Start here: what Valley fever is
Valley fever — medically, coccidioidomycosis — is a lung infection you catch by breathing dust. There is no person-to-person spread, no insect vector, no contaminated food. The reservoir is soil.
Two closely related soil fungi cause it: Coccidioides immitis and Coccidioides posadasii, which were recognized as separate species only in 2002, distinguished by DNA polymorphisms and by growth differences on high-salt media, with C. posadasii being the non-California population [1] [2]. Both are thermally dimorphic, and that two-form life cycle is the whole story of how they infect people. In soil the fungus grows as filaments (hyphae) that fragment into tough, easily airborne spores called arthroconidia. When soil is disturbed — by wind, construction, farming, digging, driving — those spores go into the air, and a person breathes them in. Inside the warm lung the fungus switches form entirely: each arthroconidium swells into a spherule, which fills with hundreds of endospores and then ruptures, releasing them to form new spherules [3] [4] [5]. That parasitic form exists only in the host, and it is why the body's usual antifungal defences work poorly against it [6].
Where it is. The endemic region is the arid and semi-arid southwestern United States — Arizona, California's Central and San Joaquin Valleys, Nevada, New Mexico, Utah — extending into parts of Mexico, Central and South America [7] [8] [9]. Risk follows dust exposure, so the burden falls on people who work outdoors and disturb soil: agricultural and construction workers, and anyone downwind of them.
What it does. Most people who inhale the spores are fine — infection is often asymptomatic or resolves on its own. Of those who get sick, the usual presentation is a community-acquired pneumonia: cough, fever, fatigue, chest pain, sometimes a rash and joint aches severe enough to have earned the old name desert rheumatism. A minority progress to severe pulmonary disease, and a smaller minority to disseminated infection, where the fungus spreads beyond the lung to skin, bone, joints, and — the most feared destination — the meninges around the brain and spinal cord [7] [10] [11].
Why it matters, and why it is in the news. Three reasons, and each is developed below.
1. It is routinely missed. In a prospective study of adults presenting with a new lower-respiratory syndrome in an endemic area, Valley fever was serologically confirmed in 16 of 55 people (29%) — and 81% of those with Valley fever had been given antibacterial drugs, which do nothing to a fungus. Symptoms alone had "insufficient predictive value… to guide clinicians without specific laboratory tests" [12]. Without deliberate testing, the true frequency is underestimated by public health statistics. 2. Case counts are rising. Reported infections have "increased dramatically over the past decade" [13], with Arizona surveillance documenting a rise from 33 to 43 cases per 100,000 population between 1998 and 2001 alone [14]. 3. The map is moving. A climate niche model projects that under a high-warming scenario, by 2100 the area of climate-limited endemicity will more than double, the number of affected states will rise from 12 to 17, and cases will increase by 50%, expanding north into Idaho, Wyoming, Montana, Nebraska, and the Dakotas [15].
Three pillars follow — measurements, medicines, and progress — with a simple simulatable climate model between the first two that explains the mechanism behind the third.
Pillar 1: measurements and diagnosis
The spectrum you are trying to place a patient on
Diagnosis in Valley fever is really two questions at once: is this Coccidioides? and how far has it gone? The IDSA guideline organizes its recommendations along exactly that spectrum, "from initial pulmonary infection, which eventually resolves whether or not antifungal therapy is administered, to a variety of pulmonary and extrapulmonary complications" [7]. That clause — resolves whether or not antifungal therapy is administered — is the single most important clinical fact about the common form of this disease, and it shapes the treatment section below.
The stages, in plain terms:
- Asymptomatic or self-limited primary infection. Most cases. Often never diagnosed.
- Primary pulmonary coccidioidomycosis. The pneumonia. Imaging shows infiltrates, and in the aftermath the infection can leave residual nodules — which look like lung cancer on a scan and are a common reason for unnecessary biopsy — or thin-walled cavities [7] [11].
- Severe or chronic pulmonary disease. Independent risk factors identified in a Kern County case-control study were diabetes, recent cigarette smoking, income below $15,000 per year, and older age [16].
- Disseminated disease. Risk factors in that same study were Black race, low income, and pregnancy [16]. Immunosuppression is a major driver — organ transplantation, HIV, and biologic therapy all raise the risk [17] [18] — and rare inherited defects in the interferon-γ/Th1 axis, such as STAT1 gain-of-function mutations, produce disseminated coccidioidomycosis, which is strong evidence that cellular immunity is what normally contains this organism [19].
