Table. Investigations into the influence on disclosure of adverse drug reactions.
| Author | Purpose | Result |
| Aursnes et al. 2008 (e12) | Assessment of extent to which ADRs of the SSRI paroxetine were known to the pharmaceutical company before application for licensing but not evaluated, by analysis of trial data from 17 RCTs (1495 patients), published in 1989, that the pharmaceutical company submitted to Norwegian health authorities with licens‧ing applications | Meta-analyses show that 19 ADRs were statistically significantly more frequent during treatment with paroxetine; however, five of these ADRs (e.g., headache) are still not mentioned in the latest product information. |
| Halpern et al. 2004 (e13) | Determination of difference in average statistical power of pharmacoepidemio‧logical studies on ADRs of antiretroviral drugs funded by profit-oriented versus non-profit organizations, by analysis of all published studies on the 15 such drugs licensed up to 1999 | Only 3 of 41 studies in which funding was specified (7%) were supported by profit-oriented organizations; therefore no comparison possible. |
| Jüni et al. 2004 (e10) | Investigation whether there was evidence for cardiovascular ADRs of rofecoxib before drug was taken off market in 2004, by meta-analysis of clinical studies | At end of year 2000, RR for myocardial infarction was 2.30 (95% CI 1.22–4.33, p = 0.01); a year later, 2.24 (95% CI 1.24–4.02, p = 0.007). |
| Nieto et al. 2007 (e3) | Evaluation of differences with regard to findings on ADRs and their interpretation between 275 studies funded by a pharmaceutical manufacturer and 229 studies not funded by a pharmaceutical manufacturer; studies published between 1993 and 2002 | Statistically significant differences for ADRs were statistically significantly less frequent in studies funded by a pharmaceutical manufacturer than in studies not funded by a pharmaceutical manufacturer (34.5% vs. 65.1%; prevalence ratio 0.53, 95% CI 0.44–0.64). After adjustment for aspects of study ‧design associated with fewer ADRs, e.g., dosage, connection was no longer significant. Authors of studies funded by a pharmaceutical manufacturer were more likely to conclude that a drug is safe than authors of studies not funded by a pharmaceutical manufacturer (prevalence ratio 3.68, 95% CI 2.14–6.33). |
| Psaty et al. 2004 (e8) | Review of connection between cerivastatin intake and risk of rhabdomyolysis, by inspection of published data and pharmaceutical company’s internal documents provided in context of legal proceedings | Pharmaceutical manufacturer aware of data indicating interaction between cerivastatin and gemfibrozil by 100 days after market launch, but only after 18 months was this included in product information. Data from clinical trials with high-dose cerivastatin and from analysis of the FDA’s ADR notification system were not publicized. |
| Psaty et al. 2008 (e9) | Presentation of data on mortality in patients with Alzheimer’s dementia or cognitive disorders being treated with rofecoxib, by inspection of published data and pharmaceutical company’s internal documents provided in context of legal proceedings | Increased mortality during rofecoxib treatment was mentioned in two publications, but data were not subjected to statistical analysis. Safety stated as “well tolerated” although internal evaluation had already shown a distinct significant increase in mortality with rofecoxib (overall mortality in intention-to-treat analysis: RR 2.56; p = 0.001). This analysis was communicated neither to FDA nor to general public in timely fashion. FDA was given an analysis with shorter observation time and thus lower increase in mortality. Company answered FDA enquiries evasively. |
| Ross et al. 2009 (e11) | Investigation of whether and when analysis of published and unpublished studies would have revealed the cardiovascular risk associated with intake of rofecoxib, by person-related analysis of all the manufacturer’s trial data before September 2004 | By December 2000, 21 of a total of 30 studies had been concluded and risk of cardiovascular thromboembolic ADRs or death was higher in persons who took rofecoxib than in those who did not take rofecoxib (RR 2.18, 95% CI 0.93–5.81; p = 0.07). Statistically significantly (p = 0.05) increased risk discernible from June 2001. |
ADRs, adverse drug reactions; SSRIs, selective serotonin reuptake inhibitors; RCTs, randomized controlled trials; RR, relative risk; 95% CI, 95% confidence interval; FDA, US Food and Drug Administration