Q fever is a zoonotic disease caused by infection with the intracellular bacterium Coxiella burnetii. This gram-negative bacterium has a nearly worldwide distribution and infects a wide variety of animals, including mammals, birds, reptiles, and arthropods [1]. The most common reservoir species leading to human exposure are thought to be domesticated livestock, specifically goats, sheep, and cattle. Livestock animals do not commonly suffer from acute disease when infected, but C. burnetii can replicate to high density in the placenta and lead to reproductive failure in these animals [2]. When C. burnetii is introduced into previously naïve sheep flocks and goat herds, “abortion storms” can occur and these are often the first indication of the presence of C. burnetii in the animals. During birth, large numbers of C. burnetii can be released into the environment and this often leads to human infections [3].

C. burnetii is transmitted to humans by inhalation of aerosols that contain the bacteria. These aerosols are often derived from dried animal waste products or birthing fluids. Inhalation is thought to result in infection of monocytic cells in the lung [4]. C. burnetii is taken up through the endocytic pathway and remains in an acidic vacuole where it replicates slowly [5]. Over time, large numbers of C. burnetii can grow in infected vacuoles, and cell egress can release infectious bacteria for further rounds of infection. Humans will typically show symptoms 7-21 days after initial infection, with higher doses having shorter incubation times [6]. It has been estimated that >50% of human infections could be asymptomatic, with concomitant seroconversion [7].

Symptomatic infections can present in both an acute and a chronic form. Acute Q fever is a febrile illness that is often accompanied by fatigue, headache, and myalgia [8]. The disease can resolve without treatment and the median duration of fever in untreated patients is 9-14 days [6,9]. More serious presentations that can occur with acute Q fever are pneumonia and hepatitis [10]. It is fatal in rare cases. Q fever can also present in a chronic form that can take place months or years after the initial infection. Patients with chronic Q fever may or may not recall an episode of symptomatic acute Q fever. Chronic Q fever requires growth of C. burnetii in an opportune location in the body. The most common presentation for chronic Q fever is culture negative infectious endocarditis, but vascular infections in chronic Q fever can also be a common manifestation, with fatal aortic aneurysms possible [11,12]. The presence of a pre-existing valvulopathy is the most common risk factor associated with chronic Q fever endocarditis [13,14].

Because of the ability of C. burnetii to replicate to high density in the placenta, there are a special set of risks associated with Q fever during pregnancy. Premature birth, stillbirth, intrauterine growth retardation, and low birth weight have all been described as adverse pregnancy outcomes due to Q fever [15-18]. Infection during pregnancy can result in a chronic infection where the bacteria are not cleared and persist in the placenta throughout the pregnancy. The bacteria may also remain after the pregnancy is completed, and could renew replication in subsequent pregnancies [19]. Delivery of babies from C. burnetii infected mothers can also create a risk for health care workers, as the generation of C. burnetii-containing aerosols is possible [20].

Although doxycycline is effective for Q fever treatment, there are reasons to seek alternative therapies. Coxiella burnetii is considered a category B agent of bioterrorism by the CDC. It is possible to produce large quantities of organisms and these would be quite stable in storage and transport. The organisms are highly infectious and transmitted by inhalation. Dispersal of concentrated aerosols over a large area would be expected to infect large numbers of people [43]. For these reasons, the US biological weapons program tested C. burnetii on human subjects in the 1950’s and 1960’s as part of “Project Whitecoat” [6,44]. In these studies, C. burnetii was found to be able to infect humans at a very low dose of less than 10 organisms, and aerosol dispersal was effective for human infection [6]. Although use of C. burnetii as a weapon would not be expected to cause widespread fatalities, it is thought that significant impairment of physical activity could be achieved, along with associated alarm in the population and significant long-term morbidity for a subset of victims [43]. Recent advances in microbiological techniques make it conceivable that C. burnetii could be modified for greater virulence or made to have antibiotic resistance [45].

