The long-term results for patients with recurrent ovarian cancer (ROC) are poor. There is a need to optimize treatment strategies to improve outcome by avoiding ineffective regimens which are often associated with exacerbated side-effects. Individualized chemotherapy regimens guided by a chemosensitivity assay (ATP-tumor chemosensitivity assay) have already been used successfully to direct chemotherapy. Taking the results of this assay into account, application of drug combinations appears more advisable. Here we present a systematic evaluation of toxicities seen with individualized chemotherapy for ROC. A total of 62 patients who received 314 cycles of antineoplastic therapies were evaluated. Three single agents (topotecan, paclitaxel and gemcitabine) and five combinations (cisplatin/gemcitabine, carbopatin/gemcitabine, gemcitabine/treosulfan, mitoxantrone/paclitaxel and carboplatin/paclitaxel) were examined. With respect to myelotoxicity, most single agents except topotecan revealed favorable results in comparison to drug combinations. However, this observation lacks statistical significance. Generally, severe myelosuppression was rare. The highest incidence of leukopenia was seen in regimens with mitoxantrone/paclitaxel or gemcitabine/treosulfan, respectively. Thrombocytopenia accompanied most commonly a topotecan therapy. In the present study combination regimens tend to be more toxic than monotherapies. When response rates are comparable, empirically chosen treatment combination therapies should only be practiced in carefully planned clinical studies.