Reanalysis of tenofovir study changes some findings

Last Updated: 2019-05-13

By Will Boggs MD

NEW YORK (Reuters Health) - A reanalysis of a study comparing outcomes with tenofovir and entecavir for chronic hepatitis B has yielded different results, with some findings no longer statistically significant, researchers from South Korea report.

The original study (https://bit.ly/2IG5Na8), published in September 2018 in JAMA Oncology and covered shortly afterward by Reuters Health, concluded that tenofovir treatment of chronic hepatitis B was associated with superior survival and lower rates of hepatocellular carcinoma (HCC), compared with entecavir treatment.

In response to several letters, Dr. Young-Suk Lim of the University of Ulsan College of Medicine, in Seoul, conducted an extensive and thorough review of their data, during which they discovered inadvertent errors resulting from the miscalculation of follow-up duration in their nationwide cohort.

Specifically, the number of reported mortality or transplant cases in the entecavir treatment group declined from 281 to 269, while the number in the tenofovir treatment group declined from 228 to 190.

As a result, the hazard ratio for mortality or transplant in the tenofovir treatment group increased from the significant 0.77 to a no longer statistically significant 0.89, relative to entecavir.

On the other hand, the hazard ratios for HCC in the tenofovir treatment group increased from 0.61 to 0.68 in the entire cohort and from 0.62 to 0.68 in the propensity-score matched cohort, with the corrected hazard ratios remaining statistically significant in favor of tenofovir.

There were no other errors in the article, Dr. Lim and coauthor Dr. Min Jung Ko reported in JAMA Oncology, online April 25.

Dr. Lim told Reuters Health by email, "Basically, there were no changes in the interpretation of our finding after reanalysis, particularly the risk of HCC between tenofovir and entecavir, which was the primary outcome of our study. However, the risk of death or liver transplantation in the nationwide cohort did not remain any more statistically significant after reanalysis. These revised results were more consistent with findings from the hospital cohort, which was used for validation."

"The interpretation of our finding should be cautious until further study mounts more evidence consistent with our findings," he said. "I think it is premature to choose one of these two drugs based on our report only. Still, other factors, such as comorbidities, renal disease, cost, and accessibility, should be considered together in the choice of these two drugs."

In a separate letter, Dr. Lim and colleagues addressed concerns raised in two other letters. One questioned why the researchers did not include family history of HCC as a significant risk factor in their analysis. Dr. Lim and colleagues cited potential measurement biases as their reason for including only objective covariates in their analysis of risk factors for HCC.

The second letter questioned, among other things, the researchers' characterization of entecavir as potentially carcinogenic. Dr. Lim and colleagues cited previous studies that raised concerns about the carcinogenic potential of entecavir "even at clinical doses during long-term treatment, especially in patients with cirrhosis, who have increased chromosomal instability in the hepatocytes."

SOURCE: https://bit.ly/2Q0dUBW, https://bit.ly/2LC6KVX, https://bit.ly/2Ji5nd6, https://bit.ly/2Jjo5Rm and https://bit.ly/2JgFPgj

JAMA Oncol 2019.