Human infections with highly pathogenic avian influenza A(H5N1) viruses in the United States from March 2024 to May 2025

Abstract

Between March 2024 and October 2024, 46 human cases of highly pathogenic avian influenza (HPAI) A(H5N1) had been detected in the United States. The persistent panzootic spread of HPAI A(H5N1) viruses and continued detection of human cases presents an ongoing threat to public health. Between November 2024 and May 2025, an additional 24 cases have been reported for a total of 70 human cases of HPAI A(H5N1): 41 were exposed to dairy cows, 24 to commercial poultry, 2 to backyard poultry, and 3 had an unidentified source of exposure. All sequenced viruses were clade 2.3.4.4b. Overall, 62 cases (89%) reported eye redness, 32 (46%) fever, and 29 (41%) respiratory symptoms; 54 of 67 cases (81%) reported receiving antiviral treatment. Most illnesses were mild; however, four patients were hospitalized. Of the hospitalized patients, three had pneumonia and one died. No human-to-human transmission was detected. Occupational exposure to infected animals was a risk factor for HPAI A(H5N1) virus infection and the risk to the general population remains low; however, the two cases exposed to infected backyard poultry and three cases with unidentified exposures highlight that ongoing vigilance is warranted.

Editor summary:

In an updated report on 70 laboratory-confirmed highly pathogenic avian influenza A H5N1 cases in the United States between March 2024 and May 2025, the absence of human to human transmission to date was confirmed, with no known mutations of resistance to neuroaminidase inhibitors.

Background

Highly pathogenic avian influenza (HPAI) A(H5N1) viruses, which have been circulating widely in wild birds for many years, have caused extensive outbreaks in U.S. domestic poultry^1–4. In March 2024, HPAI A(H5N1) virus infections were detected for the first time in dairy cows in the United States and a human case was detected in a dairy farm worker in Texas^5–7. State and local health departments, in close collaboration and coordination with the U.S. Centers for Disease Control and Prevention (CDC), monitor for and investigate human cases of novel influenza A virus infections, including HPAI A(H5N1) virus infections. Through these monitoring and surveillance efforts, human cases of influenza A(H5N1) continue to be detected, contributing to a growing understanding of the clinical spectrum and epidemiologic patterns associated with these cases.

Forty-six human cases of influenza A(H5N1), with symptom onsets from late March 2024 through October 2024, in the United States were previously described⁸. At that time most case patients had occupational exposure to infected dairy cows or commercial poultry and all but one had mild illness. Since October 2024, 24 additional human cases of A(H5N1) have been detected, including three hospitalizations with one death⁹. This report describes the epidemiologic, virologic, and clinical characteristics of the 70 human cases of influenza A(H5N1) that have been detected in the United States since March 2024. Our objective was to elucidate patterns and changes in these characteristics to inform ongoing public health actions and education to prevent further cases.

Results

Demographic characteristics and exposures

During March 2024 to October 2024, 46 human cases of influenza A(H5N1) had been reported in the United States, during November 2024 to May 2025 an additional 24 human cases were identified, for a total of 70 human cases of influenza A(H5N1) reported from March 2024 to May 2025 from 13 states (Figure 1, and Extended Data Figure 1). The last case detected during this period was in February 2025. Sixty-four case patients were detected through targeted surveillance of persons exposed to infected animals and six were detected through routine and enhanced influenza surveillance or clinician suspicion of HPAI A(H5N1) virus exposure. Four case patients (5.7%) were hospitalized during their illness, and one died.

Figure 1.

Detections of influenza A(H5N1) viruses in humans in the United States, March 2024 to May 2025. Figure includes detections of the first 46 human cases of A(H5N1) augmented with 24 additional detections. Publication of the first 46 detections from New England Journal of Medicine, Garg, et al., “Highly Pathogenic Avian Influenza A(H5N1) Virus Infections in Humans”, Volume No. 392, Page No. 9, Copyright © 2025 Massachusetts Medical Society. Reprinted and amended with permission.

Forty-one case patients (59%) had exposure to dairy cows; these cases occurred sporadically during March to December 2024, with another case detected in February 2025, when most virus detections in dairy herds were reported in the United States (Extended Data Figure 2). Twenty-four case patients (34%) were exposed to infected commercial poultry, mostly through depopulation activities; 20 of these were detected in two clusters, with illness during July and October 2024, with additional cases detected during December 2024 and February 2025. Two case patients (3%) reported exposure to infected backyard poultry, with illness during November 2024 and February 2025. Three additional case patients (4%) had unknown exposures to HPAI A(H5N1) viruses despite comprehensive investigations.

Sixty-eight case patients were adults and two were children aged <18 years (Table 1). The median age of case patients exposed to dairy cows and commercial poultry (42 years and 28 years, respectively) was younger than the median age of the two case patients exposed to backyard poultry (73 years). Fifty-eight case patients (91%) reported Hispanic/Latino ethnicity.

Table 1:

Demographic characteristics, signs, symptoms, and antiviral treatment among human cases of HPAI A(H5N1) by type of exposure, United States, March 2024 to May 2025. ^*

