Comparative complete scheme and booster effectiveness of COVID-19 vaccines in preventing SARS-CoV-2 infections with SARS-CoV-2 Omicron (BA.1) and Delta (B.1.617.2) variants

Background

The Omicron (BA.1) SARS-CoV-2 variant first reported in South Africa on 24 November 2021(1) has been classified by the World Health Organization(2) as a variant of concern (VOC), as it presents several mutations associated with increased transmissibility, immune escape, and higher risk of reinfection(3,4). The Omicron (BA.1) variant yields an S gene target failure (SGTF) signal (due to the deletion Δ69–70 in the Spike protein) in some PCR tests (e.g. TaqPath COVID-19, ThermoFisher, Waltham, MA, United States), which can be used as a proxy for Omicron detection and differentiation from the circulating Delta (B.1.617.2 and sub-lineages) variant (in which the Δ69–70 is rarely detected)(5).

The Omicron had a swift rise in Europe becoming dominant in a few weeks in England, Scotland and Denmark(6), and the European Center for Disease Prevention and Control (ECDC) risk assessment(7) referred that would become dominant in early January 2022 in all European Union member states. Considering COVID-19 vaccination, most European countries have ongoing mass population vaccination, with considerably high vaccine coverage in some countries. Hence, it is fundamental to understand how this novel VOC impacts the transmission dynamics in highly vaccinated populations. The first studies on neutralization assays revealed an extensive but incomplete escape of Cominarty BNT162b2 elicited neutralization(8), but booster dose increased the neutralization(9). These preliminary in-vitro results were confirmed by the first vaccine effectiveness studies in England and Scotland(10,11) for symptomatic infections and hospitalizations for both complete vaccination schemes and booster doses. UK study estimated a reduction of vaccine effectiveness (VE) for symptomatic infection with the Omicron with no effect for two ChAdOx1 (Vaxzevria) doses and 34% and 37% from 15 weeks post dose 2 of BNT162b2 (Cominarty)(10). Another study in Denmark found similar results with reduced VE for Omicron infection compared with Delta for two doses of BNT162b2 (Cominarty) and mRNA-1273 (Spikevax), with an increase in VE after a booster dose(12). It is still unclear the reasons for differences in VE of different vaccines and if the decreased VE translate for other countries with different vaccination coverage.

To shed some light on those questions we aim to replicate a study previously performed to estimate comparative vaccine effectiveness of mRNA vaccines between Delta (B.1.617.2 and sub-lineages) and the Alpha (B.1.1.7) VOC(13). Our objective is to measure comparative vaccine effectiveness (any vaccine) between Omicron and Delta VOC cases according to the complete primary vaccine specific scheme, time since primary vaccine scheme and booster dose, and also compare Cycle threshold (Ct) values between Omicron and Delta VOC according to the vaccination status of the cases. Additionally, we intend to translate comparative vaccine effectiveness estimates between Omicron versus Delta VOC into vaccine effectiveness estimates against Omicron variant infection using published estimates of COVID-19 vaccine effectiveness against Delta.

Methods

Results

Of a total of 15,001 sample collected during the study period for population aged 12 or more years old, 13,134 were included in the main analysis (Figure1). Of those 4898 (37.3%) were classified as Omicron. Distribution of Omicron cases differed from the distribution of Delta cases by all considered covariates. (Table 1)

Discussion

For the Portuguese population aged 12 or more years old, our results show higher odds of complete primary vaccination in cases infected by Omicron (BA.1) VOC compared to Delta (B.1.617.2 and sub-lineages) VOC cases (OR=2.1). This pattern was observed for all time intervals since completion of primary vaccination scheme according to the product characteristics (OR=2.3 for complete vaccination less than 113 days (16 weeks), OR=2.0 for 113-168 days since complete vaccination and OR=1.9 for more than 169 days (25 weeks) since complete vaccination). This finding on consistent difference between Omicron and Delta regardless time since vaccination was similar to primary VE reduction against Omicron VOC observed in UK study(10). Translation of case-case OR estimates to vaccine effectiveness estimates against Omicron VOC infection resulted in substantially lower and even null COVID-19 vaccine effect for complete primary vaccination against Omicron (VE ranged between -7.1% and 28.1%). These results are indicative of low or null protection of complete primary vaccination against SARS-CoV-2 infection also observed in other countries(10–12)

For the booster dose, available at the time of the study for population aged 50 or more years old, we estimated an OR of 5.2 (CI95: 3.1 to 8.8) that also indicated a marked reduction of protection against infection (regardless presence of symptoms) with Omicron compared to Delta. However, since booster has been shown to be highly effective against Delta VOC infection (VE: 94.0 (93.4 to 94.6)), the translation of case-case OR estimates in vaccine effectiveness estimates against Omicron VOC infection, led to estimates ranging from 64.2 to 68.8%. These results were consistent with previous findings from UK study that reported moderate vaccine effectiveness of the booster dose against symptomatic infections with Omicron up to 10 weeks after the booster uptake(10). However, the protection induced by booster dose may be of short term, since considerable VE waning, in particular against SARS-CoV-2 infection, has been previous reported in literature for primary vaccination. So further studies are required to monitor booster VE against Omicron infection along time.

