The Changing Epidemiology of Coccidioidomycosis in Los Angeles (LA) County, California, 1973–2011

Coccidioidomycosis, also known as Valley Fever, is often thought of as an endemic disease of central California exclusive of Los Angeles County. The fungus that causes Valley Fever, Coccidioides spp., grows in previously undisturbed soil of semi-arid and arid environments of certain areas of the Americas. LA County has a few large areas with such environments, particularly the Antelope Valley which has been having substantial land development. Coccidioidomycosis that is both clinically- and laboratory-confirmed is a mandated reportable disease in LA County. Population surveillance data for 1973–2011 reveals an annual rate increase from 0.87 to 3.2 cases per 100,000 population (n = 61 to 306 annual cases). In 2004, case frequency started substantially increasing with notable epidemiologic changes such as a rising 2.1 to 5.7 male-to-female case ratio stabilizing to 1.4–2.2. Additionally, new building construction in Antelope Valley greatly rose in 2003 and displayed a strong correlation (R = 0.92, Pearson p<0.0001) with overall LA County incidence rates for 1996–2007. Of the 24 LA County health districts, 19 had a 100%-1500% increase in cases when comparing 2000–2003 to 2008–2011. Case residents of endemic areas had stronger odds of local exposures, but cases from areas not known to be endemic had greater mortality (14% versus 9%) with notably more deaths during 2008–2011. Compared to the 57 other California counties during 2001–2011, LA County had the third highest average annual number of cases and Antelope Valley had a higher incidence rate than all but six counties. With the large number of reported coccidioidomycosis cases, multi-agency and community partnering is recommended to develop effective education and prevention strategies to protect residents and travelers.


Introduction
Coccidioidomycosis, also commonly known as Valley Fever, is a fungal disease normally caused by the inhalation of airborne spores of Coccidioides spp. which grow in the soil of certain areas of the Americas, particularly in routinely hot, arid to semi-arid environments [1,2]. While an estimated 60% of infected people develop mild to no symptoms [3], the remaining infected present with various symptoms and conditions that can include weeks to months of fatigue, shortness of breath, cough, fever, night sweats, loss of appetite or weight, chest pain, headache, body aches, skin rash, and pneumonia [4]. Less than five percent of infected people develop disseminated disease which is when the fungus spreads beyond the lungs to infect any other body site such as skin, lymph nodes, bones, joints, and brain [4]. Disseminated disease can lead to life-long complications and death. An estimated 150,000 new infections occur each year in the United States [4], but reported cases from 2009 to 2013 ranged from 9,438 to 22,641 [5]. In Arizona, 75% of reported cases miss work or school due to illness, and 40% of cases require hospitalization [6]. In California, the median cost of hospitalization alone is estimated to be U.S. $55,062 per patient [7]. As symptoms are nonspecific and disease awareness is low among primary healthcare providers [6,8], disease detection and timely treatment are major challenges.
At one time coccidioidomycosis was generally thought be acquired only in Central California of the United States and was commonly called San Joaquin Valley Fever [9]. Endemic areas have been recognized in parts of Central America, South America, and throughout the United States' Southwest including Arizona, California, New Mexico, Nevada, Texas, and Utah [2,6,7,10]; however, awareness of coccidioidomycosis remains low, even in endemic areas [6,8]. For example, pockets of endemicity in San Fernando Valley of LA County have been documented since the 1950s [11][12][13], yet LA County remains an under-recognized source of Valley Fever.
LA County's diverse geography of 4,084 square miles includes 70 miles of coastline, several deep valleys, mountains peaking over 10,000 feet, and a high desert area (Fig 1). With 10 million residents, a quarter of California's population, LA County is the most populous county in the United States and more populous than most of the 50 U.S. states [14]. Most residents live in the lower elevation areas south of the mountains where urbanization is heaviest. Among the 24 county health districts, West Valley, San Fernando, and Antelope Valley are distinguished by large mountain valleys typically with higher summer temperatures, stronger winds, many more coccidioidomycosis cases, and higher incidence rates than the rest of the county. These three health districts are considered endemic for coccidioidomycosis while the other 21 health districts are not known to be endemic. Antelope Valley, a high desert area of 1,600 square miles at 2,270-3,500 feet above sea level [15] with very strong winds and dust storms, has the greatest potential in the county for land development projects such as housing, agriculture, and solar farms. Strong winds, dust storms, construction work, agriculture, solar farms, archaeological digs, and other soil disturbing activities have been previously associated with coccidioidomycosis cases and outbreaks [10,[16][17][18][19][20][21][22][23][24][25][26]. Antelope Valley is directly south of Kern County, which traditionally has the greatest number of cases in California, and West Valley is southeast of Ventura County which experienced a coccidioidomycosis outbreak after a 1994 earthquake [27].
This study examines population-based surveillance data of 1973-2011 to present the epidemiology of coccidioidomycosis in LA County, California. Epidemiologic changes are analyzed together with urban development. Finally, LA County case numbers and rates are related to those of other California counties.

