Early assessment of the clinical severity of the SARS-CoV-2 omicron variant in South Africa: a data linkage study

Summary Background The SARS-CoV-2 omicron variant of concern was identified in South Africa in November, 2021, and was associated with an increase in COVID-19 cases. We aimed to assess the clinical severity of infections with the omicron variant using S gene target failure (SGTF) on the Thermo Fisher Scientific TaqPath COVID-19 PCR test as a proxy. Methods We did data linkages for national, South African COVID-19 case data, SARS-CoV-2 laboratory test data, SARS-CoV-2 genome data, and COVID-19 hospital admissions data. For individuals diagnosed with COVID-19 via TaqPath PCR tests, infections were designated as either SGTF or non-SGTF. The delta variant was identified by genome sequencing. Using multivariable logistic regression models, we assessed disease severity and hospitalisations by comparing individuals with SGTF versus non-SGTF infections diagnosed between Oct 1 and Nov 30, 2021, and we further assessed disease severity by comparing SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, with delta variant-infected individuals diagnosed between April 1 and Nov 9, 2021. Findings From Oct 1 (week 39), 2021, to Dec 6 (week 49), 2021, 161 328 cases of COVID-19 were reported in South Africa. 38 282 people were diagnosed via TaqPath PCR tests and 29 721 SGTF infections and 1412 non-SGTF infections were identified. The proportion of SGTF infections increased from two (3·2%) of 63 in week 39 to 21 978 (97·9%) of 22 455 in week 48. After controlling for factors associated with hospitalisation, individuals with SGTF infections had significantly lower odds of admission than did those with non-SGTF infections (256 [2·4%] of 10 547 vs 121 [12·8%] of 948; adjusted odds ratio [aOR] 0·2, 95% CI 0·1–0·3). After controlling for factors associated with disease severity, the odds of severe disease were similar between hospitalised individuals with SGTF versus non-SGTF infections (42 [21%] of 204 vs 45 [40%] of 113; aOR 0·7, 95% CI 0·3–1·4). Compared with individuals with earlier delta variant infections, SGTF-infected individuals had a significantly lower odds of severe disease (496 [62·5%] of 793 vs 57 [23·4%] of 244; aOR 0·3, 95% CI 0·2–0·5), after controlling for factors associated with disease severity. Interpretation Our early analyses suggest a significantly reduced odds of hospitalisation among individuals with SGTF versus non-SGTF infections diagnosed during the same time period. SGTF-infected individuals had a significantly reduced odds of severe disease compared with individuals infected earlier with the delta variant. Some of this reduced severity is probably a result of previous immunity. Funding The South African Medical Research Council, the South African National Department of Health, US Centers for Disease Control and Prevention, the African Society of Laboratory Medicine, Africa Centers for Disease Control and Prevention, the Bill & Melinda Gates Foundation, the Wellcome Trust, and the Fleming Fund.


Introduction
Since the first case of SARS-CoV-2 in March, 2020, South Africa has had three epidemic waves, with the beta (B.1.351) and delta (B.1.617.2) variants of concern dominating the second and third waves, respectively. On Nov 24, 2021, the Network for Genomics Surveillance in South Africa (NGS-SA) reported a new variant of SARS-CoV-2, which had been detected from specimens collected on Nov 14, 2021, 3 the fifth variant of concern. Concomitantly, there was a rapid increase in COVID-19 cases in Gauteng province, 2 South Africa, from the week beginning Nov 15, 2021, with SARS-CoV-2 testing laboratories reporting an increase in the number of samples with S gene target failure (SGTF) when tested on the TaqPath COVID-19 PCR test (Thermo Fisher Scientific; Waltham, MA, USA). Subsequently, increases in COVID-19 cases and samples with SGTF were observed in other provinces in South Africa, precipitating

