Coexposure to Borrelia burgdorferi and Babesia microti Does Not Worsen the Long-Term Outcome of Lyme Disease

[Guest guest]

Clinical Infectious Diseases 2000;31:1149-1154© 2000 by the Infectious Diseases Society of America. All rights reserved.1058-4838/2000/3105-0007$03.00----------------------------------------------------------------------------Coexposure to Borrelia burgdorferi and Babesia microti Does Not Worsen theLong-Term Outcome of Lyme Disease J. Wang,1,2 H. Liang,1 Oliver Sangha,1 Charlotte B.,1 A. Lew,1 A. ,1 Victor Berardi,3,a Anne H.Fossel,1 and A. Shadick11Division of Rheumatology, Immunology, and Allergy, Brigham and Women'sHospital, and 2Department of Medicine, Massachusetts General Hospital,Boston; and 3Imugen, Norwood, MassachusettsReceived 10 December 1999; revised 4 April 2000; electronically published 6November 2000.Previous studies suggest that concurrent Lyme disease and babesiosis producea more sever illness than either disease alone. The majority of babesiosisinfections, however, are subclinical. Our objective was to characterize onthe basis of a total-population survey of Nantucket Island, Massachusetts,whether coexposure to Lyme disease and babesiosis causes more severe illnessor poorer long-term outcomes than Lyme disease alone. In this retrospectivecohort study, residents indicating a history of Lyme disease were comparedwith randomly selected population controls on a standardized medicalhistory, blinded physical examination, and serological studies for Borreliaburgdorferi and Babesia microti. Serological evidence of exposure tobabesiosis was not associated with increased severity of acute Lyme disease.The groups did not differ with regard to the prevalence of constitutional,musculoskeletal, or neurological symptoms a mean of 6 years after acute Lymedisease. Prior Lyme disease and serological exposure to B. microti are notassociated with poorer long-term outcomes or more persistent symptoms Lymedisease alone.---------------------------------------------------------------------------- Financial support: National Institutes of Health (grants AR36308 andAR02033) and the American Lyme Disease Foundation. Dr. Shadick is supportedby an Arthritis Foundation Investigator Award and a Mentored ClinicalScientist Award. a Present affiliation: Imugen, Reference Diagnostic Division, 315Norwood Park South, Norwood, MA 02062. Reprints or correspondence: Dr. A. Shadick, Division ofRheumatology, Immunology, and Allergy, PBB-B2, Brigham and Women's Hospital,75 Francis St., Boston, MA 02115 (nashadick@...).---------------------------------------------------------------------------- Lyme disease and babesiosis are infections that are transmitted by thesame Ixodes scapularis tick species. Since its first description in theUnited States in 1975, Lyme disease has become the most commonly reportedvectorborne disease in the United States. Nearly 80,000 cases have beenreported to the Centers for Disease Control and Prevention (CDC) since 1982[1]. Human babesiosis, recognized in North America since 1969, is emergingin areas where Lyme disease is endemic as well. First reported fromNantucket Island, Massachusetts, babesiosis has also been described asoccurring in New York, Connecticut, Rhode Island, Minnesota, Wisconsin,Georgia, California, and Washington [25]. The infectious agent in theNortheast is Babesia microti, a protozoan parasite that infectserythrocytes. Acute babesiosis can vary in presentation from subclinical tofatal disease; advanced age and immunosuppression are believed to be factorsrelated to a poor prognosis [2, 6, 7]. The syndrome includes fever, chills,malaise, headache, hemolytic anemia, and hepatosplenomegaly. Because Borrelia burgdorferi and B. microti are transmitted by the sametick vector, concurrent infection may occur. There is a relative paucity ofinformation regarding concurrent infection, possibly because of the highrate of subclinical infection with B. microti. The majority of babesialinfections are asymptomatic [4, 6, 8]; in serological studies on NantucketIsland and Block Island, RI, only 6 of 19 and 6 of 34 seropositive subjects,respectively, had medical histories indicating prior babesiosis infection.It is not known whether subclinical babesiosis predisposes individuals toexperience more severe episodes of Lyme disease or increases the likelihoodof poor outcomes because of inadequate treatment for babesiosis. Somepatients with "atypical" Lyme disease (for instance, Lyme disease withouterythema migrans) may actually have babesiosis instead. In 1993 we initiated a population-based study on Nantucket Island tostudy the long-term outcomes of Lyme disease [9]. This large cohort was alsoevaluated for acute Lyme disease and concurrent or previous