ARTICLE

Pediatric Encephalitis: What Is the Role of Mycoplasma pneumoniae?

Laura J. Christie, MDa, Somayeh Honarmand, MSa, Deborah F. Talkington, PhDb, Shilpa S. Gavali, MPHa, Chris Preas, BAa, Chao-Yang Pan, MPHa, Shigeo Yagi, PhDa, Carol A. Glaser, MD, DVMa
aCalifornia Department of Health Services, Viral and Rickettsial Disease Laboratory, Richmond, California; bCenters for Disease Control and Prevention, National Center for Zoonotic, Vectorborne, and Enteric Diseases, Division of Foodborne, Bacterial and Mycotic Diseases, Enteric Diseases Laboratory Preparedness Branch, Atlanta, Georgia

The authors have indicated they have no nancial relationships relevant to this article to disclose.

ABSTRACT
BACKGROUND. Encephalitis is a complex, debilitating, and sometimes fatal neurologic condition to which children are especially prone. Mycoplasma pneumoniae, a common respiratory pathogen, has been implicated as an etiology of encephalitis. Evidence for recent or acute M pneumoniae infection has been demonstrated in limited studies of both pediatric and adult patients with encephalitis. PATIENTS AND METHODS. Unexplained encephalitis cases are referred to the California

www.pediatrics.org/cgi/doi/10.1542/ peds.2007-0240 doi:10.1542/peds.2007-0240

Key Words Mycoplasma pneumoniae, encephalitis, pediatric Abbreviations HSV herpes simplex virus CNS central nervous system EIA enzyme immunoassay PCRpolymerase chain reaction CEPCalifornia Encephalitis Project CSF cerebrospinal uid EEG electroencephalogram EBVEpstein-Barr virus VZVvaricella-zoster virus IgMimmunoglobulin M IgGimmunoglobulin G WBCwhite blood cell CT computed tomography
Accepted for publication Mar 16, 2007 Address correspondence to Laura J. Christie, MD, Viral and Rickettsial Disease Laboratory, California Department of Health Services, 850 Marina Bay Pkwy, Richmond, CA 94804. E-mail: lchristi@dhs.ca.gov PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275). Copyright 2007 by the American Academy of Pediatrics

Encephalitis Project for diagnostic testing. Serum, cerebrospinal uid, and respiratory specimens are tested by polymerase chain reaction and serology methods for the presence of multiple pathogens, including M pneumoniae. M pneumonia associated cases of encephalitis were compared with other bacterial agents, herpes simplex virus 1, and enterovirus.
RESULTS. Of 1988 patients referred to the California Encephalitis Project, evidence of

acute M pneumoniae infection was found in 111 patients, of which 84 (76%) were pediatric patients. Eighty percent of the 84 patients were positive for M pneumoniae by serology alone. Cerebrospinal uid polymerase chain reaction for M pneumoniae was rarely positive (2%). Patients with M pneumoniaeassociated pediatric encephalitis were a median of 11 years old, progressed rapidly (median: 2 days from onset to hospitalization), and were often in the ICU (55%). Symptoms included fever (70%), lethargy (68%), and altered consciousness (58%). Gastrointestinal (45%) and respiratory (44%) symptoms were less common. Compared with patients with other bacterial as well as viral agents, patients with M pneumoniaeassociated encephalitis had fewer seizures and less-severe hospital courses.
CONCLUSIONS. M pneumoniae is the most common agent implicated in the California

Encephalitis Project. Patients with M pneumoniaeassociated encephalitis are predominantly pediatric, and their presentations are clinically similar to enterovirus encephalitis, although they frequently require intensive care with prolonged hospitalizations. Given that M pneumoniae infection is found more than any other pathogen, increased emphasis should be placed on elucidating the role and mechanism of M pneumoniae in encephalitis.
PEDIATRICS Volume 120, Number 2, August 2007 305

characterized by altered mental status, seizures, or focal neurologic signs. Although overall rates of encephalitis are low, 0.3 to 0.5 per 100 000 individuals in the United States per year,1 pediatric patients are overrepresented with an annual incidence of 10.5 per 100 000 child-years.2 Many etiologies, both infectious and noninfectious, can lead to a similar encephalitic presentation, with viruses such as enterovirus and herpes simplex virus (HSV) being the most common proven agents.1 Other agents, including bacteria such as Mycoplasma pneumoniae, fungi, and parasites also are potential etiologies. M pneumoniae is a ubiquitous pathogen, frequently causing respiratory disease such as pharyngitis, pneumonia, and tracheobronchitis, especially in children and the elderly.3 However, extrapulmonary complications are not uncommon and may affect nearly any organ system. One of the most common extrapulmonary manifestations is central nervous system (CNS) complications with encephalitis as the most common pediatric manifestation,4,5 but dermatologic, musculoskeletal, cardiac, hematologic, and renal disease are also described.3,5 Many of these nonrespiratory disorders are postulated to be a consequence of autoimmune reactions, as well as direct invasion.3 CNS conditions that have been associated with M pneumoniae include acute disseminated encephalomyelitis, Guillain-Barre syndrome, and transverse myelitis,6,7 as well as encephalitis and meningoencephalitis. In some studies of pediatric encephalitis, M pneumoniae is the most common agent identied.8,9 Although evidence of M pneumoniae infection was found in cases of encephalitis and was proposed as an etiology, difculty in diagnosis, the presence of M pneumoniae in the respiratory tract of healthy patients, and a lack of understanding of its precise role in CNS disease presents a challenge in categorizing M pneumoniae as a denitive cause. Lack of standardization of diagnostic tests is a considerable challenge and contributes to this controversy. Many previous studies relied on complement xation of patient sera to diagnose M pneumoniae infection; more recent studies of M pneumoniae infection in patients with encephalitis base identication on serology by using enzyme immunoassays (EIA) or on direct detection by culture or polymerase chain reaction (PCR). Direct detection of M pneumoniae is typically from the respiratory tract, less frequently from the CNS.8 The California Encephalitis Project (CEP) was initiated in 1998 to better understand the etiologies, risk factors, and clinical features of human encephalitis. Of nearly 2000 cases tested in the CEP, M pneumoniae was found to be the most frequently identied agent among the 14 agents tested. To our knowledge, this study describes the largest single group of encephalitis patients with laboratory evidence of M pneumoniae infection. This study denes the epidemiologic, clinical, laboratory, and
306 CHRISTIE et al