- Coccidioidal meningitis. The most serious form, and the one that changes treatment permanently.
Serology: the test that carries most of the weight
Because culture is slow and hazardous to laboratory staff and biopsy is invasive, the workhorse is antibody testing, and its logic has been stable for decades [3] [20].
The two antibody classes answer different questions:
- IgM appears early and is the marker of acute primary infection. It is detected by tube precipitin, the corresponding immunodiffusion test, or latex agglutination [3].
- IgG appears later, outlasts IgM, and persists in chronic disease. It is detected by complement fixation (CF) and by immunodiffusion [3].
In practice a sensitive enzyme immunoassay (EIA) is used to screen, with immunodiffusion and CF used to confirm and characterize [20].
The feature that makes CF uniquely valuable is that it is quantitative. The result is a titre — the greatest serum dilution at which the reaction still occurs, reported as 1:2, 1:4, 1:8, 1:16 and so on, each step a doubling. Because titre tends to rise with the burden and extent of infection, it functions as a severity and progression marker rather than a yes/no answer: a rising titre on serial testing signals progressive or disseminated disease, a falling titre signals response to treatment, and a high titre prompts a search for dissemination. That prognostic use is why serology has "assisted in the diagnosis and prognosis of coccidioidomycosis for a half-century" [3] [7].
The important caveat is that serology depends on a working immune system. In a review of 1,797 serologic tests from 298 immunocompetent and 62 immunosuppressed patients, immunosuppressed people had lower seropositivity rates for every test type during the first year — though they can mount a response, and combining multiple test methods improved detection in that group [20]. The patients at highest risk of severe disease are therefore the ones most likely to test falsely negative.
Antigen detection helps where antibodies fail. A Coccidioides antigen EIA detected antigenuria in 70.8% of patients with more-severe coccidioidomycosis while remaining absent in 99.4% of healthy individuals and patients with non-fungal infections, with cross-reaction against other endemic mycoses in 10.7% — useful specifically for rapid diagnosis of severe forms [21]. Culture and molecular methods remain definitive, and molecular identification of thermally dimorphic fungi has moved steadily from classical to nucleic-acid-based approaches [22] [23].
Why so much is missed
Underdiagnosis is not a subtle effect. A cohort study of community-acquired pneumonia patients in southern California found that "limited and delayed testing probably leads to underdiagnosis among non-Hispanic black, Filipino, or Hispanic patients and among high-risk groups, including persons in whom antimicrobial drug therapy has failed" [24]. Combined with the finding that roughly a third of endemic-area pneumonias may be Valley fever [12], the practical recommendation follows directly: if a patient who lives in or has travelled to a Coccidioides-endemic region develops pneumonia, the diagnostic evaluation should routinely include testing for this organism [12].
Centerpiece: a simple simulatable climate-incidence model
Everything about the resurgence and the expanding map rests on a claim about climate driving incidence, and that claim has been modelled explicitly.
The mechanism has a nickname: "grow and blow." Wet conditions let the fungus grow in the soil and build biomass; a later dry, dusty, windy period aerosolises the arthroconidia and delivers them to human lungs. Because those are two separate steps separated in time, the relationship between rainfall and cases is lagged.
Write incidence in season t as the sum of an antecedent-growth term and a concurrent-dispersal term:
I(t) = β₀ + β_grow · P(t − L) + β_blow · D(t)
where P is antecedent precipitation, D is a concurrent dry/dusty dispersal index, and L is the lag.
Grounding. The model form, the lag, and the relative importance of the two terms are all taken from the climate–coccidioidomycosis modelling literature rather than invented here. Analysing Pima County, Arizona, and explicitly testing hypotheses "linking climate and dust to fungal growth and dispersion," Comrie found that precipitation during the normally arid foresummer 1.5–2 years before the season of exposure is the dominant predictor of the disease in all seasons, accounting for half of the overall variance, while dispersion-related conditions predicted incidence in fall, winter and the arid foresummer — and that "cross-validated models combining antecedent and concurrent conditions explain 80% of the variance" [25]. A follow-up assessment confirmed that roughly 80% of the variance in seasonal incidence can be explained by "precipitation and dust-related climate scenarios prior to and concurrent with outbreaks" [26]. Earlier Arizona work using Poisson regression on surveillance data had already found combinations of climatic and environmental factors highly correlated with seasonal case counts [14].
