Adenovirus Types in US Children Hospitalized or Seen in the Emergency Department With Acute Respiratory Illness, 2016–2019

Abstract

Background

Human adenovirus (HAdV) is a common cause of pediatric acute respiratory illness (ARI), contributing to 5–13% of cases worldwide. Clinical manifestations vary by HAdV species and type; therefore, delineating type-specific disease presentations and understanding severity of specific HAdV types' disease may help develop targeted interventions.

Methods

We conducted a multicenter, prospective study within the New Vaccine Surveillance Network to characterize HAdV types. Children <18 years old with ARI were enrolled in the emergency department or inpatient setting at 7 US children's hospitals from 1 December 2016 to 30 November 2019. Respiratory specimens were collected and tested for HAdV and other viruses. Subsequently, typing was conducted on HAdV specimens using single-plex real-time PCR assays targeting sequences in the hexon gene. Comparisons between HAdV types were performed to determine differences in characteristics and outcomes. Generalized linear mixed effects models were used to evaluate severity.

Results

Among the 1843 HAdV-positive cases, 1402 specimens (76%) were typed. The most common types detected were HAdV-C1 (n = 439), HAdV-C2 (n = 393), and HAdV-B3 (n = 221). Children with HAdV-B7 (n = 78) had higher odds of severe outcomes compared to those with other HAdV types (aOR = 2.05; 95% CI: 1.24, 3.40). Symptom presentation varied across types within species B, C, and E; while all species had high frequency of upper respiratory symptoms, species B cases presented with a higher frequency of non-respiratory manifestations.

Conclusions

Among children with HAdV-positive ARI, those with HAdV-B7 had higher odds of severe outcomes. These findings suggest heterogeneity in clinical presentation and severity among HAdV types, emphasizing the importance of HAdV type in future prevention and treatment strategies.

In a large, three-year, multicenter surveillance study of US children with acute respiratory illness, the authors observed that human adenovirus (HAdV) type B7 was less common than other types while associated with higher odds of severe outcomes. Their findings demonstrate clinical heterogeneity of pediatric HAdV infections and highlight the importance of type-specific surveillance to inform prevention and treatment strategies.

Human adenovirus (HAdV) is a non-enveloped, double-stranded DNA virus able to infect a wide range of tissues. HAdV is classified into 7 species (A-G) based on biochemical, genetic, and hemagglutination properties [1–3]. Sequencing has further refined this classification using hypervariable regions within the hexon, fiber, and penton base genes. Over 100 distinct genotypes have been identified, which exhibit notable clinical heterogeneity, including acute respiratory illness (ARI), conjunctivitis, or diarrheal illness [4, 5].

Worldwide, HAdV causes 5–13% of respiratory tract infections in pediatric patients, contributing to ARI-associated morbidity and mortality [6–14]. Most HAdV ARI cases are caused by species B (types 3, 7, 11, 14, 16, 21), C (types 1, 2, 5, 6), and E (type 4). HAdV-associated ARI is generally self-limiting in healthy children; however, younger children and individuals with underlying medical conditions, including immunocompromise, are at increased risk of severe HAdV-associated illness, sometimes resulting in death [15–17]. Numerous studies have investigated the epidemiological, clinical, and genomic characteristics of respiratory HAdV species [6–13, 18–23]. Among these, HAdV-B (specifically types 3 and 7) has been associated with more severe illness compared to HAdV-C. HAdV-B7 infections, in particular, are associated with prolonged fever and require longer hospital stays compared to other HAdV types [20, 24].

Gaps remain in understanding the specific contributions of individual HAdV types to variations in clinical severity, presentation, and outcomes, as comprehensive, multi-center studies are limited. Therefore, we aimed to utilize data from a multi-center pediatric ARI surveillance study conducted in 7 US sites over 3 respiratory seasons to describe the demographic characteristics, clinical presentations, and outcomes (including risk factors for hospitalization and oxygen use) associated with various HAdV types. This study builds upon our previous work that examined HAdV species in the same cohort and provides higher-resolution insights into the epidemiology of HAdV in pediatric respiratory illness [25].