	Overall N = 70	Dairy Cows N = 41	Commercial Poultry N = 24	Backyard Poultry N = 2	Unidentified N = 3
Age, n/N (%)
<18 years	2 / 70 (3)	0 / 41 (0)	0 / 24 (0)	0 / 2 (0)	2 / 3 (67)
18–49 years	51 / 70 (73)	31 / 41 (76)	20 / 24 (83)	0 / 2 (0)	0 / 3 (0)
50+ years	16 / 70 (23)	10 / 41 (24)	3 / 24 (13)	2 / 2 (100)	1 / 3 (33)
Unknown	1 / 70 (1)	0 / 41 (0)	1 / 24 (4)	0 / 2 (0)	0 / 3 (0)
Male, n/N (%)	56 / 70 (80)	41 / 41 (100)	13 / 24 (54)	--	--
Race/ethnicity, n/N (%)†
Hispanic/Latino	58 / 64 (91)	39 / 40 (98)	18 / 20 (90)	--	--
White, non-Hispanic/Latino	6 / 64 (9)	1 / 40 (3)	2 / 20 (10)	--	--
Receipt of seasonal influenza vaccination in past 12 months, n/N (%)‡	16 / 62 (26)	8 / 38 (21)	7 / 21 (33)	0 / 2 (0)	1 / 1 (100)
At least one underlying condition, n/N (%)§	17 / 64 (27)	9 / 38 (24)	4 / 22 (18)	2 / 2 (100)	2 / 2 (100)
Signs and symptoms ¶
Conjunctivitis, n/N (%)	62 / 70 (89)	41 / 41 (100)	20 / 24 (83)	0 / 2 (0)	1 / 3 (33)
Fever, n/N (%)	32 / 70 (46)	13 / 41 (32)	15 / 24 (63)	2 / 2 (100)	2 / 3 (67)
Any respiratory symptom, n/N (%)	29 / 70 (41)	13 / 41 (32)	12 / 24 (50)	2 / 2 (100)	2 / 3 (67)
Cough, n/N (%)	17 / 68 (25)	8 / 40 (20)	6 / 23 (26)	1 / 2 (50)	2 / 3 (67)
Sore throat, n/N (%)	21 / 69 (30)	11 / 40 (28)	8 / 24 (33)	1 / 2 (50)	1 / 3 (33)
Shortness of breath, n/N (%)	14 / 68 (21)	6 / 40 (15)	5 / 23 (22)	2 / 2 (100)	1 / 3 (33)
Myalgias, n/N (%)	29 / 69 (42)	14 / 41 (34)	13 / 24 (54)	0 / 1 (0)	2 / 3 (67)
Headache, n/N (%)	29 / 69 (42)	12 / 41 (29)	13 / 23 (57)	2 / 2 (100)	2 / 3 (67)
Fatigue, n/N (%)	21 / 66 (32)	10 / 39 (26)	7 / 22 (32)	2 / 2 (100)	2 / 3 (67)
Nausea/Vomiting, n/N (%)	11 / 68 (16)	3 / 40 (8)	6 / 23 (26)	0 / 2 (0)	2 / 3 (67)
Diarrhea, n/N (%)	5 / 68 (7)	2 / 40 (5)	2 / 23 (9)	0 / 2 (0)	1 / 3 (33)
Clinical constellation, n/N (%)
Conjunctivitis only	22 / 70 (31)	18 / 41 (44)	4 / 24 (17)	0 / 2 (0)	0 / 3 (0)
Conjunctivitis plus non-respiratory symptoms	17 / 70 (24)	10 / 41 (24)	7 / 24 (29)	0 / 2 (0)	0 / 3 (0)
Conjunctivitis plus respiratory symptoms	23 / 70 (33)	13 / 41 (32)	9 / 24 (38)	0 / 2 (0)	1 / 3 (33)
Other symptoms, no conjunctivitis	8 / 70 (11)	0 / 41 (0)	4 / 24 (17)	2 / 2 (100)	2 / 3 (67)
Influenza antiviral treatment
Received influenza antiviral treatment, n/N (%)#	54 / 67 (81)	30 / 40 (75)	21 / 24 (88)	2 / 2 (100)	1 / 1 (100)
Days from symptom onset to start of influenza antiviral treatment, median (range)**	2 (−1, 11)	2 (0, 8)	2 (−1, 11)	9 (7, 10)	2 (--)
Started influenza antiviral treatment within 2 days of symptom onset, n/N (%)**	28 / 49 (57)	15 / 28 (54)	12 / 18 (67)	0 / 2 (0)	1 / 1 (100)

^*Some data are not presented to protect case patient privacy.

^†Race/ethnicity was missing for 6 case patients.

^‡Seasonal influenza vaccination information was missing for 8 case patients.

^§Underlying medical conditions were unknown for 6 case patients.

^¶The presence of certain signs and symptoms were unknown for some case patients. The number of total respondents is provided for each sign or symptom.

^#Receipt of influenza antiviral treatment was unknown for 3 case patients.

^**Among case patients who reported receiving influenza antiviral treatment. Start date of influenza antiviral treatment was unknown for 1 case patients and symptom onset date was unknown for 4 case patients. Two case patients started influenza antiviral treatment 1 day before reported symptom onset.

Use of personal protective equipment

Case patients exposed to dairy cows (n = 41) or who reported depopulating infected commercial poultry (n = 21) were asked about their use of personal protective equipment (PPE) when interacting with the animals. In general, use of PPE was greater among those who reported depopulating commercial poultry than among those exposed to dairy cows (Extended Data Figure 3); use of eye protection and facemask or respirator was reported by 54% of case patients depopulating commercial poultry compared to 22% of case patients exposed to dairy cows. Reported use of eye and respiratory protection among the case patients exposed to dairy cows or who were involved in depopulating commercial poultry did not appear to increase or decrease over time (Extended Data Figure 4).

Signs, symptoms, and influenza antiviral treatment

One third of all case patients (n = 22) reported symptoms of conjunctivitis only (Table 1 and Supplemental Table 1). Among the case patients exposed to dairy cows, all (100%) reported symptoms of conjunctivitis, 32% reported fever or feeling feverish, and 32% reported having respiratory symptoms such as cough, sore throat, or shortness of breath. Among case patients exposed to commercial poultry, 83% reported symptoms of conjunctivitis and 50% reported respiratory symptoms. One third to one half of case patients with exposure to infected dairy cows or commercial poultry reported constitutional symptoms such as headache, fatigue, and myalgias; gastrointestinal symptoms were also reported, though less commonly. The constellations of symptoms appeared to be consistent over time for those exposed to dairy cows or commercial poultry (Figure 2).

Figure 2.

Reported use of personal protective equipment among case patients infected with influenza A(H5N1) viruses that were exposed to cows (n=41) or commercial poultry (n=24), United States, March 2024 to May 2025. The following personal protective equipment is recommended to reduce exposure to avian influenza A viruses from sick animals or contaminated environments in high exposure settings: NIOSH Approved^® particulate respirator, fluid-resistant coveralls, safety goggles, boot covers or boots, head cover or hair cover, and disposable gloves⁷⁰. In medium risk settings, the following PPE are recommended: NIOSH Approved^® particulate respirator, safety goggles, and disposable gloves. Reported symptoms among case patients infected with influenza A(H5N1) viruses by type of exposure, United States, March 2024 to May 2025.

