key: cord-0610584-anmt0pnu authors: Barletta, William A. title: The influence of SARS-CoV-2 variants on national case fatality rates date: 2021-08-15 journal: nan DOI: nan sha: d95025354296a923ffeb1f742069165637c916fb doc_id: 610584 cord_uid: anmt0pnu Background: During 2021 several new variants of the SARS-CoV-2 virus appeared with both increased levels of transmissibility and virulence with respect to the original wild variant. The Delta (B.1.617.2) variation, first seen in India, dominates COVID-19 infections in several large countries including the United States and India. Most recently, the Lambda variant of interest with increased resistance to vaccines has spread through much of South America. Objective: This research explores the degree to which new variants of concern 1) generate spikes and waves of fluctuations in the daily case fatality rates (CFR) across countries in several regions in the face of increasing levels of vaccination of national populations and 2) may increase the vulnerability of persons with certain comorbidities. Methods: This study uses new, openly available, epidemiological statistics reported to the relevant national and international authorities for countries across the Americas, Europe, Africa, Asia and the Middle East. Daily CFRs and correlations of fatal COVID-19 infections with potential cofactors are computed for the first half of 2021 that has been dominated by the wide spread of several variants of concern as denoted by the World Health Organization. Results: The analysis yields a new quantitative measure of the temporal dynamics of mortality due to SARS-CoV-2 infections in the form of variations of a proxy case fatality rate compared on a country to-country basis in the same region. It also finds minimal variation of correlation between the cofactors based on WHO data and on the average apparent case fatality rate. The period from mid-November, 2020 to the present has seen the rapid spread of several "variants of concern" of the SARS-CoV-2 virus [1] that have appeared across many highly populated nations with both increased levels of transmissibility and virulence. Most recently the Delta (B.1.617.2) variation has become the dominant cause of COVID-whether the new variants increase the susceptibility for severe consequences for people with common comorbidities. Reference [2] examined that question with respect to the original stain of the virus using a cutoff date of statistics of 30 December 2020. Since then the number of reported cases of COVID-19 has risen from 82.9 million to 193. 4 Over that same period the number of deaths rose from 1.81 million to 4.15 million. [3, 4] During 2021, more than 3.7 billion people have received at least one dose of an anti-COVID vaccine and more that 1 billion are considered fully vaccinated. [5] The starting point of the analysis is an examination of the time behavior of the pandemic averaged CFR. Figure 1 that shows the pandemic averaged CFR for the period from November 1 to the cutoff date of this study (July 18, 2021) when variants of concern in addition to Alpha were widespread. hypothesis. The fluctuations in the average rates for Australia, Japan, Korea, and the U.S. are not readily explainable from these pandemic averaged data. As the pandemic averaged data are suggestive but far from dispositive, further analysis requires one to introduce a proxy measure of CFR that is more sensitive to temporal variations in virulence but is far less sensitive to the irregularities in the timing of government reports of fatalities ascribed to COVID-19 infections, an example of which appears in Figure 2 . The manifest irregularities in reporting with far fewer reports on the weekends and the statistical noise necessitate introducing a proxy for estimating the daily case fatality rate. Using an appropriate proxy rate, one can explore whether the daily CFR in several countries shows evidence of more virulent variants taking hold and/or whether a vigorous program of COVID-19 testing and vaccination decreases the mortality rate of the disease. A second question is whether these new variants of concern have a different susceptibility to co-morbidities and economic factors that did the original wild variant of the SARS-CoV-2 virus. To answer this question one can analyze the correlations of potential contributing co-factors during the period of January1, 2021 to July 18, the data cutoff for this research, over the same set of countries studied in [2] . To explore the influence of variants of concern one introduces a credible proxy for daily case fatality rates, pCFR, which will be sensitive to the spread of variant throughout a country. The definition of pCFR and the subsequent analysis of it time distribution on a country-by-country basis are presented in Sections 4 and 4.1 respectively. To evaluate any changes in the susceptibility to co-factors one can follow the method of reference [2] , in which the input data are based only on national epidemiological statistics reported to the relevant national and international authorities. For consistency of analysis, this study considers the same sample of countries as reference [2] . The regions and their constituent countries are given in the Appendix. The analysis of this paper evaluates 1) the linear correlation between a proxy CFR for country pairs and 2) the linear correlation of the pandemic averaged CFR and potential cofactors using the usual Pearson "product moment correlation" of Equation (1) . The resulting values of the correlations show minimal differences between those with a data cutoff of December 30, 2020 and the data with a cutoff of July19, 2021. The one exception is the jump by a factor >2 in the correlation between the apparent CFR and number of COVID-19 cases per 1 million persons. That jump persons may be due to the increased virulence of the variants of the concern that have spread