serologicalevidence of babesiosis. The population-based design permits us tocharacterize the epidemiology of concurrent infection and the subsequentcourse of Lyme disease.Materials and Methods Participants. From 1993 through 1995, surveys were mailed to all6046 adult (aged 17 years) permanent residents of Nantucket Island. Theisland, located off the southeastern coast of Massachusetts, has the highestreported incidence of Lyme disease in the country [10]. Adults wereidentified by census files. The survey inquired about prior diagnoses ofLyme disease or babesiosis, Lyme-related symptoms, and tick exposure. Twofollow-up mailings were sent to nonrespondents. All respondents who reported a clinician's diagnosis of Lyme disease ora history of a positive Lyme serological test were invited for clinicalevaluation. In addition, a random sample of the remaining respondents,balanced by age tertile and gender, were invited as potential controls. Anoutline of the sampling scheme for this study evaluation has been detailedelsewhere [9]. Informed consent was obtained. The study was approved by theInstitutional Review Board of the Brigham and Women's Hospital, Boston. Clinical evaluation. The clinical evaluation included acomprehensive medical history and physical examination. The medical historyincluded information on symptoms during any prior episodes of Lyme disease,comorbidities, and current constitutional, musculoskeletal, and neurologicalsymptoms. The physical examination was performed by an investigator blindedto disease status and included the American College of Rheumatology Glossaryjoint examination [11] and a standardized neurological examination of musclestrength, deep tendon reflexes, and sensation to light touch and vibration. Serological studies. All subjects had blood drawn for serologicaltesting. Serum samples were tested for IgM, IgG, and IgA antibodies to B.burgdorferi, by means of antibody-capture EIA [12], and for the pattern ofIgG reactivity to polypeptides of B. burgdorferi by means of Westernblotting [13]. In addition, serum samples were tested for IgG antibodies toB. microti by indirect immunofluorescence [14]. Confirmation of Lyme disease and babesiosis. We classified subjectsas having Lyme disease on the basis of the CDC surveillance case definition,which requires (1) the presence of erythema migrans or (2) serologicalconfirmation of infection along with evidence of 1 late manifestation ofdisease [15]. Late manifestations include arthritis, meningitis, cranialneuritis, radiculoneuropathy, encephalomyelitis, myocarditis, and acutesecond- or third-degree atrioventricular block. Subjects were said to have acute concurrent infection if they reporteda history of simultaneous Lyme disease and babesiosis, met the CDC criteriafor Lyme disease, and had documented laboratory confirmation of acutebabesial infection. Laboratory confirmation consisted of identification ofintraerythrocytic parasites on a peripheral blood smear or an acute, 4-foldrise in titer of antibody to B. microti during the period of acute Lymedisease symptoms. Cases of coexposure were those in which patients met theCDC criteria for Lyme disease and had antibodies to B. microti but did notreport an episode of acute concurrent infection. Confirmatory andsupplementary information was sought from the ambulatory-care and hospitalrecords for subjects who reported babesiosis. Statistical analysis. All data were analyzed with the SASstatistical package on a SUN Sparcstation 2000 system (Sun Microsystems,Inc., Palo Alto, CA) [16]. Differences between groups were evaluated withthe 2 and Student's t-test for categorical and continuous outcomes,respectively. The Fisher's exact test was used to analyze categorical datain instances where 1 cells had expected counts <5. All P values were2-tailed and were considered significant if <.05. We estimated the population prevalence of babesiosis from our surveydata and results of the clinical evaluation. Our sampling technique, whichwas designed to maximize case-finding, sampled 5 subjects who reported ahistory of Lyme disease and 2 subjects who reported no history of Lymedisease (1 of which had symptoms suggestive of possible infection), so weused weights to correct for this sampling. We assigned 1 of 3 weights toeach observation (which reflected the relative frequency of these responsesin the postal survey), according to group. These weighted values were thensummed across the 3 survey groups to approximate the prevalence of exposureto Babesia species on Nantucket Island. For example, there were 3 classes of subjects from the initial postalsurvey (3703 subjects): history of Lyme disease, no history of Lyme disease,and possibly infected with Lyme disease (see Methods, Participants). Theproportions