NCEPHALITIS IS AN inammation of the brain, often

imaging ndings of patients with encephalitis with evidence of M pneumoniae and examines other complementary testing methodologies for M pneumoniae diagnosis. PATIENTS AND METHODS Referrals to the CEP are received statewide from clinicians seeking diagnostic testing for immunocompetent patients 6 months old who meet the case denition: encephalopathy (depressed or altered level of consciousness for 24 hours, lethargy, or a change in personality) or ataxia, and have 1 of the following: fever (temperature 38C), seizure(s), focal neurologic ndings, cerebrospinal uid (CSF) pleocytosis, or abnormal electroencephalogram (EEG) or neuroimaging study. Demographic, clinical, and laboratory data are collected by a standardized case history form that accompanies clinical samples including acute and/or convalescent serum, CSF, and respiratory (eg, nasopharyngeal, endotracheal tube aspirates, sputum) samples. Testing for at least 14 agents is performed on all cases, with additional testing guided by exposures or clinical information. Current core testing includes testing of acute serum for St Louis encephalitis, West Nile virus, Western Equine encephalitis, measles, Epstein-Barr virus (EBV), and M pneumoniae. CSF is tested by PCR for HSV-1 and HSV-2, varicella-zoster virus (VZV), and enterovirus, as well as by serology for measles. HSV and VZV serology is performed on CSF if indicated. The respiratory swab is processed for viral isolation and is tested for enterovirus and M pneumoniae by PCR, as well as seasonally for inuenza A and B. When available, paired sera are tested for antibodies to St Louis encephalitis, Western Equine encephalitis, West Nile virus, M pneumoniae, HSV, VZV, adenovirus, Chlamydia spp., and seasonally for inuenza A and B.10 Additional testing (eg, Mycobacterium tuberculosis culture, bacterial culture, Bartonella spp. serology) may be performed by the Viral and Rickettsial Disease Laboratory or by the local hospital on the basis of individual clinical information. Testing has been modied over time with changes in available assays and technology. Glaser et al11 summarized CEP testing from 1998 to 2005. In accord with Title 17 of the California Code of Regulations, 2500 (j),(1) 26412643, encephalitis is a reportable communicable disease in California. As such, informed consent is not needed for routine diagnostic testing performed on patients specimens submitted through the CEP by the Viral and Rickettsial Disease Laboratory of the State of California. Routine M pneumoniae testing consists of EIA testing of serum for M pneumoniae immunoglobulin M (IgM) with the Meridian ImmunoCard Mycoplasma (Meridian Bioscience, Inc, Cincinnati, OH)12 and an in-house immunoglobulin G (IgG) EIA13 on both acute and convalescent serum when available. During the rst 2 years of the CEP, CSF samples from patients with serologic evi-

dence of M pneumoniae infection were sent to the Centers for Disease Control and Prevention for M pneumoniae PCR testing. From January 2001 to August 2003, M pneumoniae PCR testing was performed on all CSF samples at the California Department of Health Services.14 Given the low yield of positive CSF PCR results for M pneumoniae, CSF PCR testing was limited after August 2003 to only those patients in whom other testing (serologic, respiratory PCR) for M pneumoniae was positive. A case of M pneumoniaeassociated encephalitis is dened as a patient with evidence of acute M pneumoniae infection by a positive IgM, signicant rise in IgG titers between acute and convalescent specimens, or a positive respiratory or CSF PCR. We present data on all pediatric (18 years old) and adult patients with evidence of recent or acute M pneumoniae infection enrolled in the CEP from June 1, 1998, through July 31, 2006. Statistical analysis was performed by using the Kruskal-Wallis test, Fishers exact test, and 2 test as appropriate (statistical signicance set at .05). RESULTS Demographics and Clinical Characteristics During the study period of June 1998 through July 2006, 1718 acute sera were tested for M pneumoniae IgM, and 753 acute and convalescent pairs were tested for both IgM and IgG. From October 2001 to present, 1025 respiratory samples were routinely tested for M pneumoniae by PCR.14 During the study period, 600 CSF samples were tested for M pneumoniae by PCR. Among the agents identied, M pneumoniae was the most frequently found pathogen. One hundred eleven (5.6%) of 1988 patients enrolled in the CEP had evidence of recent or current M pneumoniae infection (Table 1). Of these, 84 (76%) of 111 were in patients 18 years old. Of note, only 46% of overall patients enrolled CEP are in the pediatric age group (906 of 1988). We further analyzed the pediatric subset, comparing them with 27 adult patients who had evidence of recent or acute M pneumoniae infection based on the same criteria.

