Backing up our gains
As we eagerly await the start of the most ambitious vaccination campaign in our nation’s history questions remain. Australia has adopted a hybrid approach with half the adult population slated for a highly efficacious vaccine in Pfizer and the other half a moderately efficacious one in Astra Zeneca (AZ). One group can expect a breakthrough rate of mild to moderate COVID-19 infections of 5% while the other will contend with a higher breakthrough infection rate of 40% when the virus starts to re-circulate as we eventually relax our international borders. If we get the more transmissible South African variant, this could worsen the breakthrough rate for the AZ vaccine based on preliminary data from South Africa possibly up to 90%. This crude calculation of breakthrough infections does not account for a subset of the population who will remain vulnerable because they either decline vaccination (e.g. personal beliefs, pregnancy) or have impaired immune systems from chemotherapy or transplantation (4% of the Australian population are severely immunocompromised and nearly 20% moderately so i.e. 6 million in total). For Australia’s adults and children with impaired immune systems, COVID-19 is potentially life threatening because vaccine induced immunity may be partial if at all. Preliminary data showed that the Novavax vaccine was 60% effective against the South African variant but 49% in HIV infected people demonstrating that people with impaired immune systems do not respond as well to vaccination. Susceptible populations including Australia’s 5.6 million children will need the “protection of the herd”. Breakthrough infections in vaccinated people may not lead to hospitalisation but may still result in days off work and lost productivity. Whether “long COVID” complicates breakthrough infections is unknown, noting that it currently affects 10% of infections mainly on the milder end of the spectrum.
What then are the implications of this hybrid approach beyond the worthy goals of saving lives and preventing hospitalisations (major losses Australian’s have been largely shielded from)? Will we be able to open up our bars, restaurants and music venues? Will international travel resume providing a much needed boost to our tourism, hospitality and higher education sectors but critical also to commerce and trade? At the minimum, will this spell the end of precipitous border closures that have caused social and economic chaos? Many have rightly argued for prioritising health above economy but this downplays the human cost of economic downturns manifested as job losses, suicidality, relationship breakdowns, drug and alcohol misuse and child abuse- a hidden pandemic within a pandemic.
Forecasting is hard, but we only need to search our recent history for answers. The custodians of public health are our state governments and over the past 12 months we have seen our premiers transition from “flattening the curve” to elimination, taking a zero tolerance approach to COVID-19. They have no appetite for COVID-19 and will trigger restrictions based on local perception of risk. The rising threat of variants has seen an intensification of restrictions when viral escape from hotel quarantine occurs. This does not portend well for a moderately efficacious vaccine like Astra Zeneca, that may not stand up to the more resistant variants harbouring the E484K mutation (i.e. glutamic acid (E) swapped for lysine (K) at location 484 on the viral RNA code).
So far, a handful of variants have caught our attention but they have worrisome implications. The B.1.1.7 variant initially recognised in September 2020 in the UK, is associated with a 40% to 75% increased transmissibility which helped it quickly outcompete other variants to become dominant in the UK. It contains an unusually large number of mutations (n=24 including the N501Y mutation which is shared by both the UK and South African variants) located in the receptor binding domain of the spike protein and a site for neutralising antibodies. These mutations increase binding to the ACE2 receptor found on human cells and viral transmission. A survey on Jan 27 by the Office of National Statistics UK, revealed that people with this variant are more likely to have a cough, sore throat, fatigue, fever or muscle pains but less likely to report loss of smell or taste than those infected with other variants with the greatest difference seen in coughing, perhaps explaining why the B.1.1.7 variant is more contagious. Less clear is whether the B.1.1.7 variant is associated with an increased risk of death. A study matching 14,939 people with the B.1.1.7 variant to 15,555 comparable infected people without the variant, identified more deaths (104 vs 65 i.e. 0.2% vs 0.1%) within 28 days from diagnosis providing an increased the risk of death of 65% from baseline (relative risk 1.65, 95% CI 1.21-2.25). An association of the B.1.1.7 variant with increased mortality was corroborated by three independent groups including the London School of Tropical Medicine (Hazard ratio 1.35, 95% confidence interval 1.08 to 1.68) , Imperial College (mean ratio of case fatality rate in B.1.1.7 infected people versus non-B.1.1.7 was 1.36 (1.18 to 1.56), and Exeter University (Hazard ratio 1.91, 95%CI 1.35 to 2.71). Putting this into perspective, the UK’s Chief Medical Officer, Patrick Vallance, suggested that in every 1000 men aged 60 years who were infected with the new variant 13 or 14 might be expected to die, compared with 10 in 1000 infected with the original variant.