Among the two case patients exposed to backyard poultry, the reported signs and symptoms were similar to influenza-like illness; both initially reported fever, headache, fatigue, and respiratory symptoms. Among the three case patients for whom the exposure was unidentified, one reported symptoms of conjunctivitis, two reported fever, and two reported respiratory symptoms.

Eighty-one percent of all case patients received influenza antiviral treatment (78% among those who were not hospitalized), and 55% received antivirals within 2 days of symptom onset. All treated case patients received oseltamivir and two of the hospitalized patients received additional antiviral treatment.

Hospitalizations

Of the four hospitalized case patients, one had illness in August 2024, one had illness in November 2024, and two had illness in February 2025. One hospitalized case was exposed to commercial poultry infected with influenza A(H5N1) viruses, two were exposed to sick and dead infected backyard poultry, and one had an unknown exposure. The median age of the hospitalized case patients was 68 years, which was significantly older than case patients who were not hospitalized (median 39 years, Wilcoxon p-value = 0.015). Three of the hospitalized case patients reported having at least one underlying medical condition (Table 2). The hospitalized case patients were admitted a median of 3 days after illness onset (range: 1–10 days).

Table 2.

Hospitalized case patients with highly pathogenic avian influenza A(H5N1), United States, March 2024 to May 2025.

	Hospitalized case patients N = 4
Median patient age (years)	68
Presence of at least one underlying condition, n	3
Days from symptom onset to hospital admission, median (range)	3 (1, 10)
Treatments and clinical interventions
Received oseltamivir treatment, n	4
Received combination antiviral treatment, n*	2
Days from symptom onset to start of influenza antiviral treatment, median (range)	6 (2, 10)
Received respiratory support, n	3
Received antibiotics, n	3
Received systemic corticosteroids, n	2
Hospital outcomes
Pneumonia, n	3
Other complications, n	2
Admitted to intensive care unit, n	3
Death, n	1
Days in hospital, median (range) †	11 (3, 24)

^*One patient received oseltamivir, baloxavir, amantadine; one patient received oseltamivir (replaced by peramivir), baloxavir, rimantadine.

^†Days from hospital admission to discharge or death.

All four hospitalized case patients were treated with influenza antivirals starting at a median of 6 days after symptom onset (range: 2–10 days): two received oseltamivir monotherapy and two received combination antiviral treatment (one received oseltamivir, baloxavir, amantadine; and one received oseltamivir, baloxavir, rimantadine, with peramivir replacing oseltamivir). Respiratory support was provided to three patients; two patients received high-flow nasal cannula, and one patient received a higher level of support. Pneumonia was diagnosed in three patients and more severe complications were noted in two patients. The median length of hospitalization, to discharge or death, was 11 days (range 3–24 days). One hospitalized patient died.

One hospitalized patient, without pneumonia or lower respiratory tract disease, had influenza A(H5N1) virus confirmed from an upper respiratory tract specimen collected 2 days after illness onset (Extended Data Figure 5). The three hospitalized patients with lower respiratory tract disease had initial upper respiratory tract specimens (including nasopharyngeal or combined nasal-throat swabs) collected at a median of 3 days (range: 1–7 days) after illness onset that all tested negative for influenza viruses. Additional specimens from the upper respiratory tract and, in some cases, lower respiratory tract (e.g., induced sputum, bronchoalveolar lavage fluid, or, in intubated patients, an endotracheal aspirate), were collected that led to diagnosis of influenza A(H5N1) (Extended Data Figure 5 and Extended Data Table 1).

Virologic and Serologic Characterization

Sequencing of viruses identified from all 70 case patients was attempted; 27 complete genomes and 29 partial genomes were recovered. Fifty-six viruses were confirmed by hemagglutinin (HA) sequence to belong to HPAI A(H5N1) clade 2.3.4.4b (Figure 3). For the 27 viruses with complete genomes, one of three genotypes were identified: 22 viruses (76%) were B3.13 genotype, 4 (14%) were D1.1 genotype, and 1 (3%) was D1.3 genotype. If no sequence or only partial genomes were recovered from case patient specimens, then the virus genotype was inferred based on sequence similarities to other viruses identified in animals to which the case patient was exposed or identified in the same setting (i.e., depopulation operation) where other case patients had complete genomes analyzed (Extended Data Table 2). Of the hospitalized cases, two case patients had genotype D1.1 A(H5N1) viruses detected (with backyard poultry exposures), one had genotype B3.13 (with unidentified exposure), and one had genotype D1.3 detected (with exposure to commercial poultry). Phylogenetic analysis of the HA gene sequences obtained from human cases clustered closely with sequences obtained from dairy cattle or poultry, which coincided with the reported source of exposure of the human cases. Sequences from the two human cases from California and one from Missouri that did not report a known exposure source clustered closely with B3.13 genotype viruses detected primarily in dairy cattle (Figure 3).

Figure 3.

Phylogenetic trees of the hemagglutinin genes of clade 2.3.4.4b avian influenza A(H5N1) viruses. Sequences obtained from human cases in the United States belonging to the B3.13 (A) and D1.1 or D1.3 (B) genotypes are shown in green font.

The viruses that were grouped as B3.13 genotype had the PB2 M631L substitution, which is found in B3.13 viruses from dairy cattle and has been associated with mammalian adaptation potentially as a mechanism to enhance polymerase activity in mammalian cells¹⁰. All B3.13 viruses sequenced from humans had PB2 631L, which suggests this mutation was present in viruses in dairy cattle and did not occur after human infection. Viruses from 10 case patients had other molecular markers of interest (Table 3): three with amino acid substitutions PB2 E627K or D701N that are associated with mammalian adaptation, one with amino acid substitutions HA P136S and A156T that may impact antigenicity, one specimen contained virus with low frequency minor variants in the HA (A134V, N182K and E186D) that are associated with increased binding to α2–6 receptors, three with a substitution NA S247N that may reduce susceptibility to oseltamivir based on laboratory tests, one with a substitution PA I38M that is associated with decreased susceptibility to baloxavir, and one with a substitution M2 S31N which is associated with cross-resistance to amantadine and rimantadine.

