To push or not to push the panic button? In the case of swine flu this year in India, that seems to be the question. Ever since two researchers from MIT raised concerns about India’s system of influenza surveillance in a US journal on March 11, and pointed out that a strain of swine flu (H1N1) from 2014 may have mutated into a more virulent form, the media has been looking to India’s National Institute of Virology (NIV) – which monitors the virus in India – for answers. Nearly a month before that, the NIV’s deputy director indicated that a study of the virus was in progress, and that no significant mutations had been found in the current strain of swine flu circulating in the country, which was the same as the strain found during the pandemic of 2009.
But the NIV is yet to make the results of this study, conducted along with the National Centre for Disease Control (NCDC), public. The NIV also said that it had studied the 2014 strains the MIT researchers referred to, and did not find the mutations they pointed out. In their paper, published in the ‘commentary’ section of Cell Host & Microbe, the MIT researchers speculated that a particular mutation in one of the strains from 2014 – which has been linked to increased virulence – might explain why swine flu was still in circulation, even though swine flu cases normally drop in summer.
Meanwhile, in the first three months of this year, India saw over 2,000 swine flu-related deaths – according to official records, that’s more than the number of people killed in either 2009 or 2010, when the effects of the 2009 swine flu pandemic were seen. Is this a secret government conspiracy, in the manner of Hollywood summer movies, or is the media exaggerating the threat, also in the manner of Hollywood summer movies?
* * *
Here’s the thing: the data the MIT scientists,Kannan Tharakaraman and Ram Sasisekharan, studied is available in a public database where information on swine flu, or H1N1, is stored – anyone at an institution registered with the database can access it for free. To break the tie, so to speak, Grist Media asked Bianca Habermann, a computational biologist at the Max Planck Institute in Germany, to look at the data available to verify if the mutations specified in the MIT paper actually took place. Read on for her findings.
Viruses are unpredictable – that’s why they need to be monitored carefully. And the H1N1 virus is only one of a gamut of diseases that circulates in India during flu season (which has two peaks in India: the onset monsoon, and the onset of winter).In 2009, H1N1 caused a pandemic that tapered off in mid-2010. According to the World Health Organisation (WHO), there were over 18,000 swine flu-related deaths reported during the pandemic. The Centers for Disease Control and Prevention (CDC) Influenza Division estimates that this figure is actually over 15 times higher. Officially, India saw 981 deaths in 2009 and 1,763 in 2010.
There’s also H3N2 – the predominant influenza virus in other countries this season –another influenza A virus, and influenza B viruses that need attending to. These regular, seasonal flu viruses can also kill. Experts agree that while the swine H1N1 (2009) virus may be more infectious, resulting in a larger number of cases, the percentage of deaths it causes shows no significant difference from the percentage of deaths caused by regular flu.On March 28 this year, 33,761 people were reported to have been infected with H1N1, while 2,035 people were reported to have died of it – that’s a mortality rate of around 6 percent. Last year, there were 937 reported cases, with a mortality rate of around 23 percent. (Experts caution that these numbers across years are likely skewed, thanks to underreporting of the total infected population – only severe cases with a higher chance of mortality were likely to have been tested. So with potentially higher numbers of infections, the mortality rate may actually be much lower.)
Influenza viruses mutate all the time – it’s just what they do. Many mutations in a virus may never see the light of day because they can be detrimental to the virus itself, causing it to self-destruct, as Habermann points out. And not all mutations lead to increased virulence.
“Mutation is a totally random event,” says Santasabuj Das, scientist at the National Institute of Cholera and Enteric Diseases (Indian Council of Medical Research), Kolkata – “it doesn’t have to build over successive years.” That’s why different years can see different strains that aren’t necessarily more virulent than they were the previous year, and mutations present in a strain one year don’t have to be present in a strain the next year.
