The Spectrum and Presentation of Disseminated Coccidioidomycosis

CLINICAL RESEARCH STUDY

The Spectrum and Presentation of Disseminated Coccidioidomycosis Rodney D. Adam,a Sean P. Elliott,b Mihra S. Taljanovicc a Departments of Medicine and Immunobiology, bDepartment of Pediatrics, and cDepartment of Radiology, University of Arizona College of Medicine, Tucson.

ABSTRACT PURPOSE: Extrapulmonary dissemination of Coccidioides species is associated with significant morbidity and mortality. The clinical manifestations vary widely according to the host, the severity of illness, and location of dissemination. The morbidity and mortality can be reduced by early recognition and treatment, which in turn depends on understanding the spectrum and presentation of disease. METHODS: We performed a retrospective analysis of 150 cases with extrapulmonary nonmeningeal disease seen from 1996 to 2007 at a referral medical center in an endemic region. RESULTS: Hematogenous dissemination was associated with high mortality and occurred primarily in immunocompromised patients, but only 30% of patients with more limited forms of dissemination were immunocompromised. In keeping with prior studies, there was a preponderance of males (nearly 2:1) and people of African or Asian (especially Pacific Islanders) descent. In contrast, Hispanics and diabetics were not at increased risk. Serology was frequently negative in immunocompromised patients, but the diagnosis could be established by isolation of the organism in culture, or in histologic or cytologic specimens. CONCLUSIONS: Although coccidioidomycosis is a great imitator, the diagnosis can usually be made readily if a high level of suspicion is maintained and appropriate diagnostic testing is performed. In most patients, that will include serologic testing in addition to cultures and histology or cytology of appropriate samples. © 2009 Elsevier Inc. All rights reserved. • The American Journal of Medicine (2009) 122, 770-777 KEYWORDS: Cardiac; Cytology; Diagnosis; Ethnicity; Immunocompromised; Peritonitis; Prognosis; Skeletal; Vertebral

Coccidioidomycosis is an invasive fungal disease caused by a species of Coccidioides and is common in the Sonoran desert region of southern Arizona and northern Mexico and the San Joaquin valley region of California. It also is found in other desert or semidesert regions of the western hemisphere. Infection rates are estimated at 3% per year in southern Arizona.1,2 Approximately one third of infections are symptomatic,3 usually presenting with a systemic or respiratory illness that resolves spontaneously over a period of weeks to months. About 30% of cases of communityacquired pneumonia in Tucson are caused by Coccidioides Funding: None. Conflict of Interest: None. Authorship: All authors had access to the data and had a role in writing the manuscript. Requests for reprints should be addressed to Rodney D. Adam, MD, Infectious Disease Section, University of Arizona College of Medicine, 1501 N. Campbell Ave., Tucson, AZ 85724-5039. E-mail address: [email protected]

0002-9343/$ -see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2008.12.024

spp.4 Whether or not the primary infection is treated, a minority of cases are complicated by persistent symptomatic pulmonary disease or by extrathoracic (disseminated) disease. In contrast to primary infection, disseminated infections seldom resolve spontaneously and are frequently fatal in the absence of effective systemic antifungal therapy. Coccidioidomycosis is a reportable disease in Arizona and, in addition to an increase in number of cases coincident with increased reporting requirements in 1997, there has continued to be an increase in numbers of reported cases. The growing elderly population of Arizona that is fueled by people moving to southern Arizona for retirement may be part of the reason, adding people who may have waning immunity due to age or medical problems. In addition, advances in medical therapy that result in impaired cellmediated immunity, such as tumor necrosis factor-alpha blockade5 or organ transplantation,6,7 may increase the number of people who are at risk. Coccidioides infections may involve most systems of the body, so awareness of