METHODS

Study Design

Prospective, active, population-based ARI surveillance was conducted in the Centers for Disease Control and Prevention (CDC)-sponsored New Vaccine Surveillance Network (NVSN). Enrollment occurred in 7 pediatric medical centers (see Supplementary materials) from December 2016 through November 2019. Enrollment restrictions in the emergency department (ED) to children under 5 years old were implemented at 3 sites during specific periods between December 2016-November 2019 (Seattle, WA: December 2016-April 2017, November 2017-April 2018; Pittsburgh: December 2016-April 2018; Kansas City: December 2016-April 2017, November 2017-April 2018). The study was reviewed and approved by CDC and study site institutional review boards (see 45 C.F.R. part 46.114; 21 C.F.R. part 56.114.). Parents or legal guardians provided written informed consent, and eligible children provided assent for participation, according to local regulations. This analysis was conducted using the same cohort and detection methods described in Probst et al [25] ensuring consistency in participant selection, specimen collection, and viral detection techniques.

Study Population and Data Collection

ARI Cases

Detailed enrollment criteria have been previously reported [26]. Briefly, children <18 years old who presented to the ED (restrained to <5 years old for specific sites as detailed above) or were admitted to the hospital were enrolled if they met eligibility criteria for ARI (ie, they resided in the catchment area of the site and had fever and/or an ARI symptom for <14 days).

Asymptomatic Controls

Control enrollment was targeted based on frequency matching by date of enrollment for ARI subjects or using a range such as date of enrollment +/− 2 weeks. Healthy controls (HC) 15 days to <18 years old were enrolled at each site's outpatient clinic(s). Patients were included as HC if they had no acute gastroenteritis symptoms within 14 days of enrollment or any ARI symptoms within 3 days of enrollment, were a resident of the catchment area, and were not already enrolled as a HC during the current surveillance season.

Data Collection

A standardized parent/guardian questionnaire was used to collect sociodemographic, clinical, and health information, including age, sex, race and ethnicity, smoking in the household, childcare or school attendance, other children living in the same household, insurance status, and underlying medical conditions (UMCs). For ARI cases, medical chart reviews were systematically conducted to record physical examination findings, UMCs, and disease severity outcomes such as hospitalization, length of stay in the hospital, oxygen support, intubation, and intensive care unit (ICU) admission. Symptoms at presentation were also collected and the frequency of symptoms by HAdV types are presented in Figure 1.

Figure 1.

Frequency (percent) of symptom presentations for each HAdV type, grouped by HAdV species (excluding HAdV types with n < 10), excluding cases with respiratory viral co-detection, from 1 December 2016 through 30 November 2019, New Vaccine Surveillance Network. The color gradient represents a scale of the frequency of reported symptoms; orange to red colors indicate more frequent symptoms (50–100%) and light green to dark green indicate less frequent symptoms (0–49%). Abbreviation: HAdV, human adenovirus.

Specimen Collection

Mid-turbinate nasal and, in some cases, oropharyngeal swab(s) were collected for research testing at each site. If a research specimen was unobtainable, alternative respiratory specimens such as a tracheal aspirate from intubated patients or salvaged respiratory specimens collected per standard of care during the same healthcare visit (eg, nasal wash, nasopharyngeal swab, or mid-turbinate nasal swab) were used. The research or clinical laboratory at each enrollment site employed CDC-approved, validated molecular methods to test specimens for HAdV and other respiratory viruses, collectively including influenza A, B, and C; respiratory syncytial virus (RSV); parainfluenza virus (PIV) types 1–4; human metapneumovirus (HMPV); seasonal human coronaviruses (HCoVs) types 229E, NL63, OC43, and HKU11; and human rhinoviruses/enteroviruses (RV/EV) [26]. HAdV-positive specimens were sent to Vanderbilt University Medical Center (VUMC), where typing for species C (1, 2, 5, 6), B (3, 7, 11, 14, 21), and E (4) was conducted using CDC-developed single-plex quantitative polymerase chain reaction (qPCR) assays targeting the hexon gene [27, 28]. The assay only tested for the specific aforementioned HAdV types.

