Clinical Features and Complications of Coxiella burnetii Infections From the French National Reference Center for Q Fever

Key Points español 中文 (chinese)

Question  What are the characteristics and clinical presentations of Coxiella burnetii infection using 21st-century–clarified definitions?

Finding  In a cohort study of 2434 patients with Q fever, the following new critical Q fever foci were identified: acute endocarditis, lymphadenitis, and bone marrow involvement in hemophagocytic syndrome. Lymphadenitis is a risk factor for lymphoma, and the elevation of IgG anticardiolipin antibody titers in acute Q fever is associated with complications.

Meaning  Screening for anticardiolipin antibodies may help prevent acute Q fever complications; the use of transthoracic echocardiography in acute Q fever and positron emission tomographic scanning in suspected persistent focalized infection is justified to improve the care of patients with Q fever.

Importance  Q fever remains widespread throughout the world; the disease is serious and causes outbreaks and deaths when complications are not detected. The diagnosis of Q fever requires the demonstration of the presence of Coxiella burnetii and the identification of an organic lesion.

Objective  To describe the hitherto neglected clinical characteristics of Q fever and identifying risk factors for complications and death.

Design, Setting, and Participants  This prospective cohort study conducted from January 1, 1991, through December 31, 2016, included patients treated at the French National Reference Center for Q fever with serologic findings positive for C burnetii and clinical data consistent with C burnetii infection. Clinical data were prospectively collected by telephone. Patients with unavailable clinical data or an unidentified infectious focus were excluded.

Main Outcomes and Measures  Q fever complications and mortality.

Results  Of the 180 483 patients undergoing testing, 2918 had positive findings for C burnetii and 2434 (68.8% men) presented with clinical data consistent with a C burnetii infection. Mean (SD) age was 51.8 (17.4) years, and the ratio of men to women was 2.2. At the time of inclusion, 1806 patients presented with acute Q fever, including 138 with acute Q fever that progressed to persistent C burnetii infection, and 766 had persistent focalized C burnetii infection. Rare and hitherto neglected foci of infections included lymphadenitis (97 [4.0%]), acute Q fever endocarditis (50 [2.1%]), hemophagocytic syndrome (9 [0.4%]), and alithiasic cholecystitis (11 [0.4%]). Vascular infection (hazard ratio [HR], 3.1; 95% CI, 1.7-5.7; P < .001) and endocarditis (HR, 2.4; 95% CI, 1.1-5.1; P = .02) were associated with an increased risk of death. Independent indicators of lymphoma were lymphadenitis (HR, 77.4; 95% CI, 21.2-281.8; P < .001) and hemophagocytic syndrome (HR, 19.1; 95% CI, 3.4-108.6; P < .001). The presence of anticardiolipin antibodies during acute Q fever has been associated with several complications, including hepatitis, cholecystitis, endocarditis, thrombosis, hemophagocytic syndrome, meningitis, and progression to persistent endocarditis.

Conclusions and Relevance  Previously neglected foci of C burnetii infection include the lymphatic system (ie, bone marrow, lymphadenitis) with a risk of lymphoma. Cardiovascular infections were the main fatal complications, highlighting the importance of routine screening for valvular heart disease and vascular anomalies during acute Q fever. Routine screening for anticardiolopin antibodies during acute Q fever can help prevent complications. Positron emission tomographic scanning could be proposed for all patients with suspected persistent focused infection to rapidly diagnose vascular and lymphatic infections associated with death and lymphoma, respectively.

