This report reviews the pulmonary and extrapulmonary manifestation of infections due to Coxiella burnetii.
Q fever, a zoonosis, is due to infection with C. burnetii. This spore-forming microorganism is a small Gram-negative coccobacillus that is an obligate intracellular parasite. The most common animal reservoirs are goats, cattle, sheep, cats, and occasionally dogs. The organism reaches high concentrations in the placenta of infected animals. Aerosolisation occurs at the time of parturition and infection follows inhalation of this aerosol. There are three distinct clinical syndromes of the acute form of the illness: nonspecific febrile illness, pneumonia, and hepatitis. The chronic form of Q fever is almost always endocarditis, but occasionally it is manifest as hepatitis, osteomyelitis or endovascular infection.
The pneumonic form of the illness can range from very mild-to-severe pneumonia requiring assisted ventilation. Multiple round opacities are a common finding on chest radiography. Treatment with doxycycline or a fluoroquinolone is preferred. Susceptibility to macrolides is variable.
In conclusion, Coxiella burnetii pneumonia should be considered when there is a suitable exposure history and when outbreaks of a pneumonic illness are being investigated.
In August 1935, E.H. Derrick, a pathologist, who was Director of the Laboratory of Microbiology and Pathology at the Queensland Health Department, Brisbane, Australia, was contacted by the Director General of Health and Medical Services for Queensland and instructed to investigate an outbreak of an undiagnosed febrile illness among workers at the Cannon Hill abattoir in Brisbane 1.
Derrick 2 noted that this illness lasted 7–24 days and was characterised by fever, headache, malaise, anorexia and myalgia. Blood cultures were negative and serum samples had no antibodies to influenza, typhus, leptospirosis, typhoid and paratyphoid. Derrick 2 named the illness Q (for query) fever.
Subsequent investigations in Australia and in the USA resulted in the isolation of the aetiological agent of Q fever. It was eventually named Coxiella burnetii in honour of Burnet and Cox, the two scientists who played an important part in its discovery.
Early on there was no indication that C. burnetii was a respiratory pathogen.
C. burnetii is a pleomorphic coccobacillus with a Gram-negative cell wall that measures 0.2×0.7 µm and is an obligate intracellular microorganism 3. C. burnetii undergoes phase variation, which is akin to the smooth-to-rough transition of lipopolysaccharide (LPS) of Gram-negative bacteria 4. In experimentally infected animals the first antibody produced is to C. burnetii protein (phase II antigen), and later, antibody is produced to C. burnetii LPS (phase I antigen). There is also a phase intermediate between phase I and phase II 5, 6. In infected humans the predominant antibody response in acute Q fever is to phase II and in chronic Q fever it is to phase I antigen. There is no morphological difference between phase I and phase II cells, although they do differ in the sugar composition of their LPS 7, their buoyant density in caesium chloride, and in their affinity for basic dyes. The LPS of C. burnetii is nontoxic to chick embryos at doses of >80 µg·embryo−1 in contrast to Salmonella typhimurium LPS which is toxic in nanogram amounts 6.
Spore-like formation explains why C. burnetii is so successful as a pathogen. It can survive for 7–10 months on walls at 15–20°C, for >1 month on meat in cold storage and for >40 months in skimmed milk at room temperature 8.
Q fever is a zoonosis and direct or indirect contact with animals is important in its epidemiology. Cattle, sheep and goats are the primary reservoirs of Q fever for man; however, many different species of animals in different countries are infected with C. burnetii 9. C. burnetii has been identified in arthropods, fish, birds, rodents, marsupials and livestock 3. Indeed, it naturally infects >40 species (including 12 genera) of ticks found on five continents 3. Lice, mites and parasitic flies are also infected 10. C. burnetii localises to the uterus and mammary glands of infected animals 9. Infected cows can shed C. burnetii in milk for up to 32 months 11. Large concentrations of C. burnetii are present in the infected placenta and aerosols are created during parturition 12. Inhalation of these contaminated aerosols by susceptible humans results in Q fever. This explains why, in many areas, annual outbreaks of Q fever occur around the time of livestock kidding 13, 14. Pets, including cats, dogs, and rabbits are a new source of C. burnetii infection 15–19. In a recent study, the wild brown rat was implicated as a part of the link in Q fever between farm animals and cats 20. A family outbreak of Q fever in France was due to C. burnetii contaminated pigeon faeces 21.
In some countries, infection of domestic or wild animals results in considerable infection among humans in contact with these animals, whereas in other areas little if any transmission to man occurs 22, 23.
