Methods

Case finding and study populationThe study population included all patients in whom Ad infection was identified from 1 November 2006 through 31 July 2007 by any of the 4 Oregon laboratories that conduct viral isolation or polymerase chain reaction (PCR) testing for Ad. In addition, we asked clinicians and infection-control practitioners from Oregon and southwest Washington State to report cases

Definitions of a case patient and severe illnessWe defined a case patient as a person with a clinical illness and a positive Ad laboratory test result from 1 November 2006 through 31 July 2007; cases originally detected by direct florescent antibody testing alone were confirmed by viral isolation or PCR assay. We defined severe illness as a case that resulted in the admission of the patient to an intensive care unit (ICU) or death

Definitions of normal physiological signs, by ageBecause the case patients with Ad infection ranged in age from the very young (newborns) to the very old (elderly persons), we used the following specific definitions for normal physiological signs [10]. Elevated respiratory rate was defined as >60, 40, 30, 25, and 20 breaths per minute for ages <6 weeks, 6 weeks to <6 months, 6 months to <3 years, 3 years to 6 years, and >6 years, respectively. Similarly, decreased systolic blood pressure was defined as <50, <70, <80, <80, and <90 mm Hg for the same age groups, respectively. Decreased oxygen saturation level was defined as <93% for all age groups

Study designWe conducted a retrospective descriptive study of the clinical, laboratory, and radiographic characteristics of patients infected with Ad during the study period. All available medical records were reviewed by 1 of 4 clinicians, using a standardized data collection tool. We attempted to gather information on medical history and possible exposures through medical record review and interview of patients or their caregivers when possible. The medical record review was conducted under Oregon’s special studies statute for issues of public health significance. In addition, a convenience sample of all available archived Ad isolates from 1993 through 2007 stored at the Oregon State Public Health Laboratory (OSPHL) and the Providence Portland Infectious Disease Laboratory was submitted to the Centers for Disease Control and Prevention (CDC) for Ad typing

Ad typingAd isolates or PCR-positive clinical specimens were typed on the basis of sequences of a partial region of the hexon gene, as described elsewhere [11]. During the outbreak, an Ad14-specific real-time PCR assay was developed at the CDC and was used by the OSPHL to facilitate rapid testing of clinical specimens. Primer and probe sequences were designed from alignments of the hexon gene from all currently recognized Ad types. Ad14 primer and probe sequences and locations (see GenBank accession number AY803294) were as follows: forward primer, 5′-GAAAATCATGGTGTGGAAGATGAA-3′ (nt 19443–19466); reverse primer, 5′-CAAGCTTGGTCTCCATTTAACTGA-3′ (nt 19552–19521); and probe, 5′-ACGGCATCGGTCCGCGAACA-3′ (nucleotides [nt] 19492–19591). The probe was labeled at the 5′ end with 6-carboxy-fluorescein and was quenched at the 3′ end with Black Hole Quencher–1 (Biosearch Technologies). The Ad14 real-time PCR was performed using iQ Supermix (Bio-Rad), with each 25-μL reaction containing 0.5 μmol/L forward and reverse primer, 0.1 μmol/L probe, and 5 μL of nucleic acid extract. Thermocycling was performed using either an iCycler iQ Real-Time Detection System (Bio-Rad), an ABI PRISM 7000 Sequence Detection System (Applied Biosystems), or an ABI 7500 Fast Real-Time PCR (Applied Biosystems), with the following thermocycling conditions: 3 min at 95°C for activation of the iTaq DNA polymerase, 45 cycles of 15 s at 95°C, and 1 min at 60°C. Each run included appropriate positive and no-template controls, and findings for select samples with positive results were confirmed by hexon gene sequencing, as described above

Statistical analysisMedical record and interview data were stored in an Access database (Microsoft) and were analyzed using SAS software (version 8.0; SAS Institute). With the exception of age, for which we reported the median and range in years, we reported the median and interquartile range for continuous variables. The Wilcoxon&amp;rank sum test was used to compare the distributions of continuous variables between those with and those without severe disease. Categorical variables are reported as percentages in each category. The χ² test was used to compare categorical variables. Logistic regression was used to assess associations between severe disease and clinical factors in univariate and multivariate analyses

