Conversation with Mircea Sofonea: "Today, the spread of the virus is exponential".

What's the latest on the epidemic in our country? What are we to make of the measures announced on Thursday March 18 by Prime Minister Jean Castex? Answers from Mircea Sofonea, Senior Lecturer in Epidemiology and Evolution of Infectious Diseases at the University of Montpellier.

Mircea T. Sofonea, University of Montpellier

The Conversation: You recently studied the V1 variants (from the B.1.1.7 lineage initially detected in the UK in September), V2 (from the B.1.351 lineage detected in South Africa in October), and V3 (from the P.1 lineage detected in Brazil and Japan in January), which are currently circulating in France. What have you learned from this work?

Mircea Sofonea: We have analyzed the proportion of variants by age, by analyzing 40,000 RT-PCR data specifically targeting certain sites which allow us to determine whether we are dealing with one of these three variants, or with the "historical" strain, which circulated in our country prior to their importation. These analyses were carried out on samples from the Cerba Group's network of laboratories, as well as from the Montpellier University Hospital, enabling us to obtain information from all over France.

Analyses show that, since February 16, these three variants have probably been responsible for more than half of all infections in most French regions. Although these RT-PCR data do not allow us to distinguish between the variant of Brazilian origin and that of South African origin (as the sites targeted by the analyses are identical for these two variants), we note - as we saw in England - that these new variants have a propensity to be more present in younger people: in the data analyzed, the proportion of infections caused by the variants gradually decreased with age, reaching a factor of two between the ages of 5 and 80.

We were also able to calculate the effective breeding numbers of the new variants, as well as those of the historical strain, for the months of January-February.

(Editor's note: The reproduction number is an estimate, over the last 14 days, of the average number of individuals contaminated by an infected person. The basic reproduction number (or R0) is used at the start of an epidemic, in the absence of transmission control measures and when the population is entirely susceptible to the virus. Over the course of the epidemic, this number changes: we speak of the effective or temporal reproduction number (Rt). If it is less than 1, the epidemic regresses; if it is greater than 1, it progresses).

If we consider that the reproduction number of the historical strain is 1, then that of the three variants as a whole is between 1.37 and 1.64 (95% confidence interval). We are currently carrying out more detailed analyses on each variant, as the V2 and V3 variants (identified in South Africa and Brazil respectively) are likely to be less contagious than the V1 variant (identified in England).

This means that if these variants had been the ones circulating at the start of the epidemic, before the introduction of preventive health measures, their reproduction number would not have been three, like that of the historical strain, but at least four.

TC: How do you respond to those who claim that these variants have no influence on the dynamics of the epidemic, since it stagnated in France and even declined in some countries in February?

MS: We did observe a stable plateau, which can be explained simply: at that time, the historical strain was more affected by curfew measures. It was in decline, which may have given the impression that the epidemic was slowing down. But at the same time, new variants, particularly V1, were gaining strength. This slowdown, at a time when more transmissible viruses were circulating, may seem paradoxical on the surface, but the breakdown of incidence between a historical epidemic in decline and a new one on the rise helps to resolve this paradox.

To a certain extent, this is reminiscent of the scenario at the end of the summer, when we saw a sharp rise in the number of contaminations, without any impact on hospital dynamics: this was because the people who became contaminated were younger. This rejuvenation in the number of contaminations, linked to the relaxation of barrier measures among young people during the summer, generated a kind of statistical illusion when we were only interested in the number of cases. This illusion disappeared, however, when age groups were taken into account. We then saw what happened in October: the virus spread to older people, and hospitalizations began to rise again.

That's the difficulty of studying this pandemic: we have to remain focused on quantitative analysis, without invoking hypotheses that have no proven biological support - unlike some "reassuranceists" who were betting on a drop in virulence, on a plethora of false positives, or even on a supposed "natural virus cycle". But neither should we reason by pure analogy: at the start of the pandemic, the scientific community based itself on what it knew of the SARS epidemic of 2002-2003, which led national and international observers to downplay its seriousness. In fact, people infected with SARS were not contagious at the time of symptom onset, and sometimes only became so five days later. With SARS-CoV-2, people are contagious before symptoms develop, and there are also asymptomatic or paucisymptomatic people...