C. burnetii has been an extraordinarily successful pathogen. By 1955, C. burnetii was found in 51 countries on five continents 24. In the 1990s, New Zealand was one of the few countries that was free of C. burnetii infection 25. However, major differences occur in the manifestations of Q fever from country to country. In Nova Scotia, Canada, and in the Basque region of Spain, pneumonia is the predominant manifestation of Q fever 26, 27, while in the Canary Islands (southern Spain) it is fever and hepatitis 28. In contrast, in the south of France both hepatitis and pneumonia are observed but hepatitis is more frequent than pneumonia 29. The reasons for these differences are not currently understood. It is noteworthy that in an outbreak of Q fever in Bonavista, Newfoundland that was associated with exposure to infected goats 30, a nonspecific febrile illness was the major manifestation of the infection. However, in cat-related outbreaks in nearby Nova Scotia, pneumonia (whether it is associated with exposure to infected cats, dogs, rabbits) is the exclusive manifestation of C. burnetii infection 15–17. In addition, Q fever in a geographic area may be endemic or epidemic and shift back and forth between these two extremes.
A review of Q fever in Germany from 1947–1999 revealed a cyclical incidence pattern with peaks occurring every 5–10 yrs 32. The mean annual incidence ranged from 0.1–3.1 per million in various parts of the country 32. Forty outbreaks were identified since 1947. Sheep were the source in 24 outbreaks while cattle were implicated in four community outbreaks and two abattoir outbreaks 32.
Some aspects of the epidemiology of Q fever seem to be unique to Europe. British residents who lived along a road over which farm vehicles travelled, developed Q fever as a result of exposure to contaminated straw, manure, or dust from farm vehicles 35. In a Swiss valley, 415 residents who lived along a road over which sheep travelled to and from mountain pastures developed Q fever 36, 37. Those persons who lived in six villages close to the road had high rates of infection, ranging from 11.8–35.8% (mean 21.1%) while those who lived in villages off the road had significantly lower rates of infection (range 2.1–6.8% (mean 2.9%)). An outbreak of Q fever involving 58 people in Northern Italy was associated with three flocks of sheep which passed through the area between late May and early June 38. The prevalence of C. burnetii antibodies in these flocks ranged from 45–53%.
While the aerosol route is the major one whereby humans are infected, rarely is there person-to-person transmission 40–42 and infection via contaminated blood or the percutaneous route 43, 44. Person-to-person transmission is so uncommon that isolation is not recommended for patients who are admitted to hospital for treatment of acute Q fever.
There is a suggestion from epidemiological studies that ingestion of contaminated milk is a risk factor for Q fever infection 45, 46. However, evidence from experiments where contaminated milk was fed to volunteers is contradictory 47–49. In a case control study, Hatchette et al. 30 found that ingestion of pasteurised cheese and tobacco smoking were both risk factors for acquisition of Q fever during an outbreak of Q fever on a caprine cooperative in Newfoundland.
The present authors feel that the route of infection may explain the difference in the manifestations of Q fever in some countries e.g. pneumonia in Nova Scotia, Canada versus hepatitis in Marseille, France. The present authors used five different strains of C. burnetii to infect mice via the intraperitoneal or intranasal route. Those infected intranasally developed pneumonia only, while those infected intraperitoneally developed hepatitis, splenomegaly and pneumonia. Bronchiolar changes were seen only in mice inoculated intranasally 50. These data have been reproduced in a guinea pig model of Q fever 51.
Vertical transmission rarely occurs 52, 53 but increased surveillance may reveal additional cases of vertical transmission. Indeed, in the town of Martigues in Southern France, Q fever complicated ≥1 in 540 pregnancies 54.
Sexual transmission of Q fever has been demonstrated in mice 55 and viable C. burnetii has been demonstrated in bull semen 56. There is also a suggestion that Q fever can be transmitted sexually in humans 57. Milazzo et al. 58 reported the case of a 53-yr-old male who developed orchitis as a complication of Q fever. The orchitis had its onset 3 days after the patient had intercourse (29 days after onset of the patient's illness). Fifteen days later the patient's spouse developed Q fever. C. burnetii deoxyribonucleic acid (DNA) was identified by polymerase chain reaction (PCR) in the semen of the index case 4 and 15 months after onset of the acute illness
Clinical manifestations of Coxiella burnetii infection
The infections due to C. burnetii can be divided into the acute and chronic varieties. Chronic Q fever almost always means endocarditis or rarely hepatitis. Chronic Q fever will not be discussed further in this article.
