Prof Salim Abdool Karim Weekly COVID-19 UPDATES

24 April 2022
The cases have risen dramatically in South Africa (Slide 2 – red circle) and the 7-day moving average of cases has just passed the threshold that I use for defining the start and end of a wave (ie. 5 cases per 100,000 pop as a daily average).  Also, test positivity has shot up to well above 10% (Slide 6 shows that it is nearing 20%), ie. other criteria used to define the start of a wave are also being met. As a result, the current trends in cases in South Africa are shown in this week’s slides as the 5th wave (Slide 4 – wave 5 is in red) and the same has been done for provincial slides where the threshold of 5 cases per 100,000 as a daily average has passed. 
I know that it is still April and that I previously said that I expected the 5th wave to start in early May, based an inter-wave period of about 13 weeks (3 months). The current increase is occurring 11 weeks following the end of the 4th wave. But the current increase is not all that it seems at first glance as there is a high level of uncertainty as to what exactly is driving the current surge in cases in South Africa. As Sarah Zhang headlined her article in The Atlantic that I shared with you in last week’s email, “The Coronavirus will surprise us again!”
But has the 5th wave really started in South Africa this week?
And I previously explained that a new wave would likely need to be driven by a new variant (likely to be named Pi) that has greater transmissibility and some level of immune escape. Does South Africa have a new virus with higher transmissibility and immune escape spreading in the country, ie. is there a virus to drive a new wave? 
It is too soon to give definitive answers to both these questions at this time. But there are several clues that can help us get closer to an answer. Let’s start with the elementary question.  The current drastic rise in cases could very well be one of 2 situations:
  1. The rise in cases over this last week could reflect a short-lived uptick in cases due to a new sub-variant of omicron. The omicron BA.1 wave started in South Africa on 2 December 2021 and was well on its decline when the opening of schools in January saw a brief uptick in cases due to omicron BA.2 before the original wave continued its decline. So, we know that omicron sub-variants can lead to short-term increases in cases – BA.2 has been causing similar upticks in many parts of the world that have completed their BA.1 waves. Since the January uptick in cases by BA.2, this sub-variant quickly became the dominant variant most likely due to its higher transmissibility (which has not been proven yet but has been estimated through modelling).  But BA.2 did not have sufficient escape from BA.1 immunity, especially when BA.1 antibody titres are high so soon after the wave, and so it has continued spreading at a low level to the few susceptibles available.

But now, 2 new omicron sub-variants BA.4 and BA.5 are spreading in South Africa (Slide 11). Omicron BA.1 was dominant in December 2021, but by February, omicron BA.2 had taken over as the dominant circulating sub-variant and by April, BA.4 and BA.5 are on their way to becoming dominant, reducing BA.2 to about half of the circulating viruses. Note how BA.1 has been almost entirely squeezed out in April – comprising just 3% of circulating sequenced viruses (Slide 11) – barely 3 months after it caused a massive wave in SA (and in most of the world). So, the current increase could be a second short uptick in cases due to omicron sub-variants BA.4 and BA.5. Why the uptick now? It is now just over a week since Easter, Ramadan and Passover took place (an uncommon overlap). The Easter long weekend hosted many weddings, church services and gatherings that can lead to such a drastic increase in cases. In 2021, SA had an increase in cases about 2 weeks after Easter. Personally, I am aware of 2 Easter weekend weddings this year where most of the guests became infected. Hence, there is a chance that this increase in cases could be related to gatherings being held with few public health measures – if it is gathering-related, it could be short-lived, especially with vaccination lowering secondary attack rates (recall my past slides on the Science & NEJM papers showing a 72% and 68% reduction in secondary attack rates due to 2 doses of Pfizer). 

​​2. Alternatively, it could indicate the start of a sustained increase creating the 5th wave due to omicron BA.4 or BA.5 or both. Several indicators increase the likelihood of this being the case – a) the increase is across all provinces (slide 22 with a magnification of the last few weeks shown in slide 23) and not just the major cities, b) SA is due for a 5th wave about now based on the past cyclical waves of infection (while we do not know for sure why the virus spreads in waves, it is thought to be related principally to waning immunity from the past wave – and since omicron causes mostly asymptomatic infection, which is associated with lower antibody levels, this early start of the wave could be due to the omicron wave immunity waning to low levels more quickly having started at a lower peak),  and c) BA.4 / BA.5 look circumstantially like they have higher transmissibility and immune escape, while we await the epidemiological data on transmissibility and lab data on immune escape.