Pediatric Diagnostic Testing In the pediatric patients with M pneumoniaeassociated encephalitis, the diagnosis was made primarily by the presence of M pneumoniae IgM (n 77 [92%]) in acute and/or convalescent serum (Table 1). Of those, 47 (61%) of 77 had a single serum specimen positive for IgM. Of the 7 pediatric patients without a positive IgM, the diagnosis was made by detecting a rise in IgG only (n 2) or by a positive respiratory PCR for M pneumoniae only (n 5). In 6 patients (15%), a positive IgM was accompanied by a rise in IgG, and in 12 patients (21%), both IgM and a respiratory PCR for M pneumoniae were positive. In a single pediatric patient, CSF was positive for M pneumoniae by PCR. Only 2 other patients within the CEP had positive CSF PCR tests for M pneumoniae; 1 was an adult patient with encephalitis, the other was a pediatric patient, who was excluded from this analysis because it was ultimately classied as meningitis rather than encephalitis. To investigate whether intrathecal antibodies were present, antibody testing for M pneumoniae was performed on a small subset of pediatric CSF specimens by sample availability. Positive M pneumoniae IgM testing was found in only 1 (3%) of 32 CSF samples; positive IgG testing was found in 2 (12%) of 17 CSF samples. Adult Diagnostic Testing In contrast with the pediatric patients and consistent with previous studies, adults were less likely to have a positive (12%) IgM (21 [78%]). Diagnosis based on an M pneumoniae IgG rise alone or a positive respiratory PCR for M pneumoniae alone was seen more often in adults than in pediatric patients (23% and 20% of adults, respectively). Only 6 adult patients (22%) were diagnosed with M pneumoniae by a single IgM-positive specimen, and a single adult patient had a positive CSF for M pneumoniae by PCR. Coinfections Apparent coinfections of M pneumoniae with another pathogen evidenced by serology, PCR, or culture were

TABLE 1 M pneumoniae Testing Results for Patients With Evidence of M pneumoniae Infection
No. Positive/No. Tested (%) IgM Only Pediatric (n 84b) Adult (n 27c) Total (n 111) 59/84 (70) 12/27 (45) 71/111 (85)

Respiratory PCR Only 5/56 (9) 2/10 (20) 7/66 (11)

Signicant IgG Change Only 2/39 (5) 5/22 (23) 7/61 (12)

CSF PCR Onlya 0/62 (0) 0/27 (0) 0/89 (0)

IgM/Respiratory PCR 12/56 (21) 0/10 (0) 12/66 (18)

IgM/Signicant IgG Change 6/39 (15) 8/22 (36) 14/61 (23)

IgM/CSF PCR 1/62 (2) 1/27 (4) 2/89 (2)

indicates positive. a CSF was tested routinely for M pneumoniae by PCR from January 2001 to August 2003; before and after this time period only samples with other evidence of M pneumoniae (ie, positive serology or respiratory PCR) were tested by CSF PCR. b One pediatric patient had positive IgM, a signicant rise in IgG, and a positive respiratory PCR and, thus, is counted twice in both the IgM/respiratory PCR and the IgM/signicant IgG change columns. c One adult patient had a signicant rise in IgG and a positive respiratory PCR and, thus, is counted twice in both the respiratory PCR only and signicant IgG change only columns.

PEDIATRICS Volume 120, Number 2, August 2007

307

TABLE 2 M pneumoniaeAssociated Encephalitis Cases With Evidence of Coinfections

Age, y 1 3 3 4 5 8 11 11 18 26 28 35 38 52 M pneumoniae IgM M pneumoniae IgG Change in M pneumoniae IgG Rise Steady M pneumoniae CSF PCR Rise Rise Rise Rise M pneumoniae NP/TH PCR Other Agentsa Enterovirus IgM Enterovirus IgM Enterovirus NP and CSF PCR Rhinovirus NP culture Inuenza B NP and ETT PCR EBV IgM, IgG EBV monospot, EBV IgM and IgG Enterovirus NP and CSF PCR B henselae IgG, parainuenza NP culture B henselae IgG with signicant rise in titer EBV CSF PCR Enterovirus CSF culture Inuenza B IgG with signicant rise in titer, and B henselae/quintana IgGb Rhinovirus NP culture

Rise

indicates positive; , negative; blank, testing was not performed; NP/TH, nasopharynx/throat; ETT, endotracheal tube. a Enterovirus IgM was tested by an in-house assay15, all other testing was performed as described by Glaser et al.11 b Unable to distinguish between B henselae and B quintanapositive IgG results.