Q fever (or Query fever) is a globally widespread zoonosis that was first described in 1937. The previous dichotomy between acute and chronic Q fever, which was based on serologic criteria, has led to the current confusion between chronic infection and post–Q fever fatigue syndrome.¹ The diagnosis of infection requires the demonstration of an organic lesion (identified as the infectious focus) and evidence of a microbial infection (as proven by serologic findings, polymerase chain reaction analysis, culture, and/or immunohistochemical analysis using anti–Coxiella burnetii antibodies).^2-5 This concept of the disease and paradigm shift was made possible by revolutionary improvements in imaging during the 21st century. The systematic and early use of transthoracic echocardiography (TTE) and positron emission tomographic (PET) scanning has allowed the identification of infectious foci that were previously undetected and that are now crucial and decisive for the diagnosis of C burnetii infection and to guide the choice of therapeutic approach. Therefore, new definition criteria have recently been published for C burnetii endocarditis, vascular infection, osteoarticular infection, lymphadenitis, and interstitial lung disease.^2-5 Consequently, the term chronic Q fever should no longer be used.²

This change is particularly important because the serologic response is strain dependent. As an example, IgG antibody titers to phase I C burnetii are higher in French Guiana, where a unique strain is endemic, than in metropolitan France.^6,7 For this reason, Q fever postinfectious syndrome is defined by an association of elevated IgG titers to phase I C burnetii with subjective symptoms only. This definition is different from that of the Netherlands team, who consider these cases to be chronic infection.¹ The 26-year experience of the French National Reference Center for Q fever, which contains data collected prospectively from patients worldwide, gave us the opportunity to reanalyze C burnetii infection using the clarified 21st century definition and to highlight new, rare, and unusual infectious foci of the disease.

The French National Reference Center for Q fever is designed by the French government to collect data from patients with Q fever as part of epidemiologic surveillance and receives serum samples from France and abroad (eFigure 1 in the Supplement). Epidemiologic, clinical, and biological data from positive cases are collected as described later. In this prospective cohort study, we report the epidemiologic, clinical, and biological data collected in the French National Reference Center from January 1, 1991, to December 31, 2016. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.⁸ According to the procedures of the French Commission for Data Protection (Commission Nationale de l’Informatique et des Libertés), collected data were anonymized. The study was approved by the local ethics committee of IHU (Institut Hospitalo-Universitaire)–Méditerranée Infection and by the French National Drugs and Health Products Agency. An oral consent was obtained from study participants by the referral physician.

For each positive C burnetii test finding, clinical data were collected by telephone from the referral physician using a standardized questionnaire (eFigure 2 in the Supplement). Patients with positive serologic test results who presented with clinical manifestations consistent with an active C burnetii infection were included. We asked the referral physician for serologic tests 3 and 6 months after diagnosis to monitor clinical and serologic markers of improvement. The duration of monitoring was set at 5 years in case of persistent focalized infections. Patients with unavailable clinical data or unidentified infectious focus were excluded from the study analysis.

Serologic testing and molecular detection were performed as previously described.^9-11 Culture and immunohistochemistry using specific anti–C burnetii antibodies and fluorescence in situ hybridization were performed as previously described.^9,12-15

Primary (acute) C burnetii infection was defined by the association of acute clinical symptoms with the following serologic criteria: IgG titers representing phase II (≥200) and IgM titers representing phase II (≥50) or seroconversion within 3 months of the primary symptoms.¹⁶ Persistent C burnetii focal infection was diagnosed using the recently updated criteria as persistence of clinical symptoms for more than 3 months in addition to the identification of an infectious focus (eTables 1-3 in the Supplement).^2-4 Immunosuppression was defined in patients with known organ deficiency (those undergoing hemodialysis or before transplant), patients who were receiving an immunosuppressive drug or who underwent splenectomy, and patients with polymetastatic cancer.

Findings for G-isotype anticardiolipin (IgG aCL) antibodies were defined as positive at greater than 22 IgG anti–phospholipid-binding units (GPLU). Since 2012, tests have been systematically performed when an active C burnetii infection has been identified.^17,18

For all patients with a positive C burnetii serologic test result since 2001, we recommend a cardiac TTE to detect known or unknown valvular defects or new valvular lesions compatible with endocarditis, because these conditions require prophylactic or curative treatment.^17,18 Since 2009, the use of a PET scan is systematically recommended, when accessible, to detect deep infectious foci when focalized persistent infection is suspected.²

