High Seroprevalence of Mycoplasma pneumoniae IgM in Acute Q Fever by Enzyme-Linked Immunosorbent Assay (ELISA)

Q fever is serologically cross-reactive with other intracellular microorganisms. However, studies of the serological status of Mycoplasma pneumoniae and Chlamydophila pneumoniae during Q fever are rare. We conducted a retrospective serological study of M. pneumoniae and C. pneumoniae by enzyme-linked immunosorbent assay (ELISA), a method widely used in clinical practice, in 102 cases of acute Q fever, 39 cases of scrub typhus, and 14 cases of murine typhus. The seropositive (57.8%, 7.7%, and 0%, p<0.001) and seroconversion rates (50.6%, 8.8%, and 0%, p<0.001) of M. pneumoniae IgM, but not M. pneumoniae IgG and C. pneumoniae IgG/IgM, in acute Q fever were significantly higher than in scrub typhus and murine typhus. Another ELISA kit also revealed a high seropositivity (49.5%) and seroconversion rate (33.3%) of M. pneumoniae IgM in acute Q fever. The temporal and age distributions of patients with positive M. pneumoniae IgM were not typical of M. pneumoniae pneumonia. Comparing acute Q fever patients who were positive for M. pneumoniae IgM (59 cases) with those who were negative (43 cases), the demographic characteristics and underlying diseases were not different. In addition, the clinical manifestations associated with atypical pneumonia, including headache (71.2% vs. 81.4%, p=0.255), sore throat (8.5% vs. 16.3%, p=0.351), cough (35.6% vs. 23.3%, p=0.199), and chest x-ray suggesting pneumonia (19.3% vs. 9.5%, p=0.258), were unchanged between the two groups. Clinicians should be aware of the high seroprevalence of M. pneumoniae IgM in acute Q fever, particularly with ELISA kits, which can lead to misdiagnosis, overestimations of the prevalence of M. pneumoniae pneumonia, and underestimations of the true prevalence of Q fever pneumonia.


Introduction
Q fever, a zoonosis distributed worldwide, is caused by infection with the obligate intracellular microorganism Coxiella burnetii [1]. The primary animal reservoirs include cattle, sheep, and goats, and the major transmission route is human inhalation of aerosols or ingestion of unpasteurized dairy products contaminated with feces, urine, or reproductive tissues of infected animals. The clinical presentation of Q fever includes acute and chronic forms [2]. Acute Q fever presents with asymptomatic infection or influenza-like symptoms with various degrees of pneumonia or hepatitis. Culture-negative infective endocarditis is the major presentation of chronic Q fever [2]. During the largest outbreak of Q fever in the Netherlands, pneumonia (61.5%) was the most common presentation, and hepatitis accounted for only 0.4% [3].
Mycoplasma pneumoniae and Chlamydophila pneumoniae are common atypical pathogens of community-acquired pneumonia [4]. Due to a laborious and expensive culture procedure, serological assessment of antibodies is widely used to diagnose M. pneumoniae [5] and C. pneumoniae [6], with microimmunofluorescence (MIF) as the recommended method. However, because MIF is a time-consuming procedure and requires an experienced operator, the enzyme-linked immunosorbent assay (ELISA) has become the most commonly used method in clinical practice.
Because culture of C. burnetii is a laborious procedure and must be performed in a biosafety level 3 laboratory, serological assessment of antibodies against C. burnetii has become the gold-standard method for clinical diagnosis of Q fever. However, serological cross-reactivity with other intracellular pathogens, including Bartonella species [7,8], Legionella micdadei [9], Rickettsia species [10], Chlamydophila species [11], Mycoplasma pneumoniae [10], cytomegalovirus [10], and Epstein-Barr virus [10], has been reported in Q fever patients. Recently, a case of acute Q fever masquerading as M. pneumoniae pneumonia was reported [12]. In clinical practice, we have found several cases of acute Q fever that were serologically positive for M. pneumoniae or C. pneumoniae IgM before the final results of acute Q fever examinations were available. This finding indicates that patients can be misdiagnosed as atypical pneumonia if they are not tested for Q fever.
The aim of this study was to investigate the seroprevalence of antibodies against M. pneumoniae and C. pneumoniae, the two most common pathogens of atypical pneumonia, in patients with acute Q fever.

