Simon Thornley

15/5/2020

New Zealand has now progressively opened up, and we are now at level two. The next item to consider is what happens if we have another cluster of cases? What if the number of cases rises steeply? The government has stated that it has achieved ‘elimination’ of the virus, although there are dribs and drabs of new cases, mostly related to known sources. Recently, outbreaks occurred in Hokkaido, Japan, after lockdown was relaxed. The Northern Japanese island locked down for a second time in response to this second wave. Should we follow suit?

Our response depends on the answer to a number of questions. These include:

• How effective are lockdowns?
• How do similar viruses behave?
• How widespread is immunity?
• What are the risks to our health?

On the questions of the efficacy of lockdowns, there is likely to be ongoing squabbles, which will inevitably spill a lot of academic ink. We now have some compelling evidence that lockdowns in Europe were not especially effective. The trajectory of the epidemic was already declining in many countries when lockdowns were implemented, and the author of the article concluded that lockdowns were unlikely to have saved lives. Other evidence, such as a comparison of US States, that are either under lockdown or not reinforces this view. This analysis has now been subject to major revisions and re-analysis, but the conclusion remains the same – per capita cases and deaths from Covid-19 in each State are not materially different under either policy. The main factor linked to cases and deaths was testing rates. The more tests that were carried out by State, the more cases were found. Closer to home, our New Zealand – Australia comparison, in which New Zealand locked down harder and tighter than our cousins over the Tasman did not support the “hard lockdown” theory.

It is hard to believe with the Covid-19 blinkers on, but there are a number of other coronaviruses that have been circulating for many years that we pay little attention to. As well as the deadly SARS that has been eliminated, there are other coronaviruses that we have been living with for many years that have escaped our gaze. These latter examples are more similar to Covid-19 than SARS and MERS. These coronaviruses, including HCoV-229E, HCoV-HKU1 and HCoV-OC43 have a history of causing fatalities in resthome populations, like our new virus on the block. These viruses have now become endemic with winter seasonal peaks. With these similarities, it is likely that we will have to learn to live with future Covid-19 cases, particularly with winter around the corner.

The risk of future waves is likely to be related to the extent of our exposure to the virus. Tests of immunity in hard hit countries are returning immunity levels of about 5%, such as in Spain. Other tests of cell mediated immunity suggest higher real levels of immunity than those obtained from antibody tests alone. In Germany, 34% of antibody negative healthy donors showed markers of cellular immunity. In New Zealand, we don’t yet know our immune status, since we haven’t tested for it. With the comparative evidence that indicates that lockdowns are not especially effective, the fall in case numbers in New Zealand strongly indicates that widespread immunity is rising. The rise, fall and now low number of cases in China, with only smaller contained outbreaks after the initial peak, suggest that immunity is sustained, at least in the medium term.

We also need to consider how much of a threat the virus poses. In even hard hit countries, for the majority of working age people, the risk of death from the virus is about the same as a daily thirty kilometer trip by car. For those under the age of forty, the mortality risk is extremely low. We now know that hospitals in Australasia were never stretched, even at the epidemic’s peak. It simply makes little sense to squirrel children and working age people away, when the economic effects of lockdowns are ruinous.

The evidence that I see simply does not add up to an endorsement of further lockdowns. The elderly, particularly those who live in rest homes, deserve the greatest protection we can afford. For the rest of us, we can safely get on with our lives and progressively open up the country, even in the face of further cases.

By Simon Thornley

19 May 2020

Why does New Zealand need a serosurvey?

New Zealand urgently needs to test for antibodies to Covid-19. The standard test for Covid-19 at present is a genetic test that only detects whether or not the virus is currently in the body. Serology is a test of a person’s immune response to the virus and persists long after the virus has disappeared. This test gives important information about who in the community has recovered from infection and is thus unlikely to get the infection and pass it on to others. The overwhelming picture from this information is that the virus is much more widespread than is shown from genetic test positive cases.

In the response to swine flu in 2009, a serosurvey provided crucial information to dial back the clamour to stamp out the virus, since that survey showed that the virus was much more widespread than initially thought. As a consequence, it also followed that the virus was much less deadly than initially believed.

We can now count a total of 18 regions or countries that have conducted serological surveys and reported results in English to determine the extent of population exposure to the virus (Table). The proportion of the population who were found to have positive Covid-19 antibodies ranged from between 0.5%  in Colorado to 25.9% in Northern France. In some studies, the rate of positivity increased substantially as the study progressed. Even in the low prevalence regions, these findings strongly indicate that the virus is widespread and unlikely to be amenable to an elimination strategy.

Table 1. Prevalence of positive antibody tests to Covid-19 in surveys from around the world.