Statistical Analyses

Demographic and clinical characteristics are presented as absolute and relative frequencies for categorical variables. Continuous variables are described using the mean (standard deviation; SD). Comparisons between HAdV types were conducted using logistic regression for binary outcome variables, multinomial logistic regression for multinomial outcome variables, and linear regression for continuous outcome variables, each including study site as a fixed effect and using cluster-robust standard errors (ie, generalized estimating equations with a working independence structure) with no other covariate adjustment to account for clustering by individuals contributing more than one observation to the data (n = 15). HAdVC-6, B11, B14, and B21 were excluded from type-specific comparisons due to their infrequent detection. For our main analyses, we compared HAdV types within the same species. Initial comparisons excluded any HAdV cases that were co-detected with other respiratory viruses. We refer to analyzed detections as “single HAdV detections.” Secondary comparisons included all HAdV detections, which may include co-detections with other respiratory viruses. A third model was performed as a sensitivity analysis, which excluded patients with non-HAdV co-detection but included patients with HAdV-type co-detection.

We used generalized linear mixed-effects models to evaluate factors associated with severity among children with HAdV-C1, C2, B3, C5, or B7. See Supplementary materials for model details. A sensitivity analysis was conducted to assess factors associated with severity in children under 5 years old, accounting for age variations in ED enrollment during specific periods of the study timeframe, as outlined in the study design. Statistical significance was assigned at the nominal α=0.05 level (2-sided, if applicable) for all analyses. Stata (version 17.0) was used for within-species comparisons and R version 4.4.0 was utilized for all other statistical procedures.

RESULTS

Enrollment and Population

From 1 December 2016, through 30 November 2019, 29 381 ARI cases were enrolled, of which 20 929 (71.2%) were positive for at least one respiratory pathogen (Supplementary Figure 1). HAdV was detected in 6.3% of all ARI enrolled and tested children (n = 1843). Of 1703 HAdV-positive specimens received by VUMC, 1402 (82.3%) specimens were successfully typed using qPCR. There were 1024 (73.0%) children with HAdV-C only, 312 (22.3%) with HAdV-B only, 18 (1.3%) with HAdV-E only, and 48 (3.4%) with a combination of HAdV-C, HAdV-B, and/or HAdV-E. Detections over time are displayed in Supplementary Figure 2. Over this same time period, 4626 HC were enrolled among all sites, and 3940 (85.2%) had a respiratory specimen collected. Of those with a specimen collected, 138 (3.5%) were identified as positive for HAdV, among whom 63 (45.7%) had samples that were successfully typed.

HAdV-type Frequency in ARI Cases

The most common type within the HAdV-C species was type 2 (n = 439, 42.9%), followed by type 1 (n = 393, 38.4%), type 5 (n = 117, 11.4%), and type 6 (n = 20, 2.0%); 55 (5.4%) children had combinations of HAdV-C1, C2, C5, or C6. The most common type within HAdV-B species was type 3 (n = 221, 70.8%), followed by type 7 (n = 78, 25.0%), type 14 (n = 7, 2.2%), type 21 (n = 2, 0.6%), and type 11 (n = 1, 0.3%), and there were 3 (1.0%) children with a combination of HAdV-B3, B7, B11, B14, or B21. HAdV-E4 was detected in 18 children. Co-detection frequencies with other respiratory viruses among HAdV types are presented in Figure 2 and Supplementary Table 1.

Figure 2.

Co-detections between HAdV types and other respiratory viruses; from 1 December 2016 through 30 November 2019, New Vaccine Surveillance Network. Abbreviations: CoV, human coronavirus; HAdV, human adenovirus; hMPV, human metapneumovirus; PIV, parainfluenza virus; RSV, respiratory syncytial virus; RV/EV, rhinovirus/enterovirus.

HAdV-type Frequency in HC

Among typed HAdV in HC samples, 60/63 (95.2%) were HAdV-C only, 1 (1.6%) was HAdV-B only, 1 (1.6%) was HAdV-E only, and 1 (1.6%) was HAdV-C and HAdV-B. The most common HAdV-C was type 2 (n = 27, 45.0%), followed by type 1 (n = 26, 43.3%), type 5 (n = 4, 6.7%), and type 6 (n = 1, 1.7%). Type 7 was detected in a single HC with HAdV-B.

HAdV Types Within HAdV Species C in Children With ARI

Within HAdV-C ARI cases, we compared children with types 1, 2, and 5, excluding those co-detected with a non-HAdV virus and those with more than one HAdV type detected (Table 1). Children with HAdV-C5 were, on average, slightly older than those with type 1 or 2 (2.4 vs 1.9 and 1.5 years old, respectively; P = .05). Otherwise, children with type 1, 2, or 5 did not differ in terms of other demographic or clinical characteristics, including hospitalization. The proportion hospitalized was 14.0%, 14.8%, and 15.6% for HAdV-C1, 2, and 5, respectively, but was higher when HAdV-C was co-detected with a non-HAdV respiratory virus (28.0%, 25.5%, and 35.9%, respectively; Supplementary Table 2).