found in both pediatric and adult patients (Table 2). Of patients with evidence of a single other pathogen in addition to M pneumoniae, the most common coinfection was with enterovirus,15 found in 5 patients. EBV was found in 3 other patients, whereas rhinovirus was identied in 2 patients. One patient each had evidence of inuenza B or Bartonella henselae in addition to M pneumoniae. Two patients had coinfections with 2 other agents in addition to M pneumoniae: 1 patient had evidence of coinfections with both B henselae/quintana and parainuenza, and a second patient had evidence of B henselae and inuenza B. Pediatric Demographics and Clinical Features The median age of the pediatric patients with M pneumoniaeassociated encephalitis was 11 years (range: 6 months to 18 years) (Table 3). There was no seasonality and no gender predominance (51% male) identied. Patients generally had a rapid onset of CNS signs (median time to admission: 2 days), although occasionally patients reported illness for several weeks to months before admission (range: 0 92 days). Pediatric patients were most commonly white, Hispanic (43%). Associated symptoms were common: 70% of pediatric patients were febrile, 45% had gastrointestinal symptoms, 44% reported upper respiratory infection symptoms, and 14% had rash. Just over half of pediatric patients (55%) required ICU care at the time of enrollment. Neurologic symptoms included lethargy (68%), altered consciousness (58%), focal neurologic signs (37%), and seizures (40%). Hallucinations were present in 18% of pediatric patients. Adult Demographics and Clinical Features Adult patients had similar demographics and clinical presentations to pediatric patients. However, adults
308 CHRISTIE et al

tended to be white, non-Hispanic more commonly than pediatric patients (52% vs 19%; P .01), and gastrointestinal symptoms were less frequent in adults (15% vs 45%; P .01). General Laboratory Findings CSF ndings were consistent with those reported in the literature for M pneumoniaeassociated encephalitis in children: mild-to-moderate pleocytosis (median: 33 white blood cells [WBCs]/mm3; range: 0 990/mm3; Table 4). The differential was predominantly mononuclear cells in 67% (35 of 52 samples with known differentials). The median protein level was 45 mg/dL (range: 7360 mg/dL), and glucose levels were generally normal (median: 62 mg/dL; range: 32172 mg/dL). Results of adult CSF panels were not statistically different. Pediatric Neuroimaging Cranial computed tomography (CT) was performed in 56 pediatric patients and was often normal (n 46 [82%]) (Table 4). In contrast, neuroimaging by cranial MRI showed abnormalities in 49%, including diffuse abnormalities or multilobar ndings (24%) and white matter abnormalities (16%). Of patients who had an EEG, most were abnormal (34 of 43 [79%]). The most common abnormality on EEG was diffuse slowing or dysfunction seen in over half (51%), whereas focal abnormalities (19%) or seizures (14%) were found less often. M pneumoniae in Comparison With Other Etiologies in the CEP Comparisons of demographic and clinical ndings were made between patients in the CEP with evidence of M pneumoniae to patients with (1) other bacterial and rickettsial agents (including Mycobacterium tuberculosis/bovis, B henselae, Neisseria meningitidis, Staphylococcus spp., Strep-

TABLE 3 Demographic and Clinical Findings in M pneumoniae Patients

n (%) Pediatric (N 84) Demographic Age, median (range), y Males Interval from CNS onset to admit, median (range), d Length of stay, median (range), d ICU admission Race/ethnicity White, Hispanic White, non-Hispanic Black Asian Other/unknown Clinical General symptoms Fever Gastrointestinal symptoms Upper respiratory infection symptoms Rash Neurologic symptoms Lethargy Altered consciousness Seizures Focal neurologic signs Stiff neck Hallucinations Coma 11 (6 mo18 y) 43 (51) 2 (092) 8 (172) 46 (55) 36 (43) 16 (19) 14 (17) 6 (7) 12 (14) Adult (N 27) 38 (2082) 15 (56) 2 (028) 10 (582) 14 (52) 5 (19) 14 (52) 2 (7) 2 (7) 4 (15)

54 (70) 38 (45) 37 (44) 12 (14) 57 (68) 49 (58) 34 (40) 31 (37) 20 (24) 15 (18) 12 (14)

20 (74) 4 (15) 8 (30) 4 (15) 20 (74) 19 (70) 7 (26) 9 (33) 10 (37) 3 (11) 6 (22)

tococcus spp., Rickettsia rickettsii, Klebsiella pneumoniae, and Tropheryma whipplei), (2) HSV-1, and (3) enterovirus (Table 5). Not surprisingly, patients with M pneumoniae and enterovirus were younger than those with HSV-1. Over half of those patients with M pneumoniae required ICU admission, signicantly less than patients with other bacterial agents. As expected, patients with evidence of M pneumoniae had respiratory symptoms although not signicantly more than for other etiologies of encephalitis. Patients with evidence of M pneumoniae clinically resembled patients with enterovirus infections, in that they had fewer seizures than those with HSV-1 or other bacterial agents. The CSF ndings of patients with evidence of M pneumoniae did not show any signicant differences compared with other etiologies. Abnormal neuroimaging was less frequently seen in those patients with recent M pneumoniae than in patients with HSV-1. DISCUSSION M pneumoniae is a widespread respiratory pathogen, increasingly recognized as an etiology of upper tract respiratory illness and pneumonia in all age groups. Nearly 80% of the population has evidence of past exposure to M pneumoniae by young adulthood,16 with seroprevalence increasing even more during epidemics, typically seen every 3 to 7 years.17,18 Extrapulmonary complications are not uncommon and have been estimated to occur in 5% to 10% in hospitalized patients.19,20 Bitnun et al8 found evi-