Materials and Methods
The County of Los Angeles Department of Public Health has used the surveillance case definition of coccidioidomycosis of the Council of State and Territorial Epidemiologists (CSTE)  [5,[28][29][30] and has required health care providers and laboratories to report any identification of the disease since July 1, 1955. In 2008, the CSTE case definition was modified so that a single IgG positive test was sufficient for laboratory confirmation. Lacking clinical or laboratory evidence of disease, already existing in the surveillance database as a case, and being a non-resident of Los Angeles County were criteria for exclusion as new cases. Long Beach and Pasadena residents were excluded as these cities have their own health departments. Other noteworthy developments were the ability of local laboratories to report disease electronically starting in February 2002 and the addition of coccidioidomycosis to the state of California's list of laboratory reportable diseases in December 2009 [31].
Case data from paper documents of annual disease summaries, which were available for 1973 and later, were compiled with electronic data since 1992 and analyzed with Microsoft Excel, Microsoft Access, and Statistical Analysis System software. Pearson correlation coefficients and Mantel-Haenszel formulas were used for statistical tests. Disease onset dates were established through medical records or case interview. Mutually exclusive race-ethnicity categories of Asian, black, Hispanic, and white were defined. Any identification of being Hispanic trumped all other race-ethnicity identifications. Designation of each case's health district was routinely conducted to process investigation and follow-up of cases. Although data on travel history, occupation, and outdoor activities involving dirt were already being collected as part of routine surveillance, starting in 2005 additional exposure questions were added to case interviews to capture information on construction, earth excavation, dust storm, and other outdoor exposures within four weeks of symptom onset. Case interviews were conducted until February 2009 but data collection of exposures continued when such information was indicated in medical chart notes. Coccidioidomycosis-related mortality was defined by indication of death in the case report.
Other data sources included the United States Census Bureau for incidence rate calculations using 1990, 2000, and 2010 decennial counts, and for the number of building permits for new residential buildings constructed during 1996-2011 to measure urban development; Los Angeles County Vital Statistics population estimates for incidence rate calculations for years between the decennial census years; Los Angeles County Location Management System 2014 for geographic informational mapping; and the California Department of Public Health for case frequency and incidence rate comparisons among all the California counties.
The Institutional Review Board of the County of Los Angeles Department of Public Health (IRB) gave verbal consent of this study as the work performed is part of routine Department functions. The IRB waived written informed consent from participants because after de-duplication patient information was anonymized and de-identified prior to analysis.

Overall trends
Los Angeles County confirmed 3,338 reported coccidioidomycosis cases between 1973 and 2011. From 1973 to 2003, the number of reported coccidioidomycosis cases typically numbered between 21 and 80 per year (Fig 2). An outbreak during 1992 to 1994 that involved strong winter wind storms, the 1994 Northridge earthquake, and a larger outbreak seen in other California counties [27,32,33] briefly brought annual cases between 95 and 106. But a change started coccidioidomycosis based on environmental conditions and history of high case numbers and incidence rates. The high desert area of Antelope Valley continues to have the greatest potential for land development for projects such as housing, agriculture, and solar farms. Map was made by E.R. using ArcGIS

Demographic Trends
Age. Demographic aspects of the change starting in 2004 included the following. All age groups showed increasing cases after 2003, with many groups reflecting the steep spikes of the overall incidence in 2004, 2005, 2010. During 1995-2003, average annual incidence rate increased with age until age 55-64 years (Fig 3). However, during 2004-2011, average annual incidence rate continued to increase with age at age 65 years. Between 1995-2003 and 2004-2011, age-specific average annual incidence rates increased 151%-318%. The greatest increases in rates occurred in the 65, 0-14, and 15-24 year age groups (318%, 234%, and 226%, respectively).
Race-Ethnicity. Regarding race-ethnicity, white and Hispanic cases rose sharply in 2004, 2005, and 2011 while increases among Asian and black cases were not as prominent. Usually 0%-15% of annual cases had missing race-ethnicity data. During 2006-2008, 36%-48% of cases were missing race-ethnicity data and the number of white cases substantially dropped.  (Fig 4). During 1995-2003, blacks, Hispanics, and whites had 1.54, 2.94, and 2.90 times higher numbers, respectively, compared to Asians. During 2004-2011, blacks, Hispanics, and whites had 1.95, 4.13, and 4.68 times higher numbers, respectively, compared to Asians. While whites and Hispanics led with much higher annual numbers on average, blacks led in average annual incidence rates. Between   . In all age groups, females had fewer cases than males (S1 Table). Within age groups, incidence trends of females generally reflected those of males (S1-S4 Figs).  districts not known to be endemic experienced a percent increase in cases that ranged between 67% and 1500% (Fig 7). Seventeen of the 21 health districts not known to be endemic had case increases of 100% or more between 2000-2003 and 2008-2011.
Of 2,021 cases with hospitalization status data, 1,408 (70%) were hospitalized. During 1998-2009, when few cases were missing hospitalization status data (one to seven per year), the annual average of percent hospitalized was 68% with a range of 63%-79%.

