Proton Pump Inhibitor Usage Reduces Sustained Viral Response Rates for Veterans With HIV/HCV Coinfection on Ledipasvir/Sofosbuvir

A Real-world Study From a Multicentre VA Cohort

Tzu-Hao Lee; Austin Chan; William Bryan; Lawrence Park; Mohamed Hashem; Mary Townsend; Cynthia Moylan; Rachel Britt; Steve Choi; Susanna Naggie


J Viral Hepat. 2021;28(4):630-636. 

In This Article


In light of inconsistent reports of the impact of PPI usage on SVR with LDV/SOF therapy, and the common occurrence of patients with HCV taking PPI therapy, we leveraged the multicentre VA CDW data to further investigate this relevant clinical question. Our study revealed that using PPI was associated with a lower SVR in this large veteran cohort. The difference is smaller in the overall cohort (95.0% vs. 96.1%) with a number needed to harm 90.9, than in the group with HIV/HCV coinfection (93.4% vs. 96.9%), with a number needed to harm 28.6.

In prior studies of people with HCV infection, the impact of concomitant PPI use SVR has been reported, but the results are somewhat inconsistent.[9,10] Terrault et al[9] analysed the data from the HCV-TARGET study (n = 1788) and found that PPI use was associated with a lower SVR rate in an age- and sex-adjusted multiple logistic model (93.5% vs. 97.2%, OR 0.41; 95% CI 0.25–0.67). The association remained significant (OR 0.57; 95% CI 0.25–0.67) using inverse probability weighting analysis.[9] However, in the study by Tapper et al[10] (n = 1979), there was no difference in SVR between PPI use and nonuse in the propensity score matching analysis (97.8% vs. 97.2%, OR=0.34, 95% CI=0.05–2.29).

There are several differences between our study and other published cohorts. Our cohort is the largest published to date (n = 9703) with a primary objective of determining the impact of PPI usage on SVR for patients taking LDV/SOF. We leveraged the pharmacy prescription system in the VA to include only acid suppression medication prescriptions that were filled immediately prior to and during the DAA course of therapy, making it more likely the veteran was actively taking the acid suppression medication. Additionally, most of the PPI were prescribed with a 90-day supply to meet the VA cost-efficiency goal, so we believed the majority of the patients in PPI group had significant PPI exposure during their DAA therapy. We only included veterans with SVR12 results available, so that the only contribution to treatment failure was relapse or re-infection versus loss to follow-up or other non-virologic outcomes. We excluded any veteran who received a course of therapy that was not consistent with the AASLD/IDSA HCV treatment guidelines, limiting other potential reasons for treatment failure. Thus, the study cohorts were uniquely defined across the different studies.

Due to the large size of this cohort, we were able to study subgroups of patients that might have unique reasons for risk of treatment failure, including veterans with HIV, cirrhosis or those who have previously failed HCV therapy. Here, we report that using a PPI is associated with a lower SVR rate in the overall cohort and the subgroup of patients with HIV/HCV coinfection. The effect is larger in the cohort with HIV/HCV coinfection, with a number needed to harm of 28.6. According to the study by German et al,[12] patients with HIV/HCV coinfection who take LDV/SOF have a lower concentration of ledipasvir compared to patients with HCV mono-infection, with geometric mean ratio 70.8%–71.9%. The true reason behind this finding is still unclear, but this may be related to elevated gastric PH, which has been reported in people with HIV.[13–15] Using PPI, which affects the absorption of the LDV, could further decrease the concentration of LDV in the blood, resulting in a lower SVR rate in patients with HIV/HCV coinfection. While Tapper et al. did include patients with HIV in their study, very few patients with HIV (n = 23) were taking PPI.[10] Furthermore, Bhattacharya et al. did not find an association of PPI use and SVR in a large veteran cohort,[11] which would have overlap with the cohort reported here. As noted above, there are significant differences in the study cohorts including our inclusion of 8, 12 and 24-week courses of therapy, the inclusion of only veterans with SVR12 viral load available, and our usage of the VA pharmacy database to specifically identify veterans in our cohort actively receiving acid suppression medications.[11]

The limitation of this study includes residual confounding between the acid suppression and the control groups because this is a retrospective study, and PPI use was not a randomized effect. We have adjusted for confounders using statistical methods, which should minimize the bias, including inverse probability weighting analysis. We did not include HIV medication in our analysis; however, there is no compelling reason that the use of HIV medication would be different in the group of PPI and non-PPI. Few HIV medications are contraindicated with PPI (atazanavir and rilpivirine), and there is no significant interaction with these medications and ledipasvir. During the study period, 8 weeks of treatment of LDV/SOF was not recommended for black patients or patients with cirrhosis, HIV or failed prior HCV treatment. In our study, we excluded patients with cirrhosis or who failed prior HCV treatment that received 8 weeks of LDV/SOF but included black patients and patients with HIV that received 8 weeks of LDV/SOF. We completed a sensitivity analysis to exclude black patients and patients with HIV that received 8 weeks of LDV/SOF with no change in the result or conclusions (Table S1–S4).

Given our study was based on a VA cohort, 96% of our patients are male with a median age of 63 years. This could affect the generalizability of our results to other patient populations. In fact, female sex is associated with a higher chance of SVR12 in our study (Table 3). The other limitation of our study is the accuracy of the record of acid suppression medication use. Some of the patients could have taken over-the-counter acid suppression medication, which is not recorded in our study, and some of the patients might not actually take the acid suppression medications despite having an active prescription. Lastly, while we tried to limit this cohort to include only veterans with SVR12 data available, and thus include only virologic failures, we cannot differentiate relapse from re-infection. We assumed any recurrent positive HCV RNA was due to viral relapse.

In this large real-world cohort of veterans receiving LDV/SOF, PPI use is associated with lower SVR with a more significant effect in the patients with HIV-HCV coinfection and a smaller number needed to harm (28.6 vs. 90.9). According to our study, acid suppression medications should be used cautiously in patients with HIV/HCV coinfection receiving LDV/SOF. Ideally, the PPI would be held or discontinued when LDV/SOF is the DAA chosen for HCV therapy in this patient population. There are other approved and recommended DAA regimens for which PPI concomitant dosing has not been shown to impact SVR, and these should also be considered when the PPI cannot be held or discontinued.