To study a virus, you look through its genetic makeup – you “sequence” its genes. To borrow a great analogy from journalist Samar Halarnkar, sequencing genes is like searching a book, one letter at a time, for spelling mistakes.
The hemagglutinin (HA) gene in a virus helps it bind to the surface of host cells and dodge antibodies; mutations in the gene can alter the ability of HA proteins to bind to cells. Researchers studying a virus look for this gene sequence in order to monitor changes in virulence.
The MIT researchers studied two HA sequences from 2014 – the only two that year deposited from India – named A/India/6427/2014, collected in March, and A/India/5964/2014, collected in May. The March strain, which they discuss in the paper, exhibited three mutations – K166Q, T200A, and D225N. D225N is linked to increased severity of the virus.
Back in March, in an official statement through the Ministry of Health and Family Welfare, the NIV said: “Our experts have carefully examined the findings mentioned in the above-mentioned publication. We found that the strain analyzed in the said publication and the sequence data of the original H1N1 virus A/India/6427/2014 as available with NIV did not show any of these mutations. Subsequent report on antigenic/genetic analysis of this H1N1 virus by CDC/WHO as communicated to NIV also did not report any oseltamivir [the drug for swine flu – sold in the market as Tamiflu] resistance or any other genetic changes in HA genes that could be virulent markers…The genetic analysis of the HA gene of the H1N1 isolates from the present 2015 outbreak do not show any such mutations as mentioned in the above publication.”
Largely, our media has pitted the MIT paper against the NIV, as if the truth must be one or the other. And given how much less chatty the NIV has been and our inherent distrust of our government institutions, many have assumed that the MIT is right.
But it’s a messy business, presenting several scenarios.
When we asked Habermann, she ran the data independently and in her analysis of the sequences, she did find all three mutations in the March strain. In the May strain, she found just the first two. But given that no sequence from the current outbreak of 2014-15 is available, she points out, it cannot be discerned whether the current strain carries any of these mutations.
Now we have three possible explanations.
One: that the MIT researchers are right about the mutations (which Habermann’s analysis of the data confirms). The facts seem clear: the HA sequence from March 2014 deposited with the GISAID database does indicate the mutation. But whether this mutation exists in the current circulating strain in India, and the impact it has, can only be proven by examining the current strain itself and sufficiently monitoring the situation. This is why the joint study by NIV and NCDC referred to in the media is important.
Two: that the NIV is also right. It is possible that the NIV hasn’t found the three mutations indicated by the MIT researchers and Habermann if there is a discrepancy between the original data with the NIV and the data finally deposited with GISAID – something the NIV appears to hint at in its press release, which talks of “sequence data of the original H1N1 virus A/India/6427/2014 as available with NIV.” DT Mourya, the institute’s director, also appears to have indicated this discrepancy to a reporter. A slip, as it were, ’twixt cup and lip.
Three: That the NIV has not examined the data carefully enough, or is attempting to downplay the situation and maintain calm by denying the mutations. If the “original” data with the NIV does show mutations in the March 2014 strain, the NIV has either made an error or deliberately misled the public about their existence. Even if the mutations do not have a significant impact on the ground, that doesn’t explain why the NIV should entirely deny their presence.
Without more information from the NIV, it’s hard to figure out which of these scenarios – or which combination of them – is most likely. The NIV may be right about the 2014 D225N mutation being irrelevant to this year’s strain, whether or not it exists. And it may just be changes in the weather that have resulted in swine flu’s continued circulation. As for the rest, we’ll just have to wait and see. The entire mutations controversy has underlined, at the very least, the glaring holes that remain in the way India tracks and responds to influenza outbreaks.
But leaving the mutation business aside, doctors in states like Haryana and Jammu & Kashmir have accused the authorities of downplaying the situation, and fears abound that the situation is far worse than is apparent. “The subject is being actively debated and discussed in all strata of the society. Already enough panic exists about the disease,” said the Delhi High Court in response to a PIL seeking immediate steps to control the swine flu outbreak, adding, “We are satisfied with the steps already taken.”