Returning to the February situation, we note that it was compatible with a decline in the historical strain and an increase in new variants, which were gradually becoming the driving force behind new epidemics. Moreover, it should be remembered that the oscillations in the number of reproductions were slight: we were still above 0.9, in other words, we were in a dynamic that did not lead to a rapid drop in incidence, which remained frozen at a high level and was not likely to prevent a possible rebound.

Today, we're back on an epidemic growth trajectory, with a reproduction rate of between 1.02 and 1.11 nationwide (calculated on critical care admissions), i.e. a 50% increase in admissions in one month. In itself, this is not explosive. However, against a backdrop of already high hospital occupancy, tension is fast becoming a problem in critical care departments in some regions.

The Conversation: What's more, as viral circulation increases, so does the risk of other variants emerging?

MS: Exactly. But this is not just a national issue: solving it would require global coordination. However, it is always best to avoid the emergence of new variants on our soil, hence the classification of the mutant detected in Lannion, Brittany, as a variant to be monitored.
This raises the question of global vaccination strategy: concentrating vaccination on certain countries, as is currently the case, will not prevent the emergence of a variant elsewhere. Instead, the idea should be to break the epidemic everywhere, as each outbreak represents an additional opportunity for SARS-CoV-2 to mutate and give rise to a new lineage that is more contagious or capable of evading vaccines...

TC: Do we know why the epidemic has started up again more quickly and more strongly in certain areas, such as the Grand Est region, Île-de-France or Hauts-de-France, even though they had already been heavily affected?

MS: We're still working on quantitative proof (we'll be submitting a scientific article on the subject shortly), but habitat structuring and population density seem to play a major role. We know that, in addition to sanitary measures and collective immunity, the dynamics of the epidemic depend on a number of factors, although we are not yet in a position to estimate precisely the contribution of each of them.

One of these factors is housing density, the distribution of the population over the territory, and the connectivity of the urban fabric. On the eastern side of our country, there are many large, well-connected conurbations. This is also where the majority of borders are located, opening the way to the rest of Europe, where the virus also circulates. The weather also plays a role. There's a correlation between temperature, humidity and virus circulation. In the East, the continental climate encourages people to stay at home. The situation is different in the West, with its milder oceanic climate.

The epidemiological history of an area also influences the way in which an epidemic develops: collective immunity varies from one place to another, as does the cumulative incidence of disease, the level of vaccination, and so on. People's behavior, based on their perception of the infectious risk, also plays a role: if the risk is perceived as high, people pay more attention to barrier gestures, comply more closely with sanitary measures... Finally, superpropagation events (gatherings, etc.) also play the role of local gas pedals, but are unpredictable.

All these questions deserve to be explored in a quantitative and rigorous way, involving specialists from the human sciences. Unfortunately, time and resources are lacking.

TC: So immunity acquired during previous waves, or ongoing vaccination, has not protected these regions?

MS: As far as vaccination is concerned, in February it didn't make much difference, since only 2% of the population had received 2 doses on average in the country. As for natural immunization, we estimate that it was below 20%.

However, the collective immunity threshold to be reached in order to limit the virus' circulation is high - over 70% if we take into account the over-contagiousness of the variants. The examples of epidemics that have spread with little or no hindrance, for example on fishing boats, on the aircraft carrier Charles de Gaulle, and above all in the Brazilian city of Manaus, are also edifying. Here, the relative cumulative incidence has reached a peak close to that predicted by theory (over 80%), yet the epidemic continues. With a considerable number of deaths. Arnaud Fontanet and Simon Cauchemez estimated that, in the absence of measures, up to 450,000 people could have died in France.