Data sources and linkage
In this data linkage study, we linked national, South African, individual-level data from four sources: (1) national COVID-19 case data for the period Jan 1-Dec 6, 2021, reported in real time to the National Institute for Communicable Diseases (NICD) Notifiable Medical Conditions Surveillance System (NMCSS); (2) SARS-CoV-2 laboratory test data (test used and PCR cycle threshold values) for the period Oct 1-Dec 6, 2021, for public sector laboratories (the National Health Laboratory Service reported all test data) and one large private sector laboratory (reported TaqPath PCR tests only); (3) SARS-CoV-2 genome data for the period March 31, 2020-Dec 6, 2021, for clinical specimens sent to the NICD from private and public diagnostic laboratories (predominantly from the Gauteng, North West, Mpumalanga, and Northern Cape provinces) and collected through the pneumonia surveillance programme 7 in five provinces (Western Cape, KwaZulu-Natal, North West, Gauteng, and Mpumalanga); and (4) DATCOV (between March 31, 2020, and Dec 21, 2021), which is an active surveillance system for COVID-19 hospital admissions, with comprehensive coverage of all hospitals in South Africa. 8 The case and hospitalisation data we used are from the aforementioned national, comprehensive, populationbased datasets including the whole of South Africa. The

Research in context
Evidence before this study On Nov 24, 2021, the Network for Genomics Surveillance in South Africa reported a new variant of SARS-CoV-2 (B.1.1.529), which had been detected from specimens collected on Nov 14, 2021, in South Africa and was later designated by WHO as the omicron variant of concern. The omicron variant of concern has a high number of mutations, some of which are concerning due to predicted immune evasion and increased infectivity. Data on the clinical severity of the omicron variant compared with previous SARS-CoV-2 variants are needed to guide public health planning and response. Using the search terms "SARS-CoV-2", "Omicron", and "severity", we searched PubMed and Google Scholar for studies published in English between Jan 1, 2020, and Dec 22, 2021, assessing how the clinical severity of infection with the omicron variant compares to the severity of infection with other SARS-CoV-2 variants. We identified a single published study from Denmark that comprised 785 participants infected by the omicron variant of SARS-CoV-2, of whom nine were hospitalised, but included an insufficient number of people to robustly evaluate disease severity.

Added value of this study
We assessed the clinical severity of infection with the SARS-CoV-2 omicron variant using S gene target failure (SGTF) on the Thermo Fisher Scientific TaqPath COVID-19 PCR test as a proxy, COVID-19 national case and admission data, SARS-CoV-2 laboratory testing data, and SARS-CoV-2 genomic sequence data. We found that individuals with SGTF versus non-SGTF infections had an 80% lower odds of being admitted to hospital, but we were not able to make any firm conclusions on the risk of severe disease among hospitalised individuals, possibly due to the low numbers of individuals included in this analysis. When compared with delta variant infections from the third wave, SGTF infections were associated with a 70% lower odds of severe disease among hospitalised individuals.

Implications of all the available evidence
Our early analyses indicate a significantly reduced odds of hospital admission among individuals with SGTF versus non-SGTF infections diagnosed between Oct 1 and Nov 30, 2021, and a significantly reduced odds of severe disease among SGTF-infected individuals compared with individuals infected earlier with the delta variant in South Africa. Immunity (due to previous infection, vaccination, or both) in individuals infected with the omicron variant could, in part, account for this reduced severity. It is unclear as to what extent the omicron variant might have reduced intrinsic virulence, which could have contributed in part to the reduced severity observed in our study. Data from other settings on the severity of the omicron variant will be important to assess the behaviour of the omicron variant in countries with different levels of previous infection and vaccination.
national COVID-19 case database (NICD NMCSS) is a laboratory-based surveillance programme that receives real-time electronic data on all laboratory-confirmed SARS-CoV-2 cases in South Africa. Reporting of cases is mandated by law and audited. Each case is investigated by public health officials and identified data errors are corrected on the database. The SARS-CoV-2 laboratory test data used in this analysis are national data. The included laboratories (all public sector laboratories and one private sector laboratory) were chosen as these laboratories were able to provide data on the PCR test method and cycle threshold values, and represented more than half of all SARS-CoV-2 tests done in South Africa between Oct 1 and Dec 6, 2021. The DATCOV surveillance programme has been described in detail previously. 8 Using DATCOV, we have included data on all hospi talised patients with laboratory-confirmed SARS-CoV-2 throughout South Africa. In public sector hospitals, data are abstracted from patient records and directly captured by dedicated data capturers, and, in private hospitals and the Western Cape province public sector, data are transferred electronically in daily bulk extracts from hospital data systems. DATCOV includes patients admitted to hospital for COVID-19 symptoms, patients with acquired nosocomial SARS-CoV-2 infection, and those who have tested positive incidentally when admitted to hospital for other reasons. Case and test data were obtained on Dec 6, 2021, and DATCOV data were obtained on Dec 21, 2021. For hospitalisation and severity analyses, cases were censored to those with a specimen collected before Dec 1, 2021.
We performed individual-level data linkage using unique identifiers, when they existed; otherwise, we used a combination of name, surname (family name), and date of birth. Following data linkage, a combination of automated and manual checks were used to ensure accuracy. Ethical approval was obtained from the Human Research Ethics Committee (Medical) of the University of the Witwatersrand for the collection of COVID-19 case and test data as part of essential communicable disease surveillance (M210752), and for the DATCOV surveillance programme (M2010108).