So is the government really hiding a deadly mutation lurking in the shadows, waiting to come and get us all? By allegedly not monitoring the swine flu situation diligently enough, has India fed its furry little mogwai after midnight, leaving it with a nasty gremlin to battle? As unlikely as this scenario seems, only time – and a ton more information – will tell.
This isn’t the first time we’re hearing of H1N1 mutations in India. In 2013, NIV’s deputy director Mandeep Chadha said that the virus had exhibited “very minor mutations” similar to ones reported in other parts of the world, including the US and Singapore, but could still be treated with Tamiflu.
To study an H1N1 virus for possible mutations, swabs collected from deep in the throat or nasal tract of patients may be sent to a laboratory where advanced genome sequencing facilities are available. In India, these can be done in government facilities like IGIB in Delhi and CCMB in Hyderabad or run by private companies like the Bangalore-based Genotypic), where the virus strain is isolated and the genetic code in its RNA is sequenced. These sequences are deposited in public databases – such as that of the National Center for Biotechnology Information (NCBI) or the Global Initiative on Sharing All Influenza Data’s (GISAID) EpiFlu database. Experts say the most expensive part of this process is creating the network needed to collect the swabs and infrastructure to store them for community-based surveillance of the disease.
A virus’ genetic code tells you about the proteins in it – hemagglutinin (HA) and neuraminidase (NA) that hold the key to predicting whether or not a strain is more virulent. The HA gene, mentioned earlier, is the target of vaccines, and cheaper to sequence than an entire genome. The NA gene allows the mobility of a virus through an infected cell, and that is what drugs like Tamiflu are meant to target.
But predicting how an outbreak of a mutated virus will play out is hard, Habermann indicates – or even how a virus will evolve in the next season. Constant mutations are also why seasonal influenza vaccines have to be reformulated every year.
More on darling D225N
The D225N mutation that the MIT researchers found in the Indian strain is similar to a mutation that was found during a study in Florida in 2014.
In an email interview, one of the authors of the Florida study, Nicole M Iovine, Assistant Professor of Medicine and Hospital Epidemiologist at University of Florida Health, said: “In 2013-2014, we at University of Florida Health noticed that patients in their 30s and 40s were becoming severely ill with H1N1, and some died. This caught our attention because usually only persons [below] 2 years of age or [above] 65 years of age develop severe influenza, or die from it. We began to sequence the influenza strains from patients who died, and found several mutations.” One of them was in HA, says Iovine. “Others have shown that this particular mutation enables the virus to bind deep in the lungs, as opposed to just the upper respiratory tract like usual influenza. This means that the virus could cause a viral pneumonia, and that is unusual and of concern.”
It’s important, though, that the limitations of the Cell Host & Microbe commentary are borne in mind. “The research paper,” says Mourya, the NIV’s director, in an email to this reporter, “is an academic commentary based on in-silico [computer based] modeling and citation of other researchers and not an actual study with H1N1 strains circulating in India. We and others around the world have also seen such changes which are not new in H1N1, but there is no data to correlate such changes with disease virulence at population level.”
Santasabuj Das is the one of the authors of a study published earlier this year by a group of Indian researchers, who compared all HA gene sequences of H1N1 available with the NCBI from 1918 (the first deadly H1N1 pandemic – Spanish influenza) to December 2014 to trace the possibility of a further H1N1 pandemic in near future. The viral sequences they studied were primarily from the Western Hemisphere, and they found that the H1N1 strain from the 2013-2014 flu season appeared to be similar to the 2009 pandemic strain – and appeared to show increased virulence. It said that a more severe global H1N1 outbreak was likely in the near future. But the limitations of this study, Das points out, are the same as those for the MIT researchers’ paper. “It was a computational study and it was a prediction, like theirs. There is certainly no experimental proof that this [mutation] has actually increased the virulence. I believe that in the laboratory, [NIV] scientists should actually evaluate whether the prediction of increased virulencestands true or not.” The fact that the researchers were only able to study two HA sequences from 2014, he points out, is “unfortunately too inadequate to come to any conclusion on whether they are representative strains or whether they are causing this outbreak.”
