How effective is Cuba’s Abdala vaccine?

Vietnam will buy 10 million doses of Cuba’s Abdala vaccine [1]. But scientific information about this vaccine is very rare. Through some preliminary information I did an analysis and found what they reported to be fairly consistent with the efficacy data.

There is very little scientific information about the Cuban Abdala vaccine because there are no articles mentioned in the literature. However, through the popular press, this is a vaccine made by recombinant protein method (ie recombinant protein like Novavax of the US). It should be added that Novavax of the US has been tested in phase III and is 90% effective. [2], and ratified by Australia.

1. Abdala vaccine effectiveness

Cuba is poor, but it is a country with experience in making vaccines and they have proven achievements in the past. During this epidemic, Cuba researched and produced its own vaccine, but scientific information about their vaccine was too rare in the literature. Searching in Pubmed, we only see articles with comments about the Abdala vaccine, and through that we know that they have been testing the Abdala vaccine since last year (7/2020) [3].

The phase III RCT trial was registered on March 18, 2021, and according to the protocol they compared 3 doses of the vaccine to a placebo. The vaccine required 3 doses on day 1, day 14, and day 28. The number of volunteers reached 124,000 people aged 19 to 80.

What were the results? According to official information from the Cuban government, based on data from 48,2900 volunteers, the effectiveness of the vaccine is 92.28%. [4]. If true, this number is even better than the efficiency of the US Novavax (90%), but of course it cannot be scientifically compared like that.

2. Abdala efficiency analysis

Let’s test whether Cuba’s reported 92.3% efficiency figure is reasonable.

Actually, like I said yesterday [5] The effectiveness of a vaccine depends only on two key numbers: the number of infections in the vaccine group and the control group.

It’s nice that in the wiki page we have those two Cuban numbers. Accordingly, the total number of infections (symptomatic and identified by PCR) is 153; of these 142 were in the control group, and only 11 were in the Abdala vaccine group.

Because of the high sample size and 1:1 distribution, we do not need to know the exact number of sample sizes for each group, since the efficacy information is in those two case numbers (11) and 142). And, since the number of infections is very low compared to the number of tests (volunteers), we can assume that the number of infections obeys the Poisson distribution. Assuming that we do not know how effective the vaccine is, Abdala, we can attribute it to a Uniform distribution. Integrating these two distributions (Poisson and Uniform) gives us the posterior probability. We can simulate 10,000 times and take the median and 95% confidence interval to see if the vaccine efficacy reported by the Cuban authorities is reliable.

Using the above method, I estimated the Abdala vaccine effectiveness to be 91.7%, and the 95% confidence intervals ranged from 85.6% to 95.6%. They used the mean (92.3%), and I used the median (91.7%), so a slight difference is understandable. Thus, their report is consistent with data of 142 infections in the control group and 11 infections in the vaccine group.

How does it compare to other vaccines? In fact, it is not possible to compare efficacy between vaccines because each trial used a different control group. However, we can look at the efficacy distribution of each currently available vaccine (see Chart below). The way to read this chart is that the vaccine has a higher peak and the shorter the horizontal axis, the higher the confidence level, the more consistent it is. Looking at this analysis, it is true that Cuba’s Abdala vaccine is as reliable as Pfizer and Moderna, even higher than AstraZeneca and Sinopharm.

Distribution of efficacy in reducing the risk of covid-19 infection of the vaccines Abdala, JJ, Novavax, Sinopharm, AstraZeneca (AZ), Moderna, and Pfizer. This analysis uses the Bayesian method.

However, the above analysis is only based on uncensored data. We have to wait for their data to be officially published in a medical journal before we can make scientific comments.







Leave a Reply