Definitions
Infections were classified as SGTF (as a proxy for the omicron variant) when an individual tested positive on the TaqPath COVID-19 PCR test with a non-detectable S gene target and a cycle threshold of 30 or less for either the ORF1ab or nucleocapsid gene targets. 9 We restricted to a cycle threshold of 30 or less to avoid incorrectly classifying infections for which the S gene was not detected (because of low viral load) as SGTF (ie, high cycle threshold values). Infections were classified as non-SGTF when an individual tested positive on the TaqPath COVID-19 PCR test with a cycle threshold of 30 or less for either the ORF1ab or nucleocapsid gene targets and had detectable S gene target. The delta variant was identified by genome sequencing.
An individual was classified as having been admitted to hospital if they linked to a case on the DATCOV database with an admission date from 7 days before to 21 days following the date of specimen collection. We included admissions with a test date before admission because clinicians might not suspect SARS-CoV-2 infection at the time of admission, initial tests might be negative, or infection might be nosocomially acquired. Severe disease was defined as a hospitalised patient meeting at least one of the following criteria: admitted to an intensive care unit (ICU); received any level of oxygen treatment; was ventilated; received extracorporeal membrane oxygenation; had acute respiratory distress syndrome; or had died. Comorbidity was defined as the presence of at least one of the following conditions: hypertension, diabetes, chronic cardiac disease, chronic kidney disease, asthma, chronic obstructive pulmonary disease, malignancy, HIV, and active or previous tuberculosis. Data on obesity and pregnancy were not included as these data were not systematically collected. Reinfection was defined by an individual having at least one positive COVID-19 test (either PCR-based or antigen-based) more than 90 days before the current positive COVID-19 test. Vaccination was defined as having at least one dose of a SARS-CoV-2 vaccine (Ad.26.COV2.S [Johnson & Johnson] or BNT162b2 [Pfizer-BioNTech]).