Results

Emergence of Ad14 in OregonWe screened 270 Ad isolates archived by Oregon laboratories from 1993 through 2007 by the Ad14 real-time PCR assay. Before 2005, none of 47 available isolates were type 14. In 2005, 2006, and 2007, 7 (17%) of 42, 29 (51%) of 57, and 68 (55%) of 124 Ad isolates were type 14, respectively, suggesting that this serotype emerged recently as the dominant local serotype in Oregon. Of the 56 non–type 14 Ads available for complete typing, 21 were species B (18 were type 3, 1 was type 7, 1 was type 11, and 1 was type 21), 27 were species C (11 were type 1, 15 were type 2, and 1 was type 6), 2 were species D (1 was type 25, and 1 was type 29), and 3 were species E. Three could not be typed but were not Ad14

During the study period from 1 November 2006 through 31 July 2007, a total of 91 specimens tested positive for Ad. Of these, 74 (81%) were available for confirmatory testing, and 7 (8%) subsequently tested negative for Ad. Of the 67 remaining specimens, 40 (60%) were identified as Ad14, and 27 (40%) were identified as other types. Most case patients (58%) became ill between February and May 2007 (figure 1)

Demographic characteristics and underlying illnessMedical records were available for review for 38 (95%) of the patients with Ad14 infection and for 24 (89%) of the patients with non–type 14 Ad infection. Although Ad14 afflicted patients across the life span, 23 (61%) of the 38 patients with Ad14 infection were >40 years of age (figure 2). In contrast, 18 (75%) of the 24 non–type 14 Ad cases occurred in children <5 years of age, and 11 (46%) occurred in patients <1 year of age. Age, sex, chronic underlying condition, history of smoking, and health outcome among patients with Ad infection are shown in table 1. A substantial majority of the patients with Ad infection, regardless of virus type, were male. Overall, nearly half (47%) of the patients with Ad14 infection reported 1 or more chronic underlying conditions; this proportion rose to 76% among the patients with Ad14 infection who had severe disease, but no specific condition predominated. Notably, 60% of adult patients with Ad14 infection reported having smoked cigarettes during the previous 30 days. Few of the mostly young patients with non–type 14 Ad infection reported having a chronic underlying condition. The remainder of this descriptive study focuses on the patients with Ad14 infection, because there were too few patients with non–type 14 Ad infection of any specific type to provide a large enough group for comparison; additionally, the dramatic difference in age distribution between the patients with Ad14 infection and those with non–type 14 Ad infection limits meaningful comparison of laboratory and underlying-condition variables

Presenting signs and symptomsSymptoms reported at the first medical encounter for patients with Ad14 infection are detailed in figure 3. Fever and cough were the most common complaints, followed by shortness of breath, vomiting, and diarrhea. Initial values for physiological signs and oxygen saturation levels were available for most hospitalized patients and are summarized in table 2. Most hospitalized patients were tachypneic, and nearly half had a documented elevated temperature, a decreased oxygen saturation level, or both on admission. Although 9 (31%) of 29 hospitalized patients ultimately required blood pressure support during the course of their illness, documented low blood pressure was uncommon at the time of admission

Radiograph findingsTwenty-six (90%) of the 29 hospitalized patients with Ad14 infection had abnormal chest radiograph findings on admission (table 2). Radiographic patterns included single lobe (14 patients [54%]) and multilobe (10 [38%]) infiltrates; interstitial infiltrates (3 [12%]) and pleural effusion (4 [15%]) were less commonly noted. Of those who initially had single lobe infiltrates, 10 (71%) of 14 progressed to multilobe involvement. Representative chest radiographs from 2 patients (one 18 and one 56 years of age) are shown in figure 4

Laboratory findingsInitial laboratory findings for patients with Ad14 infection who were >15 years of age are summarized in table 3. Lymphocyte counts of ⩽1000 cells/μL and creatinine levels of ⩾1.2 mg/dL were common, but anemia, neutropenia, thrombocytopenia, and elevated serum transaminase enzyme levels were not. Few children <15 years of age had laboratory studies performed

TreatmentAll hospitalized patients received supportive care, and 25 (86%) of 29 received broad-spectrum antibacterial agents, usually directed at suspected community-acquired pneumonia–causing pathogens. Of the 29 hospitalized patients, 6 (21%) received cidofovir, of whom 4 (67%) survived. Clinicians commented that the number of doses of cidofovir was limited for each recipient by nephrotoxicity