Acute Q fever has three major manifestations: 1) a self limited febrile illness; 2) pneumonia; and 3) hepatitis.
Q fever pneumonia
A panel of experts reviewed the literature on pneumonia and summarised the aetiology of community-acquired pneumonia (CAP) as part of the process of developing the British Thoracic Society recommendations for treatment of CAP 59. In the UK, 1.2% of 1,137 patients had C. burnetii pneumonia, none of the 236 patients in one study of pneumonia treated on an ambulatory basis and none of the 185 patients treated in intensive care unit had this diagnosis 59. In six studies involving 654 patients from continental Europe who were treated on an ambulatory basis, 0.8% had Q fever pneumonia 59. In 23 studies involving 6,026 patients requiring hospitalisation for the treatment of pneumonia in continental Europe, 0.9% had Q fever pneumonia while none of 453 CAP patients hospitalised in Australia and New Zealand had this diagnosis 59. Surprisingly 2.3% of 1,306 patients hospitalised in North America for the treatment of CAP had Q fever pneumonia 59. However, all of the cases from the North American series were from one study from Nova Scotia 60. Furthermore, serological studies were not performed for C. burnetii in all of the studies that were reviewed, thus the importance of C. burnetii as a cause of CAP may be underestimated.
Symptoms and signs
Almost all patients with Q fever pneumonia complain of fever. A variety of other symptoms are frequently present, however headache is more common than it is in patients with pneumonia due to other aetiologies 61. Indeed patients with Q fever pneumonia often state that the headache is the most severe pain that they have ever had. Table 1⇓ gives the range of symptoms and signs in patients with Q fever pneumonia as reported in a variety of studies 61, 62. Table 1⇓ indicates that approximately one-half of the patients with Q fever pneumonia have physical findings suggestive of pneumonia on examination of the chest. Of course that means that half do not, thus a high index of clinical suspicion is often necessary to make a diagnosis of pneumonia in the first instance, and later a diagnosis of Q fever.
The spectrum of illness due to Q fever pneumonia ranges from very mild to very severe. The latter is infrequent but in the author's experience with over 300 cases of Q fever pneumonia ∼2% require admission to an intensive care unit. The patient described by Oddo et al. 63 is an example of severe respiratory distress syndrome due to Q fever pneumonia that required mechanical ventilation for 21 days.
It is not uncommon for a variety of extrapulmonary manifestations to be evident at the time of presentation or to appear during the course of the illness. These include bone marrow necrosis 64, haemophagocytosis 65, haemolytic anaemia 66, lymphadenopathy mimicking lymphoma 67, transient hypoplastic anaemia 68, splenic rupture 69, and erythema nodosum 70. Neurological manifestations include confusion, meningitis, meningoencephalitis, optic neuritis, and demyelinating polyradiculoneuritis 71–79. Pericarditis and myocarditis are also manifestations of Q fever that may occur in patients with pneumonia 80.
The total white blood cell count is usually normal, with 25% of patients having an elevated count. However, lymphopaenia is common. Thrombocytopaenia may be present in 10% of the patients at the time of presentation, however, thrombocytosis is usually seen in the recovery phase of the illness. Occasionally platelet counts of 1 million×109 L−1 are seen. A low serum sodium concentration may be present, usually as a result of inappropriate secretion of antidiuretic hormone. Mild elevation of liver function tests is not uncommon. Microscopic haematuria is present in ∼50% of patients with Q fever pneumonia. A variety of autoantibodies have been described in acute Q fever including antimitochondrial antibodies 81, anticardiolipin antibodies 82, 83 and antismooth muscle antibodies 84.
Chest radiographic manifestations of Q fever pneumonia
Table 2⇓ and figures 1–3⇓⇓⇓ summarise radiological features of Q fever pneumonia. Multiple rounded opacities are very suggestive of Q fever pneumonia in some geographic areas such as Nova Scotia, Canada. Other conditions, such as septic pulmonary emboli due to tricuspid valve infective endocarditis, can also present as multiple rounded pulmonary opacities, although history and physical examination serve to distinguish the two. In many instances, however, there is nothing distinctive about Q fever pneumonia on chest radiographs.