dence of M pneumoniae in 31% of pediatric patients with encephalitis in Toronto, and M pneumoniae was strongly implicated as the etiologic cause of encephalitis in 7%. Among patients referred to the CEP, M pneumoniae infection is the most common potential etiology identied, although diagnostic methods are imperfect. Serology, although a current standard of diagnosis, is problematic. The utility of IgM antibodies for diagnosis varies with age; it is found fairly consistently in acutely infected children but less frequently in adults.21 Adults are especially difcult to diagnose serologically because they have likely been infected with this fairly ubiquitous pathogen in the past, and during an acute infection no IgM may be generated or conversely, IgM may persist for months.6 Positive IgG results are difcult to interpret in the absence of paired serum specimens, because M pneumoniae is a common pathogen, and a single positive test may merely indicate past exposure. Although theoretically promising, serologic testing for IgM and IgG antibodies in CSF seems to be of limited use. This study demonstrates the variability in testing results and that a battery of tests, including serology and PCR, is required for diagnosis. The relatively low rate of respiratory symptoms in M pneumoniaeassociated encephalitis patients was not inconsistent with other reports. Bitnun et al8 reported only 13 (65%) of 20 patients with M pneumoniaeassociated encephalitis had respiratory symptoms, and suggested
PEDIATRICS Volume 120, Number 2, August 2007 309

TABLE 4 Laboratory Testing and Imaging Results in M pneumoniae Patients

n (%) Pediatric (n 84) Laboratory ndings CSF WBC count, median (range), cells per mm3 CSF mononuclear predominance CSF protein, median (range), mg/dL CSF glucose, median (range), mg/dL Studies CT normal MRI normal Abnormal MRI ndings Diffuse/multiple lobe White matter abnormality Thalamus or basal ganglia Meningeal abnormality Edema Gray matter abnormality Single lobe abnormality Brainstem abnormality EEG normal EEG abnormalities Diffuse slowing/dysfunction Focal ndings (not seizures) Seizures
an

Adult (n 27) 25 (01320) 13 (72b) 56 (14222) 60 (38140) 12 (57d) 10 (50f) 3 (15f) 4 (20) 0 0 3 (15) 1 (5) 1 (5) 0 2 (18h) 7 (64h) 5 (45) 0

33 (0990) 35 (67a) 45 (7360) 62 (32172) 46 (82c) 27 (49e) 13 (24e) 9 (16) 6 (11) 6 (11) 6 (11) 5 (9) 5 (9) 3 (6) 9 (21g) 22 (51g) 8 (19) 6 (14)

52; b n 18; c n 56; d n 21; e n 55; f n 20; g n 43; h n 11.

that respiratory symptoms seem to be less common in cases of neuroinvasive M pneumoniae disease than those that may be immune-mediated.6 In addition, M pneumoniae infection may have no or very mild respiratory symptoms5,6 such that upper respiratory infection symptoms were not reported in patients in the CEP. The variable length of IgM persistence could indicate M pneumoniae exposure in the months before presentation, not causally related to the current encephalitis. Finally, although the EIA kit used has a reported specicity of 90%,12 false-positives may occur and could account for a small proportion of those without respiratory symptoms or those in whom a second identied pathogen may be the main etiology of disease. Given difculties with interpretation of serology, direct identication of M pneumoniae by culture or PCR may seem appealing. Several investigators have identied M pneumoniae by culture from CSF or brain tissue in patients with encephalitis,6 but in general culture is rarely attempted because of low yield and the signicant labor and expense involved. The yield of identication of M pneumoniae by PCR from CSF or brain tissue is quite variable.8 Although evidence of M pneumoniae in the respiratory tract by PCR is useful in supporting its role in encephalitis, difculty in interpreting positive results may arise because M pneumoniae can persist for months after infection in the nasopharynx.18 Given the serious limitations of currently available diagnostic testing, testing based on specic pathogenic
310 CHRISTIE et al

mechanisms has been proposed to elucidate the role of M pneumoniae in encephalitis.22 Three prevalent pathogenic theories have been put forth in recent years: (1) direct neuroinvasion, (2) infection leading to immune dysfunction, and (3) neurotoxin elaboration.20,23 It is possible that multiple mechanisms may be occurring within a single patient, and different mechanisms may be used in different patients with M pneumoniaeassociated encephalitis. Neuroinvasion would seem to occur in a minority of cases as direct isolation or PCR identication of M pneumoniae from CSF or brain tissue is rarely found.8 No denitive evidence is yet available for a neurotoxin associated with M pneumoniae disease, although evidence of a toxin that directly contributes to cytopathology of mammalian cell lines was found.24 Autoantibodies seem a plausible mechanism, however, because M pneumoniae antigens were postulated to act as molecular mimics to components of myelin. Autoantibodies to myelin antigens were identied in some patients from several small studies after M pneumoniae infection, including antigalactocerebroside,22 antiasialoganglioside,25 and antimonosialoganglioside.26 Furthermore, in 1 study, antigalactocerebroside antibodies were found in 25% of M pneumoniae infections without CNS involvement and 100% of patients with CNS involvement.22 Interestingly, evidence of coinfection with other pathogens was found in this and other studies.6,8 This may suggest that M pneumoniae could also predispose patients to invasion by other pathogens via disruption of the respiratory or gastrointestinal mucous membranes; conversely, other pathogens may predispose patients to infection with M pneumoniae. Identifying the mechanism of disease may have important therapeutic implications. The role of antibiotics in M pneumoniaeassociated encephalitis remains unclear because there have been no controlled clinical trials and reports of children improving without therapy are found. Outcomes in case reports of pediatric patients treated with antibiotics have shown mixed results; some children improve during therapy whereas others do not.6,8,19,20 This lack of consistency may reect the different pathogenic mechanisms of CNS disease. Certainly, antibiotics such as macrolides, tetracyclines, and quinolones have shown in vitro activity against M pneumoniae, as well as in vivo activity in extra-CNS disease. However, antibiotics may have limited ability to cross the bloodbrain barrier and achieve therapeutic levels within the brain and spinal uid.6 Nevertheless, several review articles have proposed antibiotic therapy in M pneumoniaeassociated encephalitis, including azithromycin, ciprooxacin, doxycycline, or chloramphenicol.6,20 Intravenous immunoglobulin and steroids have also been proposed as therapeutic options for M pneumoniaeassociated CNS disease,20,27 especially in the context of M pneumoniaeassociated white matter disease, such as acute disseminated encephalomyelitis.