The study Das was part of did not involve looking at the Indian HA sequences from 2014. When I ask why, he gives me a rather peculiar answer – the Indian sequences were in a different database, which required registration (which is not a time-consuming process, he clarifies). The NCBI database, which they used, did not.
The MIT researchers’ paper, Mourya writes, “speculated a possible implication of three mutations as the major reason for H1N1 virulence in India based on the analysis of a single H1N1 virus deposited in the GISAID database by the CDC lab from an originating virus isolated by NIV during routine surveillance and not outbreak.”
When it comes to the discrepancies between the MIT’s analysis of the Indian HA sequences from 2014 and the NIV’s, if sequencing was done incorrectly, there may be errors in the information available in the public database, says Das. “Only verification of the original data available with the originating institution can reveal this.”
In order to obtain a clearer pictures of a virus, its entire genome sequence must be studied, and not just its HA sequence. The NIV claims to have studied 6 whole genome sequences from the current strain, which it says is the same as the 2009 pandemic strain and exhibits no significant mutations.
I askSatyajit Rath, a senior scientist at the National Institute of Immunology in Delhi, about possible reasons for the NIV not revealing its data yet. “What does it even mean, having the NIV ‘show its data’?” he says. “It is not a query that can necessarily be instantly answered. It may well be the case that the government wants to maintain calm by downplaying the situation. But it may also be the case that the NIV isn't ready to reveal details of its work yet because it isn't confident that it will stand up, as yet, to the scrutiny of its peers, being preliminary. It may just be that the sequences are still being checked or double-checked, which takes time.” According to Das, it is possible that the new genome sequences with NIV have already been deposited with public databases, but the institute may be waiting to publish its own report before the data is made public. Mourya did not respond to emails seeking clarification on this.
Rath also points out that given the larger picture of India’s public health system, we might be missing the larger picture by worrying about a possible mutation. “You cannot fight a mutation alone, you fight the entire disease on the ground. In India, where providing basic public healthcare to all is a challenge, worrying over whether an influenza virus may have mutated in a certain way, is missing the woods for the trees.”
The waiting, watching game
In addition to indicating mutations in the 2014 strain, the MIT researchers highlighted the fact that India was sharing very little of the information it had on H1N1 on public databases, and called for greater surveillance of the virus. When I asked DT Mourya, NIV’s director, about India’s alleged lack of surveillance, he seemed to refute this in an email: “India is a National Influenza Center (NIC) [these centers are WHO-approved] and data representative viruses are always deposited with the Center for Disease Control and Prevention (CDC/WHO) global reference centers for purposes of vaccine formulation and genetic characterization of strains through continuous interaction and data sharing between international labs.” However, Masato Tashiro of the WHO Collaborating Centre for Reference and Research on Influenza in Tokyo told Lancet in early March, “We have not yet received detailed information in regard to the characteristics of the circulating viruses [in India].”
Satyajit Rath, scientist at the National Institute of Immunology can’t stop himself from laughing when discussing the MIT researchers’ call for better surveillance. “Biomedical researchers working in the trenches in India know better than anyone else that there's a need for surveillance – it's no surprise if their hackles were raised on hearing it from researchers overseas, whose tone some may see as patronizing,” he says on the phone from Delhi. He isn’t wrong – Mourya’s email response to my queries on the MIT researchers’ paper seemed rather grumpy at times, accusing the paper’s title of being misleading (but he didn’t clarify why this was when I mailed him subsequently).
Having the right infrastructure in place, which is expensive, is nevertheless crucial to having better surveillance in India, Das says. “Obviously [as the fact that only two HA sequences from 2014 in India are publicly available] we do not have good enough surveillance.”