To compare the distribution of continuous or dichotomous variables between 2 groups, we used the 2-sided t test or the 2-sided Fisher test, respectively. Two-sided P < .05 was considered to indicate a significant difference between 2 groups. The mortality rate was computed as the number of deaths occurring in the cohort divided by the number of person-years during the study period. The Cox proportional hazards regression model was used to determine the factors associated with mortality risk. The proportional hazards assumption was tested in the Cox regression models by examining the rescaled Schoenfeld residuals. The indicative factors for complications were determined using the logistic regression model (for evolution to a persistent focalized C burnetii infection), the Poisson regression model (in cases of rare manifestations of acute Q fever), or the Cox proportional hazards regression model (for lymphoma). All multivariate models were adjusted for sex, age, and year category at baseline (before 2009, 2009-2012, and after 2012), and interactions were also tested. We used Stata/SE software (version 14.2; StataCorp LP) for all the analyses.

From 1991 to 2016, 277 666 serum specimens were tested for antibodies to C burnetii (eFigure 1 in the Supplement). Of 180 483 patients undergoing testing, 2918 had positive findings for C burnetti. Of these, 2434 had a positive serologic result and an identified infectious focus (1674 [68.8%] men and 760 [31.2%] women) (Figure 1). A total of 2105 patients (86.5%) lived in metropolitan France, and 222 (9.1%) lived in Latin America, mostly in French Guiana (eTable 4 in the Supplement). The mean (SD) age of patients was 51.8 (17.4) years (range, 0-98 years); 58 (2.4% of the patients) were 16 years or younger. The ratio of men to women was 2.2 in adults and 0.9 in children (27 males [46%] and 31 females [54%]) (eFigure 3 in the Supplement). Mean (SD) follow-up was 16 (29) months. The medical records and cardiac TTEs at the time of diagnosis revealed that 640 patients (26.3%) presented with a valvulopathy, 91 (3.7%) were immunosuppressed (eTable 5 in the Supplement), and 36 (1.5%) were pregnant women (eTable 6 in the Supplement). Positron emission tomographic scanning was performed for 291 patients (12.0%) (eFigure 4 in the Supplement).

Among the 2434 patients included, 602 (24.7%) had a single-day follow-up. Nine hundred twenty-four of 1806 patients with acute Q fever (51.2%) were followed up for more than 3 months, and 149 of 766 patients with persistent C burnetii infection (19.5%) were followed up for more than 5 years.

In the overall cohort, 1806 patients (74.2%) had acute Q fever, and 31.5% (766 of 2434) presented with a persistent focalized infection. Among the 2434 patients, hepatitis was the most frequent clinical form of Q fever (933 [38.3%]), followed by endocarditis (533 [21.9%]), and pneumonia (618 [25.4%]) (Figure 2). Hepatitis (836 [46.3%]), pneumonia (480 [26.6%]), and flulike syndrome (350 [19.4%]) were the main clinical presentations of acute Q fever, followed by lymphadenitis (97 [5.4%]) (eTables 7 and 8 in the Supplement).

Among the 766 patients diagnosed as having persistent focalized C burnetii infection, 581 (75.8%) presented with endocarditis, 145 (18.9%) had a vascular infection, and 56 (7.3%) had an osteoarticular infection (eTables 9-11 in the Supplement). The mean (SD) age at diagnosis was 60 (17) years for these 766 patients, and the median duration of follow-up was 15.1 months (interquartile range, 1.7-45.3 months). In 91 patients (15.7% of patients with a final diagnosis of endocarditis), the use of transesophageal echography was critical to identify a valvular lesion, which was not detected with TTE. We witnessed the evolution of acute Q fever to a persistent C burnetii infection in 138 patients (7.6%) (Table 1).

Lymphadenitis was identified in 97 patients (4.0%). Lymphadenitis was concomitant with persistent focal C burnetii infection in 36 of 97 cases and was the unique infective focus in 23 cases (eTable 8 and eFigure 5 in the Supplement). Positron emission tomographic scanning enabled the identification of deep lymphadenitis in 18 of 41 cases (43.9%).

