Medtronic PLC has acknowledged that an undisclosed number of patient deaths were left out of a recent analysis of its popular paclitaxel-coated balloon for blocked leg arteries, but the company said the omission didn't affect long-term safety rates in new analyses.

Companies put paclitaxel on medical devices to make them more effective. But devices that release the drug into the legs came under worldwide scrutiny in December after a study of studies in the Journal of the American Heart Association (JAHA) documented a 93 percent increased risk of death five years after patients were randomized to get devices with paclitaxel.

Death rates were the same at one year. But nearly 15 percent of people who received paclitaxel devices for peripheral artery disease (PAD) died after five years, vs. 8 percent mortality seen in control groups.

The analysis included 28 randomized controlled trials on 13 different balloons or stents coated with paclitaxel for PAD. (Coronary devices with the drug were not part of the analysis.)

Companies that make paclitaxel devices for PAD, including Medtronic, Boston Scientific, and Cook Medical have responded with detailed analyses from their proprietary patient-level data sets, which contain more information on each patient than the studies analyzed in the JAHA report. In each case, the companies said their internal data didn't support the findings from the larger study.

On Friday, Medtronic announced two recent studies involving its In.Pact Admiral drug-coated balloon are being revised following the discovery of a "programming error" that caused mortality data from its ongoing In.Pact Global Study to be inadvertently omitted from summary tables in statistical analyses published the Journal of the American College of Cardiology (JACC) and a related journal dealing with cardiovascular interventions called JACC: IC.

Although the revisions are still being peer-reviewed, Medtronic said the revised data continue to support findings in the recent studies that there was no statistically significant difference in death rates between In.Pact Admiral patients and plain balloon angioplasty for PAD at five years.

The In.Pact Admiral drug-coated balloon is a lucrative device thought to generate more than $200 million a year in sales for Medtronic.

Dr. Konstantinos Katsanos, the Greece-based lead author of the critical analysis in JAHA, said it's too early to say whether the additional deaths in the In. Pact Global data will support his paper's findings. (In.Pact Global was not one of the 28 trials studied in JAHA.) But he noted that revisions involving nonpublic data highlight the need for companies' patient-level data to be released and independently analyzed.

"Unless all patient data is openly shared for noncommercial parties to analyze objectively, the jury will remain out," Katsanos said via e-mail Monday.