Table 1.

Bivariate Comparisons of Types Within HAdV-C and HAdV-B Species, Without Other Respiratory Viral co-detection, From 1 December 2016 Through 30 November 2019, New Vaccine Surveillance Network

	HAdV-C n = 323				HAdV-B n  = 206
Characteristic	HAdV-C1, n = 136	HAdV-C2, n = 155	HAdV-C5, n = 32	P Valuea	HAdV-B3, n  = 150	HAdV-B7, n  = 56	P Valuea
Age, years—mean (SD)	1.9 (2.2)	1.5 (1.4)	2.4 (3.3)	.05	4.9 (3.4)	4.2 (4.1)	.76
Age group—n (%)				.01	…	…	.58
<12 m	41 (30.1)	71 (45.8)	14 (43.8)		14 (9.3)	10 (17.9)
12–59 m	88 (64.7)	78 (50.3)	14 (43.8)		77 (51.3)	29 (51.8)
≥5–17 y	7 (5.1)	6 (3.9)	4 (12.5)		59 (39.3)	17 (30.4)
Male—n (%)	71 (52.2)	90 (58.1)	22 (68.8)	.26	80 (53.3)	34 (60.7)	.23
Race, ethnicity—n (%)				.33			N/A
Non-Hispanic White	25/134 (18.7)	30/154 (19.5)	3 (9.4)		34/149 (22.8)	27 (48.2)
Non-Hispanic Black	63/134 (47.0)	76/154 (49.4)	14 (43.8)		44/149 (29.5)	17 (30.4)
Non-Hispanic Other	11/134 (8.2)	11/154 (7.1)	2 (6.3)		9/149 (6.0)	6 (10.7)
Hispanic	35/134 (26.1)	37/154 (24.0)	13 (40.6)		62/149 (41.6)	6 (10.7)
Smoking in household—n (%)	38 (27.9)	44 (28.4)	8 (25.0)	.87	27 (18.0)	18 (32.1)	.29
Childcare/school attendance—n (%)	63 (46.3)	72 (46.5)	16 (50.0)	.90	102 (68.0)	36 (64.3)	.60
Other children in the home—n (%)	91 (66.9)	99 (63.9)	20 (62.5)	.70	76 (50.7)	25 (44.6)	.20
Insurance status—n (%)				N/A			N/A
Public	106/134 (79.1)	121/152 (79.6)	25 (78.1)		97/148 (65.5)	43/54 (79.6)
Private	18/134 (13.4)	20/152 (13.2)	4 (12.5)		33/148 (22.3)	10/54 (18.5)
Both	0/134 (0.0)	2/152 (1.3)	0 (0.0)		4/148 (2.7)	0/54 (0.0)
None/self-pay	10/134 (7.5)	9/152 (5.9)	3 (9.4)		14/148 (9.5)	1/54 (1.9)
Underlying medical condition—n (%)	35 (25.7)	56 (36.1)	11 (34.4)	.29	64 (42.7)	23 (41.1)	.95
Hospitalized—n (%)	19 (14.0)	23 (14.8)	5 (15.6)	.98	41 (27.3)	33 (58.9)	.001
Length of stay—mean (SD)	2.6 (2.9)	2.6 (1.8)	1.4 (1.1)	.21	2.4 (2.4)	2.6 (2.4)	.36
Oxygen support—n (%)	4/19 (21.1)	8/23 (34.8)	2/5 (40.0)	.59	13/41 (31.7)	5/33 (15.2)	.59
ICU—n (%)	2/19 (10.5)	2/23 (8.7)	0/5 (0.0)	N/A	7/41 (17.1)	4/33 (12.1)	0.72
Intubated—n (%)	2/19 (10.5)	0/23 (0.0)	0/5 (0.0)	N/A	2/41 (4.9)	1/33 (3.0)	.41

Bold values indicate statistical significance.

Abbreviations: HAdV, human adenovirus; ICU, intensive care unit; SD, standard deviation.

