1 May 2021 Episode 135 Modeling, vaccination in real world and variants

Sat, 05/01/2021 - 17:06



Vaccination alone will not end the epidemic.  Continued or intermittent non-pharmacological intervention (NPI, such as physical distance, masks, restriction of gatherings…) AND strict test-trace)isolate policies will have to be maintained.  


Ep 135-1: Alagoz medRxiv: Modeling effects of vaccination and non-pharmacological intervention (NPI) on epidemic in 3 US cities (Dane, Milwaukee and NYC) over next year;

  • Based on COVAM (COVID-19 Agent-based simulation Model), modeling interactions in those cities
  • Endpoint is not “herd immunity”, but “controllable spread date” = when number of new infections is down to a level that contact tracers will be able to track and isolate all contacts of each new infection (e.g. 20 cases in Dane).


An example: In Dane County and Milwaukee:  

  • If 0.1 % of population can be vaccinated each day, starting January 5th
  • If adherence to NPI is 65 %
  • If effectiveness of vaccine  is 90 %

Then 50% of the population is vaccinated and controlled spread could be achieved by July  and August 2021, respectively versus in March 2022 in both regions with NPI alone (without vaccine).


Clearly, adherence to NPI needs to be maintained at high level:   the “controllable spread date” with the same vaccination rate/coverage is calculated May with 70 % adherence, July with 65 % and only Sept with 60 % !!!  (There is also over 50,000 extra cases if adherence drops from 70 to 60 %). 


Clearly, therefore

  • Vaccination may have a larger impact when the history of COVID-19 infections is lower to start with.
  • Relying on vaccination alone to control the spread of COVID-19 delays the timing of controllable disease by several months, and this has important implications for the timing of lifting policies for NPIs that are currently in place.


Limitation: potentially higher transmissibility of new variants not taken into account.


Ep 135-2: Moore Lancet Infect Dis: Modeling of vaccination and NPI in UK over the next year

Central message: Although vaccination of the most vulnerable groups will allow for some relaxation of NPI, this must be done gradually to mitigate large-scale public health consequences.

  • If vaccination uptake is high (75 % in younger adults to 95 % in elderly)
  • If vaccination protects 85 % against infection and 88 % against severe disease
  • If NO NPI

Then R will be 1.58  (See Fig 1) and 21,000 extra COVID deaths.   


Conclusion: Although mass vaccination will inevitably reduce R and disease prevalence, other measures, such as intensive test, trace, and isolate strategies, will be needed

Maintaining low levels of infection is likely to be key to the success of test, trace, and, isolate strategies and in reducing the risk of vaccine escape.


Ep 135-3: Giordani models vaccination and NPI in Italy over one year

  • 4 vaccination speeds are considered: by Jan 2022: 0 – 47 - 64 or 90 % of population.
  • Also 4 levels of restrictions:
    • None with R0 = 1.27;
    • Two types of partial restrictions (with 1 month everything open and the next closed) each with R0 = 1.1 either first open then closed or vice versa
    • Full closures R0 = 0.9  (= eradication scenario)


Fig 1b shows effect on cumulated mortality by Jan 2022:

  • No NPI and no vaccines: almost 300,000 extra deaths, reduced to 89,000 with slow (47%) vaccination, 71,500 with medium (64%) and 50,700 with fast vaccination.
  • With “full-blown NPI” the death toll will be reduced to 30,000 by Jan 2022, but if fast vaccination is added it will be only 18,000.  
  • The scenarios with partial restriction are in-between of course.  Remarkable, if the second type (starting with 1 month “pre-emptive” closure is applied), the results are clearly better: reduction of mortality from 47,000 to 27,000 if the vaccination is slow.


Their conclusions:

  • Stricter restrictions curb transmission more than faster vaccine rollout.
  • NPIs remain crucial to control the epidemic, in part owing to circulation of highly transmissible variants of SARS-CoV-2.
  • Pre-emptive strategies (first Close, then Open at low case numbers) could drastically reduce hospitalizations and deaths, without aggravating socioeconomic costs, as compared to a delayed intervention (first Open, then Close to prevent ICU saturation).