In addition to the potential of engineered antibiotic resistance in C. burnetii, natural resistance to doxycycline is also a concern. Although resistance to doxycycline does not appear to be a common occurrence, doxycycline resistant isolates do exist. In 2005, 3 isolates with doxycycline minimum inhibitory concentrations (MICs) of greater than 8 μg/ml were reported [46]. This included one from an endocarditis patient that died during treatment with doxycycline, one isolate from a goat, and a third from a human acute Q fever case. The genome sequence from the resistant isolate obtained from the endocarditis patient is available [47], but no potential mechanisms of resistance have been described. The extent of doxycycline resistance is not known. It can be difficult to obtain human isolates and testing for doxycycline resistance is not commonly performed. However, advent of newer assays based on real-time PCR have allowed more testing of new isolates in recent years [48].

Pregnancy also presents a need for more alternative therapies for Q fever. The reference treatment of doxycycline is not acceptable for use in pregnancy due to risk to the fetus, and negative effects of co-trimoxazole have not been ruled out [23,49]. Not all studies of Q fever during pregnancy have found a large impact on fetal health, suggesting that strain and host factors may influence whether acute Q fever can impair fetal growth and development [41,42,50]. This raises questions about the universal need for treatment during pregnancy, and the risk/benefit calculation that must be made. Safer alternative therapies would be beneficial.

The current protocol of using doxycycline plus hydroxychloroquine to treat chronic Q fever represents a significant improvement over previous therapies. Examination of Q fever endocarditis outcomes in France from 1983 to 2006 shows that use of the combination therapy has coincided with reduced mortality and a reduced need for heart valve replacement [28]. However, relapses can still occur and the treatment recommendation of antimicrobial therapy for at least 18 months is one of the longest treatment regimens used for a bacterial infection. Alternative or adjuvant therapies would be beneficial for Q fever endocarditis with the goal of shortening the duration of therapy and reducing relapse events.

A number of different antibiotic therapies have been investigated for Q fever (Table 1). Some have been given intentionally for Q fever treatment, or given as an empiric treatment before a diagnosis was made, or investigated for efficacy against C. burnetii in vitro. Randomized, controlled studies evaluating different treatments have not been reported for acute or chronic Q fever. Current treatment recommendations are based on combinations of case studies, retrospective cohort studies, and in vitro data. Many of the in vitro studies were performed by counting the number of visually infected cells under the microscope [51,52], but more recent studies have evaluated growth using real-time PCR [48]. The quinolones have been used with some success [51]. In vitro, gatifloxacin has been shown to be highly effective, and moxifloxacin, gemifloxacin, levofloxacin, pefloxacin, and ofloxacin also have some inhibition of C. burnetii growth [52-56]. Ciprofloxacin has some antimicrobial effect, but not as strong as the other quinolones [52,54]. A more recent study using real-time PCR to get a better quantitative assessment of C. burnetii growth found very little effect for ciprofloxacin [48]. The β-lactam antibiotics, including penicillin, amoxicillin, and cephalosporins, appear to have little effect on the growth of C. burnetii [9,26,48,57].

In vivo use of quinolones has primarily been associated with Q fever endocarditis. Combinations of quinolones and doxycycline were mainly used before 1999 [28]. The drugs ofloxacin and pefloxacin showed some benefit for chronic Q fever when used in conjunction with doxycycline, but success was limited and required at least three years of treatment [29]. In 1999 a study was published that compared the combination of doxycycline and hydroxychloroquine to the combination of doxycycline and ofloxacin [33]. It was thought that this combination would be more effective due to the ability of hydroxychloroquine to raise the pH of the phagolysosomes where C. burnetii replicates and allow more effective doxycycline activity [32]. Indeed, the doxycycline/hydroxychloroquine combination allowed a reduced duration of treatment and had fewer relapses [33]. Although before 1999 doxycycline in combination with quinolones was considered a reference treatment, this combination has not been widely used since that time [28].

Rifampin has been used with some success against C. burnetii. In vitro studies have found it to be very effective [48,57]. Rifampin is usually used in combination with other drugs because the risk of developing drug resistance is high. Before the common usage of hydroxychloroquine, a combination of rifampin plus doxycycline was sometimes used for the treatment of Q fever endocarditis [28,29]. Rifampin has also been used to treat Q fever during pregnancy [58,59]. A combination of rifampin and erythromycin has resulted in a successful delivery and recovery from a human placental infection [58]. Rifampin is placed in pregnancy category C by the US FDA.

