follow‐up data were validated and processed by the DAHANCA data center. To optimize the data quality, the events recorded were crosschecked with the hospital records to ensure correct registration of the site or sites of failure and course of death" (Overgaard 1998 p 138)
Blinding of outcome assessment (detection bias) Subjective measures	Low risk	No data from subjective outcomes (e.g. function)
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data from 465 of 466 participants randomised were used within the analysis
Selective reporting (reporting bias)	Unclear risk	No reporting bias was apparent, but in the absence of prospective registration or a trial protocol, assessment must remain 'unclear'
Other bias	Unclear risk	DAHANCA 5: no other biases noted DAHANCA 6: "when the recruitment number was reached in October 1997, it became apparent that the number of patients with glottic tumours would not express enough events to secure a conclusive outcome in this groups of patients. We closed the DAHANCA 7 protocol ..." (Overgaard 2003)