Clinician Assessment for Coxiella burnetii Infection in Hospitalized Patients with Potentially Compatible Illnesses During Q Fever Outbreaks and Following a Health Alert,Montana,2011

Introduction

C oxiella burnetii is the causative agent of Q fever and an endemic bacterial pathogen in the United States (McQuiston et al. 2006). Goats, sheep, and cattle are the most common animal reservoirs (Maurin and Raoult 1999, McQuiston et al. 2006). Humans typically become infected through inhalation of contaminated aerosols and dust. Possible tickborne transmission of C. burnetii has also been reported (Ecklund et al. 1947).

The incubation period of Q fever is usually 2–3 weeks (Maurin and Raoult 1999). The most common clinical manifestation is a febrile and non-specific flu-like illness (Maurin and Raoult 1999). Pneumonia and hepatitis occur in more severe cases (Tissot-Dupont et al. 1992). Less than 5% of acutely ill patients will develop chronic Q fever, a life-threatening illness (Maurin and Raoult 1999). Acute Q fever is treated with doxycycline, which is most successful at preventing severe illness when initiated within 3 days of symptom onset (Powell et al. 1962).

In 2011, two outbreaks of Q fever occurred in Montana. One outbreak occurred during January–May, 2011, following the sale of infected goats to farms in three Montana counties (Centers for Disease Control and Prevention 2011). The second outbreak occurred during April–May, 2011, and was associated with an office building complex located in a separate county; the source for this outbreak is unknown (unpublished data). The Department of Public Health and Human Services (DPHHS) released a health alert to all Montana county and tribal jurisdictions on June 8, 2011, urging recipients of the health alert, including physicians and non-physician clinicians, to assess patients suffering illnesses compatible with Q fever for exposure to livestock and to test patients suspected of having Q fever for C. burnetii infection.

Materials and Methods

The study was performed at 3 hospitals located in 2 of the counties associated with the goat-related Q fever outbreak and 1 hospital located in the county associated with the office building–related outbreak. We conducted a retrospective medical record review of patients with one of the following International Classification of Diseases, Ninth Revision, (ICD-9) codes as their primary diagnosis at the time of discharge: Fever, unspecified (ICD-9 code 780.6); viral infection in conditions classified elsewhere and of unspecified site (viral infection not otherwise specified [NOS]; 079.0–079.99); pneumonia, organism unspecified (486); and, chest pain, unspecified (786.5) (Anderson et al. 2011). We reviewed records with a hospital admission date occurring within a 3-week period when C. burnetii transmission was likely occurring in the community and a 3-week period following the release on June 8 of the statewide health alert. The dates of record review for each hospital included January 23–February 13 and June 9–29 (Hospital A), April 18–May 8 and June 9–29 (Hospital B), and May 19–June 29 (Hospitals C and D).

Patients were excluded from analysis if their age was <1 month at time of admission, an etiology for their diagnosis was identified, or if the patient was known to have resided in an assisted-living/skilled nursing facility or did not leave their home for >6 weeks before hospitalization. Laboratory testing of patient serum collected >14 days after illness onset and with anti-C. burnetii immunoglobulin G (IgG)-specific antibody titer to C. burnetii phase II antigen ≤1:16 were considered adequate to exclude C. burnetii infection.

Results

A total of 135 patients were identified through an ICD-9 code search. Thirty-two patients were excluded from study participation. Of the remaining 103 patients, the primary discharge diagnoses included pneumonia, organism NOS (84%), chest pain, unspecified (12%), and fever, unspecified (4%) (Table 1). Only 1 patient was assessed for livestock exposure, and no patients were assessed for tick exposure. Among the 37 patients aged 18–70 years, 23 (62%) were known to have been assessed for occupation; only 1 patient had an occupation (farmer) typically associated with livestock. One patient was tested for Q fever; testing occurred after the health alert and the diagnosis was excluded. The majority of patients received antibiotic therapy. Of the 89 patients receiving antibiotic therapy, only 3 (3%) received doxycycline; this included the 1 patient tested for Q fever.

Discussion

Patients hospitalized with Q fever are often infected via close contact with infected animals or their body secretions (Tissot Dupont et al. 1992), yet in this study only 1 patient was assessed specifically for exposure to livestock animals. Clinicians did assess over 60% of persons aged 18–70 years for occupation. However, having an occupation not typically associated with livestock exposure does not exclude the possibility of contact with infected animals or risk for C. burnetii transmission. C. burnetii can become windborne and be carried for long distances exposing humans to the bacteria through inhalation (Tissot-Dupont et al. 2004).

It is recommended that clinicians treat any patient suspected of having Q fever with doxycycline for 14 days and not delay antibiotic treatment based upon laboratory results of an acute serum specimen, which is often negative pending production of measurable antibodies (Powell et al. 1962). If doxycycline is contraindicated, then an alternative antibiotic may be used, including trimethoprim-sulfamethoxazole, fluoroquinolones, and azithromycin, although these antibiotics have lower efficacy for treating Q fever (Dijkstra et al. 2011). Excluding the 1 patient in this study who was tested for Q fever and treated with doxycycline, only 2 patients received doxycycline.

This investigation has several limitations. First, the presence of antibodies to C. burnetii could not be confirmed in the 102 patients not tested for C. burnetii infection. It is likely the majority of patients did not suffer Q fever and instead had illnesses caused by more common pathogens or disease processes. Nevertheless, patients were rarely evaluated for Q fever risk factors or for C. burnetii infection; consequently, some cases of Q fever might have been missed (Tissot Dupont et al. 1992, Marrie et al. 2002). Second, the median age of the study population was older than hospitalized Q fever patients in other cohorts (Tissot Dupont et al. 1992). However, a review of hospitalized Q fever patients in France found the risk for Q fever increased in persons aged 60–69 years (Tissot Dupont, et al. 1992). Third, it is possible clinicians had assessed patients for Q fever risk factors and occupation, but did not document the assessment.

Conclusions

In summary, we found during two Q fever outbreaks and following a health alert urging the serological testing of patients with illnesses compatible with Q fever that hospitalized patients with illnesses similar to Q fever were rarely assessed for Q fever risk factors or tested for C. burnetii infection. Clinicians should assess for Q fever risk factors and consider the diagnosis in patients hospitalized with clinical illnesses similar to Q fever and without a defined etiology (Tissot Dupont et al. 1992). Further study is also needed to evaluate the effectiveness of current healthcare alert practices for zoonotic diseases.