This is the authors' abstract. We don't add key points for this paper.
(See the Major Article by Kassis et al on pages 976–8.) Coccidioidomycosis (colloquially known as San Juaquin Valley Fever) is one of the endemic fungal infections in the Western hemisphere. Infection is initiated by inhaled fungal spores that convert to spherules, the tissue form of the fungus. The etiological agents are 2 closely related dimorphic fungi that live as molds in the desert soil in the southwestern United States, from California to central Texas, and in the neighboring regions of Mexico [1]. In California, the etiological agent is Coccidioides immitis, and elsewhere it is Coccidioides posadasii [2], except for a small, newly discovered focus of C. immitis infections in eastern Washington State. Those isolates are related to, but genetically distinct from, isolates from California [3]. Coccidioidomycosis is reportable in California and Arizona, and in both states the incidence has been increasing, which makes the need for improved methods of diagnosis and for studies to establish which patients with pneumonia require treatment more urgent [4]. Although many lung infections are mild, there are enough symptomatic infections that in southern Arizona more than 25% of all cases of non–nosocomial pneumonia are coccidioidomycosis [5]. Many of those pneumonias are self-limited (as is also true of pneumococcal pneumonia), but the illness often is debilitating and the median time to full recovery was more than 4 months in 1 study [6]. Even in the endemic areas the correct diagnosis is rarely made at the initial contact with healthcare, and antibiotics are almost always prescribed prior to the correct diagnosis [7]. Those drugs are, of course, ineffective; they increase the cost of care and they put people at risk of side effects with no possible benefit. Definitive diagnosis of coccidioidomycosis is made either by culturing the organism from sputum or visualizing spherules in sputum cytology, but most patients have nonproductive coughs so diagnosis is almost always made by serology, but experience has shown that appropriate serological tests are rarely ordered initially, and there is still uncertainty about the sensitivity of serology early in the course of infection, and of the specificity of some immunoglobulin (Ig) M (IgM) enzyme-linked immunosorbent assays (ELISAs). Why is it so hard to make an etiological diagnosis of this fungal pneumonia? In actuality, physicians rarely make an etiological diagnosis of any pneumonia, and instead resort to the imprecise diagnosis of “community-acquired pneumonia.” Guidelines for that syndrome recommend treatment with antibiotics, so antibiotics are prescribed for this fungal infection because it is indeed a “community acquired” pneumonia (while most bacterial pneumonias originate from oral pharyngeal colonization), and on clinical grounds coccidioidomycosis cannot be distinguished from bacterial pneumonias unless the patient has eosinophilia. Other than sparing patients exposure to 1 or more rounds of useless antibiotics, there are 2 other possible reasons why it could be beneficial to make a prompt diagnosis of coccidioidomycosis. One is the possibility that early treatment will shorten the course of infection and/or reduce disabling fatigue, but that has not yet been established [8]. Early treatment will prevent dissemination from the lung in vulnerable individuals, such as immunosuppressed or genetically susceptible populations (eg, African-Americans and Filipinos), but this has never been established. There is an ongoing National Institutes of Health study to test that hypothesis that is enrolling participants in the endemic areas. The other is that early treatment may prevent dissemination from the lung in vulnerable individuals, such as immunosuppressed or genetically susceptible populations (e.g., African-Americans and Filipinos). There is an ongoing National Institutes of Health study of early use of fluconazole to treat pneumonia to test those hypotheses that is enrolling patients with pneumonia in highly endemic areas. Dissemination of the infection is relatively uncommon in immunocompetent individuals, but it is the most serious complication of the infection. Dissemination may become clinically apparent during the pneumonic episode or even months to years after the pneumonia has resolved. The organism has a predilection for draining lymph nodes, skin, bones, joints, and the meninges, although nearly every organ can be infected, including mucosal surfaces, retina, prostate, and uterus. Dissemination tends to be chronic, and some infections may never be cured by currently available antifungal agents. The diagnosis of disseminated infection is often made by histopathology or culture of accessible lesions, except for meningitis, which is most often diagnosed by finding antibodies in the cerebrospinal fluid (CSF), but serological tests have been the mainstay of diagnosis for pneumonia, and for some forms of disseminated disease when biopsies are not feasible. The classic work of Smith and his colleagues [9] in the decades around World War II established the time course for the development of IgM and IgG antibodies by prospectively studying the course of infection in 39 500 cases of coccidioidomycosis acquired in California. By 1 week into the illness 50% of cases of pneumonia had a measurable IgM response and by 3 weeks that rose to 90%. Since then, the specific antigens involved in this immune response have been identified and IgM-specific immunodiffusion and commercial ELISAs have been developed that are more sensitive than the original tube precipitation method that was used to measure IgM antibody, so the percentage of people who have detectable IgM antibodies during the first week of their illness is likely higher than 50%. These new serological tests measure IgG and IgM responses using more purified antigens, but no one has done a similar large-scale prospective study to determine the time course of that antibody response or the sensitivity of the