788
Persistence of Serum Antibodies to Borrelia burgdorferi in Patients Treated for Lyme Disease Henry M. Feder, Jr., Michael A. Gerber, Stephen W. Luger, and Raymond W. Ryan
From the Departments of Family Medicine, Pediatrics, and Laboratory Medicine, University ofConnecticut Health Center, Farmington; and the Old Lyme Family Practice, Old Lyme, Connecticut
Lyme disease is a multisystem disorder caused by the spirochete Borrelia burgdorferi. The clinical manifestations of this disease were originally divided into three stages but are now described in terms of early localized (erythema migrans), early disseminated (secondary erythema migrans, arthralgias, meningitis, neuritis, or carditis), and late persistent disease (arthritis or encephalomyelitis). At the onset of early localized Lyme disease, serum antibodies to B. burgdorferi are frequently not detectable, while during early disseminated and late persistent infections, antibodies to B. burgdorf eri are almost always identifiable [1-5]. Several methods have been used to measure serum antibodies to B. burgdorjeri, the first being indirect immunofluorescence assay (IFA) [6]. Later, ELISA was developed [2], and most recently an immunoblot assay has been employed [7]. When patients who have Lyme disease are treated with appropriate antibiotics, their signs and symptoms usually resolve [I, 8, 9]. Without treatment, and occasionally even with appropriate treatment, patients with Lyme disease may suffer from persistent or recurrent signs and symptoms [I, 10, II]. Little is known about the persistence of serum antibodies to B. burgdorferi in patients who are treated for Lyme disease or about the relationship between the persistence of these antibodies
Received 24 February 1992; revised 19 May 1992. This work was presented in part at the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy held on 21-24 October 1990 in Atlanta. Reprints or correspondence: Dr. Henry M. Feder, Jr., Department of Family Medicine, University of Connecticut Health Center, Farmington, Connecticut 06032. Clinical Infectious Diseases 1992;15:788-93 © 1992 by The University of Chicago. All rights reserved. 1058-4838/92/1505-0006$02.00
and the persistence or recurrence of signs and symptoms of Lyme disease. We measured serial titers ofserum antibody to B. burgdorferi in a group of patients with Lyme disease who had received appropriate antibiotic therapy to determine if (I) freezing and storage affect these antibody titers, (2) these antibodies persist after treatment, and (3) the persistence of these antibodies correlates with the persistence or recurrence of signs and symptoms of Lyme disease.
Materials and Methods Sixty serum samples collected between 1986 and 1989 from patients with Lyme disease who had been seen at the Old Lyme Family Practice (a four-physician private practice in Old Lyme, CT) had been stored at -20°C (in a non-frostfree freezer). All ofthese samples had been positive for either IgM or IgG antibodies to B. burgdorferi when examined by ELISA (positive IgM titer, ~ I: 160; positive IgG titer, ~ I:320) at the Clinical Microbiology Laboratory, University of Connecticut Health Center (Farmington), prior to freezing and storage. In addition, VORL tests of all these specimens were nonreactive. An attempt was then made to contact the patients from whom these 60 serum samples had been drawn. After informed consent was obtained, each patient's chart was reviewed. Patients who fulfilled the Lyme disease national surveillance case definition of the Centers for Disease Control (CDC; Atlanta) [12] and who were treated with appropriate antibiotics [I, II] were asked to return to the Old Lyme Family Practice. An interim history was then obtained that focused on the persistence or recurrence of the initial signs and symptoms of Lyme disease and on the appearance of any new signs or symptoms of the disease. Specifically, patients were asked about rash, flulike
Downloaded from http://cid.oxfordjournals.org/ at University of Arizona on July 10, 2015
To determine if antibodies to Borrelia burgdorferi persist after antibiotic treatment, werecalled 32 patients with Lyme diseasefrom a primary care practice a mean of 16 months after treatment and analyzed initial and follow-up serum samplesbyELISA and immunoblot assays.Of the eight patients whose initial serum specimens were positive for IgM antibody by ELISA, three had positive titers of IgM antibody at follow-up; of the 23 patients whose initial serum specimens were positive for IgG antibody by ELISA, 19 had positive titers of IgG at follow-up. Of the five patients whose initial serum specimenswere positive for IgM antibody by immunoblot, two had positive titers of IgM antibody at follow-up; of the 30 patients whose initial serum specimens were positive for IgG antibody by immunoblot, 29 had positive titers of IgG antibody at followup. The bands on the IgG immunoblot remained remarkably constant during the period from analysis of the initial specimen to that of the follow-up specimen. Nine of the 32 patients had persistent or recurrent symptoms, and ELISA and immunoblot were not helpful for identifying these nine patients.
