We defined a case-patient as a person who had 1) illness onset June 1–October 31, 2012, clinically compatible with hantavirus disease; 2) laboratory confirmation of SNV infection by detection of IgM against SNV or increasing titers of IgG in serum, SNV-specific antigen shown by immunohistochemical test (IHC), or SNV-specific nucleic acid sequences by reverse transcription PCR (RT-PCR); and 3) a history of visiting Yosemite during the 6 weeks preceding illness onset. HPS was defined as fever ≥38.3°C with bilateral diffuse interstitial edema and respiratory compromise within 72 hours of hospitalization, occurring in a previously healthy person, or an unexplained respiratory illness resulting in death, for which an autopsy demonstrated noncardiogenic pulmonary edema without an identifiable cause (15).

Finding case-patients was facilitated through health alerts issued by CDPH, NPS, and CDC to health care providers, advising them to contact state or local public health officials to obtain confirmatory testing of patients with suspected HPS who had visited Yosemite. Commercial diagnostic laboratories that detected hantavirus antibody in submitted serum specimens were asked to notify and forward specimens to state public health laboratories or CDC for confirmatory testing. State health departments that identified the death of a previously healthy person who had a history of travel to Yosemite were asked to collect tissue specimens for confirmatory testing by CDC. CDPH and NPS issued media releases recommending that persons who experienced febrile illness after travel to Yosemite consult with their health care provider about possible hantavirus infection. Notifications were sent to all visitors registered as overnight guests at Yosemite during June 1–September 12, 2012, advising them to seek medical attention if they experienced symptoms of hantavirus infection.

Acute- or convalescent-phase serum specimens were tested at CDPH’s Viral and Rickettsial Diseases Laboratory or CDC’s Viral Special Pathogens Laboratory for IgM and IgG against SNV by using the same ELISA (7). Four-fold dilutions were performed (1:100–1:6400), and titers ≥400 were considered positive. Formalin-fixed tissues from patients who died were tested for the presence of hantavirus antigen by IHC at CDC’s Infectious Diseases Pathology Branch (16). A nested RT-PCR (17) was performed by using RNA extracted from frozen tissue. A TaqMan assay (Roche Molecular Systems, Inc., Pleasanton, CA, USA), targeting the small (S) segment RNA from North American hantaviruses, was used on RNA samples extracted from fixed tissues. This assay was used following standard protocols (SuperScript III Platinum One-Step qRT-PCR Kit w/ROX; Invitrogen, Carlsbad, CA, USA) with the following primers and probes: 143F 5′-TGGACCCIGATGAYGTTAACAA-3′, 303R 5′-ATCAGGITCAAKCCCIGTTGG-3′, and NA-181P 5′-FAM-AGACGGGCIGCTGTGTCTGCATT-BQH-3′. Medical records for laboratory-confirmed HPS patients were reviewed for relevant clinical information.

Patients with confirmed cases (or a proxy family member if the patient was deceased) and their non-ill travel companions who had shared lodging with the patient were interviewed by telephone, using a standardized questionnaire. One non-ill companion was interviewed for each patient identified. By using closed and open-ended questions, patients and non-ill companions were asked about demographic characteristics, illness history, travel, and other activities in the weeks preceding illness onset. Participants were asked about the type of lodging and activities they experienced during their Yosemite visit, particularly actions (e.g., cleaning or sweeping), observations (e.g., mouse droppings), and behavior (e.g., sleeping on the floor) potentially associated with exposure to rodents and SNV. Data from questionnaires were maintained in a spreadsheet by using standard software (Excel; Microsoft, Redmond, WA, USA). The incidence of SNV infection among guests staying in a signature tent cabin was compared with that of guests staying in other lodgings in Curry Village, according to Yosemite guest registry records for June 1–August 28, 2012. A difference was considered statistically significant if p≤0.05.

Yosemite lodgings where patients had stayed and other nearby buildings were visually evaluated for evidence of active or recent rodent activity (e.g., nesting materials, feces, or urine stains) as well as structural features that might allow rodent entry. Rodent populations in and around Yosemite guest lodgings were live-trapped for evaluation. Sherman live-traps (H.B. Sherman Traps, Tallahassee, FL, USA) were baited with rolled oats and placed overnight inside and outside buildings. Rodent abundance was estimated by trap success, defined as the ratio of captured mice to the total number of traps set. Captured rodents were anesthetized, euthanized, measured, and identified to species. A minimum of 11 μL of blood was collected from the retrobulbar sinus of each mouse and sent to CDPH for ELISA testing for IgG against SNV.

Ten case-patients were identified among persons who had visited Yosemite overnight during summer 2012. The median age of the 10 patients was 44.5 years (range 12–56 years). Eight were residents of California, 1 of Pennsylvania, and 1 of West Virginia. Infection was confirmed by serologic testing for 8 patients and by IHC and RT-PCR for 2 patients. A 540-bp fragment amplified from fresh frozen tissues of 1 case-patient who died had high homology (>90%–99% identity) with the known SNV isolates Convict Creek, NM R11, and NM H10 (GenBank accession nos. AF425256.1, L37902.1, and L37901.1, respectively).

Salient case-patient clinical characteristics are displayed in the Table. Eight patients had respiratory signs or findings compatible with HPS, 3 of whom died. Seven HPS case-patients were hospitalized for a median of 7 days (range 2–30 days); 5 required intensive-level care with ventilatory support.

