Spotlight: Comprehensive Analysis of Late-Stage Trial Data of AstraZeneca’s COVID-19 Vaccine

On 8th December, AstraZeneca and the University of Oxford formally published results from a large clinical trial of its COVID-19 vaccine- ChAdOx1 nCoV-19 (AZD1222) in Lancet. This is the first peer-reviewed study that shows the induction of immune responses against spike protein using a viral vector, provides immunity against the infection. We present here some important takeaways.

Which trials were included in the interim analysis?

AstraZeneca has four ongoing COVID-19 trials; COV001 (Phase 1/2; UK), COV002 (Phase 2/3; UK), COV003 (Phase 3; Brazil), and COV005 (Phase 1/2; South Africa). While data from all four were used for safety analysis, only COV002 and COV003 contributed to the efficacy analysis. In each of the trials, participants were randomly assigned with the COVID-19 vaccine or a control (meningococcal vaccine or saline).

Why was there a difference in the dosing?

In the first study (COV001), a viral load of 5×10¹⁰ particles was used and fixed for subsequent trials. This was calculated using a spectrophotometry method. However, when the manufacturing of the vaccine started for large-scale trials (COV002) at different contract manufacturing organizations, they calculated the viral load using two different methods-spectrophotometry and qPCR. While spectrophotometer gave a value of 5×10¹⁰ viral particles, qPCR showed only half of it, i.e. 2.2×10¹⁰.

Since the first study (CVOV001) used a spectrophotometer to calculate the viral load, this method was fixed to calculate the viral load in further studies after consulting with the national regulator. But when this value of vaccine was used on patients, it only gave half the response seen in the first study. After further analysis, it was understood that an interfering agent gave the wrong values in a spectrophotometer, and thus, qPCR was set for calculating the viral load hereafter.

This led to the formation of two different groups: the first group was given a low primary dose (LD) of 2.2×10¹⁰ viral particles and the second group was given a standard primary dose (SD) of 5×10¹⁰ viral particles. The booster dose for both groups was a standard dose.

What did each of the trials reveal about vaccine efficacy?

The primary analysis was done 14 days after the booster dose in seronegative participants with a positive nucleic acid amplification test. This was a pooled analysis and included both LD/SD and SD/SD patients. Around 11,636 participants out of 24,000 were analyzed. In total, 131 cases of symptomatic COVID-19 in LD/SD or SD/SD recipients were found. Of 131, 30 cases (LD/SD-3, SD/SD-27) were found in the vaccine group and 101 in the control group. “In participants who received two standard-dose vaccines, vaccine efficacy was 62.1%, whereas, in those who received a low dose as their first dose of vaccine, the efficacy was 90.0%.

In the secondary analysis that measured the efficacy of the vaccine after the very first dose, the vaccine was found 64.1% effective. This analysis was done 21 days after the first dose of the vaccine.

As for the ability of the vaccine to control severe COVID-19 infections, ten participants were hospitalized due to the COVID-19 infection and two of them had severe disease. All ten were found in the control group. This shows that the vaccine is effective at preventing severe COVID-19 infections.

Why was there a difference in the timing of booster dose in different groups?

Initially, there was no plan of boosting the primary dose. However, after the data from the early immunogenicity cohorts came out, it revealed that boosting the primary dose increased the immunogenicity response to the virus. Boosting began on August 3rd. But a primary dose of participants in LD/SD (18-55 years) began on May 31st and of participants in SD/SD (18-55 years) began on June 9. Therefore, there was a gap in the booster dose of more than 28 days for these participants. However, a set of older participants in the age group 56-69 and above 70 were enrolled only after the booster dose was decided. For these participants, there was a gap of 28 days, as planned initially.

To better understand the variability in the dosing interval, 2741 participants in the UK trial in the LD/SD group, received a second dose at least 12 weeks after the first and only 22 received a second dose within 8 weeks of the first. The median interval between doses for the SD/SD group in the UK trial was 69 days. Conversely, the majority of participants in the Brazil trial in the SD/SD group (2493 of 4088) received a second dose within 6 weeks of the first (median 36 days).