The distributions of Ct-values in our study were similar between VOC and vaccination status. We observed no statistically significant confounding-adjusted mean differences either by vaccination exposure or VOC. Few studies have previously explored the differences in Ct-values between Omicron and Delta VOC. Existing evidence on viral loads differences between Omicron and Delta were not consistent, with studies from Denmark and Switzerland showing similar Ct-value distributions between Delta and Omicron cases(21,22), while in France significantly higher Ct values were observed for cases infected with Omicron(23). No significant differences in the Ct values between people infected with Omicron and people infected with Delta is thus not surprising and it has been corroborated by other studies(21,22). In fact, it is believed that the molecular basis underlying Omicron’s higher transmissibility likely relies on its mutational pattern involved in alternate entry mechanisms, increased host-cell binding and immune evasion, rather on higher viral loads spread through aerosolization(24,25).

Our study has several limitations. First, we were unable to differentiate between symptomatic and asymptomatic infections. Second, our study was based on RT-PCR tests, so rapid antigen test widely used in Portugal for SARS-CoV-2 diagnosis(26) were not covered by our data. Although study included cases samples from National SARS-CoV-2 Genomic Surveillance Network and three major clinical pathology laboratories, we cannot exclude that cases included in our study might not be representative of the overall infections detected in Portugal during week 49-51 2021. Hence, we compared age group, sex and region of study sample and the overall cases notified with infection in the same period. We found no significant differences in the sex and age group distribution between the universe of identified cases through rapid or RT-PCR testing and notified to SINAVE and the study sample. However, the north region tends to be overrepresented in our study sample. Third, incentives for testing had changed during the study period, negative tests regardless vaccination status became required to go to restaurants, hotels and to participate cultural events(27). This change may affect study results. However, selection of cases to the study sample was independent from the VOC type and from vaccination status. Fourth, this study did not collected data on comorbidities, considered relevant confounding variables in vaccines effectiveness research. So, we were not able to adjust for comorbidities in our models, neither estimate proportion of those with comorbidities by vaccination status. This can affect study results since individuals with comorbidities, such as immunosuppressed patients were prioritized for additional vaccination dose in Portugal(19). Finally, to translate case-case result to VE estimates we used previously published estimates of VE against symptomatic infection with Delta, while our study sample included notified infections regardless their symptoms.

Among study strengths, we should mention its large sample size due to use of surveillance data, well-established and high positive predictive value in SGTF method to classy infections as Delta or Omicron variants, and the robustness of results when we changed the sampling strategy in our sensitivity analyses (WSG-only or restriction to Ct<25).

Although direct comparisons to other studies are challenging due to differences in methodology, outcome and exposure definitions, vaccination calendar, eligibility for booster, vaccine brand-specific policies, testing patterns and other non-pharmacological interventions in place, our findings corroborate the general trend. More specifically our results support a marked reduction of primary and booster vaccination schemes effectiveness in preventing Omicron infections compared to Delta observed in other countries(10,12). Our findings suggest that complete primary vaccination may not be protective against SARS-CoV-2 infection in regions where Omicron variant is dominant. Even with lower booster vaccine effectiveness against Omicron, massive roll out and acceleration of booster vaccination campaign can contribute to reduction of number of susceptible to SARS-CoV-2 infection and its incidence in the population.

Supplementary data

• Download figure
• Open in new tab

Figure S1

Distribution of Ct values by VOC and vaccination status, population aged 12 or more

• Download figure
• Open in new tab

Figure S2

Distribution of Ct values by VOC and vaccination status, population aged 50 or more

Construction of empirical distributions and estimation of confidence intervals for COVID-19 vaccine effectiveness (VE) against Omicron using data on VE against Delta and odds ratio from case-case study design

We computed estimates of vaccine effectiveness against the Omicron for complete primary vaccination scheme and for the booster dose by combining previously published vaccine effectiveness estimates against Delta and OR estimated in this study using the following formula: where VE_Omicron represents COVID-19 vaccine effectiveness against Omicron VOC, VE_Delta COVID-19 vaccine effectiveness estimates against Delta VOC and OR_case−case the ratio of vaccination odds between Omicron cases versus Delta cases obtain through a case-case design. To estimate 95% confidence intervals for COVID-19 VE against Omicron we used Monte Carlo simulations. First, we constructed empirical distributions for all input parameters, namely of vaccine effectiveness against Delta (VE_Delta) and OR_case−case from case-case design. To construct empirical distribution of VE_Delta we assumed a normal distribution for log(1-VE_Delta), since COVID-19 VE_Delta estimates were obtained in test-negative design study as VE_Delta=1-Odds ratio(OR_Delta), estimated by logistic regression model, as OR_Delta=exp(β_{ve_delta}), where β_{ve_delta} represents coefficient from logistic regression model for vaccination exposure. We transformed VE_Delta into and VE_Delta 95% confidence interval upper and lower bounds into and , respectively. We computed a standard error , where Z_0.975 represents a quantile of standard Normal distribution, and generated pseudo-random numbers from Normal distribution: Simulated values of were transformed back to original scale using following formula: .

To construct empirical distribution of OR_case−casewe generated pseudo-random numbers from Normal distribution where represents coefficient from logistic regression model and its standard error

To return to original scale, we applied an inverse transformation: We draw 10 000 samples of OR_Delta and VE_Delta used them to construct empirical distributions of VE_Omicron The 2.5 and 97.5 percentiles of these empirical distributions were used as lower and upper limits of the 95% confidence intervals for VE_Omicron