^aComparisons between HAdV types were conducted using logistic regression for binary variables, multinomial logistic regression for multinomial variables, and linear regression for continuous variables, each including study site as a fixed effect and using cluster-robust standard errors (ie, generalized estimating equations with a working independence structure) with no other covariates included. N/A: not applicable; P-value could not be calculated due to zero cell counts.

HAdV Types Within HAdV Species B in Children With ARI

We compared children with HAdV-B3 to those with HAdV-B7, excluding any co-detection. Children with HAdV-B7 were more likely to be hospitalized than those with HAdV-B3 (58.9% vs 27.3%, respectively; P = .001; Table 1). There were no differences between the 2 HAdV-B types in terms of other clinical outcomes, such as oxygen support or length of hospital stay. When comparing all cases of HAdV-B3 and HAdV-B7, including co-detection with a non-HAdV virus, hospitalization remained higher among those with HAdV-B7 compared to HAdV-B3 (59.0% vs 30.8%; Supplementary Table 2).

Healthy Controls Compared to ARI Cases

Table 2 presents demographics, HAdV type frequency, and co-detections among ARI cases and HC. Relative to cases, HC were younger and were less likely to have co-detection with another virus (P < .001 for both). Furthermore, HAdV-C1 was detected in 41.3% specimens from HCs and 28.0% from patients with ARI; HAdV-C2 was detected in 42.9% of specimens form HC and 31.5% from patients with ARI (a P-value could not be calculated for this comparison). No HCs had HAdV-B3.

Table 2.

Bivariate Comparisons of Demographics and HAdV Type Detection Between Children With ARI and HC, From 1 December 2016 Through 30 November 2019, New Vaccine Surveillance Network

Characteristic	ARI Cases, n = 1402	HC, n = 63	P Valuea
Age (y)—mean (SD)	2.3 (2.6)	1.4 (0.9)	<.001
Male—n (%)	793 (56.6)	28 (44.4)	.06
Race and Hispanic origin—n (%)			.31
Hispanic	347/1385 (25.1)	10 (15.9)
Non-Hispanic Black	544/1385 (39.3)	30 (47.6)
Non-Hispanic other	139/1385 (10.0)	8 (12.7)
Non-Hispanic White	355/1385 (25.6)	15 (23.8)
Insurance—n (%)			N/A
Public	1007/1383 (72.8)	47/57 (82.5)
Private	267/1383 (19.3)	10/57 (17.5)
Both	16/1383 (1.2)	0/57 (0.0)
None, self-pay	93/1383 (6.7)	0/57 (0.0)
Daycare attendance—n (%)	787/1399 (56.3)	26 (41.3)	.02
HAdV Species—n (%)			<.001
B	312 (22.3)	1 (1.6)
C	1024 (73.0)	60 (95.2)
E	18 (1.3)	1 (1.6)
Co-detection	48 (3.4)	1 (1.6)
HAdV Type—n (%)			N/A
HAdV-C1	392 (28.0)	26 (41.3)
HAdV-C2	441 (31.5)	27 (42.9)
HAdV-B3	221 (15.8)	0 (0.0)
HAdV-E4	18 (1.3)	1 (1.6)
HAdV-C5	116 (8.3)	4 (6.3)
HAdV-C6	20 (1.4)	1 (1.6)
HAdV-B7	78 (5.6)	1 (1.6)
HAdV-B11	1 (0.1)	0 (0.0)
HAdV-B14	7 (0.5)	0 (0.0)
HAdV-B21	2 (0.1)	0 (0.0)
Co-detection	106 (7.6)	3 (4.8)
Pathogen co-detection—n (%)	784 (55.9)	20 (31.7)	<.001

Bold values indicate statistical significance.

Abbreviations: ARI, acute respiratory illness; HAdV, human adenovirus; HC, healthy controls; SD, standard deviation.

^aComparisons conducted using logistic regression for binary variables, multinomial logistic regression for multinomial variables, and linear regression for continuous variables, each including study site as a fixed effect and using cluster-robust standard errors with no other covariates included. N/A: not applicable; P-value could not be calculated due to zero cell counts.