A concerning scenario is looming in the UK with the possibility of a more transmissible, deadly strain harbouring vaccine escape potential. The spike protein mutation E484K (located on the viral receptor binding domain which locks onto the ACE2 receptor on human cells) also found in the B.1.351 (South African) and the P1 (Brazilian) variants, was found in a very limited number of UK samples reported on 26 Jan (11 of 214,159) but may be under detected because only a fraction (7-10%) of isolates are sequenced in the UK. These variants are likely to become the dominant global strains in 2021, just like the D614G variant overtook the original Wuhan strain in 2020. Indeed, the South African variant is now found in 30 countries, the UK variant in 70 countries and the Brazilian variant in at least 5. These estimates represent the tip of an iceberg because few countries perform mutational tracking.
A consistent reduction in vaccine efficacy has been seen in clinical trials where these variants are prevalent. On January 28, Novavax reported preliminary results from 15,000 people in its phase 3 clinical trial. Efficacy in preventing symptomatic disease was 95% against the original SARS-CoV-2 strain, 85% against the UK variant (good) but fell to 60% and then 49.4% in 4400 South Africans without HIV and inclusive of HIV (6% of 4400) respectively. Among South Africans the majority of sequenced infections (92.5%) were the B1.351 variant. In response, Novavax indicated that they are planning clinical tests in the second quarter (April, May, June) of 2021 with a new candidate vaccine targeting the B1.351 variant. Preliminary findings from J&J on January 29 of its viral vector single shot vaccine (using Adenovirus 26, similar to its Ebola vaccine) in 44,325 adults from South America, the US and South Africa showed a similar reduction in efficacy with the variants. Vaccine efficacy using a slightly different combined endpoint of moderate to severe disease 28 days after vaccination varied across regions being 72% in the US, 66% in Latin America and 57% in South Africa for a pooled estimate of 66% overall. Concerning results of a phase 1/2 AZ trial in South Africa on Feb 8 showing no reduction in all types of infection among 1,749 predominantly young HIV negative people (median age 31 years, 68% black, 14% white, 16% mixed race) prompted the South African government to pause the planned roll out of the AZ vaccine and change to the J&J vaccine (9 million doses) and Pfizer (20 million doses). These efficacy estimates have great uncertainty around them but were predicted from lab studies of blood drawn from vaccinated people, showing a marked reduction in the potency of antibodies induced by the AZ vaccine against the South African variant (public presentation by SA chief investigator).
Moderna and Pfizer completed their trials before these variants became prevalent but have subsequently performed lab based studies to assess the potency of their vaccines. Moderna tested the potency of neutralizing antibodies from people in its clinical trial against the full spectrum of spike protein mutations associated with the B.1.1.7 and B1.351 variants and found that it retained potency against B.1.1.7 but had a 6.4 fold reduction against B1.351. This has prompted them to develop a booster candidate against B1.351 and consideration for a third standard booster 6-12 months later. Pfizer similarly tested the activity of neutralizing antibodies from 20 trial participants against three engineered viruses carrying key mutations found on the SA and UK variants including N501Y, E848K and D614G. They found that the loss of potency was mild at 0.81 to 1.46-fold compared to the parent virus, with the greatest loss seen with the combination of E484K + N501Y + D614G mutation, still showing that their vaccine retained neutralising activity.