In patients diagnosed with Q fever, 16 were diagnosed as having a lymphoma. Fourteen (87.5%) were men, 14 had B-cell non–Hodgkin lymphoma, and 2 had T-cell lymphoma (6 diffuse large B-cell lymphomas, 2 follicular lymphomas, 1 gastric lymphoma, 1 mucosa-associated lymphoid tissue lymphoma, 2 marginal zone lymphomas, 1 mantle cell lymphoma, and 1 lymphoplasmacytic lymphoma) (eFigure 6 in the Supplement). Fourteen patients with lymphoma presented lymphadenitis. One patient had concomitant hemophagocytic syndrome and no lymphadenitis.

Acute Q fever endocarditis (50 patients [2.1%]) was a risk factor for evolution to persistent focal C burnetii infection. Acute Q fever endocarditis occurred with hepatitis in 25 cases (50.0%) and with pneumonia in 13 (26.0%).

Thrombosis was diagnosed in 23 patients of our cohort. Thrombosis was concomitant with persistent focalized C burnetii infection in 12 cases, including 8 with endocarditis, 5 with vascular infections, and 1 with osteoarticular infection (3 patients presented with endocarditis and vascular infection).

Among the 2434 patients included in the analysis, neurologic involvement was diagnosed in 32 (1.3%). Fifteen patients presented with meningoencephalitis; 11, with meningitis; and 6, with encephalitis. In 25 of these patients, the neurologic infection occurred as a part of acute Q fever; in 7, it occurred as part of persistent endocarditis. Three patients who developed encephalitis as a complication of septic emboli died.

Thirty-one patients of the cohort (1.3%) had pericarditis. Twenty-three had acute Q fever, 11 had a persistent C burnetii infection, and 3 had both. Among the 7 patients with myocarditis, 2 developed severe complications, including 1 conduction defect and 1 case of endomyocardial fibrosis.

Eye involvement (n = 10) was the unique possible persistent focus of C burnetii infection in 7 cases and was concomitant with persistent lymphadenitis, pneumonia, and endocarditis in 1 patient each. Uveitis was the main observed manifestation (n = 7), followed by papillitis (n = 2), chorioretinitis (n = 1), and optic neuritis (n = 1). One patient presented with uveitis and papillitis.

Alithiasic cholecystitis was diagnosed in 11 patients (0.4%). In 4 of the 5 patients for whom IgG aCL was measured, these antibody titers were elevated (Table 2). For 2 patients, histologic analysis of the gallbladder showed inflammatory infiltrates, but immunohistochemical analysis yielded negative findings in both cases.

Hemophagocytic syndrome was diagnosed in 9 patients (0.4%), all with acute Q fever. In 1 patient, evolution to endocarditis occurred; 1 had evolution to vascular C burnetii infection; and 1 presented with a marginal B-cell lymphoma diagnosed based on a splenic biopsy. All 9 patients had elevated IgG aCL titers (>22 GPLU) (Table 2).

Seven patients with persistent focal C burnetii infection presented with interstitial lung disease. All patients had severe and advanced fibrotic lung lesions.¹⁹ A pseudotumor of the lung was detected in 3 patients. One had persistent endocarditis, 1 had persistent lymphadenitis, and 1 had acute Q fever.¹⁸

Seven patients with positive findings for C burnetii infection presented with giant cell arteritis, 5 had acute Q fever, and 1 had vascular infection. Half of the patients had positive IgG aCL titers.

In French Guiana, the ratio of men to women was 1.5, and acute pneumonia represented clinical presentation in 154 of 220 (70.0%). Only 13 patients (5.9%) presented with elevated aCL antibodies in the acute phase of the disease, which is much lower than that observed in metropolitan France (1115 of 2105 [53.0%]) (P < .001).

Ninety-one patients were immunocompromised. Among these, 52 were receiving immunosuppressive therapy. Sixty-six patients (72.5%) had acute Q fever, and 31 (34.1%) presented with persistent focal infection (eTable 5 in the Supplement).