Association of HAdV Types and Severity

We assessed the association between HAdV type and severe outcomes in children, initially modeling all HAdV cases, including those with respiratory virus co-detection. Our analysis focused on HAdV types C1, C2, B3, C5, and B7 to determine whether severe outcomes varied across the most frequently detected types. In the primary model (Table 3), HAdV-B7 was significantly associated with severe outcomes compared to children with other HAdV types (aOR 2.05; 95% CI 1.24–3.40). Additionally, those with underlying conditions had more than twice the odds of severe outcomes compared to those without (aOR 2.53; 95% CI 1.99–3.22). The sensitivity analysis of children under 5 showed that HAdV-C2 was associated with lower odds of severe outcomes (aOR 0.60; 95% CI .39–.93) whereas HAdV-B7 was associated with higher odds of severe outcomes (aOR 1.91; 95% CI 1.08–3.38) compared to children with other HAdV types (Supplementary Table 4). Similar to the primary model, having an underlying medical condition was also associated with severe outcomes (aOR 2.53; 95% CI 1.95–3.29).

Table 3.

Association Between HAdV Types and Severe outcomes ^a , Including Cases With Respiratory Viral co-detection, Among Children Enrolled From 1 December 2016 Through 30 November 2019, New Vaccine Surveillance Network (n = 1402) ^b

	aORc	95% CI	P Value
HAdV-types
HAdV-C1	0.84	.57, 1.23	.38
HAdV-C2	0.75	.51,1.10	.14
HAdV-B3	0.87	.55, 1.40	.57
HAdV-C5	1.01	.65, 1.55	.98
HAdV-B7	2.05	1.24, 3.40	.005
Age, y	1.00	.95, 1.06	.97
Sex, male	1.02	.80, 1.31	.84
Underlying medical condition	2.53	1.99, 3.22	<.001
Co-detection	0.40	.30, .52	<.001
Study site
Rochester	Reference
Houston	2.60	1.51, 4.50	.001
Seattle	0.55	.31, .96	.04
Cincinnati	0.62	.36, 1.05	.08
Kansas City	0.16	.08, .30	<.001
Pittsburgh	1.41	.90, 2.20	.14
Nashville	0.48	.29, .79	.004

Bold values indicate statistical significance.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; HAdV, human adenovirus.

^aSevere outcome defined as a 6-level ordinal outcome. 0: discharged from the emergency department; 1: hospitalization with no oxygen use; 2: hospitalization with oxygen use; 3: hospitalization with intensive care admission; 4: hospitalization with intensive care admission and intubation; 5: death.

^bIndividual patient identifier included as a random effect.

^cThe aOR indicates the odds of a one level increase on the ordinal scale. The aOR was estimated from a generalized linear mixed effects model with a 6-level ordinal outcome.^a Covariates in the model included age (in years), sex, a binary indicator of one or more underlying medical condition(s), study site (as a random effect), and co-detection with a non-HAdV virus.

In contrast, co-detection with another non-HAdV virus was associated with lower odds of severe outcomes (aOR 0.40; 95% CI .30–.52). RV/EV accounted for 56% of co-detections with a non-HAdV virus. To assess the robustness of these findings, we conducted secondary analyses (Table 4 and Supplementary Table 3). These results did not indicate association of specific HAdV type detections with hospitalization.

Table 4.

Association Between HAdV Types and Severe outcomes ^a , Excluding Cases With Respiratory Viral Co-detection or Co-detection of Multiple HAdV Types, Among Children Enrolled From 1 December 2016 Through 30 November 2019, New Vaccine Surveillance Network (n = 553) ^b

	aORc	95% CI	P Value
HAdV-types
HAdV-C1	0.80	.17, 3.78	.78
HAdV-C2	0.65	.14, 3.10	.59
HAdV-B3	1.18	.27, 5.20	.83
HAdV-C5	0.77	.13, 4.52	.77
HAdV-B7	3.34	.71, 15.63	.13
Age, years	1.01	.93, 1.09	.83
Sex, male	1.44	.92, 2.27	.11
Underlying medical condition	3.03	1.98, 4.64	<.001
Study site
Rochester	Reference
Houston	2.27	1.03, 4.99	.04
Seattle	0.38	.13, 1.17	.09
Cincinnati	0.39	.15, 1.03	.06
Kansas City	0.15	.04, .58	.005
Pittsburgh	1.42	.68, 2.96	.35
Nashville	0.51	.022, 1.17	.11

Bold values indicate statistical significance.

Abbreviations: aOR, adjusted odds ratio; CI, confidence interval; HAdV, human adenovirus.

^aSevere outcome defined as a 6-level ordinal outcome. 0: discharged from the emergency department; 1: hospitalization with no oxygen use; 2: hospitalization with oxygen use; 3: hospitalization with intensive care admission; 4: hospitalization with intensive care admission and intubation; 5: death.

