This review found that adjuvant carmustine wafers conferred no survival advantage in patients with Grade III or IV tumours. With temozolomide there was limited evidence of a small but significant advantage in patients with Grade IV and III tumours in overall survival and progression free survival. The authors' conclusions seem appropriate given the strength of the evidence presented.

To assess the clinical and cost effectiveness of carmustine wafers (BCNU-W) temozolomide (TMZ) as an adjuvant to surgery and radiotherapy compared with radiotherapy and surgery alone (only clinical effectiveness is presented in this abstract).

MEDLINE, The Cochrane Library, EMBASE, DARE, NHS EED and the HTA database, ISI Web of Science and ISI Proceedings were searched from inception to August 2005 for studies published in English (abstracts were excluded). Search terms were reported. Bibliographies and the US Food and Drugs Administration website and the National Research Register were searched.

For the BCNU-W review, studies of BCNU-W adjunctive to surgery with subsequent radiotherapy (with or without standard systemic chemotherapy) were eligible for inclusion. The comparator had to be either placebo wafer with or without radiotherapy, or surgery with or without radiotherapy, chemotherapy and standard antineoplastic agents (other than those in the intervention). For the review of temozolomide, studies of surgery followed by radiotherapy with concomitant temozolomide, then an adjuvant course of temozolomide compared with surgery followed by radiotherapy, with or without systemic chemotherapy with standard antineoplastic agents (other than those in the intervention) were eligible for inclusion.

Systematic reviews, randomised controlled trials (RCTs), non-randomised controlled studies (where they gave best estimates of a required parameter or RCT data was scanty or uninformative) of children and adults with newly diagnosed grade III or IV primary gliomas were eligible for inclusion for both reviews.

BCNU-W: The included studies were of patients aged between 21 and 86. Most exhibited Grade IV tumours and Karnofsky performance status ranging from 40 to 100. The treatments used were intraoperative placement of greater than or equal to eight wafers (61.6mg BCNU) with radiotherapy (regimens and doses varied where reported). The included RCTs excluded patients over 65 years old. Follow-up ranged from 12 to more than 24 months in the RCTs and from seven (mean) to 16.8 (median) months in the observational studies.

Temozolomide: The included studies were of patients from 23 to 75 years with Grade IV tumours (where reported). In the included studies the dose of temozolomide was 75mg/m²/day for six to seven weeks in most studies; adjuvant temozolomide was 150 to 200mg/m2/day in various regimes with 60 Gy radiation therapy for six weeks. Follow up ranged from a median of 11 to 18 months.

Studies were selected independently by two reviewers. Disagreements were resolved by discussion.

RCTs were assessed with regard to sample size, selection bias, performance bias, attrition bias and intention to treat analysis and detection bias. Potential conflicts of interest were noted. The case series were assessed using the same criteria where applicable and also whether the study was prospective and whether it enrolled consecutive patients.

One reviewer performed the validity assessment and this was checked by a second reviewer.

Overall survival, periodic survival rates, 12-month survival rates and progression free survival (PFS) and where appropriate post-progression survival were extracted, as actual numbers where possible, into a data extraction form.

Data were extracted by one reviewer and checked by another.

Absolute survival data were combined at a fixed time point using a random-effects model, where data were available. Studies were grouped by study design and outcome and discussed narratively in the text and presented in tables.

BCNU-W: Four studies were included in the review of BCNU-W (n=339): two RCTs (n=240 and n=32) and two case series (n=67). Both RCTs appeared to use adequate allocation concealment and randomisation (although both had imbalances in Grade III tumours between study arms). Blinding may have been compromised. The larger RCT rigorously used intention to treat analyses, but used questionable statistical methods. All studies allowed treatment at the investigator's discretion post-study, which may have confounded survival rates.

BCNU-W was associated with improved overall survival in both RCTs (median 2.3 and 4.2 months gained). Analysis of the the larger RCT found this was not statistically significant, based on protocol specified unstratified data (hazard ratio 0.77, 95% CI: 0.57 to 1.03, p=0.08).

There was no significant improvement in progression-free survival in all patients or overall survival in the overall population or in the Grade IV glioma subgroup in either RCT. Twelve-month survival was significantly better with BCNU-W in the smaller RCT (absolute survival of 62.5% BCNU-W patients versus 49.2, p=0.029), but not the larger RCT.

Intracranial hypertension was the only adverse event significantly more common with BCNU-W (one RCT, n=240), 9.2% versus 1.7%, p=0.019.