Adam et al

Disseminated Coccidioidomycosis

771

disseminated coccidioidomycosis requires knowledge of pitin (immunoglobulin M) or complement-fixing (immunopotential presenting signs and symptoms and maintenance globulin G) antibody tests;10 a positive result, even in an of a high degree of suspicion in patients with a compatible undiluted sample, is highly specific. Positive immunodiffuillness. We have performed a review of the cases of exsion tests were sent to a reference laboratory for completrapulmonary coccidioidomycosis seen at our institution ment fixation titer (1996 to 2000) or for immunodiffusion from 1996 through 2007, focusing (ID) titer (2000 to 2007, performed on the cases that included involveat the Southern Arizona VA Health ment of organ systems other than Care system). The ID and compleCLINICAL SIGNIFICANCE the central nervous system. ment fixation results correlate well with each other, allowing direct ● Hematogenous disease occurs primarily comparison of the titers.11 in immunocompromised patients and

METHODS

has a high mortality. Serologies are freWe retrospectively reviewed cases quently negative, but the diagnosis can RESULTS of disseminated coccidioidomycooften be made by bronchoscopy with sis seen at the University Medical Types of Center (UMC), the primary teachbronchoalveolar lavage. Coccidioidomycosis ing hospital for the University of ● Males and people of African and possiArizona College of Medicine in We identified 207 patients with bly Pacific Isles descent are at markedly Tucson, Arizona, and certain affildisseminated coccidioidomycosis, increased risk, but Hispanics are not, iated outpatient clinics. UMC has including 136 cases with only exnor are diabetics. approximately 300 beds and is a tra-central nervous system (central referral center for all of Arizona, nervous system), 57 with only ● Coccidioidomycosis involving a joint is especially including trauma, cancentral nervous system disseminatypically associated with adjacent ossecer, and solid-organ transplantation, and 14 with both. The 150 ous disease. tion. Metropolitan Tucson comcases with extra-central nervous prises most of the population of system dissemination were classified as shown in Table 1. CoccidPima County and is located in ioides species commonly cross tissue boundaries, so clinical southern Arizona within an area of endemic transmission of cases often overlap categories. Therefore, we established a Coccidioides species. Eighty-five percent of the patients hierarchical system based on decreasing severity as shown with known county of origin were from Pima County. by the descending rows in the table (hematogenous ⬎ mulPatients were considered for inclusion if they received at tisystem ⬎ axial ⬎ peripheral skeletal ⬎ soft tissue ⬎ skin). least a portion of their care for coccidioidomycosis at the Cases of visceral coccidioidomycosis did not overlap with hospital or one of its affiliated clinics and sufficient inforthe other groups. mation was available for analysis. Patients were considered to have disseminated coccidioidomycosis if they had evidence of extrathoracic infecTable 1 Types of Coccidioidomycosis tion. In addition, patients were included if they had diffuse nodular pulmonary disease even if there was no other eviSystemic dence of extrapulmonary infection, because this manifestaClassification Numbers Male Risk Factor* Death† tion usually results from hematogenous dissemination.8,9 Hematogenous 41 22 36 14 Patients with central nervous system disease were included Multisystem 15 10 5 1 in this report only if there also was evidence of disseminaAxial skeleton 28 21 6/26 tion to other parts of the body. Peripheral 26 18 11 Records of infectious disease consultations, culture, cyskeleton Soft tissue 15 13 4/14 tology, and histologic results, and discharge diagnoses were (muscle or reviewed to identify patients with extrapulmonary coccidlymph node) ioidomycosis. Each potential case was reviewed individuSkin 16 9 2/15 ally to determine whether it fit the criteria for disseminated Visceral organ 9 4 3 coccidioidomycosis. Totals 150 97 67/147 15 (8 male; 14 had Fungi grown from clinical specimens that were susrisk factors) pected to be Coccidioides immitis/posadaseii were submit*Risk factors included corticosteroid therapy (19), tumor necrosis facted to ARUP Laboratories in Salt Lake City, Utah, for tor-alpha blockade (2), organ transplantation (11: 5 kidney, including 1 confirmation as Coccidioides species by DNA testing. The with pancreas, 2 lung, 2 liver, 2 heart), HemLymph malignancy (7), chronic presence of spherules with endospores in histologic or cyrenal disease (4), chronic liver disease (4), autoimmune; no steroids (3), pregnancy (2), Ifn-gamma receptor def (1), unknown (3), none (80). tologic preparations from clinical specimens is diagnostic of †Death occurring during the initial hospitalization for disseminated coccidioidomycosis. Screening serologic tests consisted of disease. qualitative immunodiffusion tests mimicking tube-preci-