Table 3.

Molecular markers of interest detected in influenza A(H5N1) viruses collected from ten case patients, United States, March 2024 to May 2025. ^*

Virus	Genotype	Gene	Change	Effect
A/Texas/37/2024	B3.13	PB2	E627K	Associated with mammalian adaptation 57–59
A/Washington/251/2024	D1.1	NA	S247N	May slightly reduce susceptibility to oseltamivir in laboratory tests 61,62
A/Washington/252/2024	D1.1	NA	S247N	May slightly reduce susceptibility to oseltamivir in laboratory tests 61,62
A/Washington/253/2024	D1.1	NA	S247N	May slightly reduce susceptibility to oseltamivir in laboratory tests 61,62
A/Missouri/121/2024	B3.13	HA	P136S A156T	Changes are at antigenic sites and may affect antigenicity and replication fitness 63
A/California/150/2024	B3.13	PA	I38M	Decreased susceptibility to baloxavir in laboratory tests50
A/Louisiana/12/2024*	D1.1	HA	A134A/V [Alanine 88%, Valine 12%] N182N/K [Asparagine 65%, Lysine 35%] E186E/D [Glutamic acid 92%, Aspartic Acid 8%]	May result in increased virus binding to α2–6 cell receptors in upper respiratory tract of humans64–66
A/Iowa/124/2024	D1.1	M2	S31N	Associated with cross-resistance to amantadine and rimantadine 46
A/Nevada/10/2025	D1.1	PB2	D701N	Associated with mammalian adaptation 67,68
A/Wyoming/01/2025	D1.1	PB2	E627K	Associated with mammalian adaptation 57–59

^*Public database accession numbers for the sequences are provided in Extended Data Table 3.

^†Genetic changes were seen in deep sequencing analysis from a nasopharyngeal/oropharyngeal specimen ⁶⁹. These changes were not observed in deep sequencing analysis from a nasopharyngeal specimen from the same patient. The low frequency changes represent a small proportion of the total virus population in the sample and likely emerged during the patient’s infection rather than primarily transmitted from animals at the time of infection.

Nine of the 70 reported case patients, including three of the hospitalized patients, had convalescent blood specimens for serologic testing collected at a median of 31 days (range: 15–33 days) after their symptom onset. Four (44% of 9) were seropositive. For both seropositive (n = 4) and seronegative (n = 5) cases, convalescent blood specimens were collected a median of 31 days after symptom onset. Two of the hospitalized case patients were seropositive, both had lower respiratory tract disease.

Contact investigations

State and local health departments conducted investigations into all cases. A total of 180 household contacts (median 3 per case patient) were reported from 56 case patients who provided household contact information. None of the household contacts tested positive for influenza A(H5) viruses. Serum specimens collected from 15 contacts were also tested using serologic assays (including household, healthcare workers, and other close contacts from two cases with unknown exposure); none were seropositive.

Discussion

Since the first detections of HPAI A(H5N1) viruses in dairy cows in Kansas and Texas in March 2024, the viruses have been detected in dairy cows in 17 states and in poultry in all 50 states^11,12. Through May 2025, there have been 70 sporadic human A(H5N1) cases detected in the United States with the last symptom onset in February 2025. Four case patients were hospitalized, one of whom died, but 66 (94%) case patients had mild illness. There continues to be no human-to-human transmission identified, as determined by contact investigations and routine and enhanced surveillance. Though no additional human cases have been identified in the United States since February 2025, the persistent panzootic of HPAI A(H5N1) viruses and continued detection of human cases in other countries presents an ongoing threat to public health that warrants continued vigilance^13,14. CDC and state and local public health agencies continue to monitor for human influenza A(H5N1) virus infections and signals that suggest the virus may be changing in ways that impact antiviral treatment, clinical severity, or the ability to transmit effectively between humans.

Since the previous report on the first 46 human cases in the United States, we continue to observe that most human infections with HPAI A(H5N1) viruses in the United States have occurred after direct exposure or contact with infected animals or their products⁸. This observation supports previous risk assessments and the ongoing recommendations to avoid unprotected contact with animals (including wild animals, companion animals, or livestock) suspected to be infected or surfaces or animal products that are suspected to be contaminated^15–17. Following the detection of the initial human cases in the United States, state and local departments of health and agriculture, CDC, and the United States Department of Agriculture engaged with the public and groups at higher risk of infection to provide information on HPAI A(H5N1) viruses, infections in animals, the risk to people, and ways to protect against infection (Extended Data Figure 6 for CDC-supported activities) ^15–17.

While most initial cases detected in the United States were associated with occupational exposures to infected animals, two case patients, described in this report, were infected with influenza A(H5N1) following exposure to sick and dead infected poultry in their backyards. Both case patients were hospitalized, and one infection was fatal. These cases serve as a reminder that infections with these viruses can lead to severe illness and outcomes. Although it is not clear whether these cases were severe because of baseline health status, older age, the duration, dose, or route of exposure to the viruses, or a combination of factors, the pattern of exposure and greater illness severity more closely resembles what has been seen in other countries where unprotected close contact with backyard poultry is common^18–22.

This may suggest that while clade 2.3.4.4b influenza A(H5N1) viruses can cause severe illness and death, factors related to who is exposed, how they are exposed, and when they may seek care and get treatment could play a greater role in modulating clinical manifestation. Globally, where mortality from influenza A(H5N1) virus infections averages around 50%, backyard flock owners are a well-established risk group and there may be important lessons for risk communication and education that we can learn from other settings and apply locally^23–27. In the United States, where backyard poultry farming is seeing a growth in popularity, we need a better understanding of the characteristics of backyard poultry owners and their flocks to better target public health messaging and education²⁸.

We also report three case patients where the source of exposure to HPAI A(H5N1) viruses was not identified^29–31. Even with comprehensive investigation, it is not uncommon that a clear exposure to the virus may not be determined, possibly due to the limits of recall or lack of a clear exposure. Globally, roughly 5% of human cases with avian influenza virus infection have not had a known exposure to an infected animal³². These cases remind us of the importance of ongoing surveillance, and specifically virus subtyping, in broader populations without recognized occupational exposures, particularly when there is widespread circulation of avian influenza viruses in animals.