Today, according to our models, the population immunization rate is around 14%. Those of the Institut Pasteur are closer to 17%. In both cases, we're at less than 20%, and there are regional disparities... In less-affected regions like Brittany and Nouvelle-Aquitaine, we can see that it's mainly local determinants that will influence the circulation of the virus, rather than immunity, which is too low.

According to the Institut Pasteur, however, vaccination currently enables us to reduce hospitalizations by a fifth compared to a situation without vaccination coverage.

TC: At his press conference on March 18, Prime Minister Jean Castex stated that ruling out the option of generalized containment at the end of January was "the right decision, because if we had had to contain then (...) we would have had to inflict on the country a containment of probably three months". What do you think?

MS: Of course not. Just as in a car, the slower the vehicle's speed, the shorter the braking distance, an earlier health response would have enabled a quicker return to a low incidence, which can be more effectively controlled by the screening/tracking/isolation triptych, as a relay to restrictive measures.

This would have provided better medium-term visibility for the population, hospital services, the economic sector and scientists. All other things being equal, our model suggests that if the number of reproductions had been brought back to its November level between January 15 and February 15, by mid-March there would have been fewer than 1,500 COVID patients in critical care departments (instead of the 4,269 counted on March 18).

While other countries took more draconian measures, France contented itself with a curfew. With only relative success: the epidemic did indeed freeze, but at a high level of virus circulation, which for weeks has resulted in several hundred deaths a day, to which must be added morbidity, people who will be left with after-effects of the infection, long forms of Covid...

It's important to understand that the higher the initial incidence, the greater the effort required to maintain an epidemic in a stationary state. Indeed, with constant resources, the effectiveness of screening, tracing and isolation measures implemented by primary care physicians, ARS and the Assurance Maladie decreases when the number of transmission chains becomes too high.

TC: What do you think of the measures? Isn't it paradoxical to "confine" while extending the curfew? Trying to "curb the virus without locking us in"?

MS: No, it's a good idea to encourage people to live outdoors, as long as barrier measures are always respected, and it's not an excuse to increase contacts outside the family circle. It should be stressed that such a configuration is unprecedented, and relies even more heavily on collective responsibility. We'll have to wait another two weeks before assessing its effectiveness.

It should be emphasized that measures to contain the epidemic are all the more effective the earlier they are taken. If the aim is to achieve a low level of circulation, this can be achieved more quickly by implementing strict measures, then relaxing them after two weeks when the effects are seen. If we wait too long and implement insufficiently effective measures, we tire the population, and run the risk of losing support for the measures. In fact, poorly enforced containment would be the worst possible solution, since we'd have to pay the high socio-economic costs without reaping the health benefits.

In Germany, the authorities have emphasized in their communications that we can't wait for the hospital situation to deteriorate before taking action. They have also set clear targets, with a timetable, which has won public support. In France, in December, an arbitrary limit of 5,000 new cases per day was set, which was ultimately not respected. There's still no end in sight in our country: today, the spread of the virus may be slower than in October, but it's once again exponential, and there's still no reason to expect a general relaxation.

TC: What do you think of the "race against time" towards spring and vaccine coverage mentioned by Prime Minister Jean Castex and Health Minister Olivier Véran?

MS: Between now and mid-April, the slowdown in the epidemic will depend on the measures announced this evening. Thereafter, vaccination coverage will probably be sufficient to stem the epidemic, in conjunction with the continuation of the measures in place since spring 2020. However, the situation could become fragile again in certain territories if these measures are relaxed too quickly.

On the other hand, we may well wonder why we haven't yet heard of occasional relaxations in certain untouched regions, such as the South-West. The territorialization and earliness of measures should be considered in both directions.

Mircea T. Sofonea, Senior Lecturer in Epidemiology and Evolution of Infectious Diseases, MIVEGEC Laboratory, University of Montpellier

This article is republished from The Conversation under a Creative Commons license. Read theoriginal article.