Investigation of and Response to 2 Plague Cases, Yosemite National Park, California, USA, 2015

Rapid interagency investigation and public health response probably reduced risk for transmission to other Yosemite visitors and staff.

the patients and environmental samples indicated that the patients had been exposed in different locations and that at least 2 distinct strains of Y. pestis were circulating among vector-host populations in the area. Public education efforts and insecticide applications in select areas to control rodent fleas probably reduced the risk for plague transmission to park visitors and staff. P lague is a zoonotic disease caused by the gram-negative bacterium Yersinia pestis; the organism's reservoir is rodents and the vectors are fleas (1,2). Transmission to humans can occur through bites by infected fleas or through handling Y. pestis-infected rodents (1,2). Epidemics of plague still occur on the continents of Africa, Asia, and North and South America (3). Plague was introduced to California in 1900 (1,(4)(5)(6), where over the next 25 years it caused occasional outbreaks in rats commensally residing with humans in urban areas (2,4,6). Shortly after its introduction, Y. pestis moved into wild rodent populations, establishing a sylvatic transmission cycle (7,8). In subsequent decades, plague spread across California and other western states (9) periodically affecting humans (4)(5)(6)(10)(11)(12)(13).
The human risk of contracting plague is higher during epizootic transmission when Y. pestis is amplified among susceptible rodent hosts (2), such as the California ground squirrel (Otospermophilus beecheyi), the golden-mantled ground squirrel (Callospermophilus lateralis), and certain chipmunk species (Tamias spp.) (2,3,14,15). Higher mortality rates among these animals lead to the release of infectious fleas into the environment (2). The California ground squirrel plays a major role in human exposure in California because its predominant flea species, Oropsylla montana, is a competent Y. pestis vector (1,2) that is often abundant on this rodent and in its burrows (16) and will readily bite humans (1,11). Since the 1980s, evidence of Y. pestis transmission in rodents in the Sierra Nevada mountains has been generally restricted to locations at elevations >1,200 meters (California Department of Public Health, unpub. data, 1983Health, unpub. data, -2015. Despite ongoing sylvatic transmission, human plague remains rare in the western United States (17)(18)(19), including in California, where no cases have been confirmed since 2006 (20,21).
During the summer of 2015, the Los Angeles County Department of Public Health (LACDPH) and the Georgia Department of Public Health reported 2 cases of plague in persons who had recently travelled to Yosemite National Park (Yosemite). The California Department of Public Health (CDPH), in collaboration with the US Centers for Disease Control and Prevention (CDC) and the National Park Service (NPS), investigated the increased Y. pestis transmission in Yosemite. We summarize the epidemiologic, laboratory, and environmental findings and the public health response.

Epidemiologic and Laboratory Investigation
We defined a case of plague as clinically compatible illness and isolation of Y. pestis from a person with a history of travel to Yosemite during the 7 days before illness onset. Clinically compatible illness included fever, headache, chills, and malaise in conjunction with regional lymphadenitis, septicemia, or pneumonia (22). Patients were identified by their county or state health department and reported to CDPH or CDC.
Diagnosis of plague was made after PCR testing of clinical specimens, including blood and bubo aspirates; Laboratory Response Network assays and culture were used. Recovered isolates were confirmed as Y. pestis by bacteriophage lysis (23). For whole-genome multilocus sequence typing (MLST), DNA extracted from Y. pestis isolates was sequenced by using the PacBio RS II platform and sequence reads were assembled by using a hierarchal genome assembly process (Pacific Biosciences, Menlo Park, CA, USA). Allele calls for 3,979 Y. pestis open reading frames (ORFs) (4,046,060 bp) and cluster analyses were performed as described (24).
Local and state public health officials interviewed patients with confirmed cases and their family members who had traveled with them. Respondents were asked about their illness history, travel, activities, and interactions with rodents in and around the Yosemite area during the week before illness onset.