Among 58 children, 4 were neonates. Fourteen children presented with persistent endocarditis, whereas lymphadenitis was the unique clinical presentation in 2 children (eTable 12 in the Supplement).

Thirty-six included patients were pregnant women (mean [SD] age, 30 [6] years). Infection occurred mostly during the 6 first months of pregnancy. Pregnancy complications were identified in 22 of these patients (61.1%) (eTable 6 in the Supplement).

Being male was associated with an increased risk of vascular infection independent of age (odds ratio [OR], 3.4; 95% CI, 2.0-5.7; P < .001). Men presented with higher IgG aCL titers than women in the primary phase of the disease (OR, 1.6; 95% CI, 1.3-2.1; P < .001), independent of age. Twenty-seven of 58 children (46.6%) were boys; 1654 of 2376 adults (69.6%) were men (P < .001).

Fifty-eight (2.4%) of the 2434 patients diagnosed with Q fever died. Among these, 46 were men (79.3%), the mean (SD) age at the time of Q fever diagnosis was 65.8 (12.9) years, and the mean (SD) age at the time of death was 69 (12) years. Among the 1806 patients with acute Q fever, 3 died (2 of complications of a solid tumor, and 1 of fulminant Q fever hepatitis). Among 766 patients with persistent focalized infection, 55 died, 43 had endocarditis (4 with concomitant vascular infection), and 16 had a vascular infection (of whom 4 had concomitant spondylodiscitis) (Figure 3).

Among the 2434 patients with Q fever (accounting for a total duration of follow-up of 3276 person-years, with a median follow-up duration of 4.6 months [interquartile range, 0.1-18.1 months]), the mortality rate was 1.8 (95% CI, 1.4-2.3) per 100 person-years. The mortality rate was 2.4 (95% CI, 1.8-3.3) per 100 person-years among patients with C burnetii persistent endocarditis, and 6.2 (95% CI, 3.8-10.1) per 100 person-years among patients with C burnetii vascular infection.

All Cox proportional hazards regression risk models used to test the association between each condition and mortality were adjusted for sex, age, and calendar period. Persistent focal C burnetii infections were associated with an increased risk of death (hazard ratio [HR], 10.9; 95% CI, 3.2-37.1; P < .001). Persistent endocarditis (HR, 2.4; 95% CI, 1.1-5.1; P = .02) and vascular infection (HR, 3.1; 95% CI, 1.7-5.7; P < .001) were associated with an increased risk of death. Meningitis (HR, 4.0; 95% CI, 1.4-11.6; P = .009) and spondylodiscitis (HR, 8.3; 95% CI, 3.3-20.9; P < .001), when associated with cardiovascular infection, were also associated with a higher risk of death.

According to the Cox proportional hazards regression model with lymphoma as the dependent variable, after adjustment for sex, age, and year category at baseline, we found that the presence of lymphadenitis was associated with a higher risk of lymphoma (HR, 77.4; 95% CI, 21.2-281.8; P < .001) as was the presence of hemophagocytic syndrome (HR, 19.1; 95% CI, 3.4-108.6; P < .001). Valvulopathy, thrombosis, lymphadenitis, maximum high IgG titer (>800), acute Q fever endocarditis, and male sex were identified as factors associated with evolution to persistent focal C burnetii infection (Table 1).

IgG anticardiolipin titers were available for 1328 patients (54.6%). Positive aCL antibody findings were indicative of acute Q fever endocarditis (incidence rate ratio, 3.9; 95% CI, 2.0-7.5; P < .001) and hemophagocytic syndrome (all patients in this group were positive for aCL antibodies) (Table 2). Immunosuppression was not indicative of any complication. In case of acute Q fever, the receiver operating characteristics analysis showed that the presence of aCL antibodies were significantly associated with acute Q fever complications such as acute Q fever endocarditis (area under the curve [AUC], 0.67; 95% CI, 0.58-0.76; P < .001), thrombosis (AUC, 0.72; 95% CI, 0.60-0.85; P = .002), hemophagocytic syndrome (AUC, 0.78; 95% CI, 0.67-0.89; P = .003), meningitis (AUC, 0.68; 95% CI, 0.56-0.79; P = .01), and alithiasic cholecystitis (AUC, 0.75; 95% CI, 0.60-0.90; P = .05) (eTable 13 in the Supplement).