Macrolides can have an effect on C. burnetii growth but the results of their use has been somewhat mixed. Some in vitro studies have found very little inhibition of C. burnetii growth in the presence of erythromycin or azithromycin [46,52]. However, in 13 C. burnetii isolates that were resistant to erythromycin in vitro, growth inhibition was observed in the presence of telithromycin [46]. Other studies have found inhibition of C. burnetii growth by clarithromycin in vitro, although it was generally not as effective as doxycycline or some quinolones [53,54,60]. Inhibition of growth at a low MIC using azithromycin in vitro has also been reported [60].

In spite of these mixed results using macrolides in vitro, studies of Q fever patients treated with macrolides suggest that these drugs can have some benefit. Acute Q fever pneumonia patients given treatment with macrolides have been reported. A study of 11 acute Q fever patients from Spain that were given erythromycin found that all 11 had rapid improvement and were afebrile by the fourth day of treatment [61]. Only 2 of 8 patients on other antibiotics improved during this time frame. Patients not responding to other antibiotics promptly improved after beginning erythromycin treatment [61]. A second study from Spain found that 25 acute Q fever pneumonia patients improved rapidly after treatment with erythromycin, although not as quickly as 23 patients receiving doxycycline treatment [27]. A larger study of 113 Greek acute Q fever patients found that erythromycin, clarithromycin, and roxithromycin were all effective at reducing the duration of fever [26]. Although not as effective as doxycycline, the macrolides were much more effective than β-lactams. Macrolides are an attractive potential alternative therapy for Q fever. Some macrolides are in pregnancy category B, and could be a safer alternative for treatment in some cases. However, it is clear that macrolides are not as effective as doxycycline, and many strains of C. burnetii appear to be resistant to erythromycin in vitro [46]. This suggests that macrolides should not be a primary choice for Q fever treatment. However, newer macrolides such as clarithromycin and azithromycin could be explored further and represent a possible alternative therapy, especially during pregnancy.

A common contemporary alternative to doxycycline is co-trimoxazole. This drug is a mixture of trimethoprim and sulfamethaxzole at a 1 to 5 ratio. It functions as a folic acid antagonist and blocks DNA synthesis to reduce bacterial replication. Co-trimoxazole is the recommended treatment for pregnant women as it has activity against C. burnetii, and when supplemented with folic acid, has a better safety profile in pregnancy than doxycycline [23]. The use of co-trimoxazole for Q fever during pregnancy is based primarily on a study published in 2007 that found benefit with co-trimoxazole therapy for at least 5 weeks during pregnancy [49]. Q fever patients that were put on co-trimoxazole long-term during pregnancy developed fewer obstetric problems than patients that did not receive this treatment. Pregnant patients treated with co-trimoxazole were also less likely to develop a chronic Q fever serologic profile [49]. Although other studies have found that the benefit of long term treatment during pregnancy may not be as large as that found in the 2007 study [41], and questions about the risk of using co-trimoxazole during pregnancy remain, current recommendations are for long-term co-trimoxazole treatment for pregnant Q fever patients [23,49]. Co-trimoxazole has been used sporadically for chronic Q fever, with variable results. Successful clearance of the bacteria has been reported, but also failures. It is recommended that pregnant women with a chronic serologic profile be placed on doxycycline plus hydroxychloroquine after delivery [23,49].

Doxycycline remains the first choice of treatment for acute Q fever, and doxycycline plus hydroxychloroquine is the first choice for chronic Q fever. No alternatives have the efficacy of doxycycline, and it has a long track record in successful treatment of C. burnetii infections [28]. A problem with the use of doxycycline for acute Q fever is the diagnostic delay. Q fever is rarely suspected in acute febrile illnesses, although studies have shown that it may be a fairly common cause of febrile disease [62,63]. If diagnostics are ordered, it usually consists of a serologic assay which provides a single titer. The prevalence of Q fever antibodies in the general population (3.1% in the US) makes it impossible to distinguish between a current infection and past infection based on one titer result [62]. In addition, titers may not be present until 7-10 days after symptom onset. For these reasons, a confirmed diagnosis usually depends on comparison between acute and convalescent serum samples, with a four-fold rise in IgG against Phase 2 C. burnetii confirming a diagnosis. This is of course not useful for a treatment decision, as the convalescent sample will be taken in most cases after symptoms have resolved and treatment is not recommended for asymptomatic patients. Thus, the diagnostic delay is an obstacle to timely treatment with doxycycline. The use of PCR can provide a much more timely diagnosis. C. burnetii DNA can be detected both in whole blood and serum in the time before IgG antibodies are present in the blood. PCR can therefore provide a positive diagnosis while patients are symptomatic and guide treatment decisions. In the recent Q fever epidemic in the Netherlands, PCR was considered an essential aspect of the diagnostic protocol [21]. Increased use of PCR for Q fever diagnostics could make administration of doxycycline for symptomatic patients more realistic.

