Case series and case reports are a cornerstone of drug safety research; however, the characteristics of case series published in the literature remain poorly examined. A narrative review of case series addressing drug safety, published in the literature between 1 January 2003 and 15 July 2009, and identified through a PubMed search, was conducted in order to determine their characteristics and quality according to the criteria found in the US FDA Pharmacovigilance Guidance 2005. Of 130 publications that met the search criteria, 11.5% included an analytical component and 88.5% were descriptive. The median number of cases included in a given case series was 7 (range 2-2195) and the median time period for recruitment of the cases was 23 months (range 0.5-96). Overall, 43.1% of case series consisted of individual case reports, while 24.6% originated from cohorts and 21% from pharmacovigilance databases. Of the case series, 65.1% concerned adults (age ≥18 years), 11.6% elderly (age ≥65 years) and 8.5% youth (<18 years). Adverse effects involved mainly the skin (18.5%) and the circulatory system (13.8%). The main suspected drug classes (Anatomical Therapeutic Chemical classification) were nervous system drugs (23.1%) and antineoplastic and immunomodulating agents (20.0%). On average, six out of the possible nine US FDA Pharmacovigilance Guidance Criteria were fulfilled, with 27% of publications fulfilling at least seven criteria. Only 10% reported data on co-morbidity. In conclusion, this review highlights the reporting gaps and heterogeneity in published case series with respect to size, recruitment period and quality.