Exposures 1-4 Weeks Prior
Among measured exposures in the one to four weeks before illness, being in an area in sight of construction and being in an area in sight of earth excavation had the strongest associations with cases residing in endemic health districts (Table 1). Case residents of endemic areas had 5.48 (95% confidence interval: 3.87-7.75) times greater odds of being in an area in sight of construction and 5.46 (95% CI: 3.67-8.13) times greater odds of being in an area in sight of earth excavation within four weeks of illness onset than cases that resided in health districts not known to be endemic. After these two exposures, in order of diminishing magnitude, being in a dust storm, participating in outdoor activities involving recreational vehicles such as   motorcycles and dirt bikes, participating in any outdoor recreation, participating in outdoor activities involving work with dirt, and having a job in an endemic health district had statistically stronger associations with cases that resided in endemic health districts. Conversely, travelling to an endemic area outside of Los Angeles County within four weeks before illness was more closely associated with cases that resided in health districts not known to be endemic. When focusing on cases without travel to endemic areas outside of LA County within four weeks of illness (n = 853), exposures regarding construction, earth excavation, dust storms, outdoor activity with dirt, and jobs in endemic areas in LA County became more strongly associated with case residents of endemic areas, and exposure by outdoor activities involving recreational vehicles became more associated with residents of health districts not known to be endemic.