We present a comprehensive description of a 26-year cohort of patients with Q fever from the French National Reference Center for Q fever. Endocarditis was the second infectious C burnetii focus identified.

New complications identified included acute Q fever endocarditis, hemophagocytic syndrome, thrombosis, lymphadenitis, and lymphoma. Anticardiolipin antibodies (IgG aCL) during acute Q fever were indicative of hepatitis, cholecystitis, endocarditis, thrombosis, hemophagocytic syndrome, meningitis, and progression to persistent endocarditis. Meningitis and spondylodiscitis complications were associated with an increased risk of death.²⁰

In the Netherlands, where Q fever has been responsible of 4000 cases in 4 years, physicians used a definitive criterion based on a serologic cutoff rather than the clinical features. Consequently, Q fever complication and unusual presentation of the diseases are probably underestimated.^21-23 In the Netherlands, the mortality rate varies from 1% in cases of acute Q fever to 13% in cases of persistent focalized C burnetii infection (9% for endocarditis and 21% for vascular infection).^22,24 To our knowledge, no study on aCL associated with Q fever has been published from the Netherlands. In addition, TTE is not performed in case of acute Q fever.²⁵ Therefore, comparison of the clinical features observed herein with those observed in the Netherlands remains difficult.

By contrast, in French Guiana, clinical manifestations of the disease presented some peculiarities.²⁶ Strikingly, the disease affects men and women almost equally. Although the Guiana strain has been described as a highly virulent strain in vivo (unpublished data; C.M., Aurelia Caputo, PhD, Yassina Bechah, PhD, et al; June 2018), no elevation of aCL antibody levels was observed in this region. Regarding lymphoma, a prospective study needs to be performed with the definition criteria used herein.

In pregnant women, the highest proportion of complications (61.1%) corroborates previous dramatic reports on Q fever during pregnancy.²⁷ Placentitis and microthrombi have been described.²³ In children, Q fever has been marked by an age-related increase in incidence.²⁸ The imbalance in the sex ratio distribution of the disease occurred after puberty, and males were most affected by vascular infections and presented with a higher secretion of aCL antibodies during acute Q fever. Thus, sex hormones likely influence the host’s response during Q fever, and further investigation is warranted to determine the mechanism involved.²⁹

In 2015, Melenotte et al¹³ described an association between C burnetii infection and non–Hodgkin lymphoma. In that study, C burnetii lymphadenitis and hemophagocytic syndrome were identified as risk factors of lymphoma. To determine how bone marrow and lymph nodes could influence lymphomagenesis, further investigations are warranted. In any event, because lymphadenitis was the unique infective focus of C burnetii in 23 patients (0.9%) in our cohort, and because lymphadenitis was identified in 43.9% of the cases with PET scan imaging, the latter is justified to identify C burnetii lymphadenitis as a prelymphomatous stage.³

First described as a subacute disease and later considered a fatal chronic disease, heart valvular injury is now an acute Q fever clinical entity.^30,31 Acute Q fever endocarditis is associated with evolution to persistent C burnetii endocarditis and must be seriously considered with immediate and systematic TTE.¹⁷