Globally, since 2022, there have been 85 human cases reported with confirmed or presumed infection with HPAI A(H5N1) clade 2.3.4.4b viruses, including 10 hospitalizations and three deaths^4,13,33,34. During previous periods when wild birds and commercial poultry were greatly impacted by clade 2.3.4.4b HPAI A(H5N1) viruses in the United States, only one human case had been detected in 2022, which was mild and self-limiting³⁵. With 70 human cases identified in the United States since March 2024, we have seen a greater range of clinical illness, but the clinical spectrum is still milder than would have been expected given global data. Although, the reasons for this are not fully known, milder clinical illness among the U.S. cases might be explained, in part, by the affected population, who are generally healthy farm workers, and that active monitoring in this population resulted in rapid virus detection and early access to influenza antivirals. The hospitalized cases in the United States were older and diagnosed and received antiviral treatment later in the course of their illness than those not hospitalized. The route of exposure (ocular versus respiratory), modality (contact to conjunctivae), as well as the virus dose and duration of exposure, could also play a role in clinical severity, but more work is needed to further characterize exposures and transmission between infected animals and humans. We are also learning more about the role of pre-existing cross-neutralizing immunity in clinical severity of clade 2.3.4.4b A(H5N1) virus infections in animal models, and we cannot rule out the influence of other host or viral factors^36–38.

Upper respiratory specimens from three of the hospitalized patients were initially negative for influenza viruses. Seasonal influenza viruses attach preferentially to cells with α2,6-linked sialic acid receptors, which are prevalent in the upper respiratory tract in humans and can be readily detected in upper respiratory specimens during infection; whereas, HPAI A(H5N1) viruses bind preferentially to α2,3-linked sialic acid receptors, which are prevalent in the lower respiratory tract and found on conjunctivae in humans^39–42. While one case in this report had low frequency mutations at residues A134V, N182K, E186D that may result in increased virus binding to α2–6 cell receptors, these changes represented a small proportion of the total virus population identified in the sample analyzed. Given these changes have not been observed in sequences from circulating animal viruses, it is likely the mutations were generated by replication of this virus in the patient with advanced disease rather than primarily transmitted at the time of infection. Thus for hospitalized patients with lower respiratory tract disease and who are under investigation for HPAI A(H5N1), specimens should be collected from both the upper and lower respiratory tract (such as induced sputum or, in intubated patients, an endotracheal aspirate or bronchoalveolar lavage fluid) for influenza testing to increase the chance that influenza viruses could be detected. Also for hospitalized patients, empiric antiviral treatment with oseltamivir should be started as soon as possible¹⁵. Consideration should be given to use of combination antiviral treatment such as was administered to two hospitalized case patients in the United States and a critically ill patient in Canada^33,43.

Markers of mammalian adaptation detected in the polymerase genes of sequences obtained from human cases (i.e., PB2 627K and 701N) were identified in dairy cattle sequences only sporadically and at very low frequencies. Given this and that only a few human cases were detected with these changes, we hypothesize that these mutations likely emerged during the course of viral replication after human infection. The exception may be the mutation of PB2 D701N found in the sequence from the human case in Nevada, which was a mutation also identified in multiple dairy cattle in Nevada⁴⁴. No known molecular markers of resistance to neuraminidase inhibitors were detected in viruses sequenced from human infections from the United States⁴⁵. One virus had M2 S31N, a marker of resistance to amantadine and rimantadine, and another had a substitution PA I38M, previously associated with reduced susceptibility to baloxavir^46–48. Both case patients had clinically mild illness and only received oseltamivir treatment, suggesting that these mutations were likely present in the transmitted viruses. In addition, three genotype D1.1 viruses from humans had the NA S247N substitution, which is associated with slightly reduced susceptibility to oseltamivir in a laboratory assay^49,50. While the frequency of mutations associated with reduced susceptibility to available influenza antivirals is currently low, it is important to continue monitoring and characterizing antiviral susceptibility in viruses in animals and infected humans to inform clinical treatment guidance.

Since the first report of human infections in the United States, additional laboratory studies have been completed, providing insights into the serologic response to infection and potential virologic adaptations. From serologic assessment of nine cases, four cases were seropositive. We may be seeing lower seropositivity because of the dose or site of inoculum (e.g., ocular infection may generate a more localized immune response than respiratory infection), lower antibody response to mild illness^51,52, prior immunity from seasonal influenza vaccination, which is not intended to prevent A(H5N1), or seasonal influenza virus infection that may be cross-reactive and impact the serologic response to HPAI A(H5N1) viruses^36,53,54. It is also possible that a lack of serologic response in some case patients indicates that infection may not have occurred, but that HPAI A(H5N1) virus genetic material was detected incidentally because the exposure setting was highly contaminated⁵⁵. Serologic sampling of the cases in the United States has been too limited to draw general conclusions, and more investigations are needed into the serologic response to these viruses.

From ongoing genetic characterization, we have observed gene reassortment events with avian influenza A viruses circulating in wild birds resulting in numerous genotypes detected in dairy cattle and poultry⁵⁶. Whether reassortment has increased the likelihood of HPAI A(H5N1) cases in humans is not known and cannot be ruled out, as some of the resulting virus genotypes have molecular substitutions that are persistently detected, and which recent studies have suggested may impact mammalian adaptation. For example, all genotype B3.13 viruses share PB2 M631L substitutions, which was shown to increase influenza polymerase activity in human cells¹⁰. Additionally, we have seen the PB2 E627K substitution, associated with mammalian adaptation, detected sporadically in two viruses identified in human cases, including a genotype B3.13 virus and a genotype D1.1 virus^57–59.