^bIndividual patient identifier included as a random effect.

^cThe aOR indicates the odds of a one level increase on the ordinal scale. The aOR was estimated from a generalized linear mixed effects model with a 6-level ordinal outcome.^a Covariates in the model included age (in years), sex, a binary indicator of one or more underlying medical condition(s), and study site (as a random effect).

Symptomatology of HAdV Types

The clinical presentation of each HAdV type (single detections only) was examined (Figure 1). Fever, congestion, cough, loss of appetite, and irritability were predominant symptoms for each type. The symptom presentation varied between HAdV-B3 and HAdV-B7 cases. Upper respiratory symptoms, such as fever, congestion or runny nose, cough, fatigue, were the dominant feature, though 23–62% of cases experienced additional manifestations such as chills, headache, abdominal pain, or conjunctivitis. Some symptoms were more prevalent among patients with HAdV-B7 compared to HAdV-B3, such as cough (85.7% vs 74.7%), loss of appetite (92.9% vs 78.7%), fatigue (87.5% vs 78.7%), lethargy (23.2% vs 10.7%), and seizure (7.1% vs 4.0%).

Within species C, all evaluated types (1, 2, 5, and 6) were associated with upper respiratory symptoms, though fever was the most prevalent symptom in over 90% of cases for each type. Congestion was reported in over 81% of cases, followed by cough in ≥71%. The frequency of vomiting varied slightly among different types; vomiting was lower in HAdV-C5 cases (9.4%) compared to HAdV-C1 (27.9%) and HAdV-C2 (27.7%) cases. HAdV-E4 manifested as fever, cough, and loss of appetite. The frequency of other symptoms was low; however, the sample size was small.

DISCUSSION

In this prospective, population-based surveillance study, we characterized HAdV types and their associations with clinical presentation and severity in a large US pediatric cohort with ARI. Roughly 6% of ARI cases were HAdV-positive, with types C2, C1, and B3 most commonly detected. Our prior work using the same cohort revealed increased hospitalization risk with HAdV detection [25]. The current analysis details specific HAdV types and respective clinical associations, including asymptomatic shedding and increased severity for certain types, such as HAdV-B7. Notably, almost one-third of HAdV-infected children in our cohort were hospitalized, with an increased risk of hospitalization observed for those infected with HAdV-B, particularly HAdV-B7. Children with UMCs who tested positive for HAdV experienced more severe outcomes than children without UMCs, emphasizing the need for clinical vigilance in these vulnerable populations. Our findings highlight important aspects of HAdV infection in pediatric populations, including the prevalence and clinical severity of different HAdV species and types, type-association with distinct symptom profiles, and the impact of UMCs on disease outcomes.

HAdV-C was the most prevalent species across the 3 respiratory seasons included in the study, with HAdV-C2 and HAdV-C1 most common. These findings align with pre-COVID-19 literature. A Japanese study of outpatient children <13 years old reported HAdV-C as the most common species (37.1%), with types 1 and 2 the most frequent [29]. Similarly, a US study involving inpatient and outpatient children found that HAdV-C accounted for 50% of cases, with types 1 and 2 most common [15]. A study of outpatient children aged 6–48 months in Argentina showed that 81% of HAdV-positive specimens were species C [30]. However, more recent data suggest shifts in species and type prevalence since the emergence of SARS-CoV-2. A 2023 US study reported a rise in HAdV positivity rates and a predominance of HAdV-B3 observed during the 2020–2022 timeframe [31]. A similar shift was seen in an Italian study conducted from 2022 to 2023 [32].