Environmental Investigation
The environmental investigation was prioritized by patient travel itineraries and historical evidence of Y. pestis transmission at these locations or in similar habitats. To assess the scope of Y. pestis transmission and the potential exposure risk for visitors and park personnel, the investigation was expanded to include additional locations in Yosemite. At prioritized locations, visual risk assessments were conducted to evaluate the presence and abundance of rodents, the type of human activities in the area, and the potential for human exposure to infective fleas (25). In areas with suspected Y. pestis transmission, a 30 × 30 cm flannel cloth was used to sample fleas from rodent burrow entrances. Rodents were livetrapped for plague serologic testing and flea collection (25). For rodent trapping, Sherman (H.B. Sherman Traps, Tallahassee, FL, USA) and Tomahawk (Tomahawk Live Trap, Hazelhurst, WI, USA) live traps were baited once with a mixture of grains and opened either from overnight through the following midday or from early morning through noon. Relative rodent abundance was estimated by calculating the ratio of captured rodents to the total number of traps set and is referred to as the trap success rate. Captured rodents were anesthetized with isoflurane, identified to species, brushed to collect fleas, and subjected to collection of ≈0.1 mL of blood for Y. pestis antibody testing. Deer mice (Peromyscus maniculatus) collected near structures were euthanized; all other rodents were marked with a numbered ear tag (National Band & Tag Company, Newport, KY, USA) and released near the point of capture. Small mammal handling techniques were reviewed and approved by the CDPH Institutional Animal Care and Use Committee, protocol 2015-14. In addition to live rodent trapping, rodent carcasses reported by Yosemite staff or visitors were collected for testing.
Flea and rodent specimens were tested for Y. pestis by CDPH, CDC, and NPS. Blood samples from trapped rodents were sent to CDPH for concurrent testing by passive hemagglutination and passive hemagglutination inhibition to detect antibodies against Y. pestis F1 antigen (23). All positive passive hemagglutination titers >1:32, the lowest dilution tested, that were negative by passive hemagglutination inhibition were considered positive (23). Fleas collected by burrow swabbing, from live-captured rodents, or from rodent carcasses were sent to CDPH or CDC to be identified to species according to standard taxonomic keys (26) and to be tested for Y. pestis. Fleas of the same species from the same burrow or rodent host were sorted into pools of up to 10 fleas and then homogenized in brain-heart infusion broth by using glass beads and Mixer Mill MM301 (Retsch, Haan, Germany).
Rodent carcasses were tested at CDPH, CDC, and NPS. Spleen and liver tissues were removed; for direct fluorescent antibody testing, slide touch tissue preparations were incubated with fluorescein isothiocyanate-labeled rabbit anti-F1 antibodies, washed, and then viewed by fluorescence microscopy (23). DNA was extracted from flea homogenates and carcass tissues by using the Mag-NA Pure Compact Nucleic Acid Isolation Kit (Roche Diagnostics, Basel, Switzerland) and amplified by using TaqMan primers and probe targeting the caf1 gene (27). Animal and flea specimens positive by PCR were inoculated onto sheep blood agar plates or onto cefsulodinirgasan-novobiocin agar plates to enable isolation of Y. pestis from contaminated environmental samples (23). Isolates were confirmed as Y. pestis by bacteriophage lysis and typed by whole-genome MLST (24) with the exception that genome sequencing was performed with the MiSeq platform (Nextera XT library preparation, MiSeq Reagent Kit v2, 300 cycle; Illumina, San Diego, CA, USA). Read corrections and assemblies were generated by using SPAdes 3.6 (28). All diagnostic tests were performed by using standard negative and positive controls.

