This is a PDF file of a peer-reviewed paper that has been accepted for publication. Although unedited, the content has been subjected to preliminary formatting. Nature is providing this early version of the typeset paper as a service to our authors and readers. The text and figures will undergo copyediting and a proof review before the paper is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers apply.

Background The interaction between COVID-19, non-communicable diseases, and chronic infectious diseases such as HIV and tuberculosis is unclear, particularly in low-income and middle-income countries in Africa. South Africa has a national HIV prevalence of 19% among people aged 15-49 years and a tuberculosis prevalence of 0•7% in people of all ages. Using a nationally representative hospital surveillance system in South Africa, we aimed to investigate the factors associated with in-hospital mortality among patients with COVID-19. MethodsIn this cohort study, we used data submitted to DATCOV, a national active hospital surveillance system for COVID-19 hospital admissions, for patients admitted to hospital with laboratory-confirmed SARS-CoV-2 infection between March 5, 2020, and March 27, 2021. Age, sex, race or ethnicity, and comorbidities (hypertension, diabetes, chronic cardiac disease, chronic pulmonary disease and asthma, chronic renal disease, malignancy in the past 5 years, HIV, and past and current tuberculosis) were considered as risk factors for COVID-19-related in-hospital mortality. COVID-19 in-hospital mortality, the main outcome, was defined as a death related to COVID-19 that occurred during the hospital stay and excluded deaths that occurred because of other causes or after discharge from hospital; therefore, only patients with a known in-hospital outcome (died or discharged alive) were included. Chained equation multiple imputation was used to account for missing data and random-effects multivariable logistic regression models were used to assess the role of HIV status and underlying comorbidities on COVID-19 in-hospital mortality. FindingsAmong the 219 265 individuals admitted to hospital with laboratory-confirmed SARS-CoV-2 infection and known in-hospital outcome data, 51 037 (23•3%) died. Most commonly observed comorbidities among individuals with available data were hypertension in 61 098 (37•4%) of 163 350, diabetes in 43 885 (27•4%) of 159 932, and HIV in 13 793 (9•1%) of 151 779. Tuberculosis was reported in 5282 (3•6%) of 146 381 individuals. Increasing age was the strongest predictor of COVID-19 in-hospital mortality. Other factors associated were HIV infection (adjusted odds ratio 1•34, 95% CI 1•27-1•43), past tuberculosis (1•26, 1•15-1•38), current tuberculosis (1•42, 1•22-1•64), and both past and current tuberculosis (1•48, 1•32-1•67) compared with never tuberculosis, as well as other described risk factors for COVID-19, such as male sex; non-White race; underlying hypertension, diabetes, chronic cardiac disease, chronic renal disease, and malignancy in the past 5 years; and treatment in the public health sector. After adjusting for other factors, people with HIV not on antiretroviral therapy (ART; adjusted odds ratio 1•45, 95% CI 1•22-1•72) were more likely to die in hospital than were people with HIV on ART. Among people with HIV, the prevalence of other comorbidities was 29•2% compared with 30•8% among HIV-uninfected individuals. Increasing number of comorbidities was associated with...

SARS-CoV-2 Omicron (B.1.1.529) BA.4 and BA.5 sub-lineages, first detected in South Africa, have changes relative to Omicron BA.1 including substitutions in the spike receptor binding domain. Here we isolated live BA.4 and BA.5 viruses and measured BA.4/BA.5 neutralization elicited by BA.1 infection either in the absence or presence of previous vaccination as well as from vaccination without BA.1 infection. In BA.1-infected unvaccinated individuals, neutralization relative to BA.1 declines 7.6-fold for BA.4 and 7.5-fold for BA.5. In vaccinated individuals with subsequent BA.1 infection, neutralization relative to BA.1 decreases 3.2-fold for BA.4 and 2.6-fold for BA.5. The fold-drop versus ancestral virus neutralization in this group is 4.0-fold for BA.1, 12.9-fold for BA.4, and 10.3-fold for BA.5. In contrast, BA.4/BA.5 escape is similar to BA.1 in the absence of BA.1 elicited immunity: fold-drop relative to ancestral virus neutralization is 19.8-fold for BA.1, 19.6-fold for BA.4, and 20.9-fold for BA.5. These results show considerable escape of BA.4/BA.5 from BA.1 elicited immunity which is moderated with vaccination and may indicate that BA.4/BA.5 may have the strongest selective advantage in evading neutralization relative to BA.1 in unvaccinated, BA.1 infected individuals.