772

The American Journal of Medicine, Vol 122, No 8, August 2009

Hematogenous. Forty-one patients had hematogenous coccidioidomycosis, including 9 with fungemia. Four of the fungemic patients had diffuse nodular pulmonary disease, and 2 had diffuse bilateral pneumonia. An additional 32 patients had diffuse nodular disease on chest radiography or computed tomography scan, which most likely represents hematogenous infection.8,9 Patients with hematogenous disease frequently had a fulminant presentation, sometimes dying before a diagnosis could be established. A few had a history of disseminated coccidioidomycosis before the presentation with hematogenous disease, and a minority had dissemination to specific foci, including neck, skin, and central nervous system during or after their presentation with hematogenous disease (Figures 1 and 2). Multisystem. Multisystem infections were defined as those with 2 or more distinct systems affected. By definition, this is the only category to include central nervous system infections because patients with only central nervous system dissemination were not included in the analysis. It is inter-

esting to note that there was relatively little overlap between central nervous system disease and other sites of dissemination. Nine percent of patients with non-central nervous system dissemination had central nervous system disease, and 20% of patients with central nervous system disease had non-central nervous system dissemination, indicating that central nervous system disease is usually not associated with other sites of dissemination. Seven of the 15 patients with hematogenous or multisystem disease had central nervous system plus vertebral disease, representing 19% of the 37 with axial disease, or hematogenous or multisystem disease with axial involvement and 10% of those with central nervous system disease. It is perhaps surprising that the overlap between central nervous system and vertebral disease was not greater in view of the propensity of Coccidioides spp. to disregard tissue barriers. Skeletal Infections. The skeletal system was involved in 13 of the patients with multisystem disease and 3 of those with

Figure 1 A 17-year-old Hispanic woman presented with 8 months of illness and threatened airway obstruction. (A) A chest radiograph showed right upper lobe consolidation in addition to diffuse micronodular disease. (B) The coronal reconstructed computed tomography (CT) image of the chest with lung window shows early cavitation (arrow) within the area of right upper lobe consolidation in addition to widespread bilateral micronodular disease. (C) An axial CT image of the chest with soft tissue window shows extensive necrotic mediastinal adenopathy (arrow). (D) An axial CT image of the abdomen shows multiple low attenuation lesions within the liver and spleen consistent with Coccidioides lesions (black arrows) and necrotic lymph nodes conglomerating in the portocaval region (white arrow). A right pleural effusion also is seen. Subsequent workup demonstrated interferon-gamma receptor deficiency.

Adam et al

Disseminated Coccidioidomycosis

773

Figure 2 A 36-year-old immunocompetent Chinese man presented with a neck mass found to be a necrotic lymph node. Axial computed tomography (CT) images of the neck (A) and coronal reconstructed image of the chest (B) with soft tissue windows show partially necrotic adenopathy in both sides of the neck and mediastinum (arrows). A 28-year-old immunocompetent man from the Pacific Islands presented with 1 week of odynophagia and hoarseness. (C) A sagittal reconstructed CT image of the neck with soft tissue window shows a laryngeal abscess (arrow). (D) An axial CT image of the chest with lung window shows diffuse nodular pulmonary disease.

hematogenous infection, in addition to those with axial or peripheral skeletal disease, for a total of 70 (47%). Axial skeletal disease almost always included vertebral infection, although 1 patient with osteomyelitis of the skull also was included. Despite the lack of mortality associated with this category, there was extremely high morbidity, including the frequent requirement for multiple surgical debridements. The risk was especially high for Blacks (18/28; 64%) and males (76%), but relatively few (26%) had other risk factors for dissemination (Figure 3). Of the 26 patients with peripheral skeletal disease, 14 presented with joint disease. Eleven of those with joint infections had radiographic or magnetic resonance imaging evidence of skeletal disease in bone adjacent to the joint (Figure 4). Of the remaining 3, 1 (with knee infection) had no radiographs available and the other 2 had tenosynovitis of the wrist, but did not have disease within a joint capsule. These findings suggest that most patients presenting with intracapsular joint infection also have adjacent osseus infection providing a potential source of relapse when medical therapy is discontinued.

