Treatment of newly diagnosed glioblastoma in the elderly: a network meta-analysis

BACKGROUND

A glioblastoma is a fatal type of brain tumour for which the standard of care is maximum surgical resection followed by chemoradiotherapy, when possible. Age is an important consideration in this disease, as older age is associated with shorter survival and a higher risk of treatment-related toxicity.

OBJECTIVES

To determine the most effective and best-tolerated approaches for the treatment of elderly people with newly diagnosed glioblastoma. To summarise current evidence for the incremental resource use, utilities, costs and cost-effectiveness associated with these approaches.

SEARCH METHODS

We searched electronic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE and Embase to 3 April 2019, and the NHS Economic Evaluation Database (EED) up to database closure. We handsearched clinical trial registries and selected neuro-oncology society conference proceedings from the past five years.

SELECTION CRITERIA

Randomised trials (RCTs) of treatments for glioblastoma in elderly people. We defined 'elderly' as 70+ years but included studies defining 'elderly' as over 65+ years if so reported.

DATA COLLECTION AND ANALYSIS

We used standard Cochrane methods for study selection and data extraction. Where sufficient data were available, treatment options were compared in a network meta-analysis (NMA) using Stata software (version 15.1). For outcomes with insufficient data for NMA, pairwise meta-analysis were conducted in RevMan. The GRADE approach was used to grade the evidence.

MAIN RESULTS

We included 12 RCTs involving approximately 1818 participants. Six were conducted exclusively among elderly people (either defined as 65 years or older or 70 years or older) with newly diagnosed glioblastoma, the other six reported data for an elderly subgroup among a broader age range of participants. Most participants were capable of self-care. Study quality was commonly undermined by lack of outcome assessor blinding and attrition. NMA was only possible for overall survival; other analyses were pair-wise meta-analyses or narrative syntheses. Seven trials contributed to the NMA for overall survival, with interventions including supportive care only (one trial arm); hypofractionated radiotherapy (RT40; four trial arms); standard radiotherapy (RT60; five trial arms); temozolomide (TMZ; three trial arms); chemoradiotherapy (CRT; three trial arms); bevacizumab with chemoradiotherapy (BEV_CRT; one trial arm); and bevacizumab with radiotherapy (BEV_RT). Compared with supportive care only, NMA evidence suggested that all treatments apart from BEV_RT prolonged survival to some extent. Overall survival High-certainty evidence shows that CRT prolongs overall survival (OS) compared with RT40 (hazard ratio (HR) 0.67, 95% confidence interval (CI) 0.56 to 0.80) and low-certainty evidence suggests that CRT may prolong overall survival compared with TMZ (TMZ versus CRT: HR 1.42, 95% CI 1.01 to 1.98). Low-certainty evidence also suggests that adding BEV to CRT may make little or no difference (BEV_CRT versus CRT: HR 0.83, 95% CrI 0.48 to 1.44). We could not compare the survival effects of CRT with different radiotherapy fractionation schedules (60 Gy/30 fractions and 40 Gy/15 fractions) due to a lack of data. When treatments were ranked according to their effects on OS, CRT ranked higher than TMZ, RT and supportive care only, with the latter ranked last. BEV plus RT was the only treatment for which there was no clear benefit in OS over supportive care only.   One trial comparing tumour treating fields (TTF) plus adjuvant chemotherapy (TTF_AC) with adjuvant chemotherapy alone could not be included in the NMA as participants were randomised after receiving concomitant chemoradiotherapy, not before. Findings from the trial suggest that the intervention probably improves overall survival in this selected patient population. We were unable to perform NMA for other outcomes due to insufficient data. Pairwise analyses were conducted for the following. Quality of life Moderate-certainty narrative evidence suggests that overall, there may be little difference in QoL between TMZ and RT, except for discomfort from communication deficits, which are probably more common with RT (1 study, 306 participants, P = 0.002). Data on QoL for other comparisons were sparse, partly due to high dropout rates, and the certainty of the evidence tended to be low or very low. Progression-free survival High-certainty evidence shows that CRT increases time to disease progression compared with RT40 (HR 0.50, 95% CI 0.41 to 0.61); moderate-certainty evidence suggests that RT60 probably increases time to disease progression compared with supportive care only (HR 0.28, 95% CI 0.17 to 0.46), and that BEV_RT probably increases time to disease progression compared with RT40 alone (HR 0.46, 95% CI 0.27 to 0.78). Evidence for other treatment comparisons was of low- or very low-certainty. Severe adverse events Moderate-certainty evidence suggests that TMZ probably increases the risk of grade 3+ thromboembolic events compared with RT60 (risk ratio (RR) 2.74, 95% CI 1.26 to 5.94; participants = 373; studies = 1) and also the risk of grade 3+ neutropenia, lymphopenia, and thrombocytopenia. Moderate-certainty evidence also suggests that CRT probably increases the risk of grade 3+ neutropenia, leucopenia and thrombocytopenia compared with hypofractionated RT alone. Adding BEV to CRT probably increases the risk of thromboembolism (RR 16.63, 95% CI 1.00 to 275.42; moderate-certainty evidence). Economic evidence There is a paucity of economic evidence regarding the management of newly diagnosed glioblastoma in the elderly. Only one economic evaluation on two short course radiotherapy regimen (25 Gy versus 40 Gy) was identified and its findings were considered unreliable.

AUTHORS' CONCLUSIONS

For elderly people with glioblastoma who are self-caring, evidence suggests that CRT prolongs survival compared with RT and may prolong overall survival compared with TMZ alone. For those undergoing RT or TMZ therapy, there is probably little difference in QoL overall. Systemic anti-cancer treatments TMZ and BEV carry a higher risk of severe haematological and thromboembolic events and CRT is probably associated with a higher risk of these events. Current evidence provides little justification for using BEV in elderly patients outside a clinical trial setting. Whilst the novel TTF device appears promising, evidence on QoL and tolerability is needed in an elderly population. QoL and economic assessments of CRT versus TMZ and RT are needed. More high-quality economic evaluations are needed, in which a broader scope of costs (both direct and indirect) and outcomes should be included.