| Sequence Generation (selection bias) |
Unclear |
The investigators do not describe the sequence generation process. It is unclear from the available information whether sequence generation was random. |
| Allocation Concealment (selection bias) |
Unclear |
Participants and investigators enrolling participants could not foresee assignment because a centralized service at Merrell-National Laboratories was used to allocate the study groups. Typical baseline characteristics were not reported. |
| Blinding of participants and researchers (performance bias) |
Low |
The medications and bottles were identical in appearance. Each medication was packaged in bottles of 30 tablets and each bottle was labeled with a tear-off label. The sealed-off portion contained the identity of the contents. |
| Blinding of outcome assessment (detection bias) |
Low |
The outcome assessors (physicians from multiple centers and participants) were blinded to the intervention. It is unclear if data analysts were blinded. |
| Incomplete outcome data (attrition bias) |
High |
Apparently results were analyzed for 1599 (68%) of 2359. 51 of the 2359 participants were initially excluded from the study due to “incomplete data”. Of the remaining participants, 6% did not complete the study and were classified as lost to follow up and 30% failed to meet protocol criteria and were excluded from the study. In addition, information about adverse outcomes are missing for a small fraction of analysed participants. |
| Selective outcome reporting (reporting bias) |
High |
No outcomes were prespecified. The trial was done before outcomes were registered. Multiple outcomes are reported without identifying any as primary. |
| Other potential threats to validity |
High |
Important information about the study is not available. The FDA ordered that data from one investigator be excluded because of concerns about data integrity. The trial was apparently not completed. The results were never published. |