This analysis and our conclusions are subject to limitations. We summarized data reported on detected cases in the United States, which may not be generalizable to influenza A(H5N1) viruses circulating or cases detected globally. All cases in the United States were detected through active monitoring of exposed persons or through facility-based surveillance. CDC has encouraged expedited influenza testing and subtyping among patients with critical respiratory illness to increase the likelihood of detecting influenza A(H5) virus infection; however, we suspect that we may not have detected all cases, particularly those who had mild illness or did not present to clinical care⁶⁰. Data on the use of PPE was missing for some case patients and not collected for case patients who were exposed to but not depopulating poultry; thus, we were unable to assess use of PPE in case patients exposed to backyard poultry. This was a descriptive case series and, therefore, could not formally address questions related to risk factors for exposure or infection compared with non-cases. Finally, the number of cases detected in the United States limits the precision around our conclusions, particularly among hospitalized case patients.

As HPAI A(H5N1) virus detections in dairy cows and commercial poultry have decreased in the United States, so too have detections in humans. Occupational exposures to infected animals continue to be the primary risk factor among cases in the United States. Infection risk to the general population is low and human-to-human transmission has not been identified. Additionally, no known markers of neuraminidase inhibitor resistance have been detected in viruses from case patients. While hospitalizations and a fatality have been reported, most cases have had mild illness, and it is unclear why we have not seen more patients with severe disease and should be a focus of future investigation. Given the continued panzootic of HPAI A(H5N1) viruses, the potential for severe disease and outcomes in humans, and the threat of genetic change that could spur human-to-human transmission, continued vigilance is warranted. People should avoid unprotected contact with sick or dead animals and animal environments or animal products that could be contaminated with influenza viruses. When people need to touch or handle animals suspected to be infected or potentially contaminated materials, they should exercise precautions. Employers play a critical role in implementing engineering and administrative controls and providing the appropriate personal protective equipment to reduce the risk of exposure to workers who need to touch or handle animals suspected to be infected, or potentially contaminated materials. It is important for people who might have been exposed to monitor their health for signs and symptoms of influenza virus infection and seek prompt care and treatment if symptoms develop. CDC continues to monitor the situation and will provide updated guidance on prevention, surveillance, testing, and treatment of HPAI A(H5N1) virus infection as the current situation evolves.

Methods

Case identification and interview

This activity was reviewed by CDC, was deemed to be non-research activity, and was conducted in a manner consistent with applicable federal law and CDC policy (See e.g., 45 C.F.R. part 46, 21 C.F.R. part 56; 42 U.S.C. §241(d); 5 U.S.C. §552a; 44 U.S.C. §3501 et seq.).

Human cases of influenza A(H5N1) virus infection were detected though targeted monitoring of individuals known to have been exposed to infected animals and through routine and enhanced influenza surveillance ^1–3. For targeted monitoring, respiratory specimens, and conjunctival specimens among persons with symptoms of conjunctivitis, were recommended to be collected from any person with possible exposure to HPAI A(H5N1) viruses who met clinical criteria (signs and symptoms of acute upper or lower respiratory tract infection or conjunctivitis) or was part of a public health investigation. Individuals provided assent prior to specimen collection. For routine and enhanced public health surveillance for influenza, respiratory specimens were collected from individuals at the discretion of the clinician. Public health laboratories and some clinical and academic laboratories enhanced their surveillance for non-seasonal influenza viruses. In particular, laboratories were recommended to increase the number of influenza A positive specimens that were subtyped and to accelerate the timeline for subtyping influenza A specimens from hospitalized patients ^2,4.

Specimens testing positive for influenza A virus, but with negative results for seasonal influenza A(H1) or A(H3) targets, or classified as unsubtypeable, were sent to state and local public health laboratories or CDC for confirmatory testing using real-time reverse-transcriptase-polymerase-chain-reaction (RT-PCR). Additionally, specimens with presumptive positive test results for influenza A(H5) virus detected in laboratories were sent to CDC or the state public health laboratory for confirmatory testing using RT-PCR^5,6.

Suspected, probable, or confirmed cases of influenza A(H5N1), or their proxies, were interviewed to understand the source of infection, how exposure might have occurred, signs and symptoms experienced, clinical care received, and other people the case came into close contact with⁷. Information from the interviews was entered into standard case reporting forms and submitted to CDC. Only information from patients meeting the Council of State and Territorial Epidemiologist definition of confirmed influenza A(H5) virus infection is reported here⁷. The neuraminidase could not be confirmed by the CDC Influenza Division laboratory for all confirmed cases; however, we refer to all cases as having HPAI A(H5N1) based on the neuraminidase of viruses identified in animals to which the case patient was exposed or presumed to be exposed, or in other case patients who were exposed in the same setting (i.e., depopulation operation) as another case patient.

Laboratory Methods

Hemagglutination Inhibition Assay

Hemagglutination inhibition (HI) assays to detect antibody responses to A(H5N1) viruses were performed with a modified HI assay using horse erythrocytes optimized for avian influenza A(H5) viruses as previously described^1,2. In brief, sera were heat inactivated for 30 minutes at 56°C and then tested for non-specific agglutinins and adsorbed with horse erythrocytes, if needed. Sera were then treated with receptor-destroying enzyme for 18–20 hours at 37°C to remove any non-specific inhibitors that may have been introduced during hemadsorption, followed by heat inactivation prior to the HI assay. Sera were pre-diluted at 1:10, then serially diluted 2-fold and incubated for 30 minutes with 4 hemagglutination units per 25μL of virus, incubated with 1% horse erythrocytes for 60 minutes. HI antibody titer was defined as the reciprocal of the last dilution of serum that completely inhibited hemagglutination. Antibody titer <10 (initial dilution) was reported as 5 for calculation purpose. Multiple replicates were conducted, final titers are geometric mean titers (GMTs) from multiple replicates. HI assays were conducted at CDC in Biosafety level 3 enhanced (BSL-3E) laboratories.

Microneutralization Assay (MN)

Microneutralization (MN) assays were performed as previously described^8,9. Heat inactivated human sera were pre-diluted at 1:10, then serially diluted 2-fold and incubated with 100 50% tissue culture infection dose (TCID₅₀) of influenza viruses. The virus-sera mixture was used to infect 1.5 × 10⁴/well Madin-Darby canine kidney (MDCK) cells and incubated overnight. The plates were fixed with cold 80% acetone and the presence of viral nucleoprotein was quantified by enzyme-linked immunosorbent assay (ELISA). Neutralizing antibody titers were defined as the reciprocal of the highest serum dilution that showed 50% neutralization. Antibody titer <10 (initial dilution) was reported as 5. Multiple replicates were conducted, final titers are geometric mean titers (GMTs) from multiple replicates. MN assays were conducted in CDC BSL-3E laboratories.