“I’ve spent almost 40 years working on influenza,” says 78-year-old AK Prasad, a retired professor of respiratory virology who is president of the Indian Virological Society and Chairperson of the Indian Foundation for Influenza. “I’ve been repeating this for years – we need a better system to tackle not just influenza but other infectious diseases as well. The most important thing is that preventive measures should be in position and followed rigidly…we certainly have less resources to deal with emergencies effectively.”
“There's also a cultural difference here that must be noted,” says Rath. “There's a gap between the situation in India, where hospital beds, adequate training and expertise, and funding for research is difficult to come by, and researchers in a well-funded laboratory abroad calling for greater surveillance.”
Has alarm about the mutations grown in part because the corresponding author of the paper in Cell Host & Microbe has been much easier to access, despite being in the US, than the scientists at NIV, who haven’t satisfactorily responded to queries from the media? Rath agrees that the culture of accessibility is somewhat more prominent in the industrial economies than it is in India. However, Mukund Thattai, faculty member at the National Centre for Biological Sciences in Bangalore, believes it’s a universal problem. The relationship between scientists and journalists is a difficult one, he points out – there’s the worry that years of work may be misunderstood by journalists with their eye on deadlines, who may not take the time to understand the work they’re reporting on. “So many of us have been burned before, that scientists not primarily required by their institutions or any policies to interact with the media may just prefer not to respond to queries from a journalist. I believe that the gap between scientists and journalists needs to close, but this is going to take effort from both sides.”
What about panic stations?
Meanwhile, is it a good idea to get vaccinated? The MIT paper is an opinion piece that calls for greater surveillance, but also speaks of "widespread vaccine campaigns" as a strategy to tackle influenza. According to both Das and Prasad, while vaccines for all may not be a sensible strategy, given the limitations in India’s public health system, they are helpful in protecting at-risk groups such as the very young, the elderly, and those with asthma or respiratory diseases. “Obviously in India if we're talking of a vaccination strategy it has to be directed to high-risk populations rather than vaccinating everyone. It should be kept in mind that vaccination may not be foolproof, but it can reduce the severity of the disease, and the burden on the public health system – casualties can be less,” says Das.
At the moment, the Indian government doesn’t plan to deploy the swine flu vaccine. Vaccines only work as a preventive strategy before flu season begins, they can take 4-5 weeks to take effect, and can’t be leaned on to manage a crisis once it’s already begun – at the moment it is only being recommended for healthcare providers, who come in contact with the disease on a regular basis.
Sasisekharan, the corresponding author for the MIT paper, is a well-known scientist and professor of biological engineering at MIT. He’s also a biotech entrepreneur who founded the company Visterra, which is working on an influenza A vaccine (H1N1 is a subtype of influenza A) and, according to the MIT News Office, possibly laying the ground for a universal influenza vaccine. The MIT paper ends with a small acknowledgement that Sasisekharan is on Visterra's board. When an Indian newspaper asked him about the Indian government's hints that he had an agenda to push by suggesting H1N1 had mutated for the worse, he wrote, “It has no relevance to this study.” Sasisekharan did not respond to emails from this reporter seeking his comments on the subject.
That is not to say that Sasisekharan's point about surveillance isn't an important one – Rath and Habermann both point out that the potential conflict of interest here does not necessarily make the MIT study unreliable. “It would of course be good to have an independent study in addition,” says Habermann.
As for whether swine flu today is more terrible than ever, whether we are really fighting a gremlin here, Rath cautions that there's no way of knowing the extent of underreporting of swine flu cases, either now or during the 2009 pandemic, so any comparison of these figures has to be done keeping in mind these gaps.
So until we know more about this year's swine flu – until the NIV makes public the details of its study – it's not a bad idea to keep washing those hands.
Deepika Sarma is Assistant Editor, Grist Media.