Region, Country	Sample size	Prevalence (%)
Northern France	171	25.9%
Guilan, Iran	552	22.0%
Gangelt, Germany	NA	14.0%
New York State, USA	15000	12.3%
Barcelona, Spain	578	11.2%
Wuhan, China	1402	10.0%
Aspen Colorado, USA	198	9.9%
Miami-Dade, USA	1400	6.0%
Switzerland, Geneva	760	5.5%
Los Angeles County, USA	NA	4.1%
Finland	147	3.4%
Kobe, Japan	1000	3.3%
Moscow, Russia	1000	3.0%
Santa Clara, USA	3324	2.8%
Netherlands	7361	2.7%
Denmark	9496	1.7%
Colorado, USA	5455	0.5%

NA: not available.

What are antibodies?

Antibodies are like keys in a lock that the body makes in response to viruses and other bacteria. Antibodies only fit a specific virus or bacteria. The shape of the antibody locks on to the microbe so that the body’s immune system can more easily eliminate the virus. Once a high proportion of the community have antibodies to the virus, it becomes very difficult for the virus to spread throughout the community, since it is hard for the virus to find new susceptible people to spread to.

Antibody tests are generally not used to diagnose the infection, since there may be a delay of one to three weeks from the time of infectiousness with the virus to the time that antibodies are generated by the body. Genetic tests, such as PCR, are usually used for making the diagnosis as they are positive earlier in the course of the illness.

What sort of antibody tests are available for Covid-19?

An antibody test generally involves the collection of venous blood or a finger prick to collect capillary blood. A number of test kits have been authorised by the US Food and Drug Administration for use for Covid-19. The Center for Disease Control has developed a test which is reliable for detecting SARS-CoV-2. The test is claimed by the organisation to be 99% sensitive (low false-negative rate) and 96% specific (low false-positive rate).

At present, tests of immunity are mainly recommended for assessing the extent of infection, and what proportion of the population has had mild disease from the virus. Until more information comes to light, researchers are not certain that test-positive individuals are unable to be re-infected, although this is likely to be true.

Are there other types of immunity to Covid-19 apart from antibodies?

As well as using antibodies which come from “B” white blood cells, our immune system also has “T” cells that recognise the virus directly, without the need for antibodies. A recent study from Germany has demonstrated that 83% of genetic test positive Covid-19 cases tested positive for T cells that react to the virus. Also, 34% of healthy blood donors, who were test-negative for antibodies, had evidence of reactive T cells, but at lower levels than cases. It is likely that these T cells confer some immunity to the virus, but it is unclear to what extent such people are protected.

What is the NZ government experience of antibody tests?

A wide range are available, but none have been rigorously tested in New Zealand yet. In order to be confident that these tests are useful, media interviews suggest that the government requires local evidence of testing their accuracy, despite overseas studies showing excellent accuracy with some tests. A number of tests are now endorsed by regulatory agencies in the United States.

If someone tests positive for antibodies, does that prove immunity?

The long term immunity associated with Covid-19 antibodies is not known. It is likely that they confer partial immunity, as seen with other antibodies for coronaviruses. This depends on the dose and route of administration. For example, in an animal study, mice administered coronavirus in the nose maintained immunity for at least 12 months, however, those that had exposure to the virus by mouth had high levels of immunity at one months, but lower levels at 6 and 12 months.

Conclusion

New Zealand urgently needs to test for antibodies to determine community exposure to the virus. If antibody levels are high, then this suggests that the virus is widespread. This also means that the virus is much less deadly than we feared.

 

 

 

5 April 2020

Simon Thornley

To understand the risks of ending lockdown, it is useful to think of a worst-case scenario. What would happen if ending lockdown led to a fate like Spain, Italy and New York.

A well-known epidemiologist has calculated the answer. For individuals aged less than 65 years, even in ‘pandemic hotbeds’, the risk of dying during the outbreak in hard hit European countries, is about the same as that associated with driving a car between 15 and 100 kilometres per day, throughout the pandemic. For people aged forty or younger, the risk is almost zero. Females have a risk two to three times lower than for males. For people aged younger than 65 years, with no medical conditions, the risk of death is extremely low, with this group contributing only 1/100 of all Covid-19 deaths.

Underscoring the low risk of death, the authors of the study noted that the mean age of death is approximately equal to the average life expectancy at each center. The exact ages of cases are not given for New Zealand cases, but if these are fixed at their midpoint for those for whom only a decade is given, the mean age of death is 81.6 years. This very closely approximates New Zealand’s life expectancy of 82 years. Since the numbers are so close, it is very difficult to argue that the virus is causing early death. In fact, such a pattern is replicated in almost all countries heavily affected by Covid-19. The risk of death in people aged less than 65 years was at least 92% lower compared to their older counterparts in eleven hard hit Covid-19 regions.

This analysis must force us to ask difficult questions, such as if our population of working age are at so low risk, why are we locking down our entire population? If the risks posed by the virus are so low, what are the downsides of locking down? Why are we closing our borders, and devastating our economy due to such a threat? On the basis of such a threat, why are we so obsessed with eliminating the virus?