Serologic response

A seropositive result was defined as both neutralizing antibody titer ≥40 and HI antibody titer ≥40 against a wild type 2.3.4.4b A(H5N1) virus (A/Texas/37/2024 and/or A/Washington/240/2024). The serologic methods used were also used in a prior study examining the serologic response to HPAI H5N1 clade 2.3.4.4b virus infection in humans¹⁰. In that study, patients with clinically mild illness and localized infection, such as conjunctivitis, induced neutralizing antibody responses, demonstrating that the immune responses can be detected with the methods used in the current investigation.

Next generation sequencing

Detected influenza A(H5) viruses were further characterized using next generation sequencing conducted at CDC. Sequencing was attempted on the codon complete genome. Influenza A genome amplification was performed using M-RTPCR with the Uni12/13 Inf1/3 primers¹¹. Targeted HA and NA amplification was also performed to increase the efficiency of sequencing low viral load specimens when codon complete genome amplification was limited. Amplicons were purified using exonuclease I (Thermo Fisher), and libraries generated using quarter reactions following Illumina DNA Prep’s protocol otherwise. Specimens with low viral loads had RNA and amplicons enriched with fixed Twist Comprehensive Viral Research Panel following the Fast Hybridization methodology (Twist Bioscience, 103550). Each specimen was amplified and sequenced with replicates. All amplicon libraries and enriched libraries were uniquely barcoded and sequenced on an Illumina MiSeq using a 300-cycle kit (paired end 2×150bp). The replicate sample’s demultiplexed sequencing data was merged when necessary and assembled using IRMA v1.1.3 FLU-utr module¹². The resulting assemblies were further quality checked before being released which included, median/average segment coverage >= 100x, individual base coverage >=25, start and stop codons were within coding frame, no sequencing induced indels present, checks for regions of high minor variants which could indicate contamination, and checks of an overabundance of mixed bases (minor variant >=20%). Codon complete segment assemblies that did not meet these quality checks were withheld from downstream analyses. Segment assemblies that were only partially complete or had regions of low coverage, but otherwise passed the QC checks, had these regions redacted. The passing regions of the segment assemblies were included in the downstream analyses where applicable.

Detected viruses with codon complete genome sequences were grouped into genotypes based on eight gene segments following the GenoFLU referencing system¹³. Where only a partial genome was recovered, the genotypes of viruses were inferred based on sequence similarities to other viruses identified in animals to which the case patient was exposed or presumed to be exposed, or exposure in the same setting (i.e., depopulation operation) where other case patients had complete genomes analyzed. The genotype of viruses from case patients with partial genomes or with no sequence obtained from the case patient sample was inferred based on sequence similarities to other viruses identified in animals to which the case patient was exposed or presumed to be exposed in the same setting (i.e., depopulation operation) where other case patients had complete genomes analyzed.

Identified molecular markers of interest were shown using straight H5N1 numbering (mature HA protein numbering when referring to HA amino acid residues).

Phylogenetic trees were built for the HA gene segment. Gene segments were alignment with closely related virus sequences obtained from GISAID (http://platform.gisaid.org) and the Short Read Archive/NCBI. The sequences for each strain were aligned using the ClustalW application and Muscle algorithm. Neighbor joining phylogenetic trees were built using MEGA7.0 software and the Jukes-Cantor Model of evolution with uniform rates (www.megasoftware.net).

Data analysis

Data were described for case patients overall and by the primary exposure that each patient reported. When the exposure was not clear after thorough investigation, the case patient’s exposure was reported as unidentified. Demographic characteristics, reported symptoms, and frequency of personal protective equipment use among case patients who were exposed to cows or participated in activities to depopulate infected poultry were compared over time and by reported exposures. Some information and values were suppressed from this report to protect participants’ privacy. Data were managed and analyzed using R software and SAS software, version 9.4^14,15.

Extended Data

Extended Data Fig. 1.

States where case patients infected with influenza A(H5N1) viruses have been detected, United States, March 2024 to May 2025.

Extended Data Fig. 2.

Detections of influenza A(H5N1) viruses A) in humans with exposure to dairy cows and in dairy cow herds; B) in humans with exposure to commercial poultry and detections in commercial poultry; and C) in humans with exposure to backyard poultry and reported detections in backyard poultry, United States, March 2024 to May 2025.

From New England Journal of Medicine, Garg, et al., “Highly Pathogenic Avian Influenza A(H5N1) Virus Infections in Humans”, Volume No. 392, Page No. 9, Copyright © 2025 Massachusetts Medical Society. Reprinted and amended with permission. Data available from United States Department of Agriculture (livestock/dairy cows: https://www.aphis.usda.gov/livestock-poultry-disease/avian/avian-influenza/hpai-detections/hpai-confirmed-cases-livestock; and commercial/backyard flocks: https://www.aphis.usda.gov/livestock-poultry-disease/avian/avian-influenza/hpai-detections/commercial-backyard-flocks).

Extended Data Fig. 3.

Reported use of personal protective equipment among case patients infected with influenza A(H5N1) viruses that were exposed to cows (n=41) or commercial poultry (n=24), United States, March 2024 to May 2025.

The following personal protective equipment is recommended to reduce exposure to avian influenza A viruses from sick animals or contaminated environments in high exposure settings: NIOSH

Approved^® particulate respirator, fluid-resistant coveralls, safety goggles, boot covers or boots, head cover or hair cover, and disposable gloves (for further details, see https://www.cdc.gov/bird-flu/worker-safety/index.html [last updated: 6 May 2025]). In medium risk settings, the following PPE are recommended: NIOSH Approved^® particulate respirator, safety goggles, and disposable gloves.

Extended Data Fig. 4.

Reported use of eye protection and facemask/respira tor as personal protective equipment (PPE) among case patients infected with influenza A(H5N1) viruses that were exposed to dairy cows or involved in depopulating commercial poultry, United States, March 2024 to May 2025.