A study of the AZ vaccine (Feb 4) demonstrated diminished efficacy against the B.1.1.7 variant tested against 256 patients. The study combined the low dose/standard dose and standard/standard dosing groups and showed that the vaccine had reduced efficacy in preventing a combination of asymptomatic and symptomatic infection being 66.5% for B.1.1.7 (good) vs 80.7% for non-B1.1.7 variants. Importantly, vaccination reduced the time people shed virus by a median of one week. A similar reduction in viral shedding has been shown in a real world study of the Pfizer roll out in Israel where a single dose reduced the viral load 4 fold from day 12 corroborating the findings from the clinical trial that led to its global registration. These findings suggest that onward spread of virus may be reduced with these vaccines. Astra Zeneca announced on Feb 3 that a revised shot tailored to the new variants will be available in September-November 2021 noting that Moderna have already developed one and are progressing it to a Phase 1 study, highlighting the agility of the mRNA platform.
Overall, vaccine efficacy appears to drop off most markedly for the B1.351 (South African) variant, although the equivalent data for the Brazilian variant (which is likely to be similar) are not available. This has implications for the UK B.1.1.7 variant also as this has also acquired a mutation shared by the SA variant known as E484K which affects the receptor binding domain of the spike protein. How far afield these variants have spread is unknown but Israel is expecting the B.1.1.7 to become dominant in the next few weeks. B1.351 has spread to at least 30 countries including the United Arab Emirates, Argentina, Portugal, Austria, Norway and Japan but may be under recognised in many countries including the US, which sequences only 1% of samples. The concern is that both B.1.1.7 and B1.351 are acquiring mutations such as E848K that allow them to spread faster and evade vaccines with the exception of the Pfizer vaccine which has retained good lab activity. The limitation of lab studies of vaccine potency is that they are a proxy for real world efficacy. Thus, the roll out of the Pfizer vaccine in South Africa will provide real world data on its effectiveness against this variant and should be watched closely. The unholy trinity of a more contagious variant or variants, potentially with a higher mortality and vaccine escape potential has triggered pre-emptive action from vaccine manufacturers, regulators like the FDA and governments. On Feb 3 the British government backed a Euro150 million (AUD235 million) alliance between Britain’s GlaxoSmith Kline and German biotech Curevac to develop the next generation of multivalent mRNA vaccines as variant boosters.
Announcement on Feb 1 of increased efficacy by spacing two doses of the AZ vaccine at least 12 weeks apart generated some interest. This post hoc analysis of the primary clinical trial data should be interpreted with caution. Post hoc analyses aim to ask a question rather than answer one. Why? Because they rely on analysing a slice of the main trial data; have greater uncertainty around estimates because numbers analysed are usually smaller and are prone to bias which may translate to inadvertently finding patterns where there are none. This sub-study provided the rationale for the UK vaccine authority’s decision to delay the interval between vaccine doses to at least 12 weeks in order to vaccinate more people. However, a quarter of a year is a long time to wait in the middle of a pandemic. This sub-study actually raised more questions than answers (as post hoc analyses often do!). Why was the efficacy of a single dose of the AZ vaccine at preventing symptomatic infection higher than two standard doses at 76% versus 63% respectively? In other words, vaccine efficacy seemed to decline with a second dose. This trial update contained an additional 5,541 people from SA, UK and Brazil but overall vaccine efficacy with two standard doses remained unchanged from the first interim study published on Dec 8 which led to approvals in several countries (i.e. 63% among 14,379 people in this update vs 62% in the first analysis). Reassuringly, there were no hospitalisations after the second vaccine dose (n=0 with vaccination vs 15 with control). These results will reassure Britons that delaying the second dose by 12 weeks has some supportive evidence. However, for the rest of us, we keenly await the release of the AZ trial predominantly from the US with >32K people where close to one quarter will be over 65 years, due to report in the next month or so. In this large trial, people are receiving 2 standard doses 28 days apart- potentially the dosing schedule we will receive if approved, hence discussions on single dosing and delayed interval dosing of 12 weeks are largely academic for Australia.