Soft Tissue. Patients with soft tissue infection most often had lymphadenitis or muscle abscesses, three fourths of which involved the upper part of the body and included cervical, axillary, or supraclavicular adenopathy, or airway disease (Figure 2). Those with infection of the lower body had muscle abscesses in the thigh or pelvic areas. Visceral Organ. This group included 5 patients with gastrointestinal infection, 3 with genitourinary, and one with cardiac infection. The severity ranged from a case of incidentally discovered Coccidioides salpingitis to highly symptomatic patients. Three of the gastrointestinal cases had peritonitis, all of whom were symptomatic. The cardiac case presented 2 months after cardiac transplantation with fever and no localizing signs or symptoms. A Coccidioides ID titer was 1:64, and a granuloma with a Coccidioides spherule was identified on a heart biopsy specimen that was performed to exclude rejection. Skin. Skin involvement ranged from single cutaneous lesions to multiple subcutaneous nodules. Overall, the skin is the most common site of dissemination, so the relatively

774

The American Journal of Medicine, Vol 122, No 8, August 2009

Figure 3 A 20-year-old African-American woman presented with fatigue and had a scalp lesion thought to be eczema. She was treated with oral corticosteroids and developed left hip pain that progressed over the next 2 months and was found to have extensive pelvic and vertebral disease. Sagittal magnetic resonance images demonstrate widespread contrastenhancing lesions most pronounced in T7-8, T10, L1, and L5 with extension into the spinal canal at T7-8 (A-C). (D) A sagittal reconstructed computed tomography image of the lumbosacral junction shows a lytic lesion in the L5 vertebral body (arrow).

small number of cases with skin coccidioidomycosis is reflective of the severity of cases in the current series.

Risk Factors for Dissemination Fewer than half the patients had illnesses associated with depressed cell-mediated immunity or risk of disseminated coccidioidomycosis (Table 1). Corticosteroid therapy accounted for at least 30 (45%) of the 66 with identified risk factors (19 with corticosteroid therapy and 11 with organ transplantation). In addition, it is possible that some of those with hematologic/lymphatic malignancies or autoimmune diseases had received corticosteroids that were not noted in the records. Eighty-eight percent of patients with hematogenous coccidioidomycosis had systemic risk factors including steroids/transplant (34%) and human immunodeficiency virus (HIV) infection (27%), compared with 29% in patients with other forms of coccidioidomycosis. Diabetes is a risk factor for complicated pulmonary coccidioidomycosis and is sometimes proposed as a risk factor for dissemination. However, only 7 (5%) of our patients were diabetic, compared with a nationwide prevalence of diabetes of 7%. In addition, for the final year of the study, 2007, 15% of all patients discharged from UMC had a diagnosis of diabetes. Therefore, diabetes does not appear to

be a risk factor for dissemination and is not included in the risk factor analyses. As in prior reports, males were at increased risk for dissemination, accounting for nearly two thirds of the cases (Table 1). Disseminated disease occurred at all ages. The elderly were more likely to have risk factors (20/31 [65%] of those aged over 60 years vs 46/114 [40%] of those under 60 years) (P ⫽ .024, 2-tailed Fisher’s exact test). Dissemination is more commonly documented in certain racial groups, most notably black and Filipino.12,13 Blacks were over-represented in our series by 11-fold compared with whites in comparison to UMC admissions, and by a similar ratio when compared with the demographics of Pima County and Arizona (Table 2). The 43 black patients included 37 African-Americans, 4 West Africans, 1 from Somalia, and 1 from the Caribbean. One was of mixed Filipino and black American ancestry (counted among blacks). Asians were over-represented by 10-fold in comparison to white UMC admissions, but only by 3-fold in comparison to state or county demographics; these numbers are small enough that caution should be used in their interpretation. It is of interest that 4 of the Asians were from the Pacific Isles. Native Americans comprised approximately the number expected from their proportion of the general