The proportion of neurologic involvement (1.3%) in our cohort is consistent with that in 14 previous international publications.^2,32-34 Capillary thrombi and small perivascular hemorrhages have been described, and C burnetii has been identified by immunofluorescence in the brain.^35-37 Regarding pericarditis and myocarditis, systematic TTE and the systematic prescription of serologic findings for C burnetii have improved the diagnosis of Q fever pericarditis.³⁸ Among 22 cases of myocarditis described in the literature, myocardial necrosis has been anecdotally reported.^9,39,40 Cases of optic neuritis (n = 6) and uveitis (n = 21) reported in the literature have been considered to be inflammatory phenomena triggered by the bacterium without evidence of C burnetii in the intraocular specimen.^41-45 In the literature, 17 cases of C burnetii alithiasic cholecystitis were reported. One case was confirmed by positive polymerase chain reaction results for C burnetii in the gallbladder, and 2 were associated with antiphospholipid antibodies.^46-56 Thirteen cases of Q fever hemophagocytic syndrome associated with acute Q fever have been reported in the literature, and an increase in aCL antibodies was reported in only 1 case.^57-61Coxiella burnetii interstitial lung disease, which was first described after outbreaks in the United States and Russia, is a rare and severe persistent focal C burnetii infection with advanced fibrotic lesions and poor clinical outcome.^62-64 First described in 1983 by Janigan and Marrie,⁶⁵ pseudotumors of the lung are rare manifestations of C burnetii.⁶⁶ After resection to exclude a tumoral process, histologic findings showed macrophage infiltration with C burnetii.^65,66 Finally, large vessel vasculitis in association with C burnetii infection has been described in 5 case reports.^67-70 For one of these cases, a high IgG aCL titer was observed.⁶⁸

Some limitations of the study need to be acknowledged. Six hundred fifteen patients with acute Q fever (25.3%) were lost to follow-up because in most cases their clinical course was favorable, and they no longer consulted their referring physician. In addition, cardiovascular C burnetii infections were probably overrepresented in this cohort because, as a reference center, we are solicited for severe C burnetii infections. Conversely, the mortality rate might be underestimated because of potential loss to follow-up.

Based on the new definition criteria, hitherto neglected foci of infection include the lymphatic system (ie, bone marrow, lymphadenitis) with a risk of lymphoma. Cardiovascular infections were the main fatal complications, highlighting the importance of routine screening of valvular heart disease and vascular anomalies during acute Q fever. Routine screening for aCL antibodies during acute Q fever can help prevent complications. Further investigations are necessary to evaluate the addition of hydroxychloroquine sulfate to doxycycline in cases of elevated aCL titers. A PET scan could be performed for all patients with suspected persistent focalized infection for early diagnosis of vascular and lymphatic infections associated with death and lymphoma, respectively.

Accepted for Publication: May 31, 2018.

Published: August 24, 2018. doi:10.1001/jamanetworkopen.2018.1580

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2018 Melenotte C et al. JAMA Network Open.

Corresponding Author: Didier Raoult, MD, PhD, Microbes, Evolution, Phylogénie et Infections, IHU–Méditerranée Infection, 19-21 Bd Jean Moulin, 13005 Marseille, France (didier.raoult@gmail.com).

Author Contributions: Dr Melenotte served as principal investigator. Drs Melenotte and Protopopescu had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Melenotte, Million, Bardin, Raoult.

Acquisition, analysis, or interpretation of data: Melenotte, Protopopescu, Million, Edouard, Carrieri, Eldin, Angelakis, Djossou, Fournier, Mège, Raoult.

Drafting of the manuscript: Melenotte, Protopopescu, Carrieri.

Critical revision of the manuscript for important intellectual content: Melenotte, Million, Edouard, Carrieri, Eldin, Angelakis, Djossou, Bardin, Fournier, Mège, Raoult.

Statistical analysis: Melenotte, Protopopescu, Million, Carrieri.

Obtained funding: Melenotte, Raoult.

Administrative, technical, or material support: Melenotte, Edouard, Eldin, Angelakis, Djossou, Bardin, Mège, Raoult.

Supervision: Carrieri, Angelakis, Fournier, Raoult.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by the French Government under the contract Méditerranée Infection 10-IAHU-03 from the Investissements d’Avenir program managed by the Agence Nationale de la Recherche. This work was also supported by Région Provence Alpes Côte d’Azur and European funding FEDER PRIMMI (Fonds Européen de Développement Régional - Plateformes de Recherche et d'Innovation Mutualisées Méditerranée Infection).

Role of the Funder/Sponsor: The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.