Parameters, honestly. The lag used in the figure, 21 months, is the midpoint of that grounded 1.5–2 year window. The coefficients β₀, β_grow, β_blow and the climate series P(t) and D(t) are illustrative and flagged — deterministic sinusoids standing in for a real regional record, not data from any location. The one quantitative link back to the source is enforced rather than asserted: the coefficients are set so that the antecedent-precipitation term carries about half the variance of the predicted series, which the script computes (49.7%) and checks against the published "half of the overall variance" finding [25].
What the model explains. Three things that are otherwise puzzling.
First, why a wet winter is bad news two years later rather than immediately. The delay is the fungus growing before anything can blow.
Second, why both wet and dry matter, in that order. Regional analyses bear this out with real data: across Arizona, California, Nevada, New Mexico and Utah for 2000–2015, incidence was greater "in areas with warmer air temperatures and drier soils," the mean annual cycle "peaked following periods of low precipitation and soil moisture," but year-to-year, autumn incidence in California's San Joaquin Valley was higher "following cooler, wetter, and productive springs" [27]. Soil-moisture records make the same point directly, relating incidence in Arizona and California to soil moisture from previous summers [28].
Third, why a warming, drying climate expands the disease. If the endemic range is set by a climate envelope — warm enough, dry enough, with enough seasonal rain to grow the fungus — then shifting that envelope shifts the map, which is precisely what the niche-model projection of a doubled endemic area and 12→17 affected states encodes [15].
Limits. A two-term linear model with one lag is a teaching object. Real analyses use distributed lags across many months rather than a single L, and generalised additive models rather than a straight line [29]; the effective lag varies by region and season [27] [25]; and the model says nothing about where the fungus actually lives in the soil, which remains a genuine gap — environmental detection is difficult and patchy [13].
Pillar 2: medicines
Treatment in Valley fever is unusually stratified: for the most common presentation the right answer is often nothing, and for the most severe it is forever.
Most primary infections need no antifungal. The IDSA guideline is explicit that initial pulmonary infection "eventually resolves whether or not antifungal therapy is administered" [7]. Management for an otherwise healthy person with uncomplicated pneumonia is therefore observation with structured follow-up — serial serology and imaging to confirm resolution rather than progression — plus symptomatic care. The reason this matters is that azoles are not free: they run for months, interact with other drugs, and carry hepatic and teratogenic risk. Treating everyone would harm more people than it helps.
Azoles are the mainstay when treatment is warranted — for prolonged or severe pneumonia, for patients at high risk of dissemination, and for extrapulmonary disease. Fluconazole and itraconazole are the standard oral agents; they work by inhibiting fungal lanosterol 14α-demethylase, blocking ergosterol synthesis and so damaging the fungal cell membrane. Newer triazoles have activity against Coccidioides as well, and the treatment guidelines have been revised across two decades of accumulating azole experience [7] [30] [13] [31] [32]. Notably, in the Kern County study, oral antifungal therapy before hospitalization was associated with reduced risk of coccidioidomycosis pneumonia, suggesting that early treatment of high-risk groups may prevent severe disease [16].
Amphotericin B is reserved for severe and rapidly progressive disease. It binds ergosterol directly and punches pores in the fungal membrane — the oldest and still one of the most reliable antifungal mechanisms, sixty years on [33] [34] [35]. Its problem has always been toxicity, especially to the kidney, which is why lipid formulations were developed and are now standard where the drug is needed [36] [37] [38] [39].
Coccidioidal meningitis requires lifelong therapy. This is the sharpest statement in the whole review and it rests on direct evidence. Fluconazole works: in 50 consecutive patients with active coccidioidal meningitis treated with 400 mg daily for up to four years, 37 of 47 evaluable patients (79%) responded [40], and itraconazole showed similar utility in chronic disease [41]. But response is not cure. In a study of 18 patients in whom azole therapy was stopped after apparent remission, 14 of 18 (78%) relapsed — 15 of the relapses meningeal — and relapse occurred anywhere from 0.5 to 30 months after stopping, with no pre-stop feature (which azole, how long, CSF indices) distinguishing those who relapsed from those who did not [42]. The title of that paper is the clinical bottom line: Is it ever safe to stop azole therapy for Coccidioides immitis meningitis? The answer, in practice, is no. Where azoles fail, options include intrathecal amphotericin B and, in case reports, echinocandins [43].