There are really two choices that continue to be open to us to contain the virus, in the case of increased spread. These two combinations are ongoing lockdowns, or opening up the majority of society, returning us to work and school and protecting the vulnerable. The question of the closure of our borders continues to loom, as we consider whether we can remain cut off from the rest of the world or we work toward a sort of Australasian bubble.

The lockdown affects people of all ages, taking children away from school and workers away from their jobs. In contrast, protecting the vulnerable largely means that people over working age, past their 65th birthday are vulnerable.

The toll is starting to mount. In Queenstown, 30% of the population faces unemployment. Now, more than 100,000 kiwis are looking for mortgage relief. The true magnitude of the effect of the lockdown will take some time to be realised.

At the same time that we are dialing back the real risks posed by the virus, the downsides of putting the country in handcuffs are becoming more apparent. We urgently need to get back to work and school and do our utmost to protect the vulnerable.

1 May 2020

Simon Thornley

With much journalistic ink spilled over Covid-19 it is easy to forget that our policies of lockdown and social distancing are based on a belief about the lethality of the virus and its spread.  This belief comes from interpreting evidence. Currently, two main ideas predominate. Although the government has not stated it explicitly, their elimination policy now indicates that Covid-19 cases are contained, and that the virus can be put back in the box. This optimistic picture starkly contrasts with more convincing evidence from overseas that the virus is now well and truly out of the box, being now much more widespread than first thought. While this initially sounds terrible, it is instead good news, as it allows for a more relaxed stance toward the virus.

The government is now treating the disease in a similar fashion to how we have treated measles. Under this model, the vast majority of cases of infections have symptoms, the test is accurate, and we can contain the virus through contact tracing, quarantine and vaccination. This is a good model for infections that ‘declare themselves’ by causing unequivocal disease in cases and where vaccination is available.

The government has reused this model for Covid-19. Superficially, there is some justification, since community surveys of swab testing for coronavirus have all returned negative. The elimination strategy justifies harsh lockdowns which may be severe in the short term, but pay long term dividends. Under this model, the infection fatality proportion, that is the number of deaths divided by the number of people with infection, is (at the time of writing in NZ) 1.6%, well above seasonal flu (~0.1%), and justifies the ‘eliminate’ approach.

While this may sound attractive, several lines of evidence now indicate the virus often doesn’t ‘declare itself’ like measles, and is instead far more widespread than was initially thought. For example, in Iceland, a community survey of the population showed about 1% of the population tested positive for the virus from a nose swab, and about half showed no symptoms, despite the positive test results.

Information from antibody tests adds to the evidence that the virus has well and truly left the box and left its trail circulating in our bloodstream. After infection, the body mounts an immune response. After exposure, we produce antibodies as evidence that we have seen the virus. This blood test is quite different from the genetic nose swab that has so far dominated New Zealand’s ‘elimination’ thinking. The swab results only indicate the presence of the virus at the time the test was taken.

Antibodies give a contrasting picture from nose swab tests. Varying percentages of positive antibodies are reported, but the overwhelming picture is that many more people have recovered from the virus than has been appreciated. The percentage varies between 4% (Santa Clara, California) to 21% in New York city.

Why is this good news? Well, it indicates that exposure to the virus is about 50 to 85 times that observed from nose swab tests alone. In turn, this information dials back New Zealand fatality estimates to about 0.03% of all infected cases. This adjusted mortality rate is no greater than that for seasonal influenza. This is an important reality check on modelling figures which forecast carnage from Covid-19 equivalent to World War I deaths.

These antibody surveys are from overseas, and critics may argue that this does not apply to our New Zealand Covid-19 situation. Features of our Covid-19 cases, however, support the ‘out of the box’ idea. For example, in the recent measles outbreak, sourced from overseas, the majority of cases occurred in Auckland. For Covid-19, however, cases are much more dispersed around New Zealand, with 3% having no apparent link to overseas sources or other cases (44/1,461). In an effort to stamp the virus out, we will be hunting for needles in a very large haystack.

Since the disease is much more widespread than initially thought, then lockdowns are also unlikely to be effective at reducing spread. Recent evidence supports this idea. A comparison of US States showed that regions with social distancing were doing as well or even better on average for Covid-19 case or death rates than those that had a lockdown policy. Per capita cases and death rates were largely determined by a State’s population density – a factor New Zealand has on its side. While it is tempting to compare ourselves with New York, we have a population density more similar to Vermont, Arkansas, Oklahoma and Iowa. States with these population densities have death rates 95% lower than in New York, and are almost identical whether or not the State has locked down.

The government’s idea of a contained virus simply doesn’t gel with recent antibody surveys. The idea of elimination is scientifically unsound. The weight of evidence clearly illustrates that we are dealing with a virus that is more widespread and much less deadly than we feared. Evidence strongly supports us throwing off the lockdown shackles, safely returning to work and school, while doing our utmost to protect our most vulnerable in hospitals and rest homes.