OutcomeThe outcomes for the case patients in this series are summarized in table 1. The majority of the 38 patients with Ad14 infection required hospitalization and supplemental oxygen; many additionally required admission to an ICU and mechanical ventilation. Seven (18%) died (including 2 patients with terminal malignancy who received early palliative care). The median length of hospital stay among patients admitted for Ad14 infection was 9 days (range, 1–42 days); if critically ill, the median ICU stay was 7 days (range, 2–32 days). Of note, severe disease and death also occurred among those infected with other Ad types, including types 1, 2, 11, and 21

Risk factors for poor outcome among patients with Ad14 infectionNineteen (50%) of 38 patients with Ad14 infection met our definition for severe disease, with 17 (89%) of these occurring in persons >15 years of age. Univariate analyses of demographic, underlying-condition, and laboratory variables showed that older age (odds ratio [OR], 1.04 per year of age [95% confidence interval {CI}, 1.01–1.06]), any chronic underlying condition (OR, 6.06 [95% CI, 1.49–24.76]), lymphocyte count of <1.0×10⁹ cells/L (OR, 9.1 [95% CI, 1.95–42.89]), and serum creatinine level of >1.2 mg/dL (OR, 7.6 [95% CI, 1.37–42.71]) were associated with severe disease, but these could not be distinguished as independent risk factors in multivariate analysis

Discussion

Ad14 was first isolated from Dutch military recruits in 1955 [12], and through the 1960s it caused several outbreaks of pharyngoconjunctival fever and mild respiratory illnesses among civilians and military recruits in Europe and China [13–16]. Ad14 has since been infrequently identified in Eurasia [17, 18], but cases were only recently reported in the United States [1, 19–21]. Recent surveillance surveys of US civilian and military populations have confirmed the emergence of Ad14 in 2005 and 2006 [5, 22]. Metzgar et al. [22] documented the simultaneous appearance of Ad14 ARD in April and May 2006 at military training centers located throughout the United States, although no hospitalizations or deaths were reported. During this period, 2 young children and an unrelated civilian adult with severe Ad14 respiratory disease were identified [23]. All had underlying lung disease but were immunocompetent; the adult died, and both infants had severe, prolonged illnesses. In 2007, Ad14 was reported to be the cause of a large, sustained outbreak of febrile respiratory illness among military trainees in Texas; an outbreak in a residential care facility in Washington State; and the community-wide outbreak in Oregon described here [24]

We reviewed the medical records for all documented Ad cases during the study period in our area to systematically describe the clinical spectrum and severity of laboratory-confirmed Ad infections. We found that Ad14 often caused serious respiratory infection; more than three-quarters of the identified patients required hospitalization, 62% of these needed critical care, and nearly 20% died. Most adults with Ad14 infection had a recent history of cigarette smoking, a potentially important cofactor for acquiring disease. A predominance of males among patients with Ad infection has been noted previously [7, 25] but is of unknown significance. Although the present study focuses on Ad14, our methods captured illnesses caused by any Ad type, and we note that other types also cause critical illness and death

Nearly all patients with Ad14 infection presented with fever. Adults frequently complained of shortness of breath, cough, and gastrointestinal symptoms; no conjunctivitis was noted. None of these symptoms adequately distinguish Ad infection from the many other causes of community-acquired pneumonia. Notably, neither productive cough suggestive of bacterial pneumonia nor conjunctivitis suggestive of Ad infection was commonly noted. In addition, no single laboratory finding is likely to distinguish Ad14 infection from other causes of acute pneumonia, although we observed that many of the sicker patients had normal white blood cell counts but low lymphocyte counts

We identified a number of features strongly correlated with poor outcome. Older age and underlying disease are plausible factors that may explain greater susceptibility, lower immune response, or poor end-organ reserve. The abnormal renal function measured on admission of most severely ill patients may be a marker of preexisting renal dysfunction. Alternatively, poor renal function could be the consequence of either poor kidney perfusion during periods of low blood pressure or direct kidney infection by Ad. Depressed total lymphocyte counts were most likely a consequence of infection and, although not specific to Ad, could provide an early clue to poor prognosis