Discussion
Over 1973-2011, the epidemiology of coccidioidomycosis in Los Angeles County changed. The most notable changes started in 2004. These included significant increases in case numbers and incidence rates across various demographic categories and geographic areas, a sudden and substantial rise in female cases, collectively more annual cases in the endemic areas, especially high case numbers and incidence rates in the Antelope Valley, and increasing number of case deaths in areas not known to be endemic. The housing boom during the early and mid-2000s seems to be at least part of the initial cause for the epidemiologic changes starting in 2004. Because of the distance from large business centers, residential property in Antelope Valley is much cheaper than most parts of the county. The development of thousands of new homes in Antelope Valley was very strongly correlated to disease incidence throughout the county. With the strong high desert winds, increased construction activity in formerly undisturbed regions of the endemic area would have likely released more Coccidioides spores to infect previously unexposed people. Such people easily include construction workers, agricultural workers, real estate marketers, both longterm and new Antelope Valley residents, inmates of the three local correctional facilities, and visitors seeking new homes, work, and specific recreation ranging from kid soccer tournaments to dirt biking. The contrast between the gradual coccidioidomycosis increase among males starting in 2001 and the sudden sharp rise among females in 2004 suggests how the exposed population may have expanded from the male-dominated construction industry to young families seeking or moving into the new homes. Year 2005 was the peak year for newly constructed residential buildings and for male and female coccidioidomycosis cases in age groups 25-34 years and 35-44 years.
Other factors, may have contributed to the increase in coccidioidomycosis both during and after the housing boom. These include fugitive dust, drought, and expansion of agriculture in Antelope Valley. Increased usage of agricultural crops that required tilling of new fields instead of recycling fields, pulverization of soil so carrots could grow straight, and fugitive dust due to lack of proper dust control knowledge among new land owners and inability to enforce dust control ordinances on absentee land owners were reported. While adequate data to show relationships between such environmental factors and coccidioidomycosis incidence were unavailable, examination of dozens of satellite images between 2000 and 2011 found that specific sites of land development and agricultural expansion preceded nearby incident cases and clusters by at least a month. These findings are not shown because the satellite image quality is poor and further analysis with soil type data might provide more telling results. The observed agricultural expansion might be related to the much greater increase in cases among whites and Hispanics after 2004 as greater percentages of these populations work in agriculture compared to black and Asian populations [34]. The modification in CSTE coccidioidomycosis case definition to no longer require a convalescent test could have also caused cases to increase after 2008.
In general, reports of coccidioidomycosis increased substantially across the United States between 1998 and 2011 with reports coming from 28 states and the District of Columbia [35]. Recently, researchers detected Coccidioides in Washington State soil for the first time [36]. Other than drought and the expansion of land development and human populations into endemic areas, ecologic changes in soil and climate change might have contributed to increases in LA County and in the United States [37,38]. Between 2003 and 2005 several California counties also had an increase in cases [39]. However, none had as much of an increase as LA County during those three years. Several not-known-to-be endemic areas in LA County had much higher percent increases in coccidioidomycosis than endemic areas. Additionally, 38% of survey-responsive cases residing in areas not known to be endemic reported travelling to endemic areas outside of LA County. The collection of these findings indicates the need to educate and raise awareness of coccidioidomycosis beyond residents of known endemic areas.
Coccidioidomycosis mortality is not well-established in the literature. In 1991, Kern County conducted a clinical study of 536 coccidioidomycosis patients, 29% of whom were hospitalized and 17% had unknown outcome, and found a 3.2% one-year mortality after onset [40]. This calculation included cases who might have died from causes unrelated to the coccidioidomycosis infection, and so coccidioidomycosis mortality may be less than 3.2%. Two studies looking at non-federal hospitalizations in California estimated coccidioidomycosis mortality at 9% for 1999-2002 [41] and 8% for 2000-2011 [7]. These two studies analyzed hospitalized populations so less severe coccidioidomycosis cases were likely not included. With 70% of cases being reported as hospitalized and an 11% mortality in LA County, questions of missed diagnosis and non-reporting arise, particularly regarding disease not severe enough for hospital admission. Higher coccidioidomycosis mortality and the sharp increase in number of deaths during 2008-2011 in areas not known to be endemic also point to the need for better awareness among clinicians and the general public towards improved recognition, diagnosis, reporting, case management, and prevention.
Given the large number of reported cases, collaborative partnerships among Federal, State, and local government agencies and local community organizations are recommended to develop effective education and prevention strategies to protect residents and travelers. Education and awareness on a community level are vital to appropriately recognize, diagnose, treat, measure, and prevent coccidioidomycosis. In LA County, infectious disease clinicians seemingly have the highest level of coccidioidomycosis awareness among medical providers; however, general practitioners and emergency department clinicians, typically the first to see new cases, are largely unaware of the disease. In light of this, some clinicians have suggested requiring coccidioidomycosis education for new and continuing medical licensure in endemic states. Additionally, the general population needs effective education on the disease. For higher probabilities of success, education efforts and awareness campaigns should encompass input from cases and clinicians, and participation and leadership from city and neighborhood councils, local businesses and industries, schools, and other community organizations.
While population-based, this study was limited by available resources for epidemiologic investigation and passive surveillance. As described, analysis of race-ethnicity and mortality excluded years with large percentages of missing data. Inter-observer variability among the clinicians reporting disease and public health nurses conducting case investigations may have caused inconsistent sensitivity in detecting epidemiologic factors. Similarly, under-reporting and misdiagnosis, which are recognized problems even in the most endemic areas are unmeasured and likely present. Because LA County has had mandatory laboratory reporting for coccidioidomycosis, the 2009 requirement for laboratory reporting to the state of California is not considered a major contributing factor for the observed trend of increasing disease, especially with the trend starting in 2004. In 2009, all case interviews ended and the main source for exposure data became the medical chart. As such, exposure data before 2009 likely represents people who were reachable by telephone during work hours of week days. The true burden of disease is underestimated also because surveillance excludes previously reported cases, a few of which may be re-infections or re-activations several months or years after prior infection.
The authors present this study to help raise awareness and inform government and community planning efforts to prevent unnecessary coccidioidomycosis disease and mortality. Effective education and awareness efforts beyond medical communities of Central California and Arizona are needed. Engaging local community organizations and local government agencies to collaborate in endemic areas and areas of low or unrecognized endemicity but with a history of cases is a progressive step towards better public health. LA County serves as an example of how coccidioidomycosis can potentially increase. Land development in previously undisturbed endemic areas and increased population exposure can lead to the ongoing experience of substantially more coccidioidomycosis cases and changed epidemiologic profile. In 2012 and 2013, LA County had 327 and 362 confirmed cases, respectively, but another steep increase in cases such as in 2004 could occur if development in endemic areas extensively exposes resident, working, or visiting populations. Development can involve any soil-disturbing activity including environmental cleanup. Other than epidemiologic data, ecologic studies with soil testing and geological maps on soil type can inform development planning when soil disturbance is proposed in endemic areas and nearby areas not known to be endemic for Coccidioides [37]. Planning should involve local residential communities and businesses when developing and executing education and prevention strategies to minimize exposure of residents and traveling visitors.  Table. Number of coccidioidomycosis cases (N = 2530) by age and gender, Los Angeles County, 1992-2011. Legend. Thirteen cases missing age or gender are not included. In all age groups, females had fewer cases than males. Within age groups, incidence trends of females generally reflected those of males. During 1992-2011, the greatest number of male cases occurred in the 35-44 and 45-54 year age groups and the greatest number of female cases occurred in the 45-54 year age group. (DOCX)