of subjects in the postal survey falling into each of thesecategories were as follows: history of Lyme disease (443 [12%]), no historyof Lyme disease (1893 [51%]), and possibly infected with Lyme disease (1367[37%]). The seroprevalence of Babesia among the examined subjects (n = 336)was as follows: 16% for subjects with a history of Lyme disease in theinitial postal survey (not necessarily CDC criteriaproven cases), 0% ofsubjects who gave no history of Lyme disease, and 11% of subjects who wereidentified as having a possible history of Lyme disease. To calculate thefinal population seroprevalence of Babesia exposure, the following equationwas used: (16% × 12%) + (0% × 51%) + (11% × 37%) = 6.0% (see Results).Results A total of 3703 of 6046 individuals responded to the postalquestionnaire after a total of 3 mailings, for a response rate of 61%; 569adults were invited for clinical evaluation, and 336 (57%) of them wereevaluated for this analysis, of which 171 had prior Lyme disease. Theseroprevalence of B. burgdorferi and B. microti in subjects' serum samplesis detailed in table 1. The rate of Babesia seropositivity was higher amongpatients who had prior Lyme disease than among the controls (22% vs. 7%; P =.001).Table 1. Seroreactivity to Borrelia burgdorferi and Babesia microtiamong subjects. The estimated seroprevalence of babesiosis on Nantucket Island was 6.7%(95% CI, 4.0%, 9.4%). The prevalence of coexposure was 3.5% (95% CI, 1.5%,5.4%). Thus, 26% of individuals seropositive for Lyme disease and 52% ofindividuals seropositive for babesiosis had serological evidence ofcoexposure when the data were adjusted with population weights (see Methods,Statistical Analysis). Four patients (2%) reported acute concurrent infection with Lymedisease and babesiosis. All met CDC case criteria for Lyme disease and hadpositive smears for babesiosis or a 4-fold rise in titers of antibody to B.microti. Details of their illnesses are shown in table 2. Two subjects werehospitalized; 1 required a blood transfusion and another presented withatrial fibrillation and a low hematocrit value. All 4 recovered afterantibiotic therapy against both organisms (see table 2).Table 2. Clinical data for subjects who had acute concurrent infectionsof Lyme disease and babesiosis. In table 3, subjects who met the CDC criteria for Lyme disease areclassified according to whether they had evidence of prior infection withbabesiosiseither a positive titer of antibody to B. microti or a positiveblood smear. Subjects who reported a history of babesiosis but lackedserological evidence or chart documentation were excluded from the analysis(5 patients). Evidence of previous babesial infection (found in 22% ofsubjects with Lyme disease) was not associated with increased severity ofacute Lyme disease, as judged by the number of symptoms, hospitalization, ormissed days of work. Anorexia was the only symptom associated by trend (P =.05) with previous exposure to Babesia. The prevalence of constitutional,musculoskeletal, or neurological symptoms and physical examination findingsat the time of our evaluation (a mean of 6 years after the initialmanifestations of Lyme disease) was also similar in both groups (table 4).Table 3. Characteristics of patients with Lyme disease, as related toexposure to babesiosis.Table 4. Characteristics, at follow-up evaluation, of patients withLyme disease, as related to exposure to babesiosis. Subjects with "atypical" presentations of Lyme disease, includingsystemic symptoms without an erythema migrans rash or a self-reportedhistory of Lyme disease that did not meet CDC criteria, did not have asignificantly different rate of seropositivity for B. microti (Lyme diseasewithout erythema migrans vs. Lyme disease with erythema migrans: 20% vs.22.1%, P = NS; self-reported Lyme disease that did not meet CDC criteria vs.self-reported Lyme disease that did meet CDC criteria: 17.7% vs. 20.6%, P =NS).Discussion Most infections with B. microti appear to be subclinical. The majorityof individuals in our study and in previous studies who had positive titersof antibody to Babesia reported no history of babesiosis [6, 8]. Although B.burgdorferi and B. microti are frequently acquired concomitantly, aconcurrent diagnosis of both infections is uncommon. Human babesiosis is anopportunistic disease. Clinical babesiosis emerges with host risk factorssuch as increased age, prior splenectomy, immunosuppression, prematurity,and liver disease [1719]. Only 4 individuals in our study reported acute concurrent infection.All 4 had constitutional symptoms (i.e., high fevers) and findings of anemiaor thrombocytopenia, which are most suggestive of babesiosis; theselaboratory findings are uncharacteristic of Lyme