Detecting drug-herbal interaction using a spontaneous reporting system database: an example with benzylpenicillin and qingkailing injection.

PURPOSE: The study aims to quantify anaphylaxis signal for combined exposure of benzylpenicilin and qingkailing injection (QI) compared with individual exposure of the two drugs and the background risk based on all other exposures in SRS database. METHODS: Data used in this study were collected during 2003-2014 from China Guangdong Provincial Center of ADR Monitoring. We studied the suspected ADR reports using a case/non-case design. The cases were defined as the reactions coded by WHO-preferred terms of anaphylactic shock or anaphylactoid reaction. Reporting odds ratios (RORs) were used as a measure of disproportionality and were adjusted for age and gender to reduce confounding effects. An observed-to-expected ratio Ω was also used for interaction detection. RESULTS: The crude RORs (95 % CIs) for anaphylaxis in patients who used only benzylpenicillin or QI and those who used the two drugs concomitantly compared with patients who used neither of the two drugs were 2.50 (2.34-2.68), 1.59 (1.46-1.73), and 6.22 (3.34-11.58), respectively. The adjusted RORs (95 % CIs) were 2.48 (2.31-2.65), 1.54 (1.41-1.67), and 6.01 (3.22-11.20), respectively, after being adjusted for age and gender. The measured Ω, Ω0, Ω025, and Ω975 was 1.03, 1.09, 0.14, and 1.71, respectively. CONCLUSIONS: Case reports in the database are suggestive of a safety signal which indicates that an interaction between benzylpenicillin and QI resulting in excess risk of anaphylaxis may occur. SRS databases have a potential for signaling unknown drug-herbal interactions. More effort is needed to expand this potential.