Laboratory and Epidemiologic Findings
On August 2, 2015, LACDPH presumptively diagnosed septicemic plague for a 14-year-old male resident of Los Angeles County (patient 1) and reported the suspected case to CDPH and CDC (L. Tovar Padua, David Geffen UCLA School of Medicine, pers. comm., 2016 Jan 7). LACDPH later confirmed the diagnosis. The patient became symptomatic on July 18, after camping at Crane Flat Campground in Yosemite July 12-17 and visiting Yosemite Valley and Rainbow Pool Day Use Area (Stanislaus National Forest) during this period (Table 1; Figure 1). He reported that he fed squirrels but did not touch them (L. Tovar Padua, pers. comm., 2016 Jan 7). On August 14, the Georgia Department of Public Health presumptively diagnosed bubonic plague for an 18-year-old female resident of Georgia (patient 2) after she had become symptomatic on August 11. On the same day, CDC was notified of the suspected case and later confirmed the diagnosis. During the prior week, this patient had stayed in a rental home in Oakhurst, California, and visited Yosemite Valley, Vernal Falls, Glacier Point, and Sentinel Dome in Yosemite, as well as Nelder Grove, Lewis Creek, and Bass Lake in the adjacent Sierra National Forest (Table 1; Figure 1). Patient 2 reported having observed numerous squirrels in her vicinity at Vernal Falls and Glacier Point but did not report having had any contact with them.
Neither patient reported seeing dead rodents. Wholegenome MLST showed that the genome sequence of Y. pestis isolates from each patient (blood culture from patient 1, bubo aspirate from patient 2) differed at 21 ORFs (Table 2; Figure 2), including 18 single-nucleotide polymorphisms (SNPs). Each patient was accompanied on the trip by family members who did not become ill.

Environmental Findings
Plague risk assessments were conducted for 9 locations in Yosemite and the surrounding national forests visited by the patients (Table 1; Figure 1). Within the park, 8 more sites were also evaluated for Y. pestis transmission and potential risk areas for transmission to humans.

Sites Visited by Patient 1
On August 4, the Rainbow Pool Day Use Area (Table 1; Figure 1) was evaluated and deemed to be an area of low risk because of the lack of historical documentation of plague at this habitat and elevation, low abundance of California ground squirrels observed in the day use area, and lack of other diurnal rodent species that are known Y. pestis reservoirs in this region. Limited burrow swabbing collected no fleas. A visual evaluation of Yosemite Valley (Table 1; Figure 1) was postponed because it similarly lacked historical documentation of local Y. pestis transmission and because no reports of sick or dying rodents from this heavily visited area had been received. The next week, numerous healthy California ground squirrels were noted in Yosemite Valley. At Crane Flat Campground (Table 1; Figure 1), visual assessment and subsequent burrow swabbing suggested recent epizootic activity; California ground squirrel abundance seemed to be very low relative to the number of burrows in the campground, abandoned burrows were noted, and 134 Y. pestis-negative fleas were collected from the entrances of 29 (31.5%) of 94 burrows sampled. Carrion flies (Calliphora latifrons) were also collected or observed at several burrows. Few chipmunks were observed in the campground, but several Douglas squirrels (Tamiasciurus douglasii) were noted. Rodent trapping conducted the following day corroborated the low abundance of rodents (trap success rate 6.9%; Table 3). Y. pestis antibodies were detected in 1 (12.5%) of 8 California ground squirrels (Table 4), and a pool of 8 O. montana fleas collected from a seronegative California ground squirrel tested positive by PCR for Y. pestis (Table 3). Subsequent whole-genome MLST of Y. pestis recovered from this flea pool demonstrated 100% sequence identity across all ORFs when compared with the isolate from patient 1 (Figure 2).

Sites Visited by Patient 2
The 7 locations visited by patient 2 were evaluated on August 18 and 19 (Table 1; Figure 1). Bass Lake was not visually assessed because of the historic lack of plague activity in this area. Visual assessments and burrow swabbing at Sentinel Dome, Vernal Falls, Nelder Grove, and Lewis Creek found no obvious indications of Y. pestis transmission or increased human risk. Initial evaluation at Glacier Point revealed several abandoned California ground squirrel burrows in close proximity to pathways and picnic areas. From the entrances of the 2 burrows swabbed, 21 fleas were collected. The rodents observed in the area were habituated to humans, and several were noted coming in close proximity to visitors. On the basis of these assessments, Glacier Point was identified as a potential exposure site for patient 2. Rodents were subsequently trapped and tested (Table 3); 1 (7.1%) of 14 California ground squirrels and 2 (22.2%) of 9 lodgepole chipmunks (Tamias speciosus) were seropositive (Table 4). All 118 flea pools obtained from Glacier Point, via burrow swabbing or rodent trapping, were negative for Y. pestis by PCR.