There are so many variables in the data. Why was the data pooled despite that?

Despite variability in the doses and timing of the dosing, the regulatory authority decided to pool all the data for determining the efficacy of the vaccine. The reason being that the two different dosings generated similar levels of binding antibody. Also, these discussions happened at the time when the disease rates in the UK were low, and “in the face of the pandemic, it was agreed that pooling could increase the sample size and could provide the earliest possible read on the efficacy that could contribute to public health.”

Moreover, in the UK and Brazil study groups, which had received two standard doses but were different in the timing of the booster (in the UK, the booster was given after 12 weeks and in Brazil after 6 weeks of primary dose), similar efficacy was observed. In the UK trial, the efficacy of 60.3%, and in Brazil 64.2% was observed. This indicated that the results are generalizable and therefore, pooling of data made sense.

Did the trial measure any asymptomatic infections?

The Oxford trial is the only one of the three leading vaccine developers that monitored for asymptomatic infections in participants by collecting weekly swabs. Asymptomatic infections were detected in 69 participants of which 24 were in the LD/SD group and 45 in the SD/SD group. This means an efficacy of 58.9% and 3.8% was seen in LD/SD and SD/SD groups, respectively. This shows that the lower dose of the vaccine was effective at preventing even asymptomatic infections.

Could 90% efficacy seen in the low dose group just be a chance?

While it may seem at first that the presence of high efficacy in low dose group just be a chance, as the low dose will generate a low immune response. However, similar high efficacy was observed in the asymptomatic individuals who were given a low dose (58.9% in LD/SD vs 3.8% in SD/SD). This rules out the possibility of chance. Topping that, “similar results have been seen for other vaccines where a reduced number or type of priming dose in infancy can lead to a higher response to a booster vaccine.” In the publication, authors comment that the higher efficacy in low dose group could be due to higher levels of neutralizing antibody, lower levels of anti-vector immunity with lower vector-derived antigen content of the first dose, or differential antibody functionality or cellular immunity.

Is the vaccine safe?

Across all the four studies, the vaccine showed good safety. However, many serious events occurred throughout trials that caused havoc amongst the public. There were 175 events in total, of which 84 occurred in the vaccine group and 91 in the control group. Three of the events were considered possibly related to the experimental or control vaccine.

1. A case of hemolytic anemia occurred 10 days after administration of the control MenACWY vaccine and was found related to the intervention.
2. A case of transverse myelitis occurred 14 days after the ChAdOx1 vaccine booster dose, which is possibly related to the vaccine.
3. Another COVID-19 vaccine-related serious event occurred in a participant in South Africa who recorded high fever but recovered rapidly.
4. Two additional cases of transverse myelitis were reported, one for the COVID-19 vaccine and the other for the MenACWY vaccine but eventually found unrelated to the vaccine by an independent committee of neurological experts.
5. There were four deaths in total, one in the COVID-19 group and three in the control arm. However, all were unrelated to the vaccine.

Is vaccine efficacious in older adults?

In this analysis, it is not clear how much protection does the vaccine offer to older adults as only 12% of the total cohort constitutes the older population (56 years or older). Older adults were recruited later in the trial and data is insufficient for analysis. However, earlier studies have shown that the efficacy of the vaccine in older adults is comparable to that of younger adults, and analysis in the future with more old participants will be needed.

By Ruchi Jhonsa, Ph.D.

Related Article: Russian Vaccine Makers Nudge AstraZeneca to Combine COVID-19 Vaccine, AZD1222 with Sputnik V

References

1. https://www.astrazeneca.com/content/astraz/media-centre/press-releases/2020/azd1222-oxford-phase-iii-trials-interim-analysis-results-published-in-the-lancet.html
2. https://www.astrazeneca.com/content/astraz/media-centre/press-releases/2020/azd1222hlr.html
3. https://www.astrazeneca.com/media-centre/articles/2020/covid-19-vaccine-azd1222-is-better-tolerated-in-older-adults-than-younger-adults-with-similar-immune-responses-across-all-age-groups.html

Author

Ruchi Jhonsa