Statistical analysis
The mean cycle threshold value (a proxy for viral load) for all public sector positive COVID-19 PCR tests (any PCR test) during the early wave of omicron (weeks 46-48; Nov 14-Dec 4, 2021) was compared with that of the early wave of delta (weeks 18-20; May 2-22, 2021) using the Student's t test. Early wave periods were defined from the week before the country crossed a weekly incidence of 30 cases per 100 000 people to 2 weeks later. 10  PCR test included one or more gene target, the target with the lowest cycle threshold value was used for the analysis. We described the epidemiological characteristics of SGTF infections by comparing SGTF-infected individuals with non-SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, hospitalised SGTF-infected individuals with hospitalised non-SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, and hospitalised SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, with hospitalised delta variantinfected individuals diagnosed between April 1 and Nov 9, 2021, using multivariable logistic regression analysis.
The severity of infections with the SARS-CoV-2 omicron variant was assessed in two ways: (1) by comparing individuals with SGTF versus non-SGTF infections diagnosed between Oct 1 and Nov 30, 2021, and (2) by comparing individuals with delta variant infections diagnosed between April 1 and Nov 9, 2021, with SGTF-infected individuals who were diagnosed between Oct 1 and Nov 30, 2021. Hospitalisation and severity data were obtained for cases on Dec 21, 2021, to allow for at least 3 weeks of follow-up for admission to hospital and in-hospital outcomes. Among people with COVID-19 in the DATCOV database hospitalised between March 5, 2020, and Dec 18, 2021, the median time from admission to an in-hospital outcome was 6 days (IQR 3-11 days; n=414 149) overall, 5 days (2-11; n=94 938) for individuals who died, and 6 days (3-10; n=304 845) for individuals who were discharged alive. Therefore, we were not likely to miss many outcomes with a mininum follow-up of 3 weeks. Severity analyses were restricted to admissions that had already accumulated outcomes and all patients still in hospital were excluded because, in the DATCOV programme, detailed information on hospitalisation in public sector hospitals (eg, oxygen treatment or ICU admission) is populated at the time of discharge or death.  Two models were generated to assess hospitalisation and severe disease among hospitalised individuals as outcome variables. Multivariable logistic regression was done to evaluate the association of SGTF infection versus non-SGTF infection with hospitalisation among people with COVID-19 diagnosed between Oct 1 and Nov 30, 2021. We controlled for factors known to be associated with hospitalisation (ie, age, sex, province, and health-care sector [private vs public]) and adjusted for known previous SARS-CoV-2 infection. Data on vaccination status and comorbidities were only available for hospitalised individuals; therefore, we could not control for these variables in this analysis. Multivariable logistic regression was done to evaluate the association of SGTF infection versus non-SGTF infection with disease severity among hospitalised individuals diagnosed with COVID-19 between Oct 1 and Nov 30, 2021, restricting to individuals with a known in-hospital outcome on Dec 21, 2021 (excluding those still in hospital). We controlled for factors known to be associated with disease severity (ie, age, the presence of comorbidity, sex, province, and health-care sector) and adjusted for the number of days between the date of specimen collection and the date of hospital admission, known previous SARS-CoV-2 infection, and SARS-CoV-2 vaccination status.
We compared delta variant infections diagnosed between April 1 and Nov 9, 2021, with SGTF infections diagnosed between Oct 1 and Nov 30, 2021. Multivariable logistic regression was done to evaluate the association of SGTF infections versus delta variant infections with disease severity among hospitalised individuals, restricting to individuals with a known in-hospital outcome on Dec 21, 2021 (excluding those still in hospital). We controlled for factors associated with disease severity (ie, age, the presence of comorbidity, sex, province, and health-care sector) and adjusted for the number of days between specimen collection and hospital admission, known previous SARS-CoV-2 infection, and SARS-CoV-2 vaccination status. We did not compare the risk of hospital admission between individuals with delta variant infections and those with SGTF infections because sequenced samples for the delta variant were not population-based and represented only a small subset of delta variant infections that were sent to the NICD for sequencing.
Categorical variables were summarised by use of frequency distributions and compared by use of Pearson's χ² test. Pairwise interactions were assessed by the inclusion of product terms for all variables remaining in the final multivariable additive models. Analyses were done by use of Stata, version 14.1. p values less than 0·05 were considered statistically significant.

Role of the funding source
The funders of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the report.