If the current rise in cases is actually the start of the 5th wave, then we will know in about a week’s time (ie. next week’s email should provide some clarity). If the cases continue to rise rapidly, then SA will have over 10,000 cases daily in a week’s time and we will more confident that we are in the 5th wave.  If cases hover around their current levels and the exponential increase has not occurred, then scenario a) of a short-lived uptick is more likely to be the case.
Let’s now get into some of the more technical issues regarding the mutations and whether BA.4 or BA.5 can potentially become the next global variant.
BA.4 and BA.5 have the same spike mutations – they differ only in respect of a few mutations outside the spike – BA.5 has 3 mutations that have led to amino acid changes ie. changes in the protein outside of the spike protein and the receptor binding domain. The significance of this difference is not known – functionally, BA.5 is quite similar to BA.4 and they can be regarded as just one virus for the purposes of this explanation.
The additional mutations in BA.4 compared to BA.2 have led to BA.4 being identified on a regular PCR as it is S-gene target negative (just like the Alpha variant and the BA.1 variant). This makes it easier to track BA.4 though we will need to be careful as BA.1 is also S-gene target negative on the ThermoFisher PCR kit.
BA.4 has many different mutations compared to BA.1 and slightly fewer from BA.2. I was under the impression that BA.4 is a recombinant of BA.1 and BA.3. It turns out that this is not true. Speaking with Professor Tulio d’Oliviera this week, he explained that BA.4 is not a recombinant – it stems from a phylogenetic branch sometime in December 2021. So, BA.4 is has a common ancestor with other omicron sub-variants and came into existence through multiple mutations of an omicron virus. This fact is quite important – each variant has so far emerged as a separate variant unrelated to other variants ie. each evolved independently and not from each other (phylogenetically, this can be seen in the analysis from GISAID in Slide 13). If BA.4 becomes Pi, the next variant of concern, this will mark a major shift as it will be the first variant of concern that is derived as a sub-variant of another variant of concern ie. it means that new variants of concern can now stem from mutations of existing variants. This did not happen with Delta, which had >100 sub-variants – none of them caused any concern because they were readily neutralised by Delta immunity. For BA.4 to become Pi (the next variant of concern), it will mean that there is a second way in which dominant variants can arise with sufficient mutations to escape immunity against its parent virus. If true, it may be due to the large number of mutations in omicron. Either way, this will be a major shift in how variants of concern emerge and also means that we may see even more variants of concern in future.
So, what is it about BA.4’s mutations that could enable it to become Pi. It has 2 important mutations - in positions 484 and 486. On 9 March 2021, I did a zoom presentation at the Doherty Institute in Melbourne on the Beta variant and the importance of the mutation at position 484 and I had included some of this in my weekly emails of that time. Since many of you were not receiving my weekly emails then, I will summarise it here. Of the many papers on Coronavirus mutations, two had been particularly instructive for the March 2021 presentation (Slide 16). On slide 15, you can see how the spike protein looks (with some imagination) like a torso. On the left and right shoulders, are the key contact points between the virus and the ACE2 receptor. The ancestral virus had Glutamic Acid (indicated by the letter E) at position 484, a key position on the left shoulder that makes contact with the ACE2 receptor on the human cell. At position 484, the virus is therefore negatively charged and the ACE2 receptor is also negatively charged, leading to repulsion and poor attachment at that point. Almost all variants of concern and variants of interest have a mutation at position 484 (slide 14). Beta, Gamma, Mu, Iota, etc had a mutation that changed Glutamic Acid to Lysine(K) – (E484K is the notation for this change) making the virus positively charged now at position 484. So, the change in these electrostatic charges gives the virus a huge advantage enabling it to attach much more strongly to the human cell when the virus and the ACE2 receptor have opposite charges. 
BA.4 has a mutation at position 484 that changes Glutamic Acid to Alanine (ie. E484A), which makes the variant develop even stronger viral escape. Prof d’Oliviera is also concerned about the mutation at position 486, which is next door to 484 – he thinks that this mutation (also on the left shoulder of the spike protein) holds even more potential to escape immunity. Studies assessing this are still underway. But if he is correct (and he is seldom wrong on viral mutations), BA.4 may have the immune escape capabilities needed to become Pi. 
Current evidence of the way in which BA.4 is increasing (slide 12) suggests that it is spreading faster than BA.2 as it is rapidly replacing BA.2.  The fact that the numbers of daily cases have jumped so fast this last week – much faster than we saw even with omicron BA.1, supports this contention of more transmissible virus – though this is all circumstantial evidence.
Regarding severity, it is too early to tell for sure but the one tell-tale sign is that excess deaths have not risen sharply. In the past waves, excess deaths rise early, sometimes even before the rise in cases is evident. With excess deaths remaining low (though there is a slight increase) (Slide 9 red circle), it suggests that it is unlikely that BA.4 is leading to more severe disease.
So, BA.4 seems, at this early stage, to have faster transmissibility and has some mutations that may confer it with immune escape, with early indications suggesting that it may not be leading to many severe cases. But it is still substantially similar to BA.1/BA.2 – so, it is unclear whether it will be able to spread to most people who have immunity from recent BA.1/BA.2 infection or have had 3 doses of vaccine, in order to reach a high peak of a wave. 
It is useful to monitor developments in South Africa as it may provide some indication of what may follow the current declining global omicron wave (Slide 19), especially since the rest of the world followed with waves of omicron a few weeks after it had first hit our country. In next week’s missive, I hope to have many more clues to help us get closer to a definitive answer to the 2 questions posed above. So, watch this space!!
Salim S. Abdool Karim, FRS
Director: CAPRISA
CAPRISA Professor of Global Health: Columbia University