Expanded Investigation
On August 10, NPS was notified that 2 dead rodents were found in the Tuolumne Meadows Campground, ≈25 km from the nearest location visited by the patients (Table 1; Figure 1). During the initial assessment, NPS and CDPH staff observed normal rodent diversity and abundance for this location and no fleas were captured by burrow swabbing. Over the following month, 21 rodent carcasses were collected from the campground and adjacent locations, 17 of which were tested for Y. pestis; the remaining 4 were too decomposed for testing. The 2 golden-mantled ground squirrel carcasses collected on August 10 and 8 additional rodent carcasses collected in the campground and surrounding area were positive for Y. pestis (Table 4). Flea pools from 2 of the rodent carcasses (1 Megarthroglossus divisus flea from a Douglas squirrel, 5 Ceratophyllus ciliatus mononis fleas from a lodgepole chipmunk) were also  positive by PCR for Y. pestis (Table 4). Of the 18 lodgepole chipmunks trapped in this area (Table 3), 3 (16.7%) were positive for Y. pestis antibodies (Table 4). A flea pool (3 Peromyscopsylla hesperomys adelpha fleas) from a seronegative deer mouse also tested positive for Y. pestis (Table 4). Whole-genome MLST of 5 Y. pestis isolates recovered from carcasses from Tuolumne Meadows showed that their genome sequences shared 100% sequence identity across all ORFs, compared with the isolates recovered from patient 1 and the flea pool from Crane Flat Campground ( Figure 2). The expanded environmental investigation found evidence of Y. pestis transmission at 1 other location (Table 1; Figure 1). Two visual assessments at Tamarack Flat Campground noted a lower than expected abundance and diversity of rodents and numerous abandoned California ground squirrel burrows. Follow-up trapping (Table 3) led to detection of Y. pestis antibodies in 1 (20.0%) of 5 California ground squirrels tested (Table 4). Rodent trapping for testing was also conducted in Yosemite Valley in mid-October. None of 13 California ground squirrels and 15 Peromyscus mice tested positive for Y. pestis antibodies (Table 3). Six rodent carcasses from developed sites in Yosemite Valley and 3 from the Wawona area also tested negative for Y. pestis.

Flea Control
Sites with evidence of recent Y. pestis transmission and an increased risk for human exposure were temporarily closed, and rodent burrows were treated with insecticide to reduce flea populations and protect wildlife and human health. The following 5 areas in Yosemite were identified for insecticide treatments: Crane Flat Campground, Glacier Point, Tuolumne Meadows Campground, Tamarack Flat Campground, and the Crane Flat-NatureBridge campus. In total, 16.3 kg of 0.05% deltamethrin was used per label instructions to treat an estimated 3,700 rodent burrows. Although time and logistical constraints precluded pre-and posttreatment flea evaluations at all locations, evidence from limited sampling suggested that the insecticide applications reduced the local flea populations. Before treatment at Crane Flat Campground, 134 fleas had been collected from 94 burrows and the California ground squirrel flea index (total no. fleas on rodents/total no. rodents) was 17.5. After the insecticide application, 58 treated burrows yielded no fleas and the California ground squirrel flea index was 1.0. After insecticide application at Glacier Point, the California ground squirrel flea index declined from 8.1 to 2.7. No pretreatment rodent trapping was conducted at the Tuolumne Meadows Campground to provide comparative flea indices for rodents, but the posttreatment flea index for ground-dwelling rodents was 0.9. Before insecticide application at this site, 80 rodent burrows were marked and sampled, yielding a total of 6 fleas; after treatment, no fleas were found at those same burrows.

Public Outreach
To further reduce the plague risk for Yosemite visitors and staff, NPS and collaborating agencies initiated an aggressive public education campaign. In 2014, ≈4 million persons visited Yosemite (29), and, given that plague cases are rare in the United States, it could not be assumed that most visitors were aware of plague risk or prevention measures. The public education campaign included 3 news releases issued August 6-18, media interviews, and website alerts. The park newsletter, The Yosemite Guide, which was given to persons in every entering vehicle, included information about plague. Placards with plague information were posted at park entrances, locations with confirmed Y. pestis transmission, all campgrounds, and many day use locations and trailheads. Educational pamphlets were available to visitors at a variety of locations, including affected campgrounds.