Laboratory diagnosis of Q fever pneumonia
In most instances the laboratory diagnosis of C. burnetii infection is serological. The complement fixation 89 and indirect immunofluorescence antibody tests are available 90, 91. The latter is best. An enzyme-linked immunoassay is also available in some centres 90. A four-fold or greater rise in antibody between acute and convalescent samples is diagnostic. In general, a 2-week interval between the acute and convalescent sample is sufficient. Diagnosis based on a single serum sample is not ideal, however, a phase II immunoglobulin (Ig)M titre of >1:64 or a phase II IgG titre of >1:256 by indirect fluorescent antibody (IFA) is strong evidence of recent C. burnetii infection using the IFA test.
C. burnetii can be isolated in embryonated eggs or in tissue culture. Most laboratories are not able to work with C. burnetii because of its extreme infectiousness. The shell vial technique is useful for isolating C. burnetii and for determining antibiotic susceptibility 92. C. burnetii has been isolated from the blood of ∼15% of patients with Q fever pneumonia (using tissue culture in a shell vial technique), sampled prior to antibiotic therapy, and during the first few days of disease, and in 50% of patients with Q fever endocarditis 93.
Determination of antimicrobial susceptibility of C. burnetii has been problematic, since it is an intracellular pathogen. However, there is a long history of efforts to provide antimicrobial susceptibility data about C. burnetii. Three model systems have been used: chick embryos, guinea pigs, and cell cultures. In general tetracyclines, quinolones, rifampin, telithromycin and clarithromycin are active against C. burnetii 97. Some strains are susceptible to erythromycin, others are not 97.
Sobradillo et al. 98 carried out a prospective, randomised, double-blind study of doxycycline and erythromycin in the treatment of pneumonia presumed to be due to Q fever in the Basque region of Spain. Forty-eight patients were proven by serological studies to have Q fever; 23 received 100 mg doxycycline twice daily, and 25 received erythromycin (500 mg every 6 h) for 10 days. Fever resolution was more rapid in the doxycycline-treated group (3±1.6 days versus 4.3±2 days for erythromycin-treated patients; p=0.05). The erythromycin-treated group had more gastrointestinal adverse effects (11 versus two for the doxycycline-treated patients; p<0.01). By day 40, the chest radiograph was normal in 47 of the 48 patients. The authors concluded that doxycycline was more effective than erythromycin, but they recognised the self-limiting and benign nature of most cases of pneumonia due to Q fever. Kuzman et al. 99 studied 64 patients with Q fever pneumonia. Twenty-two patients were treated with azithromycin (total dose 1.5 g administered over 3–5 days), 15 with doxycycline (100 mg b.i.d. for 10–14 days) and 15 received a variety of other antibiotics. The mean duration of fever was 2.5 days in the azithromycin-treated group, 2 days in the doxycycline-treated group, and 3.5 days in the patients who received other antibiotics. All patients were cured. A retrospective review of 130 patients with Q fever pneumonia treated between 1989–1995 was carried out by Kofterids et al. 33. Eleven patients who were treated with tetracycline became afebrile in a mean of 3 days, the 42 patients treated with erythromycin became afebrile in a mean of 4.26 days, and the 28 patients treated with β-lactam agents required 6.8 days to become afebrile. Fifteen per cent of the clarithromycin-treated patients were still febrile at 5 days compared with 35% of the erythromycin-treated patients and none of the tetracycline-treated patients.
A retrospective review of 19 patients with Q fever pneumonia showed that 11 were treated with erythromycin and eight with β-lactam antibiotics. The erythromycin-treated group became afebrile by day 3, while only two of the β-lactam-treated group were afebrile by this time (p<0.005) 100.
The treatment of choice for Q fever pneumonia is doxycycline for 10 days. Alternative therapies are a fluoroquinolone or a macrolide plus rifampin. The latter recommendations are based on in vitro susceptibility results and anecdotal experience.
Clinicians should be aware of the prevalence of Coxiella burnetii infections in the area in which they practice. Patients who present with pneumonia should be asked about risk factors for Q fever and if any of these are present, acute and convalescent serum samples should be collected and tested for antibodies to Coxiella burnetii. If a diagnosis of Q fever is confirmed Public Health authorities should be notified so that appropriate investigations can be carried out to determine the source of the infection.
↵Previous articles in this series: No. 1: Tärnvik A, Berglund L. Tularaemia. Eur Respir J 2003; 21: 361–373. No. 2: Mabeza GF, Macfarlane J. Pulmonary actinomycosis. Eur Respir J 2003; 21: 545–551.
- Received October 30, 2002.
- Accepted November 26, 2002.
- © ERS Journals Ltd