Poliomyelitis eradication requires surveillance for acute flaccid paralysis (AFP), and in all countries children with AFP who are younger than 15 years are investigated for poliovirus in stool. However, collection of two 8-g stool samples 24-h apart and within 14 days of onset of paralysis is not easy. Samples need to be stored below 8 C documented properly, and tested in an accredited laboratory.

Individuals without adequate stool samples are examined by a neurologist with electromyography and nerve-conduction and other tests. A national expert committee1 reviews these cases, decides whether any are poliomyelitis, and labels them as compatible poliomyelitis in accordance with WHO’s recommended virological classification scheme.[1,2] The occurrence of compatible poliomyelitis suggests a failure of the surveillance system.[3]

Poliomyelitis eradication in India is a huge challenge. The National Polio Surveillance Project,[2] established in India in 1997 to guide public-health experts, has been a successful model for surveillance activities globally. In India, it was hoped that poliomyelitis would be eradicated quickly, but the virus resurged in 2006. It seems that India did everything according to the surveillance rulebook.

To avoid missing cases of paralytic poliomyelitis, the prevalence of non-poliomyelitis AFP should be at least 1 per 100 000 in children younger than age 15 years. To ensure that we identify the virus, 80% of AFP cases should have adequate stool samples. If these criteria are met and no cases of poliomyelitis are identified for 3 years consecutively, we can conclude fairly certainly that the country is free of poliovirus. However, if the frequency of adequate stool samples is 80%, then 80% of all poliomyelitis cases must be confirmed. India had certification-quality surveillance for the past 8 years. The prevalence of non-poliomyelitis AFP, which had been 2 per 100 000 since 1998, increased in 2004 to about 3 per 100 000 and approximately tripled in 2006 to 6•95 per 100 000.[2,4]

The number of confirmed poliomyelitis cases has declined since 1998, except during outbreaks in 2002 and 2006. In 2005, we hoped that we had finally conquered the virus. However, the virus made a comeback. The worst affected area was Uttar Pradesh, which consistently had the maximum number of poliomyelitis cases. Why did poliomyelitis eradication fail? A key issue is hidden in AFP data. Although India achieved the recommended surveillance indicators, there was a serious anomaly: the number of compatible poliomyelitis cases had always been more than 20% of total cases.[4] The lowest frequency (30%) of compatible poliomyelitis cases was achieved in 2002, the year with the biggest outbreak; and the proportion of compatible cases has been increasing as the number of confirmed poliomyelitis cases decreased.[4]

The 2005 data in the worst affected states of the 2006 outbreak—Uttar Pradesh and Bihar—highlight this anomaly (table).[2] The frequency of non-poliomyelitis AFP is high in Uttar Pradesh and Bihar, 13 times more than expected.[2] Although the frequency of non-Polio AFP in Uttar Pradesh was very high and the frequency of adequate stool samples was more than 80%, the proportion of compatible cases was more than 80%!. Because the number of compatible poliomyelitis cases should be less than 20%, India has missed many cases over the past several years, which has adversely affected efforts to eradicate the virus. This situation seems to be an inadvertent but tragic consequence of reliance on only two indicators to assess the quality of AFP surveillance. By contrast, surveillance data for Indonesia in 2005 show that the proportion of compatible poliomyelitis cases from the total was about 20% (75 compatible cases and 349 confirmed cases).

The main aim of AFP surveillance is to detect circulating poliovirus through identification of paralytic poliomyelitis cases. The use of only two essential criteria for maintaining the quality of AFP surveillance has failed in India. A third factor should be added to the criteria for quality surveillance—ie, the proportion of compatible poliomyelitis cases among the total poliomyelitis cases should not exceed 20%. This new criterion would keep a check on the reported frequency of adequate stool samples. However, the 20% threshold should be reduced proportionately if the prevalence of non-poliomyelitis AFP is much higher than 1 per 100 000 to avoid a false sense of increased surveillance quality.

Paul T Francis
Amrita School of Medicine, Cochin, Kerala, India 682026
paultfrancis@gmail.com, paultfrancis@aims.amrita.edu

I declare that I have no conflicts of interest.