TABLE 5 Comparison of Pediatric Mycoplasma pneumoniae Patients to Other Pediatric Patients (Other Bacterial Agents, HSV-1, and Enterovirus)
n (%) Mycoplasma Pediatric Patients (N 84) Demographic Male Age, median (range) Race White, non-Hispanic White, Hispanic Black Asian Other/unknown Clinical ICU admission Prodrome or concurrent symptom Fever Respiratory symptoms Gastrointestinal symptoms Seizure Coma Death Length of hospital stay, median (range), d Laboratory data CSF WBC median (range), WBCs per mm3 CSF protein, median (range), mg/dL CSF glucose, median (range), mg/dL Abnormal neuroimaging ndings (initial) 43 (51) 11 (6 mo18 y) 16 (19) 36 (43) 14 (17) 6 (7) 12 (14) 46 (55) 54 (70) 37 (44) 38 (45) 34 (40) 12 (14) 3 (4) 8 (172) 33 (0990) 45 (7360) 62 (32172) 46 (39c) All Mycoplasma (N 111)a Other Bacterial Agents (N 45)b 30 (67) 20 (8 mo77 y) 21 (47) 15 (33) 1 (2) 7 (16) 1 (1) 34 (83) 33 (73) 11 (26) 17 (41) 21 (47) 13 (30) 5 (11) 12 (266) 85 (013000) 110 (12961) 60 (9132) 19 (44) HSV-1 (N 40) Enterovirus (N 42)

58 (52) 11 (6 mo82 y) 30 (27) 41 (37) 16 (14) 8 (7) 16 (14) 60 (54) 79 (71) 45 (41) 42 (38) 34 (40) 18 (16) 5 (5) 9 (182) 36 (01320) 52 (7360) 61 (32172) 96 (63d)

15 (38) 54 (8 mo89 y) 22 (55) 7 (18) 0 (0) 2 (5) 9 (23) 20 (57) 35 (90) 8 (21) 18 (46) 23 (59) 7 (18) 7 (18) 13 (0738) 42 (0975) 70 (15297) 69 (39112) 30 (79)

23 (55) 14 (7 mo74 y) 13 (31) 14 (33) 3 (7) 6 (14) 6 (14) 18 (50) 31 (74) 12 (29) 17 (42) 12 (29) 2 (5) 3 (7) 5 (01124) 101 (01080) 54 (16881) 68 (38159) 23 (58)

Denominators used in the calculations may vary slightly, depending on available data. a Includes pediatric and adult patients. b Includes pyogenic bacteria (see text for specic bacteria). c n 119 imaging studies, CT or MRI. d n 152 imaging studies, CT or MRI.

Among the cases investigated by the CEP, evidence of infection with M pneumoniae implicates it as the most common pathogen. Most patients in the CEP with M pneumoniaeassociated encephalitis were 18 years old, and of those pediatric patients, M pneumoniae was found more than twice as often as enterovirus and more than 7 times as often as HSV-1. The majority of pediatric cases were diagnosed by the presence of M pneumoniae IgM in serum. Testing for the presence of M pneumoniae IgM or IgG in CSF had a low yield. Not surprisingly, respiratory samples were more commonly PCR-positive (30% of pediatric patients tested) than CSF samples, which rarely yielded a positive PCR. Manufacturers literature indicates that the positive controls for the TaqMan PCR should detect levels as low as 10 gene copies14; thus, failure to detect M pneumoniae in CSF by PCR could be because of absence of organism or degradation of the organism during specimen transport. The higher frequency of positives in respiratory samples compared with CSF samples likely also indicates a much higher organism load in the respiratory tract. Respiratory symptoms are the most commonly recognized presentation of M pneumoniae illness, and the respiratory tract was the likely entry point of M pneumoniae infection for many patients in the CEP. Interestingly,

gastrointestinal symptoms such as nausea, vomiting, and diarrhea were found in a similar proportion of patients in this series as respiratory symptoms, although they are less frequently recognized as a presenting complaint of typical M pneumoniae infection.5,28,29 Many patients were ill enough to require ICU care, although mortality was low. In comparison to other etiologies of encephalitis, features of M pneumoniaeassociated encephalitis seemed more similar to an enteroviral rather than bacterial or HSV-1 etiology with fewer seizures and less ICU admissions. CSF ndings did not differ signicantly between M pneumoniaeassociated cases and other etiologies. However, it should be noted that the bacterial pathogens with which M pneumoniaeassociated cases were compared included a wide array of organisms, such as B hensalae, which typically has a very unremarkable CSF panel30,31 as well as pyogenic bacteria (eg, Neisseria meningitidis, Staphylococcus spp., Streptococcus spp., K pneumoniae) that frequently have a signicantly abnormal CSF panel.11 This inclusion of a diverse array of bacterial pathogens could result in a loss of signicance in the comparison with M pneumoniae. Neuroimaging was abnormal in less than half the patients; most commonly abnormalities were found in the white matter or as
PEDIATRICS Volume 120, Number 2, August 2007 311