Selection of study cases
Because acute Q fever, scrub typhus (caused by Orientia tsutsugamushi), and murine typhus (caused by Rickettsia typhi) are the most common rickettsioses in Taiwan, and because they are difficult to differentiate from each other by clinical manifestations [13], cases confirmed for the 3 diseases were included in the investigation. From April 2004 to December 2009, a total of 166 cases of acute Q fever, scrub typhus, and murine typhus were diagnosed and confirmed by the Centers for Diseases Control of Taiwan (Taiwan CDC) at E-Da hospital. Among them, the sera of 160 cases were available for study. To clarify the seroprevalence of each disease, 5 previously published cases of acute Q fever and scrub typhus coinfections were excluded [14]. Finally, the sera (acute or convalescent phase) of 155 cases (102 acute Q fever, 39 scrub typhus, and 14 murine typhus) were included in the study. Among them, 135 cases (89 acute Q fever, 34 scrub typhus, and 12 murine typhus) had paired sera (acute and convalescent phase) for investigating the rates of antibody seroconversion.

Ethics Statement
This study was approved by the Ethics Committee of the E-Da Hospital (EMRP-097-117). The committee waived the need for written informed consent because the demographic information and clinical data were retrospectively recorded, and all of the data were collected anonymously.

Clinical characteristics and data collection
The demographic information, clinical manifestations, and results of laboratory and imaging examinations of the included cases were obtained retrospectively by medical chart review and were recorded using an anonymous case record form.

Confirmatory diagnosis of acute Q fever, scrub typhus, and murine typhus
Serological tests for the presence of specific antibodies against C. burnetii, O. tsutsugamushi, and R. typhi were performed using an indirect immunofluorescence antibody assay (IFA) in the contract laboratory of the Taiwan CDC, as previously described [13]. Acute Q fever was diagnosed either by an anti-phase II antigen IgG titer of ≥ 1:320 and an antiphase II antigen IgM titer of ≥ 1:80, a four-fold or greater increase of anti-phase II antigen IgG titer in paired sera, or by blood that tested positive for C. burnetii DNA by polymerase chain reaction (PCR) [15]. Scrub typhus and murine typhus were diagnosed by a specific antibody titer of IgM ≥ 1:80, a four-fold or greater rise of IgG titer in paired sera, or by blood that tested positive for O. tsutsugamushi and R. typhi DNA by PCR, respectively.

Detection of serum antibodies against Mycoplasma pneumoniae and Chlamydophila pneumoniae
Serum was the residual specimen obtained for the diagnosisrelated purposes of rickettsioses for the Taiwan CDC, and it was stored at -80°C until analysis. The serum IgG and IgM antibodies against M. pneumoniae and C. pneumoniae were detected using commercially available ELISA kits of M. pneumoniae IgG (MYCG0350)/IgM (MYCM0350) and C. pneumoniae IgG (CHLG0510)/IgM (CHLM0510) (ELISA kits, NovaLisa™, NovaTec Immundiagnostica GmbH, Germany), respectively. These kits were used for the detection of serum M. pneumoniae and C. pneumoniae IgG or IgM in clinical practice at E-Da Hospital. To validate the results, a second commercial ELISA kit for the detection of M. pneumoniae IgG (SeroMP TM IgG)/IgM (SeroMP TM IgM) and C. pneumoniae IgG (SeroCP TM IgG)/IgM (SeroCP TM IgM) (ELISA kits, Savyon Diagnostics, Ashdod, Israel) was used. All of the procedures and the interpretation of antibody determinations were performed according to the manufacturer's instructions, and each examination was performed in duplicate.

Statistical analysis
Categorical variables were analyzed using the Chi-square or Fisher's exact test where appropriate. Continuous variables were analyzed using Student's t-test. All p values were twotailed, and a p value <0.05 was considered to be statistically significant. The data were analyzed with SPSS software for Windows (Release 15.0; SPSS, Chicago, IL).

Results
Overall status of M. pneumoniae and C. pneumoniae IgG/IgM in acute Q fever, scrub typhus, and murine typhus The serum IgG or IgM results of M. pneumoniae and C. pneumoniae in the 155 cases using the first ELISA kits are shown in Table 1. In acute or convalescent phase sera, 62 (40.0%) and 43 cases (27.7%) had M. pneumoniae IgM and C. pneumoniae IgM, respectively. The monthly distribution of serum positive cases of M. pneumoniae IgM is shown in Figure  1, and they reached peaks in March, August, and December. Figure 2 illustrates the age distribution of the serum positive cases of M. pneumoniae IgM, and it was highest for those aged between 40 and 59 years old. The positivity rates of M. pneumoniae IgM in the acute (13.9%, 0%, and 0%, p=0.012), convalescent (58.9%, 8.8%, and 0%, p<0.001), and acute or convalescent phases (57.8%, 7.7%, and 0%, p<0.001) of acute Q fever were significantly higher than scrub typhus or murine typhus (Table 1). In contrast, the positivity rates of M. pneumoniae IgG and C. pneumoniae IgG/IgM were not different between the 3 diseases. To further investigate the seroconversion of M. pneumoniae and C. pneumoniae antibodies at the acute and convalescent phases, the 135 patients with available paired sera were analyzed, and the results are shown in Table 2. The seroconversion rate of M. pneumoniae IgM in acute Q fever was significantly higher than scrub typhus or murine typhus (50.6%, 8.8%, and 0%, p<0.001). However, the seroconversion rates of M. pneumoniae IgG and C. pneumoniae IgG/IgM were not different. Among the 56 acute Q fever patients with serum M. pneumoniae IgM, only 2 (3.6%) and 16 (28.6%) had 4-fold increases of IgG and IgG seroconversion, respectively (Table  3). Taken together, only the positivity and seroconversion rates of M. pneumoniae IgM in acute Q fever were significantly higher than scrub typhus or murine typhus.