disease alone [20]. Theseverity of their illnesses was moderate to severe, and 2 subjects requiredhospitalization. The conditions of all improved after antibiotic therapyagainst B. burgdorferi and B. microti; clindamycin and quinine were usedagainst the latter, and either ceftriaxone or amoxicillin against theformer. Host characteristics affect the likelihood of developing babesiosisbut also may increase the likelihood of acute concurrent infection. Two ofthe 4 patients were in their seventies. None had undergone a splenectomy. Inthe present study, when coinfection occurred, the symptoms were mostsuggestive of babesiosis and the illness was quite severe, suggesting thatcoinfection with B. burgdorferi may alter symptomatic babesiosis. The acute manifestations of concurrent infection have been described in3 case reports and a recent report of a prospective study [17, 2123]. Onepatient with concurrent Lyme disease and babesiosis had pancarditis and isthe only reported fatality with Lyme disease [23]. Another reported patienthad noncardiogenic pulmonary edema [17]. Krause et al. [21] described 26subjects with concurrent infection and found them to have more symptoms anda greater duration of illness than those with Lyme disease alone. Thelong-term course of concurrent infection, however, has not been evaluated.Krause et al. [21] identified 26 residents of Block Island with symptoms andsigns of acute concurrent infection from 1990 to 1994 and followed themuntil they became asymptomatic. Compared with patients infected with B.burgdorferi alone, the concurrently infected patients had more symptoms anda longer disease course; 13 of the 26 cases were symptomatic for 3 months.The authors suggest that impairment of host defenses by B. microti is onemechanism by which the severity of disease is increased. Coexposure to babesiosis without acute concurrent infection may haveprognostic significance. A number of reports document long-term morbidity ina subset of patients following acute Lyme disease [2427]. The mechanismsunderlying these poor outcomes have yet to be elucidated, although severalrisk factors have been suggested, including delayed therapy and earlydissemination. Coexposure has never been studied as a possible risk factorfor long-term symptoms, although the implications for diagnosis andtreatment are considerable. In contrast to patients with symptomaticconcurrent infection, however, we found that individuals with Lyme diseaseand previous evidence of babesial infection did not have more severe acuteillnesses. A mean of 6 years after the index illness, they did not have moresymptoms or abnormal physical examination findings than individuals withLyme disease alone. We examined the hypothesis that "atypical" Lyme disease is due toundiagnosed babesial infection rather than Lyme disease. This is importantbecause such presentations are common in areas of endemicity and becausestandard therapy for early Lyme disease, with amoxicillin or doxycycline,does not provide adequate coverage of B. microti. We found no evidence thatpatients who had Lyme disease without erythema migrans or who hadself-reported Lyme disease not meeting the CDC case definition had a higherrate of reactivity to B. microti. Thus, "atypical" Lyme disease does notrepresent a significant number of undiagnosed cases of acute babesiosis. Our population-based study also provides information on the extent ofexposure to Babesia in a population where it is endemic. Previous studies onNantucket Island have been based on case series or laboratory surveys andmay have been subject to selection, ascertainment, or referral bias. Ruebushet al. [6] found that 11 (2%) of 577 serum samples collected at NantucketCottage Hospital reacted to B. microti. Our findings suggest that theseroprevalence is higher, at 6%. Only 4 (8%) of 48 seropositive subjects reported a history ofbabesiosis. An additional 13 subjects gave a history of babesiosis but didnot have sera reactive to B. microti. Six of these 13 subjects had priorpositive serology confirmed by medical record review, 3 as recently as 3years before our assay. The likely explanation is that titers of antibody toBabesia decay over time. Five of 11 patients whose medical records wereavailable for review had clinical illness within 2 years of this follow-upstudy. B. burgdorferi titers tend to remain positive for a longer period oftime (i.e., 510 years), particularly if treatment against the organism isadministered at a disseminated stage [9, 24]. This serological decay oftiters of antibody to Babesia would suggest that the seroprevalenceunderstates the true percentage of residents of Nantucket Island and BlockIsland ever exposed to B. microti. We conclude that exposure to B. microtiis common