Discussion
In August 2015, these 2 cases of plague were linked to exposure in the internationally popular Yosemite National Park. The initial public health investigation and response with broad media coverage of the first case led to the rapid recognition and appropriate treatment of the second casepatient (30).
The investigation found little overlap in the travel itineraries of the 2 patients, and isolation of distinct strains of   . We were able to connect the exposure of patient 1 to epizootic transmission at the campground on the basis of the visual observations at Crane Flat Campground, the positive results for rodent serology and the pool of fleas collected there, and whole-genome MLST analysis of Y. pestis isolates from patient 1 and the flea pool. The most likely exposure site for patient 2 was Glacier Point, 20 km away, on the opposite side of Yosemite Valley. Although Y. pestis-seropositive rodents were found at this location, we did not detect active infection in rodents or fleas and were therefore unable to directly link the patients' exposure to this site by whole-genome MLST. Previous findings indicate that Y. pestis whole-genome MLST alleles are not rapidly changing and that most detected changes are caused by the more slowly evolving SNPs than by more rapidly changing variable number tandem repeats (24). Our results are consistent with those of a previous SNP-based study, which indicated that widespread plague epizootics are caused by multiple Y. pestis clones arising independently in small geographic areas (31). The environmental investigation found evidence of Y. pestis transmission in disparate locations of the park, including epizootic activity in the Tuolumne Meadows area, ≈41 and 25 km from Crane Flat and Glacier Point, respectively. Evidence of Y. pestis transmission in rodents was found at 4 of the 5 areas trapped. Of the 8 species of rodents live trapped in Yosemite, Y. pestis antibodies were detected in only 5 (15.2%) of 33 lodgepole chipmunks and 3 (7.3%) of 41 California ground squirrels (Table 5). However, Y. pestis was also isolated from golden-mantled ground squirrel and Douglas squirrel carcasses and a deer mouse flea, indicating broader zoonotic involvement.  During the environmental investigation, serum samples collected from 2 bears killed in Yosemite earlier in the summer were positive for Y. pestis antibodies. One bear, killed in July on Tioga Road ≈18 km from Crane Flat Campground, was a cub, indicating that exposure was probably recent. Bears serve as sentinels for plague distribution (32), and in recent decades ≈10% of bear blood samples from the Yosemite area have been positive for Y. pestis antibodies (CDPH, unpub. data, 1980(CDPH, unpub. data, -2015. In addition to Yosemite, in 2015, increased Y. pestis transmission was evident in other parts of the Sierra Nevada mountains (33). Rodent trapping conducted by CDPH in May and June found elevated Y. pestis seroprevalence among rodents in Tulare County. A golden-mantled ground squirrel carcass collected in August from Sequoia-Kings Canyon National Park, also in Tulare County, tested positive. In August, evidence of epizootic activity was also detected in Mono and El Dorado Counties.
The plague activity in Yosemite and other parts of California in 2015 was part of a larger regional trend. Although plague is rare in the United States (median 3 human cases/year during 2001-2012) (17)(18)(19), in 2015, the rate increased in western states (16 cases reported to CDC) (19,34). Synchronous increases in Y. pestis transmission in the western United States have been documented previously and are potentially driven by large-scale climatic trends (35).
The 2015 findings for Yosemite share some striking similarities with those associated with the only human plague case previously associated with Yosemite (36). In 1959, a teenage boy became ill after camping along Yosemite Creek trail, ≈5 km from Crane Flat Campground. Subsequent investigation by CDPH and CDC found evidence of a recent epizootic plague event that had decimated the rodent populations near the campsite. During this investigation, Y. pestis transmission was also documented in Tuolumne Meadows and at Lake Tenaya.
The rapid interagency investigation and public health response to these cases probably reduced the risk for plague among Yosemite visitors and staff. Critical risk-reduction measures included expanding the investigation to recreational sites beyond those visited by the patients and localized insecticide treatments at sites with Y. pestis transmission. Increased educational efforts informing the public about how to reduce their exposure to the cause of this potentially fatal disease contributed to the early diagnosis for patient 2 and to increased reports of finding dead rodents in the park, which led to detection of Y. pestis transmission at additional locations.