Temozolomide: Four studies were included in the review of temozolomide (n=791): two RCTs (n=573 and n=130) and two case series (n=88). Neither RCT was placebo-controlled and drop out rates were high. Re-analysis of one RCT suggested that a significant minority had Grade III tumours. One RCT was adequately powered. Neither RCT reported randomisation methods. Both RCTs were susceptible to performance and detection biases and all studies were susceptible to late performance bias. The larger RCT rigorously used intention to treat analyses.

In both RCTs, radiotherapy combined with temozolomide was associated with a significant increase in survival compared with radiotherapy alone. In the larger trial, the hazard ratio was 0.54, 95% CI: 0.45 to 0.64, p<0.001 and in the smaller trial the hazard ratio was 0.66, 95% CI not reported, p=0.003. Temozolomide treatment was associated with significant survival benefit (median months gained 2.5 and 5.7).

Both RCTs reported that a greater proportion of patients treated with temozolomide survived at six, 12, 18 and 24 months. At 24 months, 26.5% of temozolomide patients were alive compared with 10.4% of the control patients. The RCTs also found that progression-free survival was significantly increased in the radiotherapy plus temozolomide groups, which improved progression free survival by 1.8 months (95% CI: 1.4 to 2.7, p<0.001). There was little difference between arms in post-progression survival in both RCTs.

One RCT reported premature discontinuation of temozolomide in 5% of cases due to severe myelosuppression. Nausea/vomiting occurred in patients treated with temozolomide despite compulsory or discretionary antiemetic medication in all studies. Grade 2 fatigue, rash and nausea/vomiting and grade 3 or 4 fatigue, unspecified constitutional symptoms and infection were significantly more common in the radiotherapy plus temozolomide arm of one RCT.

Subgroup analyses were also reported.

No direct comparisons of BCNU-W and temozolomide were identified.

The cost of surgery and radiotherapy, with follow-up, treatment of adverse effects and end of life care was estimated at £17,000 per patient. BCNU-W added an additional £6,600. This was modelled as £6,600 per patient for 0.122 quality-adjusted life-years (QALYs). The base-case incremental cost-effectiveness ratio (ICER) was £54,500 per QALY.

The cost per patient treated with surgery and radiotherapy, including adverse effects and end of life care was approximately £17,000. Temozolomide (adjuvant and concomitant) added an additional cost of approximately £7,800. This was modelled as £7,800 per patient for 0.217 QALYs. The base-case incremental cost-effectiveness ratio was £36,000 per QALY.

BCNU-W did not not confer a significant advantage in terms of survival in patients with Grade III or Grade IV tumours. There was no increase in progression free survival.

There was limited evidence of a small but significant advantage in survival for patients with Grade IV and III (7% to 8%) tumours in overall survival and progression-free survival with temozolomide. It was not clear whether this was true of Grade IV tumours alone.

The research question was well defined and inclusion criteria were clear with regard to study design, population and intervention, but there were no inclusion criteria for outcomes. Appropriate databases were searched for published and unpublished studies. Only English-language studies were sought, which may have caused relevant studies to be missed. Study selection, validity assessment and data extraction were performed by two reviewers, which reduced the possibility of reviewer error and bias. The studies were combined narratively, which seemed appropriate given the heterogeneity between studies. The quality of primary studies was variable. The authors reported that in the BCNU-W studies there was an imbalance in Grade III tumours (which may be more responsive to chemotherapy) between treatment arms at baseline. Some grade IV tumours may have been misclassified in the primary studies of temozolomide, as assessed and acknowledged by the authors and taken into account in the conclusions. The authors also stated that the results may not be generalisable as one BCNU-W RCT excluded patients over 65 years old and in one temozolomide RCT patients with a lower performance status and patients older than 70 years old were excluded. The authors' conclusions seem appropriate given the strength of the evidence presented.

Practice: The authors did not state any implications for practice

Research: The authors stated that further research of BCNU-W was needed in specific populations. For temozolomide, further research was needed to find out if patients with confirmed grade IV tumours benefited from temozolomide. Further research was needed to refine identification of subcategories of grade III and IV tumours and explore the feasibility of using these markers to inform treatment need. Trials should seek to compare different chemotherapy regimens directly rather than against placebo and to specify and evaluate sequences of treatment. Aspects of quality of life that were important to patients and changes in quality of life during progressive and stable disease should be explored in future trials and more explicit consideration of carer views sought. The value that patients put on small absolute survival advantages compared with disadvantages of treatment requirements should also be explored.

HTA programme on behalf of NICE (project no.04/20/01)

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.