Our study established that HAdV-B, specifically type 7, was associated with greater illness severity, consistent with prior observations in pediatric patients [10]. We found higher odds of severe outcomes in HAdV-B7 infections compared to other HAdV types, though some of these associations were not statistically significant. Notably, only one of all HAdV-B7 detections was in a HC. Previous outbreaks of severe and fatal pneumonia in children caused by HAdV-B7 have been reported [23]. For instance, a study of hospitalized children <14 years old from China showed that HAdV-B7 was more severe than HAdV-B3 based on length of hospital stay, respiratory support, and ICU admission [33]. Other studies from China and the US have reported that HAdV-B7 is associated with more severe disease, such as pneumonia, toxic encephalopathy, and respiratory failure [34, 35]. While we did not observe significant differences in oxygen support or ICU admission between HAdV-B3 and HAdV-B7 cases, HAdV-B7 infections were more severe by other measures, including symptoms. However, in our secondary analysis excluding cases with co-infections (Table 4), the association was no longer significant, suggesting that co-infecting pathogens may influence severity linked to HAdV-B7. Further research is needed to clarify the independent and corporate roles of HAdV-B7 in the severity infections involving co-pathogens. A live, oral vaccine against HAdV types E4 and B7, currently used in US military recruits, has demonstrated efficacy in reducing severe adenovirus infections [36, 37]. A pediatric vaccine targeting the most clinically relevant HAdV types has the potential to decrease the incidence of symptomatic or severe infections [38]. Furthermore, anti-adenoviral treatment options are severely limited, which compounds difficulties in the prevention and management of adenoviral disease and underscores the importance of developing adenovirus vaccines for use in the pediatric and other special (eg, immunocompromised) populations [39].

We observed that common symptoms such as fever, cough, and congestion were prevalent across all HAdV types, though the frequency of other symptoms varied. HAdV-C primarily caused upper respiratory symptoms, while children with HAdV-B experienced a broader range of symptoms, including conjunctivitis. Previous studies have reported variation in presenting symptoms between species and types. For example, a pediatric study in China found that cough, rhinorrhea, congestion, and exudative tonsilitis were more frequent with HAdV-B3, while HAdV-B7 was associated with prolonged fever and lower activity level [24]. A previous study at VUMC revealed a higher likelihood of conjunctivitis in species B infections, whereas species C was differentially associated with seizures—an observation not replicated in the current study [15]. These distinct clinical presentations underscore the heterogeneity of HAdV infections and highlight the importance of specific HAdV types when considering outbreaks and public health impact. However, the high prevalence of HAdV, especially species C types, among asymptomatic children suggests that shedding is common, and detection alone does not necessarily indicate causation of illness.

Our study has limitations. First, HAdV type distribution was restricted by breadth of the qPCR panel, and less common or novel HAdV types might have gone undetected. Second, our findings may not be fully generalizable to all US children due to regional variation among sites. The differences in severe outcomes observed between study sites may reflect geographic variation in healthcare practices, population demographics, or modest differences in type prevalence (eg, a higher proportion of AdV-7 in Pittsburgh). However, type distributions were relatively consistent across sites, suggesting that prevalence alone is unlikely to account for the observed inter-site variation. However, NVSN is one of the largest pediatric surveillance platforms in the US and includes data from 7 academic children's hospitals with broad catchment areas. Third, 3 sites in the NVSN restricted ED enrollment to children <5 years of age during certain time periods; this skewed the cohort to a younger age, thus the number of cases in age groups and cases by month are not fully representative. Fourth, the possibility of asymptomatic HAdV shedding, in particular species C known for its propensity for persistent infections, could have influenced study results and interpretations. Fifth, the role of respiratory pathogen co-infection in the manifestations of HAdV infection is incompletely understood. We observed milder HAdV-associated ARI in the presence of certain co-infections, but the differential increase in overall HAdV-B7 severity failed to reach statistical significance when co-infections were excluded, though these factors were modeled separately. A previous study found that HAdV infections were associated with more frequent severe outcomes with certain coinfections [26]. However, not all studies on viral coinfections have reported such differences. For example, a meta-analysis of over 4000 patients and another separate study found no differences in severe outcomes between single viral infections and coinfections [40, 41]. These findings might point to unique interactions between particular HAdV types and other respiratory viruses in ARI pathogenesis, and further research is warranted to explore these relationships epidemiologically and mechanistically.

In conclusion, our study highlights heterogeneity in clinical presentation and severity of HAdV types detected in children. While HAdV-B7 was associated with higher hospitalization rates, it did not result in more severe outcomes in hospitalized patients, suggesting that other factors, such as immune status and medical interventions, might play a role. The distinct symptom profiles and disease outcomes observed for infections by different HAdV types suggest that tailored clinical management strategies may be useful, particularly in vulnerable populations. Continued surveillance and research into mechanisms of HAdV ARI and its interaction with other respiratory pathogens are required to improve prevention and treatment strategies in the pediatric population.

Supplementary Data

Supplementary materials are available at Open Forum Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.