The following personal protective equipment is recommended to reduce exposure to avian influenza A viruses from sick animals or contaminated environments in high exposure settings: NIOSH

Approved^® particulate respirator, fluid-resistant coveralls, safety goggles, boot covers or boots, head cover or hair cover, and disposable gloves (for further details, see https://www.cdc.gov/bird-flu/worker-safety/index.html [last updated: 6 May 2025]). In medium risk settings, the following PPE are recommended: NIOSH Approved^® particulate respirator, safety goggles, and disposable gloves. Of note, facemasks are not a recommended level of PPE against exposure to avian influenza A viruses.

Extended Data Fig. 5.

Respiratory specimens, testing results, and antiviral treatment among case patients hospitalized with influenza A(H5N1) virus infection, United States, March 2024 to May 2025.

Extended Data Fig. 6.

Timeline of educational and communication activities on influenza A(H5N1) viruses initiated by CDC, United States, March 2024 to May 2025. NCFH: National Center for Farmworker Health.

Extended Data Table. 1.

Respiratory specimens and testing results among case patients hospitalized with influenza A(H5N1) virus infection, United States, March 2024 to May 2025.

	Patient 1	Patient 2	Patient 3	Patient 4
First specimen
Collection day, post-symptom onset	2	7	3	1
Specimen type	Upper respiratory	Upper respiratory	Upper respiratory	Upper respiratory
Clinic/hospital result (test type)	Positive (PCR panel)	Negative (PCR panel)	Negative (rapid antigen)	Negative (PCR panel)
State Public Health Lab, RT-PCR result	Positive (H5)	Not tested	Not tested	Not tested
CDC Lab, RT-PCR result	Positive (H5)	Not tested	Not tested	Not tested
Second specimen
Collection day, post- symptom onset		10	4	4
Specimen type		Upper respiratory	Upper respiratory	Upper respiratory
Clinic/hospital result (test type)		Positive (rapid antigen) Positive (PCR)	Positive (rapid molecular)	Negative (PCR panel)
State Public Health Lab, RT-PCR result		Not tested	Positive (H5)	Not tested
CDC Lab, RT-PCR result		Not tested	Inconclusive	Not tested
Third specimen
Collection day, post- symptom onset		11	10	6
Specimen type		Upper respiratory	Lower respiratory	Upper respiratory
Clinic/hospital result (test type)		Not tested	Not tested	Negative (PCR panel)
State Public Health Lab, RT-PCR result		Positive (H5)	Positive (H5)	Not tested
CDC Lab, RT-PCR result		Positive (H5)	Positive (H5)	Not tested
Fourth specimen
Collection day, post- symptom onset				6
Specimen type				Lower respiratory
Clinic/hospital result (test type)				Positive (PCR panel)
State Public Health Lab, RT-PCR result				Positive (H5)
CDC Lab, RT-PCR result				Positive (H5)

Extended Data Table. 2. Confirmed and inferred genotype of influenza A(H5N1) viruses detected from case patients, United States, March 2024–May 2025.

Complete sequences of 8 gene segments are needed to fully group viruses into genotypes. Complete sequences were available for 27 viruses: 76% (22 viruses) were B3.13 genotype, 14% (4) were D1.1 genotype, and 3% (1) were D1.3 genotype. The genotype of viruses from the remaining case patients was inferred based on sequence similarities to other viruses identified in animals to which the case patient was exposed or presumed to be exposed, or exposure in the same setting (i.e., the same depopulation operation) where other case patients had complete genomes analyzed: 30 viruses were inferred as B3.13 and 13 were inferred as D1.1.

Primary exposure	B3.13 genotype	D1.1 genotype	D1.3 genotype
Cows	40 (77%)	1 (6%)	0 (0%)
Commercial Poultry	9 (17%)	14 (82%)	1 (100%)
Backyard Poultry	0 (0%)	2 (12%)	0 (0%)
Unidentified	3 (6%)	0 (0%)	0 (0%)
Total	52	17	1

Extended Data Table. 3.

Public database accessions for sequences from Table 3 with molecular markers of interest detected in influenza A(H5N1) viruses collected from ten case patients, United States, March 2024 to May 2025.

Virus Name	GISAID Accession	GenBank Accession HA	GenBank Accession NA	GenBank Accession PB2	GenBank Accession PB1	GenBank Accession PA	GenBank Accession NP	GenBank Accession MP	GenBank Accession NS
A/Texas/37/2024	EPI_ISL_19027114	PP577943	PP577945	PP577947	PP577941	PP577942	PP577946	PP577944	PP577940
A/Washington/251/2024	EPI_ISL_19531300	PQ591837	PQ591839	-	-	-	-	PQ591838	PQ591836
A/Washington/252/2024	EPI_ISL_19531301	PQ591833	PQ591832	-	-	-	-	-	-
A/Washington/253/2024	EPI_ISL_19531302	PQ591834	PQ591835	-	-	-	-	-	-
A/Missouri/121/2024	EPI_ISL_19413343	PQ327626	PQ327627	-	-	-	-	PQ327628	PQ327629
A/California/150/2024	EPI_ISL_19544645	PQ591824	PQ591831	PQ591825	PQ591827	PQ591826	PQ591830	PQ591829	PQ591828
A/Louisiana/12/2024	EPI_ISL_19634828	PQ809558	PQ809557	PQ809562	PQ809563	PQ809560	PQ809559	PQ809564	PQ809561
A/Iowa/124/2024	EPI_ISL_19669941	PQ885590	PQ885592	PQ885591	-	PQ885594	PQ885588	PQ885589	PQ885593
A/Nevada/10/2025	EPI_ISL_19726293	PV113341	PV113346	PV113343	PV113344	PV113342	PV113340	PV113345	PV113339
A/Wyoming/01/2025	EPI_ISL_19749443	PV177834	PV177836	PV177837	PV177835	PV177839	PV177841	PV177838	PV177840

Data availability

This is a small case series of human cases with HPAI A(H5N1) identified in the United States during March 2024–May 2025. To protect the privacy of the individuals included in this case series, we will not be able to provide additional patient-level data beyond what is provided in Supplemental Table 1. Aggregated case data are reported weekly and publicly available at CDC’s FluView (https://www.cdc.gov/fluview/index.html).