Results
From Oct 1 (week 39), 2021, to Dec 6 (week 49), 2021, 161 328 cases of COVID-19 were reported nationally, of which 104 529 (64·8%) were from laboratories included in our analysis (appendix pp 6-7  We compared the characteristics of all individuals with SGTF infections diagnosed between Oct 1 and Nov 30, 2021, with those of individuals diagnosed with non-SGTF infections during the same period. In multivariable analyses, in addition to the proportions of SGTF versus non-SGTF infections varying significantly by geographical province, individuals with SGTF infections were more likely to be aged 5-59 years (vs ≥60 years) and be diagnosed by the private sector (vs the public sector) and were less likely to be admitted to hospital than were individuals with non-SGTF infections (table 1). Similar results were found when comparing characteristics between hospitalised SGTFinfected individuals and hospitalised non-SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021 (appendix pp 1-3) and when comparing hospitalised SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, with hospitalised delta variant-infected individuals diagnosed between April 1 and Nov 9, 2021 (appendix pp [3][4][5]. Of the 10 547 individuals with SGTF infections diagnosed between Oct 1 and Nov 30, 2021, 256 (2·4%) were admitted to hospital, compared with 121 (12·8%) of 948 individuals with non-SGTF infections (p<0·001; table 2). In multivariable analyses, after controlling for factors associated with hospitalisation, individuals with SGTF infections had significantly lower odds of being admitted to hospital than did individuals with non-SGTF infections (adjusted odds ratio [aOR] 0·2, 95% CI 0·1-0·3; table 2). In addition to variation by geographical province, hospital admission was significantly associated with young age (<5 years; aOR 9·3) and older age (≥60 years; aOR 3·1) compared with individuals aged 19-24 years, and female sex (aOR 1·3) versus male sex (  sector were less likely to be admitted to hospital (aOR 0·8; table 2). Among the 382 hospitalised individuals diagnosed with COVID-19 between Oct 1 and Nov 30, 2021, 317 (83%) had accumulated an in-hospital outcome by Dec 21, 2021. After controlling for factors associated with severe disease, the point estimate for the odds of severe disease in individuals with SGTF versus non-SGTF infections was less than 1, but the 95% CI was wide (aOR 0·7, 95% CI 0·3-1·4; table 3). The odds of severe disease varied geographically and were significantly higher among individuals aged 60 years or older (aOR 11·5) than among individuals aged 19-24 years (table 3).
Among 1142 hospitalised individuals, 1037 (90·8%) had accumulated an in-hospital outcome by Dec 21, 2021. After controlling for factors associated with severe disease, individuals with SGTF infections diagnosed between Oct 1 and Nov 30, 2021, had a significantly lower odds of severe disease than did those with delta variant infections diagnosed between April 1 and Nov 9, 2021, (aOR 0·3, 95% CI 0·2-0·5; table 4). In addition to variation by geographical province, older age (40-59 years and ≥60 years vs 19-24 years) and having a comorbid condition were associated with an increased odds of severe disease (table 4). Individuals aged 13-18 years (vs 19-24 years) and women (vs men) had lower odds of severe disease (table 4).

Discussion
In November, 2021, the omicron SARS-CoV-2 variant emerged in South Africa, concurrent with a rapid increase in the number of reported COVID-19 cases, 2 leading to the fourth epidemic wave. Although numerous studies are underway to rapidly assess phenotypic behaviour and immune escape, it is also crucial to understand the clinical severity of infections with the omicron variant. We evaluated the severity of SGTF infections (a proxy for the omicron variant) by comparing them to non-SGTF infections diagnosed during the same period and to infections with the delta variant, which dominated the third wave in South Africa. Individuals with SGTF versus non-SGTF infections had an 80% lower odds of being admitted to hospital, but we were not able to make any firm conclusions on the risk of severe disease among hospitalised people, possibly due to the low numbers of individuals included in this analysis. When compared with delta variant infections, SGTF infections were associated with a 70% lower odds of severe disease. In addition, we found that the mean PCR cycle threshold was lower during the early wave of omicron than during the early wave of delta, which might reflect higher viral loads in individuals infected with the omicron variant versus the delta variant.

Figure 2: Number of SARS-CoV-2 delta variant and SGTF samples among hospitalised patients with COVID-19 and a known outcome in South Africa by epidemiological week and variant type
Week 13 represents April 1, 2021, and week 49 represents Dec 6, 2021. DATCOV-Gen linked real-time SARS-CoV-2 test and genome data to detailed epidemiological and clinical data on hospitalised cases. SGTF=S gene target failure.
Data from the DATCOV surveillance programme have shown a higher proportion of admissions among people younger than 20 years during the early fourth wave compared with the third wave. 11 In our study, we also found that SGTF-infected individuals were younger than those with non-SGTF or delta infections; this effect was not restricted to children as there were also more SGTF infections (vs non-SGTF or delta) in younger adult age groups (eg, 19-24 years) than in the older adult age group (≥60 years). This finding could be related to increased immunity in older adults as a result of higher rates of previous infection 12 or vaccination. 13 Our findings correlate with hospitalisation data showing that, among 1061 patients in South Africa with COVID-19 and a known hospital outcome, 324 (30·5%) had severe disease during the early fourth wave, compared with 943 (67·0%) of 1408 patients during the early third wave. 11 By November, 2021, a high proportion of the South African population had some level of SARS-CoV-2 immunity as a result of previous natural infection, vaccination, or both. After the third SARS-CoV-2 wave, 60-70% of individuals in South Africa had evidence of previous SARS-CoV-2 infection. 14 By Dec 9, 2021, 58% of South African individuals aged 60 years or older, 55% aged 50-59 years, 43% aged 35-49 years, and 24% aged 18-34 years were fully vaccinated against SARS-CoV-2 (one dose of Ad.26.