during 2021 especially in countries without robust vaccination programs. Doubtless they are also affected by the pervasiveness of national vaccination programs. To explore this question and compare experience in several countries one can look at the time series of levels of daily case fatality rates. Unfortunately, as shown in Figure 2 , actual data will be very noisy, subject to uneven reporting of both new cases and deaths attributed to COVID-19 as well as to significant statistical variations in the daily data. To mitigate these deficiencies in medically definitive data, one can introduces pCFR as a plausible proxy for the apparent daily case fatality ratio. The model then overlays those data with a rolling 14-day average of the results to suggest the actual temporal trends in (2) One could equally well average the infection rate over the period from 28 to 14 days. only weak evidence for differential rates of mortality in February and March 2021 due to the B.1.117 variant even though that variant appeared in the U.S. in January 2021 [7] , before any noticeable fraction of the population had received vaccinations. In Ref. [7] , the CDC had predicted a peaking of infections due to B.1.117 in March, 2021. That might account for the slight increase in pCFR seen in March in the U.S. An objection to the initial analysis of the previous section, is that the pandemic averaged case fatality ratios are dominated by the very high mortalities at the outset of the pandemic before adequate and appropriate isolation of the infected and modalities of treatment were understood. To mitigate that objection, one can focus only on the period of November 2020 through July 2021 that is dominated by surges of variants of concern that were described at the time of the WHO's designation to have higher transmissibility and perhaps higher virulence than the original wild strain of SARS-CoV-2. The statistics with a cutoff of July 18, 2021 yield Figure 10 , Peru itself saw a strong spike in pCFR in January and February, 2021, reaching 20%. Since then pCFR has dropped gradually to roughly 5%. By July, 2021 the rate of full vaccination of the Peruvian population had reached only slightly more than 10%. [4] In addition, in analysis of the C.37 virus, Kimura et al. [14] identified a spike structure that accounts for higher resistance to vaccine induced immunity than The South American scene is complicated by the simultaneous circulation of multiple variants of concern, particularly in its most populous country, Brazil. Figure 11 provides an example for Columbia. Figure 11 . The proxy CFR, pCFR, for Columbia since November, 2020 based on the smoothed data of Ref. [4] . The dark line is a 7-day rolling average. Comparison between plots of the temporal behavior of the proxy CFR (Figures 2 -4 and In contrast with the case of Argentina, the light purple bar over the PER/COL pairing shows that shifting the Peruvian profile 60 days later in time, introduces a much larger similarity with the Columbian profile. Also, the correlation of new cases shifts to 0.68. One may interpret these results as indicating the amount of time needed for the Lambda variant to spread widely into Columbia, where its prevalence is now high. [14] For North America one observes only a moderate correlation in pCFR between the US and Canada. That low value may be explained differences in the US and Canadian The proxy for the daily case fatality rate, pCFR, computed over a smoothed distribution of deaths attributed to COVID-19 infections provides a useful metric to track the national dynamics of the spread of SARS-CoV-2 variants of concern overlaid with that With respect to the spread of the B1.117 variant the temporal profile of pCFR for the US is too variable and the data set is too noisy to confirm the timing of the spread of that variant through the US. In Germany, however, the temporal distribution of the proxy CFR of Figure 5 shows a distinct peak with a timing that coincides with the spread of B1.117 through Europe. The apparent increases in the virulence of the variants of concern might be due to increased susceptibility to severe infection in persons with certain comorbidities. Following the method of Ref. [2] for a data set limited to November, 2020 through mid-July, 2021, one finds minimal quantitative differences with the conclusion of Ref. [2] that most commonly cited co-morbidities do not in and of themselves increase the risk of serious (and possibly fatal) consequences of SARS-CoV-2 infections. However, one cannot ignore that many persons with such conditions often suffer from multiple cofactors and have a depressed level of immune function that can worsen the effects of a COVID-19 infection. The prevalence of multiple co-existing conditions may also vary from region to regions, partially explaining the regional variations seen in Figure 7 . The author acknowledges his colleagues in the World Federation of Scientists for their encouragement to continue, expand, and report this research. The author's work has been completely self-supported, without any outside funding or other material support. None declared. SARS-CoV-2 Variant Classifications and Definitions Risk Factors of SARS-CoV-2 Infections: A Global Epidemiological Study Our World in Data COVID-19 Dataset Coronavirus (COVID-19) Vaccinations World Health Rankings Emergence of SARS-CoV-2 B.1.1.7 Lineage -United States Vaccinations in the United Kingdom, UK government dashboard The Delta Variant: What Scientists Know A coronavirus variant first found in India is now officially a 'variant of concern,' the W.H.O. said Effectiveness of Covid-19 Vaccines against the B.1.617.2 (Delta) Variant Infectivity and immune escape of the new SARS-CoV-2 variant of interest Lambda SARS-CoV-2 Lambda variant exhibits higher infectivity and immune resistance Argentina reports first local Delta variant cases in Buenos Aires