Adam et al

Disseminated Coccidioidomycosis

775

Figure 4 A 46-year-old man with liver disease presented with right knee pain and effusion. Coccidioides species grew from a knee aspirate. (A) An oblique radiograph of the right knee shows a lytic lesion adjacent to the articular surface of the lateral femoral condyle, (B) Frontal image of both legs (technetium bone scan) shows increased radiotracer uptake in the right knee lateral femoral condyle (arrow) corresponding to the lesion seen on the radiograph. Additional lesions are seen in the left ankle and more subtle elsewhere in both knees. Sagittal proton density weighted image (C), sagittal proton density weighted fat-suppressed image (D), and sagittal T1W fat-suppressed image with contrast (E) of the right knee show an intraosseous abscess (arrows) in the lateral femoral condyle corresponding to the lesion seen on the radiograph and bone scan. Note rim enhancement on panel E. Additional smaller lesions are seen at the anterior aspect of the lateral femoral condyle and patella.

population, in contrast to some studies suggesting increased risk for people of the San Carlos Apache or Pima/Tohono Oodham groups.12,14,15 Because the majority of people from these tribes live on reservations that may have different exposures or medical referral patterns, limited conclusions should be drawn from these data. Hispanics comprised 25% of UMC admissions and 21% of patients with disseminated disease, suggesting that Hispanic patients were not at elevated risk.

Antecedent Illness Consistent with Primary or Pulmonary Cocci A preceding illness consistent with primary coccidioidomycosis was noted in 80 patients, 67 of whom were seen at UMC for their initial episode of dissemination. The time from the initial episode of primary coccidioidomycosis was

0-2 months in 41 of the 67 (61%), 3-9 months in 14 (21%), 1-2 years in 3 (4%), and ⬎3 years in 9 (13%) and was as long as 20 years. These data are limited by the fact that only 80 of the 150 patients had an identified syndrome of possible primary coccidioidomycosis, but do suggest that disseminated coccidioidomycosis is frequently preceded by an episode consistent with primary coccidioidomycosis within the preceding 2 months.

Diagnosis There was sufficient information to determine how the diagnosis was established in 130 patients (Table 3). Of the 18 patients who were diagnosed by histology, cultures were done in only 3, indicating that the failure to make a culture diagnosis usually resulted from lack of culture rather than poor sensitivity. For the 57 patients where both culture and

776 Table 2

The American Journal of Medicine, Vol 122, No 8, August 2009 Racial/Ethnic Background

Race/Ethnicity White non-Hispanic Black Hispanic Native American Asian

% of Pima County‡

% of Cases (n)*

% of UMC† Admissions

% of State of Arizona

41 (59)

51.5 (%)

55

54

30 21 4 5

3.4 24.8 4.6 0.6

3 25 5 2

4 36 ⬍4㛳 2

(43)§ (30) (6) (7)¶

UMC ⫽ University Medical Center, Tucson, AZ. *The denominator in this column is the 145 patients for whom race was known. †The total in this column adds up to 85%, because 15% are unknown. The numbers are from 1998 to 2003 admissions. There was little change from year to year. ‡Eighty-five percent of UMC admissions with known county of residence were from Pima County. §The odds ratio is 10.9 for black to white patients when compared with UMC admissions (P ⬍.0001) by chi-squared analysis with Yates’ correction (95% confidence interval 6.3-15.1). 㛳Four percent are listed as “other” for county statistics and would include Native Americans. ¶The odds ratio of Asian to white patients is 10.0 (95% confidence interval 4.5-21.9; P ⬍.0001 by chi-squared analysis with Yates’ correction) and 3 (P ⬍.01) when compared with state or county numbers.

