Fathoming serosurvey data, with caution

The data from the ICMR’s Fourth National Seroprevalence Survey have clear implications for the situation in India today

Updated - August 03, 2021 12:49 pm IST

Published - August 03, 2021 12:02 am IST

NEW DELHI, 02/09/2020: Health workers take blood samples for a serological survey, in north Delhi’s Adarsh Nagar on September 02, 2020. Photo : Sushil Kumar Verma

NEW DELHI, 02/09/2020: Health workers take blood samples for a serological survey, in north Delhi’s Adarsh Nagar on September 02, 2020. Photo : Sushil Kumar Verma

The past two weeks have seen the release of the Fourth National Seroprevalence Survey for COVID-19 (https://bit.ly/3lklMzJ), as well as the reporting of data of serosurveys independently conducted by State governments such as in Tamil Nadu. The data from the Indian Council of Medical Research (ICMR) conducted national serosurvey are truly remarkable, indicating that between December-January, when the third serosurvey was conducted, and June-July 2021, at least 40% of the population of India, over 500 million people, were exposed to the novel coronavirus of which a small proportion have antibodies because they were vaccinated. The data also show for the first time that over half of all children tested have already been infected, which is expected and reassuring.

Not from infection alone

Such a high rate of infection seems incredible, and if true, needs to serve as a warning to India and the world because it shows the consequences of a combination of variants, inappropriate population behaviour and delayed implementation of public health measures. It is, of course, evident, that all of the seropositivity did not come from infection alone, but given that the seropositivity rate was 67% and the serosurvey reports that a quarter of eligible adults were vaccinated, a large proportion of the antibodies resulted from detected and undetected infections in the first or second waves, with the sharp seropositivity rise indicating that the bulk must have occurred during the second wave. With less than 32 million positive cases, of which many may be positive more than once, and the likelihood of waning antibodies leading to some of those previously infected testing negative subsequently, the seropositivity of 67% is an under-estimate. There are other caveats as well.

Also read: Coronavirus | Two-thirds of Indians have antibodies, shows ICMR survey

Although the serosurvey methodology is similar to previous rounds and sampled individuals in 70 districts in 21 States, it is not clear how representative the survey really is — given the reported two-dose vaccine coverage of 13% in adults which is much higher than the national coverage at the time, and the known high heterogeneity between and within States. This is where State-level data can be useful, and the third Tamil Nadu serosurvey, conducted in over 26,000 participants and compared to previous rounds allows comparison of at least district-level information.

Data from the States

The National Serosurvey data indicated that Madhya Pradesh, Bihar, Rajasthan and Gujarat have over 75% seropositivity, while only Maharashtra, Assam and Kerala were below 60%, with Kerala the lowest, at 44.4%. Although heterogeneity across districts is likely, the fact that the national serosurveys were repeated in the same districts allows for some comparison of seropositivity rates over time. For example, in Kerala, when the national rates were approximately 0.7%, 7% and 24%, the corresponding rates were 0.3%, 0.9% and 11.6% during the first three rounds of serosurveys, showing a consistently lower exposure than the national average, but a large increase between the third and fourth serosurveys.

If we examine the data from Tamil Nadu, where the State-led serosurveys were done in October 2020, April and July 2021, some figures jump out. There appeared to be no change from October to April at the State level, with 31% and 29% positivity, but then there was a massive increase to 66%. Comparing the ICMR survey in three districts with the 38 district surveys of the Tamil Nadu Directorate of Public Health, shows that the ICMR serosurveys found a seropositivity of 43%, 24% and 31% in the districts of Chennai, Coimbatore and Thiruvannamalai in the serosurvey in December 2020-January 2021. The corresponding figures from the State serosurvey were 41%, 22% and 36% in October 2020, 49%, 20% and 34% in April 2021, which rose to 82%, 43% and 68% in July 2021. Some districts showed marked variability, with the central and southern districts showing significant declines in antibodies between October 2020 and April 2021 followed by a large increase in July 2021. For example, Madurai went from 40% to 19% to 79% and Thoothukudi from 39% to 21% to 72%.

Explaining variation

What explains these variations and how can these data be used at the national-State levels? First, there was inconsistency in the assays used for antibody estimation, with the ICMR using an ELISA test that measured antibodies to inactivated virus in its first round and then switching to a method that measured antibodies to nucleocapsid and spike proteins. Tamil Nadu used a chemiluminescence method for all its assays, while serosurveys in Delhi, Mumbai, Pune and other local serosurveys have used different methods — which means that some of the variations may be due to the testing methods which have varying sensitivity and specificity for antibodies directed against different viral antigens. Second, the sampling methods across serosurveys have differed with the ICMR using a cross-sectional survey in 70 districts that were decided/based on reported cases with some lack of clarity on how villages/wards were chosen, while others have used population proportional to size, directed and convenience sampling. Third, the age groups have varied across surveys and with time and not all serosurveys have reported age-stratified results. Fourth, antibodies decay with time, as illustrated in the data from Tamil Nadu’s first and second serosurveys.

So, analyses need to account for the underestimation of exposure. There are many other considerations for analysis, such as the proportion of people included who have reported previous SARS-CoV-2 infection or exposure or the proportion vaccinated. But in general, serosurveys more accurately reflect the experience of the population with an infectious agent than reported cases. This is especially true in India, where without targeted effort and sometimes despite detailed search, it is not possible to discern the reasons for testing, the numbers of individuals that have contributed to cumulative testing, the locations where tests were done or the types of tests used — all of which are essential to understanding the strengths and limitations of the data on reported cases, and the test positivity rate.

Reported cases represent the tip of the iceberg for a virus that can cause asymptomatic infection or result in mild symptoms that are not distinguishable from other respiratory viral infections. Serological data has value for public health because knowing where and how many have already been exposed to the virus gives a clearer picture of how and when infection has penetrated the population. The data from the ICMR’s Fourth National Seroprevalence Survey therefore, have clear implications for the situation in India today.

The Kerala example

When Kerala is reporting the bulk of the country’s positive cases and has a reproductive rate of infection of greater than 1, how concerned should we be? Seroepidemiology shows that the State has over 18 months been able to limit the exposure of its population, and vaccination data indicates that the State is outperforming most others. In general, the health-care infrastructure has not been overwhelmed, deaths are low, and shortages of oxygen and hospital beds have not been reported. Relaxation of restrictions over Eid on July 21 are likely to have resulted in cases which should peak in the first week of August, but there is a need to ensure continued population compliance with restrictions, particularly with Onam approaching and a need to ramp up vaccination. In fact, given Kerala’s success in controlling the virus so far, a key consideration should be whether States and districts with the lowest seroprevalence should receive a greater proportion of the vaccines while supplies are limited.

The immunity gap identified in Kerala demonstrates the value of serosurveillance and the need to ensure that we continue to use this very valuable tool not just to record the history of viral circulation in India but also to inform decisions going forward. At the State and district levels, what can be changed based on serosurvey data? Increasing vaccination in areas with the lowest exposure in parallel with ensuring high testing levels and health-care system preparedness in areas with high vulnerability are immediate responses.

However, serosurveys need to be continued and the data integrated with testing, vaccination and clinical data. These are needed to understand ongoing infection rates, age distribution of infections and cases, where variations could be driven by seasonal coronavirus exposure, vaccination and the effect of waning antibodies, all of which will be essential to inform policies on the need and use of booster doses and long-term system preparedness.

Gagandeep Kang is Professor, Christian Medical College, Vellore, Tamil Nadu

0 / 0
Sign in to unlock member-only benefits!
  • Access 10 free stories every month
  • Save stories to read later
  • Access to comment on every story
  • Sign-up/manage your newsletter subscriptions with a single click
  • Get notified by email for early access to discounts & offers on our products
Sign in


Comments have to be in English, and in full sentences. They cannot be abusive or personal. Please abide by our community guidelines for posting your comments.

We have migrated to a new commenting platform. If you are already a registered user of The Hindu and logged in, you may continue to engage with our articles. If you do not have an account please register and login to post comments. Users can access their older comments by logging into their accounts on Vuukle.