newer tests. IgG (complement fixing) antibodies develop later and persist for months or even years after an uncomplicated pneumonia [10]. High titers of IgG antibodies are also a bad prognostic indicator, predicting dissemination, and are thus an indication for empiric therapy [9, 10], but the exact cutoff for predicting dissemination is not clear as the complement fixation (CF) test for IgG antibody is not standardized, so titers will vary from laboratory to laboratory, immunodiffusion tests for IgG can be made more sensitive by concentrating serum but how that relates to CF titers is variable, and ELISAs are not quantitative. Serology is, of course, an indirect way to make a diagnosis of any infectious disease and is used primarily when the pathogen is difficult or dangerous to cultivate, or is uncultivatable. Finding a fungal antigen in the blood or urine is potentially a more direct way of making the diagnosis of infection, including coccidioidomycosis. The Fungitell assay, which detects β-(1,3)-D glucan in the serum, is positive in some cases of coccidioidomycosis, but it is not specific for coccidioidomycosis and it has a sensitivity of less than 50% for primary pneumonia [11]. It is more sensitive in disseminated infections. A more specific and sensitive assay could be useful for diagnosing coccidioidomycosis. In this issue of Clinical Infectious Diseases Kassis et al [12] report the results of a retrospective analysis of the performance of an antigen detection assay for a Coccidioides antigen for the diagnosis of coccidioidomycosis in 152 cases and 486 controls with other pulmonary infections. The study was a collaboration between clinicians at Valleywise Health Medical Center in Phoenix, Arizona, and Dr Lawrence Wheat, who developed the antigen assays. They used a convenience sample of sera and urine samples from suspected cases of coccidioidomycosis that had previously been sent by physicians at the Medical Center to Mira Vista Laboratory for analysis, and they were now asking how well the assay performed compared with serology or other more direct methods of establishing the diagnosis of coccidioidomycosis. The clinicians who reviewed the charts were unaware of the antigen results and categorized the cases as coccidioidal pneumonia or disseminated coccidioidomycosis, and either definite, probable, or not coccidioidomycosis. Definite cases had a positive fungal culture or spherules were seen in pathology or cytology; probable cases had positive serologies or positive antigen tests, a compatible illness, and no alternative diagnosis. In some comparisons between antigen and antibody detection, definite and probable cases were combined and compared with negative controls. This is the largest published study of the performance of combined urine and serum antigen tests. Given the way patients were entered into the study, and the use of the antigen results to define a probable case, it is impossible to calculate the true sensitivity and specificity of the tests. Nevertheless, this study does provide information about the antigen assay that suggests it is a useful adjunct to serology for the diagnosis of coccidioidomycosis. In general, the sensitivity (true positive rate) of an antigen assay increases with the numbers of organisms present at the time of testing. Therefore, it performed best in immunocompromised patients including those with AIDS. Since patients with AIDS or other severe immunosuppression can develop severe, overwhelming infections, rapid diagnosis is important, and one would expect this test to do best in that population. Indeed, in this cohort, the median values for urine and serum antigen concentrations were highest in this subset of patients, who represented 40% of their cases. In some cases, antigen detection was more sensitive than serology, probably because some 25% of immunocompromised patients had negative serologies, probably because patients with AIDS were not able to mount an antibody response. When the serology and antigen results were combined, 93% of immunocompromised patients were correctly diagnosed. In contrast, adding antigen detection to serology did not increase the percentage of immunocompetent patients who were diagnosed correctly. If the antigen test could detect infection sooner that might be an advantage, but that information is not available from this study and only 42% of immunocompetent patients were antigen positive so it is unlikely to be the case. Combined use of the antigen detection and serology did, however, make a modest but statistically significant improvement in the percentage of those cases properly diagnosed compared with serology alone. How many of the pulmonary cases were not acute pneumonia but sequelae of prior lung infections is not clear from the paper. Further clarification is needed. This paper did not deal with 2 other uses for this assay. One is the increased sensitivity of antigen detection compared with antibody detection in the CSF for the diagnosis of coccidioidal meningitis. The quantity of antigen in CSF also can be used to better measure the response to treatment of meningitis, as it presumably directly reflects the fungal burden, whereas changes in the host responses (cell counts, protein, glucose, antibody titers) are all indirect measures of improvement [13]. It is also possible that changes in antigen concentrations in serum and urine will also be useful for assessing response to treatment in other forms of dissemination, but it is doubtful that the tests are sensitive enough to predict when it is safe to stop treatment. Antigen clearance should be a necessary but not a sufficient criterion for discontinuing treatment. Financial support. J. F. is supported by a grant from the University of California, Office of the President (grant number MRP-17-454959) and by the facilities of the VA Medical Center, San Diego. Potential conflicts of interest. The author: No reported conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Joshua Fierer (2020) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: