Table 5.
Statistical Analysis of Immunogenic Response as a Correlate of Protection Against HZ, Using Cox Model in Auto-HSCT Recipients (VZV gpELISA Response), Patients With STMc (VZV gpELISA Response and VZV IFN-γ ELISPOT Assay Response), and Patients With HM (VZV IFN-γ ELISPOT Assay Response) in the Per-Protocol Population
| gpELISA Units/mLa | Auto-HSCTb | STMc | HM | |||
|---|---|---|---|---|---|---|
| Point Estimate HR (95% CI) | P Value | Point Estimate HR (95% CI) | P Value | Point Estimate HR (95% CI) | P Value | |
| Vaccine effect on HZ without adjustment for VZV gpELISAc | 0.360 (0.25–0.51) | <.001 | 0.364 (0.22–0.59) | <.0001 | — | — |
| Vaccine effect on HZ with adjustment for VZV gpELISAd | 0.364 (0.24–0.56) | <.001 | 0.378 (0.23–0.63) | .0002 | — | — |
| Effect of VZV gpELISA (log-scale) on the risk of HZd | 1.037 (0.90–1.20) | .613 | 0.957 (0.78–1.18) | .6796 | — | — |
| IFN-γ ELISPOT Assay Count/106 PBMCse | Auto-HSCT | STMc | HM | |||
| Point Estimate HR (95% CI) | P Value | Point Estimate HR (95% CI) | P Value | Point Estimate HR (95% CI) | P Value | |
| Vaccine effect on HZ without adjustment for VZV IFN-γ ELISPOT assayf | — | — | 0.364 (0.22–0.59) | <.0001 | 0.833 (0.59–1.18) | .3035 |
| Vaccine effect on HZ with adjustment for VZV IFN-γ ELISPOT assayg | — | — | 0.376 (0.10–1.47) | .1603 | 1.337 (0.38–4.76) | .6536 |
| Effect of VZV IFN-γ ELISPOT assay (log-scale) on the risk of HZg | — | — | 1.011 (0.63–1.61) | .9622 | 0.713 (0.54–0.94) | .0171 |
Abbreviations: auto-HSCT, autologous hematopoietic stem cell transplant; gpELISA, glycoprotein enzyme-linked immunosorbent assay; HM, hematologic malignancies; HZ, herpes zoster; IFN-γ ELISPOT, interferon-γ enzyme-linked immunospot; PBMCs, peripheral blood mononuclear cells; STMc, solid tumor malignancies receiving chemotherapy.
aResults for the gpELISA are reported as concentration of antibody in gpELISA units/mL.
bFor vaccine effect on HZ incidence, the treatment-by-immunogenicity response interaction was statistically significant (P = .049); this P value for the interaction was calculated based on the likelihood ratio test.
cComputed based on a Cox regression model that included time to HZ onset as the response variable and treatment group, age stratum, and expected duration of antiviral prophylaxis (for auto-HSCT recipients) as explanatory variables.
dComputed based on a Cox regression model that included time to HZ onset as the response variable and treatment group, age stratum, expected duration of antiviral prophylaxis (for auto-HSCT recipients), and the natural log-transformed VZV gpELISA as time-varying explanatory variables.
eResults from the IFN-γ ELISPOT assay are expressed as the frequency of spot-forming cells per million PBMCs.
fComputed based on a Cox regression model that included time to HZ onset as the response variable and treatment group, age stratum, and HM immunocompromised stratum (for patients with HM) as explanatory variables.
gComputed based on a Cox regression model that included time to HZ onset as the response variable and treatment group, age stratum, and HM as explanatory variables.