and usually asymptomatic. In the present study, 22% of individuals with prior Lyme disease inNantucket had evidence of prior babesial infection. This figure exceedsestimates from studies in other areas of endemicity, a finding suggestingthat B. microti may be more prevalent on Nantucket Island. Krause et al. [4]conducted a serological survey in Connecticut and found that 9.5% ofsubjects seropositive for B. burgdorferi also had antibodies to B. microti.In a study on Block Island, 11% of patients with acute Lyme disease hadserological or parasitological evidence of concomitant babesial infection[21]. On the basis of these studies, it appears that the prevalence of B.microti is higher on Nantucket Island than in other areas where it isendemic. This study had several potential limitations that require comment.Because it was retrospective, some misclassification of disease status mayhave occurred. This was mitigated by standardized serological testing,adherence to CDC case criteria for Lyme disease, and acceptance of onlyindirect immunofluorescence reactivity to B. microti as presumptive evidenceof babesial infection. This assay for Babesia is sensitive and highlyspecific [14]. Ideally, confirmation would require identification ofintraerythrocytic organisms on a peripheral blood smear during the acuteillness. However, none of our subjects was studied in the acute phase ofillness; thus, the likelihood of detecting parasitemia would be very low. For subjects with a history of acute babesiosis, confirmatorylaboratory data were used from ambulatory care and hospital records. Onlypatients with laboratory confirmation were accepted as having had babesialinfection. Because titers of antibody to B. microti can decay over time, somepatients classified as having Lyme disease alone may have had both diseases.This would likely raise the prevalence estimate for babesiosis and (iftiters of antibody to Babesia declined faster than those to B. burgdorferi)might lead to a conservative bias in a comparison of outcomes of cases ofLyme disease alone versus those in cases of coexposure. When we repeated theanalyses, restricting the sample to patients within 3 years of the initialillness in an attempt to minimize the amount of deterioration in titers ofantibody to Babesia, we found no differences in the conclusions. Finally, because most cases of babesiosis are subclinical, thepossibility that some individuals may have actually contracted the 2infections at different times cannot be excluded. Previous serologicalstudies, however, suggest that simultaneous transmission is common in areasof endemicity [28]. Though participation rates for the survey and for the clinicalevaluation were high, selection bias may have occurred. We studied a sampleof nonrespondents and found that they did not differ from respondents withregard to history of Lyme disease or Lyme-related symptoms [9], thusconfirming the validity of our sample. Exposure to B. microti is common in patients with Lyme disease, usuallywith no history of clinical babesiosis. Coexposure to Babesia probably doesnot predispose to more severe or persistent symptoms following acquisitionof Lyme disease. Though antibiotic therapy for Lyme disease does not coverB. microti, outcomes for patients with evidence of dual infection are notpoorer than for patients with Lyme disease alone. However, in practice,because some patients concurrently infected with Lyme disease and babesiosishave fulminant courses, patients who have Lyme disease with severe oratypical symptoms should be investigated for concurrent babesiosis andtreated accordingly. High fever, anemia, and thrombocytopenia are not commonin Lyme disease and should raise clinical suspicion of concurrent infectionwith B. microti [17, 20, 22, 23]. Studies are needed to determine if there are additional risk factorsfor poor long-term outcomes after concurrent infection and to develop a moreprecise clinical prediction rule for concurrent Lyme disease and babesiosis.Further studies are also needed to assess the frequency and significance ofconcurrent infection with other organisms, such as the agent causing humangranulocytic ehrlichiosis [29].Acknowledgments We thank the medical staff of the Nantucket Cottage Hospital for itssupport and cooperation; Karin Fossel, for her expert data management; andDr. Lionel Wells, for his helpful comments. In addition, this study wouldnot have been possible without the assistance of Joyce and CherylBartlett, R.N., and the primary care physicians of Nantucket Island: Dr.Christian Briggs, Dr. Butterworth, Dr. Lepore, Dr. DianePearl, and the late Dr. Voorhees.References1. Centers for Disease Control and Prevention. Lyme disease: UnitedStates. MMWR Morb Mortal Wkly Rep 1995; 44:45962. First