Results of M. pneumoniae and C. pneumoniae IgG/IgM in acute Q fever with the second ELISA kit
The second ELISA kits (Savyon Diagnostics, Ashdod, Israel) were used to validate the high seropositivity rate of M. pneumoniae IgM found in acute Q fever cases. The results and concordant rates compared to the first ELISA kits are listed in Table 4. These revealed positivity rates of 21% (21/100), 45.3% (39/86), and 49.5% (51/103) during the acute, convalescent, and acute or convalescent phases, respectively. Among the cases with available paired sera, the seroconversion rate of M. pneumoniae IgM was 33.3% (28/84). Various degrees of concordant rates were found when compared to the first ELISA kits. The 4-fold increase of IgG and IgG seroconversion in the 45 acute Q fever patients with serum M. pneumoniae IgM were both 11.1% (Table 3).

Status of M. pneumoniae IgM and clinical manifestations of acute Q fever
To investigate the possible clinical differences between acute Q fever with and without serum M. pneumoniae IgM, the 102 cases of acute Q fever listed in Table 1 were divided into 2 groups according to the results of the M. pneumoniae IgM in acute or convalescent phase serum for comparison (59 positive cases [57.8%] and 43 negative cases [42.2%]). There were no significant differences among demographic data, underlying diseases, or clinical symptoms and signs between the 2 groups of patients (Table 5). It is noteworthy that symptoms possibly associated with atypical pneumonia, such as headache (71.2%  Tables 5 and 6 between patients with and without seroconversion of M. pneumoniae IgM (data not shown).