Practical implication:

  1. UK: strong NPIs could bring the epidemic under control before reaching population

immunity, as happened in the UK during January 2021: the highly transmissible B.1.1.7 variant was brought under control by lockdown restrictions kept in place during the first crucial phases of the vaccine rollout campaign. In the meantime, vaccine rollout in the UK has enabled subsequent gradual release of NPIs


  1. Italy (and Europe): to contain the outbreak, driven by new variants, it is important to maintain NPIs, because vaccination alone will be insufficient to control the epidemic.


Ep 135-4 is a long-term modeling exercise for UK by Sandmann also in Lancet Infect Dis, essentially comparing no measures with two vaccination scenario’s and various triggers for applying of NPI on the course of the epidemic over the next 10 years.


Both vaccination scenarios are similar in uptake (75 % initially and annual revaccination), but with different efficacy (50 % and 45 week duration versus 95 % and 3 years).  The triggers for NPI varies from 10 to 100  cases per 100,000.


Interestingly, in the absence of any measure the estimate of total cases (150,000,000) and deaths (3,100,000) over 10 years in UK corresponds the present worldwide numbers (after 1 year of pandemic).


Fig 1 p.5 shows that in the best vaccination outcome scenario, there will be some recurrence in 2022, only if the level for NPI is rather relaxed (trigger at 50-100 per 100,000), but in the worst case scenario, the virus becomes really endemic, with epidemic bouts twice a year (the magnitude of which depends on the NPI threshold).  


The authors conclude that: continued physical distancing might be needed to reduce community transmission until high population-wide coverage is achieved with vaccines that provide long-lasting protection against both disease and infection    


(The paper also contains health economics considerations, which I don’t discuss here)


Ep 135-5: Importance of continued in school testing during the next autumn and winter to uncover “silent infections” in children even if a large proportion of adults is vaccinated.

In A the modeling is proposed:

With an reproduction number Re = 1.2 and vaccination in adults is 40 % with 95 % efficacy:

In order to bring attack rate below 5 %

  • A targeted approach among children to identify 11% of silent infections within 2 days and 14% within 3 days is needed

If silent infections among children remained undetected, achieving the same attack rates would require an unrealistically high vaccination coverage (81%) of this age group, in addition to 40% vaccination coverage of adults


The comment in B clarifies that with the expected moderate vaccine coverage in adults and delayed vaccination of children (to 2022?), it is worthwhile to perform very regular testing at school, also in asymptomatic children in order to prevent schools to become breeding grounds of infection that cause important transmission in the community that could lead to new restrictions, including school closures  





Ep 135-6: Rossman Nat Med: 


End of February (2 months after start of vaccination campaign):  85% of >  60 years vaccinated with two doses of Pfizer: compared to peak values

  • 45% drop in positive test
  • 77% drop in cases
  • 68% drop in hospitalizations


Decrease in COVID-19 cases and hospitalizations follows order of vaccination prioritization:

  • in individuals older than 60 years, followed by younger age groups
  • more pronounced in early-vaccinated cities compared to late-vaccinated cities.


Although a third national lockdown was implemented concurrently with the vaccination campaign, these profound effects were not observed in the previous national lockdown.


  1. UK


Ep 135-7: Vasileiou in Lancet: Older adults (mean 65yrs)  in Scotland:


Decrease of hospital admissions one month after 1st dose

  • Pfizer: 91 %
  • Astra-Zeneca: 88 %
  • Similar in 80 + yrs : 83 %  


Ep 135-8: Jones ELife:  Fourfold reduction of asymptomatic infections > 2 weeks after single dose of Pfizer in health care workers (HCW).


Ep 145-9: Hall Lancet: Large HCW cohort vaccinated with Pfizer: protection against both symptomatic and asymptomatic infection in not previous infected subjects:

  • 70 % from 21 days after 1st dose
  • 90 % from 7 days after 2nd dose


One third was already seropositive (had been infected). The incidence of infection was similar to the fully vaccinated originally seronegatives and there was no significant further decrease with vaccination.