For many years, chronic Q fever was a very difficult disease to treat, and there were many examples of relapse and treatment failures [10,28]. For these reasons various alternative therapies have been explored. The advent of the combination of doxycycline plus hydroxychloroquine as a treatment for Q fever endocarditis in the 1990’s has been very beneficial to patients as this has reduced relapses and the need for valve replacements [33]. However, this treatment still requires 18 months of antibiotic therapy, and some patients cannot tolerate doxycycline. Hydroxychloroquine can also cause ocular toxicity and may not be tolerated due to an impaired visual field. Although not common, isolates resistant to doxycycline have been reported [46,47]. It is therefore worthwhile to search for alternative treatments for Q fever endocarditis. Quinolones, rifampin, co-trimoxazole, and telithromycin have all been used on Q fever endocarditis patients [28], but these strategies have all been replaced by the doxycycline/hydroxychloroquine protocol [33]. A suitable alternative will need to have bactericidal activity and be tolerated for fairly long-term use.

An open question for Q fever treatment is the value of treating acute fever in preventing the later development of Q fever endocarditis. Rapid diagnosis and treatment of acute Q fever may not only be important for the relief of symptoms of the acute disease but could also have a role in preventing evolution of chronic disease. There could also be a benefit for high risk patients in longer term treatment after the normal two-week course of doxycycline. A recent report has evaluated the benefit of screening acute Q fever patients for valvulopathies and treating all patients with significant valvulopathies with doxycycline/hydroxychloroquine for 12 months [64]. This study found 31/72 acute Q fever patients had a significant valvulopathy. Eighteen of the 31 patients with acute Q fever and valvulopathy received doxycycline/hydroxychloroquine for 1 year and did not develop endocarditis during the follow-up period (1-4 years). The 13 patients that had acute Q fever and valvulopathy but did not receive prophylactic antibiotic treatment all developed endocarditis within one year. This study emphasizes the risk of endocarditis for acute Q fever patients with valvulopathy [14,64], and shows that it is possible to prevent endocarditis with doxycycline/hydroxychloroquine therapy. However, in other contexts, the benefit of screening every acute fever patient for valvulopathy may not be as significant [65]. Other patient groups may have lower incidence of valvulopathy and certain strains of C. burnetii may have more or less propensity for endocarditis [14].

The need for long-term treatment for Q fever during pregnancy is based on a limited number of patients. The rationale comes primarily from a study of 53 patients in France that had high rates of obstetric complications without therapy, but a very significant reduction with long-term (5 weeks or more) co-trimoxazole treatment [49]. Other studies have also found negative implications of Q fever during pregnancy [15-18]. However, a recent study from Germany of 11 women with Q fever during pregnancy found none with obstetric complications related to the infection, even though only 3 received long-term co-trimoxazole [41]. In the Netherlands, a serosurvey of pregnant women in a region that had high incidence of Q fever during 2007-2008 found 37 women with both IgM and IgG antibodies against Phase 2 C. burnetii, suggesting a recent infection [42]. The incidence of fetal complications was the same in this seropositive group as that found in seronegative mothers from the same region, suggesting that C. burnetii infection in these cases did not lead to adverse pregnancy outcomes [42]. Reasons for differences in the incidence of adverse pregnancy outcomes related to Q fever are not currently known. One possibility is that C. burnetii strain differences can impact the effect on pregnancy [50]. Because long-term treatment with antibiotics poses some risk to the fetus, it will be important to better understand the risks of C. burnetii infection for the fetus so that risk/benefit calculations can be weighed more effectively. Evaluation of genotype-specific impact on pregnancy could shed light on this matter.