Aim: To comprehensively evaluate the toxicity spectrum of taxanes from the perspective of clinical trials and the real world. Method: Pooled analyses were performed to estimate incidences of adverse events (AEs) with random-effect models after searching databases. Reports of AEs were obtained from the US Food and Drug Administration’s Adverse Event Reporting System (FAERS) database and positive signals were quantified by conducting disproportionality analysis. Results: A total of 7 formulations were analyzed in this study with 36 clinical trials involving 10828 patients and 58835 case reports from FAERS. Leukopenia (59.69%, 95% confidence interval 41.34-75.69) and neutropenia (29.69%, 23.31-36.99) ranked first among all grades and severe AEs, respectively. Alopecia had the highest estimated incidences of non-hematological AEs regardless of grades. Paclitaxel has identified 561 positive signals while its AEs were not the most severe. Docetaxel had the least signals but alopecia and depression had quantified several signals. The estimated incidences of nab-paclitaxel were higher than other formulations, especially neutropenia (46.53%, 35.01-58.42). Conclusion: The safety of nab-paclitaxel was beyond expectation and unusual signals of alopecia and depression of docetaxel need to be paid attention to. Most common AEs in clinical trials also had positive signals in FAERS, indicating consistency between premarket and postmarket studies.