diffuse abnormalities, signicantly lower than is seen in HSV-1 encephalitis. Despite frequently normal neuroimaging, EEG studies were commonly abnormal, often with diffuse slowing without evidence of seizures. CONCLUSIONS To our knowledge, this report represents the largest single group of patients with M pneumoniaeassociated encephalitis. Although the epidemiology and clinical presentations of these patients are consistent with M pneumoniae infection, strong evidence of M pneumoniae within the CNS is often lacking. Additional elucidation of the exact role and mechanisms are urgently needed in this eld. The mechanisms of M pneumoniae pathogenesis may inuence the choice of diagnostic methods, because detection of autoantibodies may be more helpful in cases of immune-mediated disease than CNS PCR, or conversely M pneumoniae CSF PCR may be useful if the organism uses direct neuroinvasion. In addition, treatment options may need to be modied as more data are generated to support or refute a direct association between encephalitis and M pneumoniae infection and by which mechanism it is acting. Relatively little research has been performed focusing on the role of M pneumoniae in neurologic disease to date; given the nding of comparatively large numbers of encephalitis patients in this analysis with evidence of M pneumoniae, additional studies are essential. ACKNOWLEDGMENT This work was supported by the Centers for Disease Control and Prevention Emerging Infections Program grant U50/CCU915546-09. REFERENCES
1. Willoughby RE. Encephalitis, meningoencephalitis, and postinfectious encephalomyelitis. In: Long SS, Pickering LK, Prober CG, eds. Principles and Practice of Pediatric Infectious Diseases. Philadelphia, PA: Elsevier Science; 2003:291292 2. Koskiniemi M, Korppi M, Mustonen K, et al. Epidemiology of encephalitis in children: a prospective multicentre study. Eur J Pediatr. 1997;156:541545 3. Waites K, Talkington D. New developments in human diseases due to mycoplasmas. In: Blanchard A, Browning G, eds. Mycoplasmas: Molecular Biology, Pathogenicity and Strategies for Control. Norfolk, United Kingdom: Horizon Bioscience; 2005: 293294 4. Talkington DF. Mycoplasmal and ureaplasmal infections. In: Scheld WM, Whitley RJ, Marra CM, eds. Infections of the Central Nervous System. Philadelphia, PA: Lippincott, Williams & Wilkins; 2004:605 611 5. Waites KB, Talkington DF. Mycoplasma pneumoniae and its role as a human pathogen. Clin Microbiol Rev. 2004;17:697728 6. Bitnun A, Ford-Jones E, Blaser S, Richardson S. Mycoplasma pneumoniae encephalitis. Semin Pediatr Infect Dis. 2003;14: 96 107 7. Candler PM, Dale RC. Cases of central nervous system complications associated with Mycoplasma pneumoniae. Pediatr Neurol. 2004;31:133138 8. Bitnun A, Ford-Jones EL, Petric M, et al. Acute childhood

9.

10.

11.

