Serological response to three alternative series of hepatitis B revaccination (Fendrix, Twinrix, and HBVaxPro-40) in healthy non-responders: a multicentre, open-label, randomised, controlled, superiority trial

Abstract

Background: Serological non-response can be present after hepatitis B vaccination in healthy adults. We aimed to establish which of three revaccination regimens is most effective at inducing protective immunity METHODS: Healthy adults (aged 18-80 years) from 16 Dutch centres (13 public health services, two university hospitals, and one travel clinic) were included in this multicentre, parallel group, randomised, controlled, superiority trial. The inclusion criterion was vaccine non-response (hepatitis B surface antibody [anti-HBs] titre <10 IU/L) after a primary series with three doses of one type of recombinant vaccine against hepatitis B virus (either HBVaxPro-10 or Engerix-B at months 0, 1, and 6). Participants were individually randomly assigned (1:1:1:1) to a vaccination series of repeated initial vaccination (HBVaxPro 10 μg or Engerix-B 20 μg) as the control, or to Twinrix 20 μg, Fendrix 20 μg, or HBVaxPro 40 μg. We used a web-based randomisation programme, stratified by centre, with a block size of four. Participants and centres were unmasked to assignment after randomisation. Laboratory staff and investigators were masked to vaccine-group assignment. All revaccination schedules were identical, with intramuscular vaccinations at 0, 1, and 2 months. Anti-HBs was measured at 0, 1, 2, and 3 months. The primary outcome was the percentage of responders (anti-HBs titres ≥10 IU/L) at 3 months. Immunogenicity and safety analyses were based on an intention-to-vaccinate analysis, the immunogenicity analysis with last observation carried forward for missing data, and the Bonferroni and the Benjamini-Hochberg method were applied to correct for multiple testing. The trial was registered in the Dutch National Trial Register and inclusion has been stopped (identifier NL3011; EudraCT-number 2011-005627-40).

Findings: The participants were recruited between Nov 1, 2012, and Sept 1, 2017. 480 participants were randomly assigned and included in intention-to-vaccinate analyses: 124 (26%) to control, 118 (25%) to Twinrix, 114 (24%) to HBVaxPro-40, and 124 (26%) to Fendrix. At month 3 the percentage of responders was 83 (67%) of 124 (95% CI 57·9-75·1 in the control group, 94 (80%) of the 118 (71·3-86·5) in the Twinrix group, 95 (83%) of 114 (75·2-89·7) in the HBVaxPro-40 group, and 108 (87%) of 124 (79·9-92·4) in the Fendrix group. Compared with the control group, the percentage of responders was superior for the HBVaxPro-40 group (adjusted difference 21·6% [95% CI 10·4-32·7], p=0·0204 [Bonferroni corrected p value]) and the Fendrix group (26·3% [15·4-37·3], p=0·0006), but not the Twinrix group (25·0% [13·0-37·0]; p=0·0846). One serious adverse event occurred (herpes zoster ophthalmicus) in the Fendrix group, which was not attributed to the vaccine.

Interpretation: Revaccinating healthy non-responders with Fendrix or HBVaxPro-40 resulted in significantly higher proportions of responders and therefore indication for these vaccines should be expanded to enable revaccination of non-responders.

Funding: National Institute for Public Health and the Environment.