South Africa's 2026 flu season is over after an early March start, NICD says

South Africa's 2026 influenza season is over, the National Institute for Communicable Diseases (NICD) said on 21 October. The season ended in the week to 27 September, and flu activity has stayed below the seasonal threshold since the week starting 28 September. The NICD's end-of-season notice treats that drop as confirmed after activity remained below the threshold for two consecutive weeks, from the week starting 28 September.

It was an early year. The season began in the week starting 9 March, well before the usual start in mid-to-late April, and peaked in the week starting 11 May. The NICD said that was earlier than most past seasons and similar to 2025, when the season began in the week starting 24 March. A usual South African season runs from that mid-to-late April start through to late August or mid-September. This one opened in early March and was called over at the end of September, so it started early and still closed inside the usual end of the calendar.

What the surveillance found

The NICD tracks flu through three sentinel systems, not through every clinic in the country. The first is patients with flu-like illness at public primary care clinics. The second is patients with flu-like illness at private GP practices. The third is patients admitted to public hospitals with pneumonia. Together they are a sample of mild illness in clinics and of the more serious pneumonia admissions in public hospitals. They are not a count of every flu case in South Africa, and the NICD does not present them as one.

Between 29 December 2025 and 11 October 2026, influenza was found in 1,029 of 6,212 specimens tested, or 16.6%. Almost all of the positive samples, 97.1%, were typed further. That is 999 of the 1,029 positives. The untyped remainder is small enough that the strain shares below describe nearly the whole set of detections.

Influenza A(H3N2) was the most common strain, at 44.0% of detections (453). Influenza A(H1N1)pdm09 made up 32.6% (335) and influenza B 20.5% (211). Seven specimens carried two flu strains at once. Those seven dual infections are counted inside the same surveillance set. They are not a separate outbreak. They mean a small number of specimens showed more than one influenza virus.

Every influenza B sample that was tested for lineage before routine lineage testing stopped in June, 111 in all, was B/Victoria. Lineage testing then stopped, so the NICD is not claiming that every influenza B virus later in the season was Victoria. It is saying that all 111 specimens typed for lineage, up to the point routine typing ended, were B/Victoria.

The NICD's update gives no figures for hospital admissions or deaths. That absence is part of the notice. A season can be declared over, and a dominant strain can be named, without a public toll of who was admitted or who died. Anyone looking for that toll will not find it in this update.

How the season sat on the calendar

Week 11, starting 9 March, is the start. Week 20, starting 11 May, is the peak. Week 39, ending 27 September, is the end. Week 40, starting 28 September, is the first week below the threshold, and the NICD required a second week below that line before calling the season over.

Set against the usual window, March is early. Mid-to-late April is the ordinary opening. The 2025 season, which began in the week of 24 March, is the recent comparison the NICD itself offers. Two early openings in successive years are a pattern only in the weakest sense: the institute says 2026 was earlier than most past seasons and similar to 2025. It does not say the start is permanently moving.

The peak in the week of 11 May still falls in autumn, ahead of the mid-winter weeks when many people expect flu to be at its highest. Clinics that plan staffing against a late-April start would have been inside the season for several weeks before that usual opening arrived.

Flu can still circulate

The end of the season does not mean flu has disappeared. The NICD warned that sporadic infections still occur outside the season and asked clinicians to keep influenza in mind when patients have symptoms that strongly suggest it. A threshold is a surveillance line, not a switch that turns the virus off. Cases below the line are still cases. They are no longer numerous enough, in the sentinel sample, for the NICD to call it a season.

Most people recover from flu without treatment, but it can cause severe illness and death. The NICD lists the groups most at risk as pregnant women; people living with HIV; people with diabetes, lung disease, tuberculosis, heart disease, kidney disease or obesity; adults aged 65 and older; and children younger than two. The notice names those groups. It does not attach a hospitalisation rate or a death count to any of them for 2026.

The NICD does not give medical advice to individuals and refers the public to their own healthcare provider. The practical line for readers is the one the institute already uses: if symptoms strongly suggest flu, a clinician should still consider it, including now that the season has been closed. People in the risk groups above are the ones for whom that consideration matters most, because the NICD has already said they are the groups in which flu is more likely to be severe.

Weekly reports on influenza and other respiratory infections, including RSV, Covid-19 and whooping cough, remain available on the NICD website. Those reports are the place any later admission or death figures would surface. They are not in the end-of-season notice.

Source: NICD, "End of the 2026 influenza season", 21 October 2026.