China's influenza positivity peaked at 49.8% in late 2025 as A(H3N2) took over
A reconstruction of 125 weeks of national sentinel surveillance found 253,219 influenza-positive specimens, a multi-wave post-pandemic pattern, and the north leading the south by about a week.
| Group | Value (%) |
|---|---|
| A(H3N2) | 45.6 |
| A(H1N1)pdm09 | 33.8 |
| B(Victoria) | 20.6 |
Influenza surveillance in China is published week by week, which makes the longer arc of the virus hard to see. A new analysis in PLoS ONE stitches those weekly bulletins back into a single series and quantifies what the post-pandemic seasons actually looked like: multi-wave, regionally uneven, and dominated in its largest surge by influenza A(H3N2) [s1]. The researchers reconstructed 125 consecutive epidemiological weeks, from 2024/W01 to 2026/W21, from 125 publicly available aggregate reports issued by the Chinese Center for Disease Control and Prevention and the Chinese National Influenza Center [s1].
Across that window, 1,796,636 influenza-like-illness specimens were tested nationally and 253,219 were influenza-positive, a weighted positivity of 14.1% [s1]. National positivity rose and fell in distinct waves and peaked in 2025/W49 at 49.8%, when 11,138 of 22,381 tested specimens were positive [s1]. Year by year, positivity was 12.5% in 2024 (70,339 positives of 561,595 tested), 15.3% in 2025 (133,291 of 868,497), and 13.5% over the first 21 weeks of 2026 (49,589 of 366,544) [s1].
A north-south offset
China's surveillance reports split the country into Southern and Northern networks, and the two did not move in lockstep. Southern China carried a weighted positivity of 14.6% over the full period (157,732 positives of 1,078,477 tested) against 13.3% in the north (95,487 of 718,168), and its mean weekly influenza-like-illness percentage was higher too, at 4.7% versus 3.9% [s1]. Comparing the two networks across all 125 paired weeks, the south ran a mean 4.41 percentage points higher on positivity (95% CI, 2.23-6.59; P < 0.001) and 0.77 points higher on influenza-like illness (95% CI, 0.48-1.05; P < 0.001) [s1].
Timing differed as well. During the largest winter wave, the north reached its positivity peak in 2025/W48 at 48.1% and the south a week later in 2025/W49 at 54.5% [s1]. When the authors examined week-to-week changes, the strongest correlation appeared when the north led the south by one week (r = 0.781), while changes in influenza-like illness moved together with no lag (r = 0.758) [s1]. The authors present the lead-lag finding as exploratory and do not interpret it causally [s1].
Subtypes changing places
The dominant virus did not hold still. Among 253,104 explicitly reported subtype or lineage detections, A(H3N2) was the largest category over the whole series at 45.6% (115,336 detections), ahead of A(H1N1)pdm09 at 33.8% (85,518) and B(Victoria) at 20.6% (52,216); the B(Yamagata) lineage, absent from global circulation since the pandemic, was not reported at all [s1]. But the annual picture shifted sharply underneath that total. In 2024 the most-reported virus was A(H1N1)pdm09 (45.1%), followed by B(Victoria) (38.3%) and A(H3N2) (16.6%); in 2025 A(H3N2) jumped to 58.3% with A(H1N1)pdm09 at 40.3%; and in the first part of 2026 the field split between A(H3N2) (52.4%) and a resurgent B(Victoria) (47.3%) [s1].
Grouping consecutive weeks by their leading virus, the authors identified five national predominance periods [s1]. A(H3N2) led in 2024/W01 alone, B(Victoria) from 2024/W02 to 2024/W14 (13 weeks), A(H1N1)pdm09 through a long stretch from 2024/W15 to 2025/W17 (55 weeks), then A(H3N2) from 2025/W18 to 2026/W08 (43 weeks) spanning the largest winter wave, before B(Victoria) returned from 2026/W09 through 2026/W21 (13 weeks) [s1]. That A(H3N2) period alone accounted for 104,877 influenza-positive specimens and the overall positivity peak [s1].
The outbreak surge
The series also tracked reported institutional outbreaks, and here the concentration was stark. Of 8,801 influenza outbreaks reported across the whole period, 7,519 fell in 2025, against 903 in 2024 and 379 in the first 21 weeks of 2026 [s1]. Weekly outbreaks peaked in 2025/W48 at 1,541, one week ahead of the national positivity peak [s1]. A(H3N2)-associated outbreaks were the largest etiological category overall at 5,517 (62.7%), ahead of outbreaks with pending etiology (1,312; 14.9%) and A(H1N1)pdm09 (858; 9.7%); in 2025, A(H3N2) accounted for 5,419 of 7,519 outbreaks (72.1%) [s1]. The co-occurrence shows that the heaviest outbreak burden coincided with sustained A(H3N2) predominance, though the authors caution it does not establish causation [s1].
What the numbers are, and are not
This is a secondary analysis of aggregate public reports, not a fresh epidemiological survey, and the authors are explicit about the limits [s1]. Positivity was recalculated from reported positive and tested specimens; the counts are influenza-positive surveillance specimens, not population case counts [s1]. Testing volume swung widely, falling as low as 1,492 specimens in one northern week in 2024, which means raw counts have to be read alongside positivity rather than as incidence [s1]. The analysis also leans on only two broad regional strata over about 2.4 years, so the authors did not attempt province-level, climate, or mobility modelling [s1]. A recurring theme of the paper is the need for source-level quality control when official weekly tables contain internal inconsistencies [s1].
Set against the broader burden of a virus the World Health Organization estimates causes around one billion infections a year, including 3-5 million severe cases and 290,000-650,000 respiratory deaths, the value of a continuous national series is in the shape it reveals [s1]. What this one shows is a post-pandemic China still settling into an irregular, multi-wave rhythm, with the dominant strain turning over season to season and the north and south peaking days apart.
This article is informational and is not medical advice.
Sources
More on
When are you contagious with a cold or the flu? What shedding studies show
With influenza, virus shedding starts within a day of infection, peaks on day two and lasts about five days on average — often before you feel your worst. Nearly a quarter of infections cause one symptom or none at all.
Flu vaccine still cut hospitalisations despite the subclade K mismatch
WHO flagged a drifted A(H3N2) subclade K in December; the first 2025-2026 effectiveness data, from 1.4 million US veterans, put overall protection near 22% and protection against hospitalisation near 32%.
Italy reports Europe's first imported human case of H9N2 avian flu
The patient had spent months in Senegal and was found to carry the poultry-adapted virus alongside tuberculosis. WHO rates the public risk as low and traced no onward spread.
Risk nudges raised flu shots in three large US trials, AI disclosure aside
In preregistered trials of more than 90,000 patients, telling people an algorithm flagged them as high-risk lifted vaccination by a few points — and saying the algorithm was involved changed nothing.