histology were performed, the sensitivity of culture was 93%, and that of histology, 84%. Cytology was obtained in only 32 patients, but when performed, was positive in 24 (75%), including 17 patients with hematogenous disease. The sources of the cytologic specimens included bronchoalveolar (BAL) (18), vertebral (2), other skeletal (1), neck (2), and skin (1). Of the 21 BAL samples that were positive for culture or cytology, each was positive 86% of the time. Eighty-one patients who were hospitalized at our facility with their initial presentation of disseminated coccidioidomycosis had a Coccidioides serology done within the period between 3 months before and 1 month after presentation. Of those patients, 23 had negative serologies, 6 had immunoglobulin M only, and 13 had titers that were positive but ⬍8. Thus, half had tests that were either negative or were within values not typically associated with disseminated infections. False-negative serologies were seen in 38% of patients with hematogenous infection and 46% of fatal cases. The sensitivity of serology in immunocompromised patients was only 61% (30/49), though it was 88% (28/32) in immunocompetent patients (P ⫽ .012, 2-tailed Fisher’s exact test).

Prognosis Fifteen of the 91 patients with disseminated coccidioidomycosis who were hospitalized at UMC with their initial episode of disseminated disease (either directly or transferred from another facility during that initial episode) died during their hospitalization. Hematogenous coccidioidomycosis accounted for most of the mortality (56% in patients with

fungemia and 28% in those without fungemia) compared with the 1% mortality in other forms of coccidioidomycosis (Table 1). Fourteen of the 15 who died had underlying diseases that conferred increased risk for disseminated coccidioidomycosis, including 7 with corticosteroid therapy and 2 with HIV/acquired immune deficiency syndrome.

DISCUSSION Hematogenous coccidioidomycosis has emerged as one of the most common forms and is characterized by the frequency of systemic risk factors, especially corticosteroids and HIV infection as seen in this and other studies,9,16 a high frequency of false serologic tests, and a high mortality. The high rate of false-negative serologic tests in immunocompromised patients (37%) is greater than the false-negative rates of 86% to 96% previously reported in immunocompromised patients.17 Likewise, the false-negative rate of 13% in immunocompetent patients is higher than the rate of ⬍5% previously reported.17,18 Therefore, diagnostic efforts also should include histology, cytology, or cultures from appropriate clinical samples. Cytologic examinations and cultures of BAL specimens were especially useful tests in patients with hematogenous infection, giving a “same day” diagnosis without the risk of obtaining pulmonary biopsy specimens. Culture, histology, and cytology are all highyield tests for the diagnosis of disseminated disease, and complement each other so that occasionally one test will reveal the diagnosis even when another test is negative. Skeletal involvement was documented in nearly half the patients with disseminated coccidioidomycosis. Although these patients were usually seropositive, the diagnosis was typically confirmed by culture or histologic evaluation of surgical biopsy/debridement samples. The disease was often marked by local extension, requiring surgical debridement, Table 3

Diagnosis

Classification

Total with Info Culture Histology Cytology Serology

Hematogenous 40 Multisystem 12 Axial skeleton 24 Peripheral 20 skeleton Soft tissue 14 (muscle or lymph node) Skin 13 Visceral organ 7 Totals 130

35 8 19 16

1 1 2 4

12

7 3 100

(7) (8) (11) (12)

1 (17) 1 3 (4) 0 (0)

3 2 0 0

(24/39) (7/7) (18/21) (17/18)

1 (8)

1 (1)

0 (11/13)

5 (11) 4 (5) 18

1 (1) 0 (0) 7

0 (9/13) 0 (6/7) 5 (92/118)

The priority of test is from left to right; culture ⬎ histology ⬎ cytology ⬎ serology, so the initial number in the column represents the number of diagnoses made by that test. The number in parentheses is the total number positive. For the serology column, the numbers in parentheses are the total positive over the total number with serologies at presentation.

Adam et al

Disseminated Coccidioidomycosis

and was sometimes complicated by multiple relapses. Even in patients presenting solely with joint infection, the adjacent bone was usually involved, likely providing the source for relapse, which is common in bone and joint disease.19,20

Strengths and Limitations This report summarizes a 12-year experience of disseminated coccidioidomycosis at a single referral medical center and is the largest series of patients with disseminated coccidioidomycosis during the antifungal era. As such, it gives the spectrum and presentation of cases that may be seen at a referral medical center. Because of the retrospective nature of the study, some of the data were incomplete. In addition, because of the potential for referral bias, caution should be used in generalizing the results to other patient populations.

