On-site audits of the two clinical trials reported from South Africa involving high-dose chemotherapy (HDC) therapy for breast cancer

An investigator at the University of the Witwatersrand in Johannesburg has reported (J Clin Oncol 1995; 13:2483-2489, and Proc Am Soc Clin Oncol 1999, 18:2a) two clinical trials purporting to be randomized prospective evaluations of HDC in comparison with treatment administered in conventional doses. One involved metastatic breast cancer, and the other high-risk primary disease. In both trials two cycles of a regimen devised at this institution combining cyclophosphamide, mitoxantrone, and etopside were used for the HDC. A total of 90 patients were reported in the metastatic disease study, and 154 (or 151) were said to have been treated in the high-risk study. Two separate on-site audits of available patient records have been performed.

Results

Of the 154 (or 151) patients allegedly entered into the high-risk study, medical records for only 58 patients (all appearing to have received HDC) were made available for review. Of the 90 patients allegedly entered on the metastatic study (based on information provided by the investigator), records for only 61 could be found. Only 25 of these appeared to have received protocol treatment (22 receiving HDC). The remainder could not be verified to have received the purported study therapy. Many of the patients reviewed for both studies did not meet the stated eligibility criteria, and there was no evidence of any acceptable randomization process. The reported results of these two studies cannot be used as a basis for evaluating HDC for either metastatic or high-risk primary breast cancer.