A clutch of European countries (e.g. Germany, Austria, Sweden, Spain, France, Ireland, Norway, Denmark, Netherlands, Poland) have not approved the AZ vaccine for people over 65 years (nor over 55 years in Italy and Belgium) due to insufficient numbers in the first interim analysis (n=711, >55 years who received 2 standard doses vs n=7,917 for Pfizer were over 55 years vs n=3,763 for Moderna were 65 years or more). This is not cause for concern but rather indicates that regulators were not confident of the AZ vaccine’s efficacy in the elderly-the US trial should answer this question. It is the same reason many countries are not vaccinating children-because initial trials had little or no data on children <16 years old. The Swiss on Feb 4 went a step further and refused to approve the AZ vaccine for any age group citing the need “to obtain additional data on safety, efficacy and quality” hinting at their concerns with the conduct of the Oxford-AZ clinical trial. Concerns by European regulators may be grounds for the TGA to await results of the large US trial before making recommendations in the elderly.
Taking a wider lens, the virus appears to be independently evolving towards a common goal characterised by higher contagiousness and reduced vaccine potency, associated with a common set of mutations including E484K and N501Y. Thus, while infections rage beyond our borders as they are predicted to do so for years to come, we should be now pivoting to the most efficacious vaccines such as Pfizer and Novavax, under the assumption that these “vaccine escape variants” will become dominant perhaps sooner than we think. Authorities are acting. South Africa has paused roll out of the AZ vaccine pending further data (noting that there is a lot of uncertainty around the lack of efficacy); monoclonal antibodies used for mild to moderate infection, by Eli Lilly, Regeneron and GSK have been rendered impotent by these variants; booster shots targeting B1.351 or designed to augment vaccine immunity have been developed (Moderna) or in progress (AZ, Pfizer, Curevac, Novavax) and regulators like the FDA are actively working to fast track vaccine updates. Booster shots are likely to become a durable revenue stream for pharma.
Mixed dose vaccination schedules with one type of vaccine followed by another type 12 months later, must be first proven to be safe and effective, hence clinical studies of the AZ vaccine followed by a mRNA booster are being planned. Viral vector vaccines like AZ or J&J have the disadvantage of potentially losing efficacy with serial boosters as the body attacks the viral vector carrier. While anti-vector immunity may be the achilles heel of the viral vector vaccines, this is not an issue with the mRNA vaccines and unknown for protein subunit vaccines like Novavax, which do not use a viral vector. Controlled studies of mixed dose schedules are important to ensure we see no untoward effects such as a decreased efficacy of a second type of vaccine or the development of antibody dependent enhancement. In this latter scenario, instead of gumming up the virus, vaccine induced antibodies may enhance viral entry into cells worsening disease. Clearly it is preferable to start on the front foot with vaccines from the same platform that can be safely boosted, avoiding the need for costly retrofitting of Australia’s immunity.
Maximising viral suppression with highly efficacious vaccines may not be beyond reach for Australia. High efficacy vaccines like Pfizer, Moderna or an updated-Novavax will blunt the threat of vaccine escape variants because they start at a higher baseline efficacy. They may enable a return to normal rather than COVID normal in line with community expectations thanks to a zero-tolerance approach by state and federal leadership. There is also a risk that immunity associated with less efficacious vaccines may dissipate more rapidly noting that we do not know how long vaccination induced immunity lasts. It is risky to pursue a hybrid approach only to find the population still vulnerable to dominant variants circling our waters.