COV2.S [Johnson & Johnson] or two doses of BNT162b2
[Pfizer-BioNTech]). 13 It has been suggested that the omicron variant has substantially increased immune escape capabilities compared with other variants. 15 This finding, together with the high levels of pre-existing immunity in South Africa at the time of emergence of the omicron variant, means that a high proportion of omicron infections are likely to be reinfections and that this proportion might be greater for omicron than for contemporaneous infections with other variants. In South Africa, during the first and second waves of COVID-19, fewer than one in ten cases of SARS-CoV-2 were diagnosed. 16 We accounted for individuals with diagnosed and previously reported infections in our analysis, but, in many cases, it is probable that previous infection could have been undiagnosed. If reinfections are less severe than primary infections, then this fact could, in part, account for the reduced disease severity observed in individuals infected by the omicron variant. 17 It is also possible that the omicron variant has reduced intrinsic virulence compared with other variants. 18,19 With the available data, it is not possible to disentangle the relative contributions of high levels of population immunity versus lower intrinsic virulence to the observed lower disease severity of the omicron versus delta variant. Incomplete vaccination data, and the fact that many previous infections were probably not detected, resulted in incomplete adjustment for the effect of previous immunity in our analyses.   20 Although genome data for non-SGTF infections were not available, of samples sequenced in South Africa, 617 (84·2%) of 733 were delta and one (0·1%) of 733 was omicron in October, 2021, and 148 (13·7%) of 1082 were delta and 881 (81·4%) of 1082 were omicron in November, 2021, suggesting that the majority of non-SGTF infections were delta. 20 Third, our analysis was done during the early fourth wave, in which numbers of people with COVID-19 were small, patients with milder symptoms were more likely to be admitted to hospital (compared with peak wave periods), and there could have been a lag in hospitalisations and severe disease outcomes caused by the omicron variant. We accounted for this limitation by only including hospitalised patients with known outcomes, censoring cases to ensure at least 3 weeks of follow-up, and adjusting for the time interval between diagnosis and hospitalisation in the severity models. The inclusion of individuals only with accu mulated in-hospital outcomes could have biased individuals with SGTF infections towards shorter hospital stays; hence, this result should be interpreted with caution. Fourth, we compared individuals with COVID-19 during the full delta wave to those with COVID-19 during the ascending phase of the omicron wave, which could have biased comparisons if case characteristics differ between ascending and descending wave phases or if the threshold for hospitalisation changed in the different time periods. Data from the DATCOV programme suggest that the proportion of patients with severe COVID-19 does not vary substantially through the different wave phases. 11 Finally, we only had information on vaccination status for hospitalised individuals with COVID-19, which was also based on self-reporting. Despite the omicron variant being detected only recently, there has been a rapid increase in the prevalence of SGTF infections across all South African provinces during November, 2021, with almost complete replacement of the delta variant in a period of 4 weeks. Our early analyses indicate a significantly reduced risk of hospital admission among SGTF-infected individuals compared with non-SGTF-infected individuals diagnosed between Oct 1 and Nov 30, 2021, and a reduced risk of severe disease among SGTF-infected individuals compared with delta variantinfected individuals. Immunity (due to previous infection, vaccination, or both) in individuals infected with the omicron variant might, in part, account for this reduced severity. Potential reduced intrinsic virulence with regards to the omicron variant might also contribute to observed reductions in COVID-19 severity. Future studies that link data on vaccination status from vaccine registries could provide information on the attenuation of disease severity in breakthrough cases of COVID-19 following vaccination. These data are early and findings could change with additional follow-up time for the more recently diagnosed SGTF-infected individuals.