BACKGROUND: In mid-November, 2021, the SARS-CoV-2 omicron variant (B.1.1.529; BA.1 sublineage) was detected in southern Africa, prompting international travel restrictions. We aimed to investigate the spread of omicron BA.1 in Africa. METHODS: In this observational study, samples from patients infected with SARS-CoV-2 from 27 laboratories in 24 African countries, collected between June 1, 2021 and April 14, 2022, were tested for omicron BA.1 and delta (B.1.617.2) variants using real-time RT-PCR. Samples that tested positive for BA.1 by RT-PCR and were collected before estimated BA.1 emergence according to epidemiological properties were excluded from downstream analyses. The diagnostic precision of the assays was evaluated by high-throughput sequencing of samples from four countries. The observed spread of BA.1 was compared with mobility-based mathematical simulations and entries for SARS-CoV-2 in the Global Initiative on Sharing All Influenza Data (GISAID) genomic database. We estimated the effective reproduction number (Rt) at the country level considering the BA.1 fraction and the reported numbers of infections. Phylogeographical analyses were done in a Bayesian framework. FINDINGS: Through testing of 13 294 samples from patients infected with SARS-CoV-2, we established that, by November-December, 2021, omicron BA.1 had replaced the delta variant of SARS-CoV-2 in all African subregions, following a south-north gradient, with a median Rt of 2·60 (95% CI 2·46-2·71). This south-north spread, established on the basis of PCR data, was substantiated by phylogeographical reconstructions, ancestral state reconstructions, and GISAID data. PCR-based reconstructions of country-level BA.1 predominance and the availability of BA.1 genomic sequences in GISAID correlated significantly in time (p=0·0002, r=0·78). The first detections of BA.1 in high-income settings beyond Africa were predicted accurately in time by mobility-based mathematical simulations (p<0·0001). Comparing PCR-based reconstructions with mobility-based mathematical simulations suggested that SARS-CoV-2 infections in Africa were under-reported by approximately ten times. Inbound travellers infected with BA.1, departing from five continents, were identified in six African countries by early December, 2021. INTERPRETATION: Omicron BA.1 was widespread in Africa when travel bans were implemented, limiting their effectiveness. Combined with genomic surveillance and mobility-based mathematical modelling, PCR-based strategies can inform Rt and the geographical spread of emerging pathogens in a cost-effective and timely manner, and can guide evidence-based, non-pharmaceutical interventions such as travel restrictions or physical distancing. FUNDING: Bill & Melinda Gates Foundation. TRANSLATIONS: For the French, Portugese and Spanish translations of the abstract see Supplementary Materials section.
![[img]](https://researchonline.lshtm.ac.uk/style/images/fileicons/text.png)