The world is in a perilous position with the emergence of variants that are independently evolving to be more contagious and evade vaccines. Access to vaccines that are variant proof is smart. Australia stands to lose more than most other countries because of our gains in disease control. The Pfizer or Moderna vaccines are attractive given the potency, agility and versatility of the mRNA platform with development of new candidates taking 6 weeks. Concern has been raised regarding the long-term effects of the mRNA vaccines, however there is no biological plausibility to these concerns as mRNA degrades very quickly (some have said too fast to induce long term immunity) and it is not integrated into the human genome. Viral vectored vaccines have not been in widespread use either with the Ebola vaccine (J&J) using Adenovirus 26 approved by the European Medicines Agency in May 2020. The Moderna vaccine may have an edge given its storage at -20C with 30 days of fridge life and a proposal to increase doses from 10 to 15 per vial. Novavax could be made on shore by CSL but not while AZ manufacture is underway. The poor efficacy of Novavax (60% excluding HIV infected people falling to 49% with their inclusion) against the South African variant is sobering but might be overcome with proposed updates (e.g. a multivalent vaccine targeting multiple variants). If procurement of additional Moderna or Pfizer vaccines is not feasible in a reasonable timeframe, then CSL should switch to manufacturing Novavax instead of AZ to ensure a better chance of economic recovery in light of advancing variants. (Correction: on Feb 11 after my appearance on “Q and A”, I learned that CSL can not manufacture Novavax).
The AZ trial in the US is due to report soon but it would seem unlikely that efficacy will jump to 90% or more-the benchmark set by the breakthrough mRNA vaccines. There are trade-offs however, but no calculus would find that the cost of pivoting outweighs the health and economic prosperity of Australia. Indeed, in May last year, treasurer the Hon Josh Frydenburg MP estimated that every week of restrictions cost approximately $4 billion in lost economic output from a combination of reduced workforce participation, productivity and consumption. For perspective, vaccine procurement has cost a total of AUD5.2 billion (AUD3.3 billion with AUD1.9 billion committed to roll out). Breakthrough infections with dominant variants threaten our recovery and can be minimised with the most efficacious vaccines. Plummeting hospitalisations of 26% and confirmed infections of 45% in the past 2 weeks among 60+ year olds in Israel associated with the Pfizer campaign is cause for optimism.
Our strong Commonwealth and state government leadership, magnificent community cooperation, border control and government funded safety net has shielded Australia from the worst of this pandemic. Recovery is now contingent on immunisation restoring confidence while also mitigating the risks posed by the variants. In that respect the mRNA vaccines have the edge but we likely have more options than we think. These include; being able to procure more Pfizer or Moderna, tool up onshore mRNA manufacturing or await a more efficacious vaccine like Novavax while awaiting its updates. Hesitancy around vaccination is normal but Australians should understand the trade offs. Market forces and constraints beyond government control may mean that choice is an illusion. Australians should realise that delaying vaccination may carry a heavy cost to themselves, their inner circle and to opening up our economy when the drawbridge eventually comes down, as it will. Perhaps Australian’s would be more reassured if the government demonstrated that it had done everything it could to procure the highest efficacy vaccines, irrespective of cost which pales next to the price we have already paid and likely to pay in lost economic output. Consistent reductions in vaccine efficacy against the SA variant seen with Novavax, J&J, Moderna and now AZ should prompt a re-think of our vaccine mix. Pfizer indicated that efficiency gains are expected to slash vaccine batch production from 110 to 60 days and their 2 billion dose commitment includes 100 million produced by Sanofi with “fill and finish” by Novartis and an additional 75 million doses by Dermapharm (Germany). They expect to deliver 200 million doses to the US two months ahead of schedule in May instead of July 2021.
Whatever the outcome of our strategy, there is no looking away from the inferno beyond our shores. Unless this is extinguished through global citizenship with rich helping poor, embers will seed outbreaks for years, threatening our national security and acting as a drag on our pandemic recovery. Australia is a jewel, an outlier, confirmed by its ranking of 8th in the world in pandemic response by the Lowy, well ahead of the UK at 66 and the US on 94. For a nation that has done so well and come so far, we ought to be ambitious by backing up our gains with the most adaptable and potent vaccines that will help us accelerate rather than limp out of this pandemic. The best vaccines will supercharge our recovery transitioning us from living to thriving, enabling a strong Australia to pay off its debts and help its neighbours in line with our core values as a people.
A previous blog “Reflections on pivoting our vaccine strategy” can be found here: https://iser.med.unsw.edu.au/blog/reflections-pivoting-our-vaccine-strategy