citation in article| PubMed2. Ruebush TK, Juranek DD, Spielman A, Piesman J, Healy GR. Epidemiology ofhuman babesiosis on Nantucket Island. Am J Trop Med Hyg 1981; 30:93741.First citation in article | PubMed3. Pruthi RK, Marshall WF, Wiltsie JC, Persing DH. Human babesiosis. MayoClin Proc 1995; 70:85362. First citation in article | PubMed4. Krause PJ, Telford SR, R, et al. Geographical and temporaldistribution of babesial infection in Connecticut. J Clin Microbiol 1991;29:14. First citation in article | PubMed5. Filstein MR, Benach JL, White DJ, et al. Serosurvey for human babesiosisin New York. J Infect Dis 1980; 141:51821. First citation in article |PubMed6. Ruebush TK, Juranek DD, Chisholm ES, Snow PC, Healy GR, Sulzer AJ. Humanbabesiosis on Nantucket Island: evidence for self-limited and subclinicalinfections. N Engl J Med 1977; 297:8257. First citation in article | PubMed7. Ruebush TK, Cassaday PB, March HJ, et al. Human babesiosis on NantucketIsland: clinical features. Ann Intern Med 1977; 86:69. First citation inarticle | PubMed8. Krause PJ, Telford SR, Pollack RJ, et al. Babesiosis: an underdiagnoseddisease in children. Pediatrics 1992; 89:10458. First citation in article |PubMed9. Sangha O, CB, Fleischman KE, et al. Lack of cardiacmanifestations among patients with previously treated Lyme disease. AnnIntern Med 1998; 128:34653. First citation in article | PubMed10. Centers for Disease Control and Prevention. Lyme disease: UnitedStates, 1994. MMWR Morb Mortal Wkly Rep 1995; 44:45962. First citation inarticle | PubMed11. ACR Glossary Committee. Dictionary of the rheumatic diseases. New York:Contact Associates International, 1982:78. First citation in article12. Berardi VP, Weeks KE, Steere AC. Serodiagnosis of early Lyme disease:analysis of IgM and IgG antibody responses by antibody-capture enzymeimmunoassay. J Infect Dis 1988; 158:75460. First citation in article |PubMed13. Dressler F, Whalen JA, Reinhardt BN, Steere AC. Western blotting in theserodiagnosis of Lyme disease. J Infect Dis 1993; 167:392400. First citationin article | PubMed14. Krause PJ, Telford SR, R, et al. Diagnosis of babesiosis:evaluation of a serologic test for the detection of Babesia microtiantibody. J Infect Dis 1994; 169:9236. First citation in article | PubMed15. Centers for Disease Control and Prevention. Case definition for publichealth surveillance. MMWR Morb Mortal Wkly Rep 1990; 39:1921. First citationin article | PubMed16. SAS Institute. SAS/STAT user's guide. Cary, North Carolina: SASInstitute, 1990:175230. First citation in article17. Golightly L, Hirschhorn LR, Weller PF. Fever and headache in asplenectomized woman. Rev Infect Dis 1989; 11:62937. First citation inarticle | PubMed18. Falagas ME, Klempner MS. Babesiosis in patients with AIDS: a chronicinfection presenting as fever of unknown origin. Clin Infect Dis 1996;22:80912. First citation in article | PubMed19. Healy GR. Babesia infections in man. Hosp Practice 1979; 14:10711,1156. First citation in article | PubMed20. Steere AC, Bartenhagen NH, Craft JE, et al. The early clinicalmanifestations of Lyme disease. Ann Intern Med 1983; 99:7682. First citationin article | PubMed21. Krause PJ, Telford SR, Spielman A, et al. Concurrent Lyme disease andbabesiosis: evidence for increased severity and duration of illness. JAMA1996; 275:165760. First citation in article | PubMed22. Grunwaldt E, Barbour AG, Benach JL. Simultaneous occurrence ofbabesiosis and Lyme disease. N Engl J Med 1983; 308:1166. First citation inarticle | PubMed23. Marcus LC, Steere AC, Duray PH, AE, Mahoney EB. Fatalpancarditis in a patient with coexistent Lyme disease and babesiosis. AnnIntern Med 1985; 103:3746. First citation in article | PubMed24. Shadick NA, CB, Logigian EL, et al. The long-term clinicaloutcomes of Lyme disease: a population-based retrospective cohort study. AnnIntern Med 1994; 121:5607. First citation in article | PubMed25. Logigian EL, Kaplan RF, Steere AC. Chronic neurologic manifestations ofLyme disease. N Engl J Med 1990; 323:143844. First citation in article |PubMed26. Szer IS, E, Steere AC. The long-term course of Lyme arthritis inchildren. N Engl J Med 1991; 325:15963. First citation in article | PubMed27. Asch ES, Bujak DI, Weiss M, MG, Weinstein A. Lyme disease: aninfectious and postinfectious syndrome. J Rheumatol 1994; 21:45461. Firstcitation in article | PubMed28. Benach JL, JL, Habicht GS, Mac A, Grunwaldt E, Giron JA.Serological evidence for simultaneous occurrences of Lyme disease andbabesiosis. J Infect Dis 1985; 152:4737. First citation in article | PubMed29. Nadelman RB, Horowitz HW, Hsieh T, et al. Simultaneous humangranulocytic ehrlichiosis and Lyme borreliosis. N Engl J Med 1997; 337:279.First citation in article | PubMed