Discussion
Various degrees of sensitivity, specificity, and cross-reactivity were reported in several commercially available kits of M. pneumoniae [16][17][18][19] and C. pneumoniae [20,21], but the sera from rickettsioses were rarely used as controls for the evaluation of specificity [19]. In the present study, 57.8% and 25.5% of acute Q fever, 7.7% and 28.2% of scrub typhus, and 0% and 42.9% of murine typhus cases were serum positive for M. pneumoniae IgM and C. pneumoniae IgM, respectively ( Table 1). The seropositive rates of M. pneumoniae IgM in acute Q fever were much higher than the acute infection rates of M. pneumoniae as determined by serological studies in healthy adolescents (6.0%) [22], adults with respiratory symptoms (3.3%) [23], and community-acquired pneumonia (CAP) (14.3%-20.0%) [24,25] in Taiwan. The seropositivity rates of C. pneumoniae IgM were also higher than the percentage of C. pneumoniae pneumonia in CAP in Taiwan (7.1%-13.0%) [24,25]. Accordingly, the high seropositivity rates of M. pneumoniae and C. pneumoniae IgM in the present study did not result from background seroprevalence. These results had two important impacts. First, in clinical practice, rickettsioses may be misdiagnosed as atypical pneumonia if only M. pneumoniae and C. pneumoniae antibodies are tested. Second, in the investigation of M. pneumoniae and C. pneumoniae ELISA kits, the sera from rickettsioses should be included for evaluating specificity and possible false-positivity of the ELISA kits.
Both Q fever pneumonia and M. pneumoniae pneumonia belong to atypical pneumonia, but serological studies for one in the other are rare [10,12,19]. A case of Q fever presented with acalculous cholecystitis was reported to have serum M. pneumoniae IgM (SeroMP TM ) [12]. Vardi [19]. In the clinical setting,  [26,27]. In an 11-year study conducted in the Canary Islands, 10% of CAP cases were mixed infections. C. burnetii was the most frequently isolated co-pathogen, but the pathogens were not clearly described [26]. In a study in northern Israel, Shibli et al. identified Q fever as the microbiological etiology of 8 CAP cases in which 3 (37.5%) and 2 (25.0%) were combined infections with C. pneumoniae and M. pneumoniae, respectively [27]. In clinical interpretations, the seropositivity and seroconversion of M. pneumoniae IgM suggests a recent infection, which indicates that as many as 50-60% acute Q fever cases could be diagnosed as M. pneumoniae infection without testing for Q fever. From the epidemiological perspective, this would underestimate and overestimate the true prevalence of Q fever and M. pneumoniae infections, respectively.
Acute Q fever, followed by scrub typhus and murine typhus, is the most common rickettsiosis, and it is an emerging and endemic disease in southern Taiwan [13]. In Taiwan, hepatitis rather than pneumonia is the predominant presentation of acute Q fever [28,29]. In our previous study, although all of the cases of acute Q fever had hepatitis, 19% of them had abnormal chest x-ray findings, and 9.5% had unilateral and 9.5% bilateral infiltration [29]. This result indicated that nearly 20% of acute Q fever cases might be diagnosed and treated as pneumonia based on chest x-ray findings without testing for Q fever. In the present study, 57.8% and 25.5% of acute Q fever cases had serum M. pneumoniae IgM and C. pneumoniae, respectively (Table 1), and 19.3% of cases with M. pneumoniae IgM had abnormalities suggestive of pneumonia by chest x-ray (Table 6). These findings highlight that, before the confirmatory results of Q fever are available or without testing for Q fever, approximately 60% and 25% of cases may be misdiagnosed as M. pneumoniae and C. pneumoniae infection, respectively. Conversely, acute Q fever may account for a portion of serologically diagnosed M. pneumoniae and C. pneumoniae pneumonia, and the true prevalence of acute Q fever presenting with pneumonia in southern Taiwan may be underestimated.
Several points suggest that the high seropositivity rate of M. pneumoniae IgM in acute Q fever may be due to serological cross-reactivity rather than true M. pneumoniae infection. First, in addition to developing auto-antibodies [10,30,31], serological cross-reactivity with other pathogens is common in Q fever, including M. pneumoniae [7][8][9][10][11]. Second, the risk of dual infections of non-zootic (M. pneumoniae) and zoonotic (C. burnetii) atypical pneumonia is low [32]. Third, the monthly and age distributions (Figures 1 and 2) were not typical for M. pneumoniae pneumonia, which generally predominates in autumn among children and young adults [33]. Fourth, the clinical symptoms and chest-x ray findings possibly associated with M. pneumoniae pneumonia were not different between acute Q fever patients with and without serum M. pneumoniae IgM (Tables 5 and 6). Fifth, it is unreasonable that patients with acute Q fever rather than scrub typhus or murine typhus tend to contract dual infections with M. pneumoniae. From the viewpoint of the immune response to M. pneumoniae, despite high seropositivity (57.8%) (  6% and 11.1%) and IgG seroconversion (28.6% and 11.1%) as measured by the two ELISA kits ( Table 3), indicating that the presence of IgM did not result from true infection. However, the hypothesis of cross-reactivity between acute Q fever serum and M. pneumoniae IgM ELISA kits requires determination by crossadsorption studies and Western blot analysis, which require culture and large quantities of C. burnetii and M. pneumoniae antigens in a professional research laboratory [7,9,34].  Table 3. Four-fold increases and seroconversion of M. pneumoniae IgG in acute Q fever with serum M. pneumoniae IgM by 2 ELISA kits. This retrospective clinical and serological study has several limitations. We only tested 2 ELISA kits commonly used in Taiwan for the serological diagnosis of M. pneumoniae and C. pneumoniae infections. Whether a similar phenomenon exists with other kits needs further investigation. The studied cases were mainly distributed across southern Taiwan, and the results require a large-scale study involving other countries to be confirmed. The hypothesis of serological cross-reactivity derived from clinical observations needs further investigation by basic medical research in a professional laboratory.
In conclusion, nearly 60% of acute Q fever cases were serum positive for M. pneumoniae IgM by ELISA. However, their clinical manifestations were not different from those that were serum negative for M. pneumoniae IgM. In addition, approximately 25% of acute Q fever cases were serum positive for C. pneumoniae IgM. Clinicians should be aware of the high seroprevalence of M. pneumoniae IgM, particularly using ELISA kits, in acute Q fever. This observation could lead to misdiagnosis, overestimation of the prevalence of M. pneumoniae infection, and underestimations of the true prevalence of Q fever pneumonia.   Table 6. Differences among imaging findings, laboratory examinations, and responses to treatment between acute Q fever patients whose sera were positive and negative for Mycoplasma pneumoniae IgM a .