This retrospective study does not allow any conclusions about the efficacy of treatment of severe Ad infection with cidofovir. Clinicians used this agent only in critically ill patients and typically associated its use with worsening renal function. Renal insufficiency on admission among the critically ill patients in this series was common, and, as mentioned above, the causes were likely multifactorial but could be worsened by using nephrotoxic medications, such as cidofovir

Large community and regional outbreaks of ARD with an increased incidence of severe and fatal outcomes caused by a related species B Ad, type 7, have been described in Europe and Asia [17, 26, 27]. In these outbreaks, a preponderance of cases occurred among young children. In the United States, outbreaks of Ad ARD have been common among unvaccinated military training populations but have been rarely detected among otherwise healthy adults in the civilian sector. In 1995, Ad35 caused a cluster of severe respiratory illnesses among adult residents and staff of a Rhode Island chronic psychiatric facility; nosocomial transmission was suspected. Patients presented with fever, cough, lobar infiltrates, and low-to-normal white blood cell counts, similar to our group of critically ill patients with Ad14 infection [9]. Of the 24 case patients, 18 required hospitalization, and 6 became critically ill. Of the 6 critically ill patients, 4 had chronic underlying illness, including 3 with pulmonary disease who recovered and 1 with renal failure who did not. In March 1997, an outbreak of Ad11 ARD occurred among students and staff at a South Dakota high school [7]. Of 146 students given a diagnosis of ARD, 43 had lower respiratory tract infections; of these, 5 were hospitalized, and 1 required intensive care

In contrast to these outbreaks, in which the institutional settings may have contributed to the exposure and spread of the virus, our cluster occurred among unrelated community-dwelling individuals over a 9-month period, predominantly in late winter and early spring. Given the concurrent outbreak reported at a military base in Texas, we sought links to military bases or contact with military personnel among our case patients but failed to identify any. We also failed to identify new links or exposures that have been implicated in previously reported outbreaks of Ad, such as day care or contaminated bodies of water. Taken together, these findings suggest that transmission of Ad14 was widespread in Oregon during the study period

Despite the notorious resistance of Ad to disinfection and the high rate of hospitalization, we identified only 1 possible instance of health care–associated transmission, from a patient to an ICU employee. Once Oregon clinicians were informed of this cluster of severe Ad illness, suspected hospitalized patients were typically cared for with both droplet and contact precautions to prevent further spread

Until recently, Ad14 had not been associated with severe disease, and we hypothesize that the currently circulating strain has either become more transmissible, more virulent, or both. Although the more-widespread availability of sensitive virologic testing using PCR techniques may contribute to increased detection of this and other respiratory pathogens, the historical data presented here distinctly show that this virus emerged in Oregon in 2005. Multiple sequences of Ad14 isolates from Oregon and geographically distant regions were identical but differed from the 1955 reference strain; these differences may lead to the identification of virulence factors responsible for our observations (D. Erdman, personal communication). Ad14 has persisted at a slightly lower frequency in Oregon after the conclusion of this study, accounting for 29 (29%) of 101 patients with Ad infection tested between November 2007 and July 2008

The present study has at least 3 significant limitations. First, one institution reported 49% of the Ad cases overall and 66% of the Ad14 cases in adults, possibly because of a laboratory protocol that reflexively tested respiratory specimens for multiple viral pathogens. In addition, clinicians at this particular institution became aware of the cluster of severe Ad infection before others in the community and thus may have ordered viral testing more often. Therefore, the cases we reviewed are not likely to be representative of all cases of Ad14 infection occurring in Oregon during the study period. Second, the frequency of critical illness reported in this series may reflect a bias toward more-extensive laboratory testing among patients with more-severe illness; we suspect that most Ad cases are never tested. Third, this retrospective medical record review did not allow us to ascertain all possible risk factors or to obtain standardized testing in every case

In conclusion, we report the emergence of Ad14 as the dominant Ad type detected in Oregon beginning in 2005. During late 2006 through mid-2007, we identified a community-based cluster of acute febrile respiratory illness caused by Ad14, in which disease was frequently complicated by pneumonia and respiratory failure. Efforts are needed to better define the risk factors for severe Ad disease, to develop consistent protocols to trigger viral testing, to study the therapeutic options for severe Ad infections, and to support surveillance for emerging viral pathogens