Ferretti et al. (2006)

Ferretti et al. (2006)
Design: Systematic review of RCTs (therapy), evidence level: 1−
Country: Italy
Inclusion criteria:
Studies: Phase III RCTs comparing AIs with tamoxifen as 1st line therapy for MBC and published as original papers in peer-reviewed journals (there was one exception to this - an abstract of a large RCT presented at the 2004 ASCO conference).

Patients: Postmenopausal women with MBC relapsing on adjuvant therapy.
Exclusion criteria:
Phase II, non-randomised studies, articles or letters
Number of patients = 2787
Analysis by the following sub-groups. All AIs were compared with tamoxifen as first line therapy for MBC:

Group AI: included non-steroidal AI (nsAI): fadrozole, letrozole, anastrozole and vorozole and steroidal AI: formestane and exemestane. Excluded the abstract of Paridaens et al (see tgAI below)

Group nsAI: included only fadrozole, letrozole, anastrozole and vorozole

Group tgAI: all third generation AIs (including an abstract excluded from other analyses which made a comparison between exemestane and TAM n = 371)
Primary outcomes: overall response rate (ORR - complete response + partial response), time to progression (TTP)

Secondary outcomes: overall survival (OS), clinical benefit (CB - defined as either CR, PR or stable disease for > 6 months).

Follow up:
Median follow-up across all trials ranged from 5.1–36 months.
7 trials were included in the meta-analysis: 1,615 patients were randomised to receive AIs and 1,623 patients received tamoxifen (TAM).

TTP in the AI patients ranged from 7.1–18 months
TTP in the TAM patients ranged from 5.6–9.8 months

OS in the AI patients ranged from 17.4–39.2 months
OS in the TAM patients ranged from 16–40 months

CB in the AI patients ranged from 50–83%
CB in the TAM patients ranged from 38–75.7%

Meta analysis by sub group:
Group AI - fixed effects model:
ORR (n = 2787): RR = 1.13 (95%CI: 1.0–1.28) P = 0.042 (AI)
TTP (n = 2549): RR = 0.88 (95%CI: 0.8–0.96) P = 0.007 (AI)
CB (n = 2787): RR = 1.11 (95%CI: 1.04–1.19) P = 0.001 (AI)
OS (n = 2787): RR = 0.97 (95%CI: 0.79–1.18) nsd

Group AI - random effects model:
ORR (n = 2787): RR = 1.11 (95%CI: 0.89–1.37) nsd
TTP (n = 2549): RR = 0.92 (95%CI: 0.68–1.26) nsd
CB (n = 2787): RR = 1.13 (95%CI: 0.96–1.33) nsd

Group AI - between studies heterogeneity:
ORR: P = 0.03
TTP: P < 0.0001
CB: P < 0.0001
OS: nsd
The significant effect of (steroidal and non-steroidal) AIs on ORR, TTP and CB seen in the fixed effects model are not confirmed in the random effects model but OS is nsd both models.

Group nsAI - fixed effects model:
ORR (n = 2166): RR = 1.23 (95%CI: 1.07–1.42) P = 0.003 (AI)
TTP (n = 1928): RR = 0.77 (95%CI: 0.69–0.86) P < 0.0001 (AI)
CB (n = 2166): RR = 1.21 (95%CI: 1.12–1.31) P < 0.0001 (AI)
OS (n = 2166): RR = 0.94 (95%CI: 0.75–1.78) nsd

Group nsAI - random effects model:
TTP (n = 1928): RR = 1.25 (95%CI: 1.12–1.31) nsd
CB (n = 2166): RR = 1.13 (95%CI: 0.96–1.33) P = 0.018

Group nsAI - between studies heterogeneity:
ORR: nsd
TTP: P = 0.002
CB: P = 0.005
OS: nsd
Only the significant positive effect of non-steroidal AIs on TTP was confirmed between fixed and random effects models.

Group tgAI - fixed effects model:
ORR (n = 2537): RR = 1.28 (95%CI: 1.13–1.44) P < 0.0001 (AI)
TTP (n = 2299): RR = 0.76 (95%CI: 0.69–0.84) P < 0.0001 (AI)
CB (n = 2537): RR = 1.23 (95%CI: 1.14–1.32) P < 0.0001 (AI)
OS (n = 2537): RR = 0.93 (95%CI: 0.76–1.15) nsd

Group tgAI - random effects model:
TTP (n = 2299): RR = 0.74 (95%CI: 0.58–0.94) P = 0.015
CB (n = 2537): RR = 1.26 (95%CI: 1.09–1.46) P = 0.0002

Group tgAI - between studies heterogeneity:
ORR: nsd
TTP: P = 0.004
CB: P = 0.008
OS: nsd
The significant positive effect of third generation AIs on TTP and CB was confirmed between fixed and random effects models. Significance was maintained despite between-study heterogeneity. OS was still not significantly different between AI and TAM patients.

The incidences of thromboembolic events and vaginal bleeding were significantly less with AI therapy than TAM in all three sub-groups analyses. All studies were non-significant for between- studies heterogeneity for these parameters.
General comments:
This paper described a meta-analysis of seven RCTs (including one abstract) which compared AIs with tamoxifen as a first line therapy for MBC. Both fixed and random effects models were applied to the data which was analysed in sub-groups according to the AI type.

The between-study heterogeneity rendered some of the efficacy results non-significant even though the relative risk results appeared to suggest otherwise. The Q and I squared figures are not given so it is not possible to state the degree of heterogeneity other than as a P value.

Median survival and progression values were not extracted from the included studies but OS and TTP were estimated (probably from Kaplan-Meier curves) at 6 months and these values were used in the meta-analyses. This must be kept in mind when comparing the results of this review with those of other similar reviews that have reported survival outcomes using median values. For example, in the review by Mauri et al. (2006) the median OS values from 25 RCTs range from 15.7–40.1 months.

Across all sub-group analyses, OS was not significantly different between AIs and TAM. Only third generation AIs showed a significant advantage in terms of both TTP and CB.

The main weakness of this review is that there is no formal assessment by the authors of the included studies. This means that it is not possible to know whether the studies are good enough to answer to question. Ordinarily such a negative point would downgrade a systematic review but, in this case, four of the studies have been reviewed elsewhere and were found to be of good quality. It therefore seems likely that the authors were rigorous in their selection but did not describe their methodology.

From: Chapter 4, Systemic disease-modifying therapy

Cover of Advanced Breast Cancer
Advanced Breast Cancer: Diagnosis and Treatment.
NICE Clinical Guidelines, No. 81.
National Collaborating Centre for Cancer (UK).
Copyright © 2009, National Collaborating Centre for Cancer.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licenses issued by the Copyright Licensing Agency in the UK. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.