  1. United States


Ep 135-10:  Thompson in MMWR: Pfizer or Moderna effectiveness against infection in HCW and other frontline workers

  • full immunization (≥14 days after second dose) was 90%
  • partial immunization (≥14 days after first dose but before second dose) was 80%.


Ep 135-11: Teran MMWR 30 April: Breakthroughs after Pfizer or Moderna vaccines: 647 out of > 13,000 vaccinated staff and residents in skilled nursing facilities around Chicago. Of these 447 (71 %) in unvaccinated; 145 (22 %) after 1 dose and before 2nd; 13 (2 %) after within 14 days after 2nd dose and 22 (3.5 %) in fully vaccinated subjects. 

These breakthroughs were less symptomatic in non-vaccinated subjects; but a similar proportion was hospitalized (20 %) and died (4 %) than in unvaccinated subjects.   


Ep 135-12: Jacobson in medRxiv 24 April:   Breakthroughs after Pfizer or Moderna vaccines: 189 out of > 22,700 HCW in Acalifornia.  Of these 60 % between two doses; 26 % within 2 weeks after 2nd dose and only 14 % after full vaccination.  There was a non-significant trend for the partly and fully vaccinated to be infected with the California, variant, carrying L452R.


HOW DANGEROUS ARE THE MUTANTS to undermine vaccines?


I will be brief, since the topic has been treated extensively in previous episodes.


Ep 135-13 and -14: very nice recent reviews, discussing the variants until April.


Ep 135-15 is on the Californian variant


Ep 135-16 and -17 on the new Indian variant, which behaves very aggressively and has “overgrown” all other variants in no time….


Financial Times India’s Covid wards are like scenes from Dante’s ‘Inferno’ https://www.ft.com/content/ad200d93-3247-409a-8afb-482234b4655c


At present, it is not very clear to what extent those variants will escape neutralization by the present generation of vaccines. But there is reason for concern, especially with the E484K/Q South-African, Indian and Brazilian variants:  


  • In vitro neutralization tests with sera from (mostly from mRNA) vaccinated subjects generally show reduced sensitivity of variants, but only about one log: hence variants are still sensitive to neutralization.
  • Several vaccination trials (e.g. adenoviral J&J and Astra-Zeneca, inactivated Coronavac and Novavax S protein) showed a lower protection in the South-African and/or Brazilian arm as compared to the UK or US arms.
  • There is not enough in vitro and in vivo data to conclude whether or not the vaccine platform matters with regard to protection against variants:
    • Are mRNA vaccine, based on the non-variant virus, intrinsically  more protective against the variants?  Possibly so, but not really clear from published data.
    • Results of Astra-Zeneca in the South-African trial was, however, clearly worse than J&J and Novavax (but the trials were not done in parallel).
  • An encouraging observation is that convalescent sera from individuals who recovered from the South-African variant showed a broad “cross-neutralizing” activity, including against the British variant.  This should certainly also be checked for the Brazilian and Indian variants   


At present, the tentative and provisional conclusion is that most of the present generation of vaccines (except Astra-Zeneca?) will offer some protection against the variants, but that this protection will wane when neutralizing antibody titers decrease. 




  1. Real world data are very encouraging, but breakthrough infections do occur
  2. It is likely and in fact evident that the South-African, Brazilian and Indian variant are less sensitive to the present generation of vaccines and therefore (descendants of) these variants are likely to re-emerge in vaccinated subjects, when their antibody levels wane.  This will happen  first in elderly and subjects with severe disease and immune suppression.
  3. Independent modeling in UK / US / Italy predicts that, besides mass vaccination, testing-tracing-isolation and period NPI will be needed to contain the epidemic.  Especially the re-installation of NPI will be very difficult to be accepted.
  4. Obviously, vaccination of adults in high income countries only will not suffice.  Roll-out over the world and in children will be needed.  


Best wishes,





More news