Conclusions Although dissemination after primary coccidioidomycosis is infrequent, when it occurs, it commonly does so within months after the primary infection. Thus, it is important to diagnose primary Coccidioides pneumonia so that complications can be identified early.4 A high level of suspicion for coccidioidomycosis should be maintained for patients who are or have been in endemic areas and who present with compatible illnesses, with the expectation that an early diagnosis will reduce the morbidity and mortality of disseminated disease.

ACKNOWLEDGEMENTS We thank John Galgiani for helpful comments on the manuscript.

References 1. Dodge RR, Lebowitz MD, Barbee R, Burrows B. Estimates of C. immitis infection by skin test reactivity in an endemic community. Am J Public Health. 1985;75:863-865. 2. Larwood TR. Coccidioidin skin testing in Kern County, California: decrease in infection rate over 58 years. Clin Infect Dis. 2000;30:612613.

777 3. Smith CE, Beard RR. Varieties of coccidioidal infection in relation to the epidemiology and control of the diseases. Am J Public Health Nations Health. 1946;36:1394-1402. 4. Valdivia L, Nix D, Wright M, et al. Coccidioidomycosis as a common cause of community-acquired pneumonia. Emerg Infect Dis. 2006;12: 958-962. 5. Bergstrom L, Yocum DE, Ampel NM, et al. Increased risk of coccidioidomycosis in patients treated with tumor necrosis factor alpha antagonists. Arthritis Rheum. 2004;50:1959-1966. 6. Blair JE, Logan JL. Coccidioidomycosis in solid organ transplantation. Clin Infect Dis. 2001;33:1536-1544. 7. Blair JE. Coccidioidomycosis in patients who have undergone transplantation. Ann N Y Acad Sci. 2007;1111:365-376. 8. Bronnimann DA, Adam RD, Galgiani JN, et al. Coccidioidomycosis in the acquired immunodeficiency syndrome. Ann Intern Med. 1987;106: 372-379. 9. Ampel NM, Ryan KJ, Carry PJ, et al. Fungemia due to Coccidioides immitis. An analysis of 16 episodes in 15 patients and a review of the literature. Medicine (Baltimore). 1986;65:312-321. 10. Pappagianis D, Zimmer BL. Serology of coccidioidomycosis. Clin Microbiol Rev. 1990;3:247-268. 11. Wieden MA, Galgiani JN, Pappagianis D. Comparison of immunodiffusion techniques with standard complement fixation assay for quantitation of coccidioidal antibodies. J Clin Microbiol. 1983;18:529-534. 12. Pappagianis D. Epidemiology of coccidioidomycosis. Curr Top Med Mycol. 1988;2:199-238. 13. Louie L, Ng S, Hajjeh R, et al. Influence of host genetics on the severity of coccidioidomycosis. Emerg Infect Dis. 1999;5:672-680. 14. Sievers ML. Disseminated coccidioidomycosis among southwestern American Indians. Am Rev Respir Dis. 1974;109:602-612. 15. Sievers ML, Fisher JR. Decreasing incidence of disseminated coccidioidomycosis among Piman and San Carlos Apache Indians. A probable environmental basis. Chest. 1982;82:455-460. 16. Rempe S, Sachdev MS, Bhakta R, et al. Coccidioides immitis fungemia: clinical features and survival in 33 adult patients. Heart Lung. 2007;36:64-71. 17. Pappagianis D. Serologic studies in coccidioidomycosis. Semin Respir Infect. 2001;16:242-250. 18. Pappagianis D. Current and future approaches to the diagnosis of coccidioidomycosis. In: Einstein HE, Catanzaro A, eds. Coccidioidomycosis: Proceedings of the 5th International Conference. Bethesda, MD: National Foundation for Infectious Diseases; 1996:116-128. 19. Bried JM, Galgiani JN. Coccidioides immitis infections in bones and joints. Clin Orthop Relat Res. 1986;211:235-243. 20. Kushwaha VP, Shaw BA, Gerardi JA, Oppenheim WL. Musculoskeletal coccidioidomycosis. A review of 25 cases. Clin Orthop Relat Res. 1996;332:190-199.