There is no licensed human vaccine. This is the largest gap in the field, and it has been recognized as such for decades — a review two decades ago argued that a coccidioidomycosis vaccine "is justified and attainable" [44], and the case-control work on severe disease ends by noting that high-risk persons "may benefit from vaccination once an effective CM vaccine is available" [16]. It remains unavailable. Progress and obstacles are covered below.
Pillar 3: progress — climate, expansion, and what is being built
The range is expanding, and the driver is climate
The projection is the headline: under a high-warming scenario, by 2100 the climate-limited endemic area more than doubles, affected states go from 12 to 17, and cases rise by 50%, with expansion north into Idaho, Wyoming, Montana, Nebraska, South Dakota and North Dakota — while precipitation limits spread further east [15]. That projection is a climate niche model fitted to contemporary climate and incidence data and then driven with Earth system model output, so it inherits both the strengths and the uncertainties of that approach.
Drought turns out to be a distinct driver, not just "dryness." The most rigorous recent analysis used California census-tract surveillance data from 2000 to 2020, generalised additive models with distributed monthly lags on precipitation and temperature, and an ensemble prediction algorithm to estimate the counterfactual incidence that would have occurred without drought. Across 81,448 reported California cases, it estimated 1,467 excess cases in the two years following the 2007–2009 drought and 2,649 drought-attributable excess cases following the later drought — set against the context that California is in "the driest multi-decadal period since 800 CE, exacerbated by anthropogenic warming" [29]. Drought followed by rain is the grow-and-blow cycle operating on a multi-year scale.
The dust side has hardened too. Reconstructing long-term dust climatology from continuous aerosol observations across the western United States for 1988–2011, investigators reported "direct evidence of rapid intensification of dust storm activity over American deserts," with the frequency of windblown dust storms increasing 240% from the 1990s to the 2000s — in contrast to declining trends in Asia and Africa — linked to Pacific sea-surface temperature variability, and then examined the relationship between that dust trend and Valley fever [45] [46] [47]. Coccidioidomycosis is now routinely cited among the diseases whose burden climate change aggravates, and among the fungal pathogens whose emergence climate is driving [48] [49] [50].
Vaccines: real progress, still no product
The immunology is encouraging and the endpoint remains out of reach. Protection against Coccidioides is cell-mediated, and vaccine-induced protection against the three systemic mycoses endemic to North America was shown in mice to require Th17 cells [51] — a mechanistic target rather than a guess. Host-response and vaccine-development work has been reviewed repeatedly across two decades [52] [53] [13], and recombinant antigen candidates have been evaluated as monovalent, divalent and chimeric constructs [54]. New vaccine platform technologies developed for other pathogens are an obvious opportunity here [55] [56]. But the honest summary is the one the field itself gives: the search for a preventative vaccine is ongoing [13].
Diagnostics and drugs in the pipeline
Diagnostics are moving toward faster, less antibody-dependent methods: antigen EIA for severe disease [21], and molecular identification of thermally dimorphic fungi shifting from classical to nucleic-acid-based approaches [22] [23] [57]. The broader challenge of diagnosing invasive fungal infection quickly is an active field in its own right [58].
Drugs. Beyond the azoles, nikkomycin Z — a chitin synthase inhibitor, attacking a target absent from human cells — has been the long-standing hope, with a recent assessment asking directly whether it is "ready to meet the promise" [59] [60] and experimental work evaluating frequent oral dosing in a central-nervous-system coccidioidomycosis model, aimed squarely at the meningitis problem [61].
Genomics underpins much of this. Comparative genomic analysis of Coccidioides and its relatives [62], population genomic sequencing revealing recent hybridization and transposon control [63], and genome-wide association approaches now being applied to the fungus itself [64] are building the target list. Work on the developmental switch — the genes governing spherule remodelling and endospore formation, and regulators such as Ryp1 — targets the transition that makes this organism pathogenic in the first place [5] [4] [65].
Dig deeper in lmmol
- The health reviews index collects the other conditions in this series. Valley fever shares its central logic with black lung: an environmental exposure measured in dust, a dose that depends on what the air is carrying, and prevention that means controlling the exposure rather than curing the damage.
- STAT1 does not yet have a static page in this graph, but it is the entity to look up next: gain-of-function mutations in it cause disseminated coccidioidomycosis, which is the clearest human evidence that Th1/interferon-γ immunity is what contains this fungus [19].
- For entities without a linked static page here, use lmmol's graph index, all diseases, or all proteins rather than guessing an entity URL.