Background: The incidence of tuberculous pericarditis has increased in Africa as a result of the human immunodeficiency virus (HIV) epidemic. However, the effect of HIV co-infection on clinical

The SARS-CoV-2 Omicron (B.1.1.529) variant first emerged as the BA.1 sub-lineage, with extensive escape from neutralizing immunity elicited by previous infection with other variants, vaccines, or combinations of both1,2. Two new sub-lineages, BA.4 and BA.5, are now emerging in South Africa with changes relative to BA.1, including L452R and F486V mutations in the spike receptor binding domain. We isolated live BA.4 and BA.5 viruses and tested them against neutralizing immunity elicited to BA.1 infection in participants who were Omicron/BA.1 infected but unvaccinated (n=24) and participants vaccinated with Pfizer BNT162b2 or Johnson and Johnson Ad26.CoV.2S with breakthrough Omicron/BA.1 infection (n=15). In unvaccinated individuals, FRNT50, the inverse of the dilution for 50% neutralization, declined from 275 for BA.1 to 36 for BA.4 and 37 for BA.5, a 7.6 and 7.5-fold drop, respectively. In vaccinated BA.1 breakthroughs, FRNT50 declined from 507 for BA.1 to 158 for BA.4 (3.2-fold) and 198 for BA.5 (2.6-fold). Absolute BA.4 and BA.5 neutralization levels were about 5-fold higher in this group versus unvaccinated BA.1 infected participants. The observed escape of BA.4 and BA.5 from BA.1 elicited immunity is more moderate than of BA.1 against previous immunity1,3. However, the low absolute neutralization levels for BA.4 and BA.5, particularly in the unvaccinated group, are unlikely to protect well against symptomatic infection4.This may indicate that, based on neutralization escape, BA.4 and BA.5 have potential to result in a new infection wave.

No abstract

The extent to which Omicron infection1–9, with or without previous vaccination, elicits protection against the previously dominant Delta (B.1.617.2) variant is unclear. Here we measured the neutralization capacity against variants of severe acute respiratory syndrome coronavirus 2 in 39 individuals in South Africa infected with the Omicron sublineage BA.1 starting at a median of 6 (interquartile range 3–9) days post symptom onset and continuing until last follow-up sample available, a median of 23 (interquartile range 19–27) days post symptoms to allow BA.1-elicited neutralizing immunity time to develop. Fifteen participants were vaccinated with Pfizer's BNT162b2 or Johnson & Johnson's Ad26.CoV2.S and had BA.1 breakthrough infections, and 24 were unvaccinated. BA.1 neutralization increased from a geometric mean 50% focus reduction neutralization test titre of 42 at enrolment to 575 at the last follow-up time point (13.6-fold) in vaccinated participants and from 46 to 272 (6.0-fold) in unvaccinated participants. Delta virus neutralization also increased, from 192 to 1,091 (5.7-fold) in vaccinated participants and from 28 to 91 (3.0-fold) in unvaccinated participants. At the last time point, unvaccinated individuals infected with BA.1 had low absolute levels of neutralization for the non-BA.1 viruses and 2.2-fold lower BA.1 neutralization, 12.0-fold lower Delta neutralization, 9.6-fold lower Beta variant neutralization, 17.9-fold lower ancestral virus neutralization and 4.8-fold lower Omicron sublineage BA.2 neutralization relative to vaccinated individuals infected with BA.1. These results indicate that hybrid immunity formed by vaccination and Omicron BA.1 infection should be protective against Delta and other variants. By contrast, infection with Omicron BA.1 alone offers limited cross-protection despite moderate enhancement.

Omicron has been shown to be highly transmissible and have extensive evasion of neutralizing antibody immunity elicited by vaccination and previous SARS-CoV-2 infection. Omicron infections are rapidly expanding worldwide often in the face of high levels of Delta infections. Here we characterized developing immunity to Omicron and investigated whether neutralizing immunity elicited by Omicron also enhances neutralizing immunity of the Delta variant. We enrolled both previously vaccinated and unvaccinated individuals who were infected with SARS-CoV-2 in the Omicron infection wave in South Africa soon after symptom onset. We then measured their ability to neutralize both Omicron and Delta virus at enrollment versus a median of 14 days after enrollment. Neutralization of Omicron increased 14-fold over this time, showing a developing antibody response to the variant. Importantly, there was an enhancement of Delta virus neutralization, which increased 4.4-fold. The increase in Delta variant neutralization in individuals infected with Omicron may result in decreased ability of Delta to re-infect those individuals. Along with emerging data indicating that Omicron, at this time in the pandemic, is less pathogenic than Delta, such an outcome may have positive implications in terms of decreasing the Covid-19 burden of severe disease.