12. 13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

encephalitis and Mycoplasma pneumoniae. Clin Infect Dis. 2001;32:1674 1684 Kolski H, Ford-Jones EL, Richardson S, et al. Etiology of acute childhood encephalitis at the Hospital for Sick Children, Toronto, 1994 1995. Clin Infect Dis. 1998;26:398 409 California Emerging Infections Program, California Encephalitis Project, Enhanced Diagnostic Testing and Epidemiology. Available at: www.ceip.us/encephalitis.htm. Accessed December 12, 2006 Glaser CA, Honarmand S, Anderson LJ, et al. Beyond viruses: clinical proles and etiologies associated with encephalitis. Clin Infect Dis. 2006;43:15651577 Meridian ImmunoCard Mycoplasma [package insert]. Cincinnati, OH: Meridian Bioscience, Inc; 2002 Cremer NE, Cossen CK, Shell G, Diggs J, Gallo D, Schmidt NJ. Enzyme immunoassay versus plaque neutralization and other methods for determination of immune status to measles and varicella-zoster viruses and versus complement xation for serodiagnosis of infections with those viruses. J Clin Microbiol. 1985;21:869 874 Association of Public Health Laboratories. Real-time (TaqMan) PCR for respiratory bacterial pathogens. Available at: www. aphl.org. Accessed January 10, 2007 Schnurr D, Yagi S, Devlin R, Mohle-Boetani J, Dondero M. IgA and IgM ELISA for the study of an echovirus 30 outbreak in California. Presented at: Tenth Annual Clinical Virology Symposium; April 24, 1994; Clearwater, FL Nir-Paz R, Michael-Gayego A, Ron M, Block C. Evaluation of eight commercial tests for Mycoplasma pneumoniae antibodies in the absence of acute infection. Clin Microbiol Infect. 2006;12: 672 694 Weiner LB, McMillan JA. Mycoplasma pneumoniae. In: Long SS, Pickering LK, Prober CG, eds. Principles and Practice of Pediatric Infectious Diseases. Philadelphia, PA: Elsevier Science; 2003:1006 Daxboeck F, Krause R, Wenisch C. Laboratory diagnosis of Mycoplasma pneumoniae infection. Clin Microbiol Infect. 2003;9: 263273 Daxboeck F, Blacky A, Seidl R, Krause R, Assadian O. Diagnosis, treatment and prognosis of Mycoplasma pneumoniae childhood encephalitis: systematic review of 58 cases. J Child Neurol. 2004;19:865 871 Guleria R, Nisar N, Chawla TC, Biswas NR. Mycoplasma pneumoniae and central nervous system complications: a review. J Lab Clin Med. 2005;146:55 63 Granstrom M, Holme T, Sjogren AM, Ortqvist A, Kalin M. The role of IgA determination by ELISA in the early serodiagnosis of Mycoplasma pneumoniae infection, in relation to IgG and -capture IgM methods. J Med Microbiol. 1994;40: 288 292 Nishimura M, Saida T, Kuroki S, et al. Post-infectious encephalitis with anti-galactocerebroside antibody subsequent to Mycoplasma pneumoniae infection. J Neurol Sci. 1996;140: 9195 Fernandez CV, Bortolussi R, Gordon K, Lee SHS, Gatien JG, Shahdrabadi MS. Mycoplasma pneumoniae infection associated with central nervous system complications. J Child Neurol. 1993;8:2731 Kannan TR, Baseman JB. ADP-ribosylating and vacuolating cytotoxin of Mycoplasma pneumoniae represents unique virulence determinant among bacterial pathogens. Proc Natl Acad Sci U S A. 2006;103:6724 6729 Kusonoki S, Shiina M, Kanazawa I. Anti-Gal-C antibodies in GBS subsequent to Mycoplasma infection: evidence of molecular mimicry. Neurology. 2001;57:736 738 Susuki K, Odaka M, Mori M, Hirata K, Yuki N. Acute motor axonal neuropathy after Mycoplasma infection: evidence of molecular mimicry. Neurology. 2004;62:949 956

312

CHRISTIE et al

27. Sakoulas G. Brainstem and striatal encephalitis complicated Mycoplasma pneumoniae pneumonia: possible benet of intravenous immunoglobulin. Pediatr Infect Dis J. 2001;20:543545 28. Azimi PH, Chase PA, Petru AM. Mycoplasmas: their role in pediatric disease. Curr Probl Pediatr. 1984;14:1 46 29. Murray HW, Masur H, Sentert LB, Roberts RB. The protean

manifestations of Mycoplasma pneumoniae infection in adults. Am J Med. 1975;58:229 242 30. Glaser CA, Gilliam S, Schnurr D, et al. In search of encephalitis etiologies: diagnostic challenges in the California Encephalitis Project, 1998 2000. Clin Infect Dis. 2003;36:731742 31. Lewis P, Glaser CA. Encephalitis. Pediatr Rev. 2005;26:353363

EVOLUTION AT WORK: WATCHING BACTERIA GROW DRUG RESISTANT Day by day, the doctors unwittingly helped the bacteria infecting their young heart patient to evolve. The more intensively they treated his afiction with antibiotics, the more the microbes resisted the therapy. . . . Last month, an international team of 11 scientists, led by biologists at Rockefeller University in New York, for the rst time identied the genetic changes that occurred as Staph bacteria developed resistance to successive antibiotics, step by step, in the living test tube of a sick man. To document events inside these virulent cells, Rockefeller University biologist Michael Mwangi and his colleagues analyzed the infections genetic code as it changed in a series of blood samples taken during the patients stay. Their work, reported in the Proceedings of the National Academy of Sciences, details how the molecular mechanics of survival are strengthening many deadly diseases. Patient X died in October 2000 after a 12-week hospital stay. His case comes to light now because researchers only recently developed the computational techniques needed to sequence generations of bacteria. The hospital, which also wasnt identied, gave the patients Staph samples to the Rockefeller team for research purposes. The techniques still are too slow and expensive for clinical use. . . . When Patient X was admitted to the hospital, he was already suffering from a Staphylococcus aureus infection, but it was still vulnerable to antibiotics. During treatment, however, the bacteria quickly developed stronger resistance to four antibiotics, including vancomycin, the drug of last resort for intractable infections, the scientists reported. As living bacteria, the Staph were driven to survive. Every time the patient took his medicine, the antibiotics killed the weakest bacteria in his bloodstream. Any cell that had developed a protective mutation to defend itself against the drug survived, passing on its special trait to descendants. . . . These resistant microbes, all disease-producing organisms spawned by the original infection, quickly accumulated 35 useful mutations. Each one altered a molecular sensor or production of a protein. Researchers then matched these gradual genetic changes to increasing levels of drug resistance, shocked that it took so little to undermine the foundation of modern infectious-disease control. We have now really looked into the belly of the beast and seen the mechanism, said Rockefeller microbiologist Alexander Tomasz. . . . The death of Patient X highlights the speed of natural selection in fostering antibiotic resistance. When you talk about the evolution of an arm or an eye or a species, you might be talking about millions of years. You can get bacteria resistant in a week, Dr Mwangi said.
Hotz RL. Wall Street Journal. June 8, 2007 Noted by JFL, MD

PEDIATRICS Volume 120, Number 2, August 2007

313