The illnesses of 2 case-patients did not fulfill the complete HPS definition. One patient had been examined in an emergency department for fever, cough, and shortness of breath and was discharged and recovered. That patient’s chest radiograph and computed tomographic scan revealed no evidence of pulmonary infiltrates or edema. The second patient did not report fever or pulmonary symptoms but had severe headache, nausea, vomiting, and diarrhea. Both patients were tested for SNV retrospectively after they heard about the Yosemite outbreak in the news and had detectable antibodies against SNV.

The onset of illness for the 10 case-patients occurred from July 2 through August 16, 2012 (Figure 1). They had stayed overnight at Yosemite during June 3–July 23, 2012; the median incubation period was 30.5 days (range 20–49 days). From interviews with case-patients (or their proxy and non-ill companions), we found that case-patients reported engaging in activities and behavior similar to those of non-ill companions while at Yosemite, and no single activity typically associated with risk for HPS was common among patients.

Nine case-patients had lodged 1–6 nights in a signature tent cabin in the Boystown area of Curry Village. One of these patients stayed in the same cabin in which another patient had lodged 13 days earlier. Staying overnight in a signature tent cabin was significantly associated with SNV infection: 9 of 10,193 guests who registered to stay in a signature tent cabin during June 1–August 28, 2012, had confirmed SNV infection, compared with 0 of 40,288 registered guests of other cabins in Curry Village during the same period (p<0.001). One case-patient had not lodged in Curry Village but had stayed in 4 different regular tent cabins in the Tuolumne Meadows area of Yosemite, ≈16 miles northeast of Curry Village and >4,700 feet higher in elevation.

Of all guests who registered to stay in a signature tent cabin during this time, ≈50% were from California, 30% were from other states, and 20% were from other countries. No SNV infections were reported among international visitors to Yosemite. No hantavirus infections were reported among ≈2,800 NPS and contract employees who worked in Yosemite during the outbreak period.

Guest lodgings in Curry Village, at an elevation of ≈4,000 feet above sea level, consisted of 319 regular tent cabins, 91 signature tent cabins, 18 standard hotel rooms, and 70 hard-sided wood cabins on ≈23 acres. All signature tent cabins were located in the single 2.5-acre Boystown part of Curry Village. Both regular tent cabins and signature tent cabins are constructed of an exterior heavy canvas affixed over a wood frame elevated off the ground on a wood base (Figures 2, panels A, B). The distinct characteristics of signature tent cabins are shown in Figure 2, panel C, and Figure 3, panel B.

In August 2012, direct (e.g., feces) and indirect (e.g., gnaw holes) evidence of rodent activity was observed in both signature and regular tent cabins. Opportunities for mouse entry were also observed for both types of cabins. Regular tent cabins evaluated often had gaps between the door and threshold and gaps between the canvas and the frame or foundation. Most signature tent cabins evaluated had gaps between the door and threshold, gaps between the outer canvas tent and inner insulated wall, and holes in the exterior walls or floors, particularly around conduits for heaters. In late August 2012, when the cabins were being inspected and retrofitted to close these gaps, evidence of active rodent infestation (tunneling and nesting and live mice) was observed in the wall foam insulation of most signature tent cabins (Figure 4).

In the Tuolumne Meadows area, ≈8,725 feet above sea level, the regular tent cabins were similar to those in Curry Village: heavy canvas was affixed over a wood frame with either a wood or a cement stone, and stone foundation. Structural points of potential rodent entry were similar to those observed in Curry Village.

Rodent surveillance was conducted over 2 successive nights each in August and September 2012. On August 21–22, a total of 185 traps were placed around the signature tent cabins and other buildings in Curry Village. A total of 95 Peromyscus mice (trap success 51%) were collected: 73 P. maniculatus mice, 16 P. boylii mice, and 6 mice of Peromyscus species that could not be identified to species because raccoons had preyed on the mice in the traps. Serum antibodies against SNV were detected in 10 (14%) of 73 P. maniculatus and 0 of 16 P. boylii mice. On September 4–5, a total of 133 traps were placed in Curry Village, primarily in Boystown, and 89 traps were placed in Tuolumne Meadows. In Curry Village, 10 P. maniculatus mice and 9 P. boylii mice were collected (trap success 14%); no SNV serum antibodies were detected. In Tuolumne Meadows, 39 Peromyscus mice (all P. maniculatus), were collected (trap success 44%); SNV serum antibodies were detected in 2 (5%).

During August 27–September 17, 2012, Yosemite contacted ≈10,000 guests who had stayed in signature tent cabins, ≈30,000 who had stayed in regular tent cabins, and ≈230,000 guests who had stayed in other park lodging, representing visitors from ≈77 different countries. The Yosemite public hotline received >4,800 calls. Hantavirus educational materials were handed to every park visitor, and educational messages were posted in public areas, in lodgings, and on the park Internet site. Additional hantavirus awareness and prevention training sessions were provided to employees and volunteers within the park and NPS-wide. CDC developed an Internet page specific to the outbreak and added staff to its toll-free hantavirus hotline.

On August 28, 2012, Yosemite closed all 91 signature tent cabins indefinitely when it was confirmed that 4 of the 6 patients identified at that time had lodged in 1 of the cabins and when mouse infestations in the cabin walls were noted. The cabins were subsequently dismantled. Approximately 1,300 buildings throughout the park were inspected, and rodent exclusion was performed where needed. Enhanced housekeeping, inspection, and documentation and response protocols were implemented. Additionally, park management began a rodent population surveillance program and implemented trapping in highly developed areas of Yosemite Valley. NPS-wide efforts include enhanced education and training opportunities, integration of public health review into building design and modification reviews, and enhanced guidelines for routine rodent exclusion inspections in all park structures.