|
Meta-analysis of randomized trials testing the biochemical modulation of fluorouracil by methotrexate in metastatic colorectal cancer |
Piedbois P, Buyse M, Blijham G, Glimelius B, Herrmann R B, Valone F, Carlson R, Machiavelli M, Delfino C, Abad A, Petrelli N |
|
|
Authors' objectives To assess the benefit of 5-fluorouracil (5FU) with methotrexate (MTX) over 5FU alone for tumour response rate and overall survival in metastatic colorectal cancer.
Searching MEDLINE and the proceedings of major congresses were searched over the 5 years preceding 1991. Other trials were identified in the Proceedings of the American Society of Clinical Oncology and through personal contact with the investigators.
Study selection Study designs of evaluations included in the reviewAll included trials were randomised controlled trials (RCTs) which compared 5FU alone with 5FU combinations.
Specific interventions included in the reviewThe treatments studied were of 5FU alone, or combined with methotrexate (5FU-MTX) or leucovirin (5FU-LV), and different combinations of these interventions. The 5FU alone was estimated to have a cumulative dosage ranging from 6,900 to 17,760 mg/m2 given cyclically as intravenous boluses.
In the 5FU-MTX combination arm, MTX was given orally or intravenously in a dosage ranging from 40 to 500 mg/m2, prior to 5FU in an estimated cumulative dosage ranging from 3,000-17,760 mg/m2 given cyclically and intravenously. In some trials, patients in the 5FU-MTX arm also received LV.
Participants included in the reviewThe participants in the trials were patients with colorectal carcinoma, with the following characteristics: median age was 62 years; 59% were male; performance status was greater than 2 on the Eastern Cooperative Oncology Group (ECOG) scale in 83% of all patients; primary tumour site was colon in 58% of patients; metastasis were limited to the liver in 34% and to the lung in 7% of patients.
Outcomes assessed in the reviewThe main outcomes assessed were: complete response, defined as the disappearance of all detectable tumour; partial response, defined as a 50% reduction in the sum of the product of the largest perpendicular diameters of all measurable disease without new lesions; and duration of survival, calculated from the date of randomisation to the date of death, whatever the cause of death.
How were decisions on the relevance of primary studies made?The authors do not state how the papers were selected for the review, or how many of the authors performed the selection.
Assessment of study quality The quality of the studies was not formally assessed.
Data extraction The protocol for the meta-analysis was sent to all principal investigators asking for the following data on all randomised patients: patient identification, date of randomisation, eligibility, treatment assigned by randomisation, age at randomisation, sex, performance status according to the ECOG scale, primary tumour site (colon or rectum), prior treatment (chemotherapy or radiotherapy in metastatic areas), localisation of metastasis, overall response status with the first allocated treatment, crossover to another treatment arm, second treatment arm in case of crossover, date of death or last visit, survival status, and cause of death if applicable.
All received data were checked and discussed with collaborators at a meeting.
Methods of synthesis How were the studies combined?The tumour response rates were analysed using a Mantel-Haenszel test, and survival data were stratified by trial using a log rank test. Treatment effects and survival were calculated as odds ratios (ORs) with 95% confidence intervals (CIs). A logistic regression was used to predict response using the following factors as covariates: randomised treatment, age, sex, prior chemotherapy, site, prior radiotherapy, and location of metastasis. A Cox regression model, stratified for trial, was fitted to the survival time with the same covariates as the logistic regression.
How were differences between studies investigated?Heterogeneity between trials was assessed using a chi-squared test. A priori grouping of trials was performed according to the following factors: the dose of 5FU, the dose of MTX in the 5FU-MTX arm, interval between MTX and 5FU, and the dose of LV in the F5U-MTX arm. The analysis was repeated using data from these groupings.
Results of the review Results from 8 RCTs involving 1,178 patients were analysed.
Overall tumour response: the OR of having an event in the 5FU-MTX group, as compared to the 5FU group, was 0.51 (95% CI: 0.37, 0.70, P<0.0001); chi-squared test for heterogeneity, 13.24 (2P=0.07).
The logistic regression used to predict tumour response indicated that randomised treatment (P<0.0001) and performance status (P=0.01) were significant predictors of response. After adjustment for performance status the OR was 0.49.
Overall survival: median survival time was 9.1 months (95% CI: 8.5, 9.8) in the 5FU alone group, compared to 10.7 months (95% CI: 9.7, 11.7) in the 5FU-MTX group. The OR in favour of 5FU-MTX was 0.87 (95% CI: 0.77, 0.98, P=0.024); chi-squared test for heterogeneity, 5.76 (2P=0.57).
The Cox regression mode, stratified for trial, showed that performance status (P<1E-10) and randomised treatment (P=0.01) were significant predictors of survival.
The analysis according to the grouping of trials, whether by dosage or by interval, showed no statistically-significant differences between groups.
Authors' conclusions The modulation of 5FU by MTX doubled the response rate to 5FU and yielded a small but statistically-significant improvement in survival.
CRD commentary This was a well-written and clearly-presented review with extensive analysis of the retrieved data.
The literature search was limited and it was unclear from the text which years of the MEDLINE database were actually searched. There was no formal assessment of the quality of the retrieved studies, and there was minimal data on the incidence of side-effects experienced by patients. Any comparison of drug therapy is incomplete without consideration of the adverse reactions and quality of life of the patients participating, and unfortunately, these factors were omitted from the review. More information on adverse reactions may have been available from the primary studies had it been sought. Significant heterogeneity was evident but possible sources of heterogeneity between trials were not discussed. The meta-analysis did not include data from one RCT, even though data from this trial were mentioned in the discussion.
Implications of the review for practice and research No comparison can be made of 5FU against 5FU-MTX without information on the quality of life and side-effects experienced by patients receiving the therapies reviewed.
Bibliographic details Piedbois P, Buyse M, Blijham G, Glimelius B, Herrmann R B, Valone F, Carlson R, Machiavelli M, Delfino C, Abad A, Petrelli N. Meta-analysis of randomized trials testing the biochemical modulation of fluorouracil by methotrexate in metastatic colorectal cancer. Journal of Clinical Oncology 1994; 12(5): 960-969 Indexing Status Subject indexing assigned by NLM MeSH Aged; Antineoplastic Combined Chemotherapy Protocols /therapeutic use; Colorectal Neoplasms /drug therapy /pathology; Female; Fluorouracil /administration & Humans; Liver Neoplasms /drug therapy /secondary; Lung Neoplasms /drug therapy /secondary; Male; Methotrexate /administration & Middle Aged; Randomized Controlled Trials as Topic; Regression Analysis; Survival Analysis; dosage; dosage /therapeutic use AccessionNumber 11994000114 Date bibliographic record published 31/01/1998 Date abstract record published 31/01/1998 Record Status 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. |
|
|
|