CID 1992; 15 (November)
Persistence of Antibodies to B. burgdorferi
Results Of the 60 patients, 50 were successfully contacted and 41 agreed to participate in the study. Nine of the 41 patients were eliminated as they did not fulfill the CDC's Lyme disease national surveillance case definition [12]. These nine patients experienced only nonspecific symptoms without ever having had erythema migrans, arthritis, objective neurological signs, or heart block. The remaining 32 patients, who are the subjects of this investigation, all lived in Old Lyme, Connecticut, or in one of the adjacent towns. Their ages ranged from 10 to 84 years (mean, 54 years), and 18 were female. The interval from the initial specimen to the followup serum specimen ranged from 2 to 36 months (mean, 15.7 months) (table 1). The predominant clinical findings for these 32 patients are summarized in table 1. Patients were treated with appropriate antibiotic regimens, including ceftriaxone (1 g twice a day for 14-28 days), doxycycline (100 mg twice a day for
14-28 days), or amoxicillin (500 mg three times a day for 14-28 days). The results of the ELISA for IgM did not change with respect to positive vs. negative following freezing and storage for 2 to 36 months; however, specific IgM titers were not determined and, therefore, could not be compared. The results of the ELISA for IgG changed little with freezing and storage. The titers measured before and after freezing showed a significant (two dilution) decrease for only one patient (3%). Five patients had antibody titers that decreased from borderline positive (l: 160) to negative (~l :80) (table 1). The results of the ELISAs performed on the initial and follow-up serum specimens are presented in table 1. The antibody titers for the initial samples are those obtained after freezing and storage. Of the eight patients for whom results of IgM ELISA of initial specimens were positive, three had positive results at follow-up. For two of the three patients, results were still positive after short intervals of 2 and 4 months, respectively. For the third patient (who presented with Bell's palsy), the results of the IgM ELISA remained positive for a period of at least 26 months, despite the fact that the Bell's palsy had resolved within 2 weeks of treatment, and he had remained asymptomatic. Of the 23 patients who had positive or borderline positive results of IgG ELISA of their initial specimens, 19 had persistently positive or borderline positive results at follow-up. The results of the immunoblots performed on the initial and follow-up serum specimens are presented in table 1. The initial immunoblot results are those obtained after freezing and storage. Immunoblots had not been performed before freezing and storage. Of the five patients for whom results of IgM immunoblots oftheir initial specimen were positive, two had persistently positive results (one after an interval of 28 months and the other after an interval of 10 months). Of the 30 patients whose results ofIgG immunoblots of their initial specimen were positive, 29 had persistently positive results. The results of IgG immunoblots converted to negative (by losing one band) for only one patient. Of the four patients whose results ofIgG ELISA converted from positive (or borderline positive) to negative, all four had positive results of IgG immunoblots of their initial specimens, which remained positive at follow-up. For example, an IgG titer determined by ELISA decreased from 1:2,560 to 1:160 for one patient, while the results of IgG immunoblots remained unchanged. The bands on the immunoblots (particularly those identified with IgG antibodies) remained remarkably constant during the interval from the initial serum specimen to the follow-up serum specimen. There was a total of 48 bands on the IgM immunoblots of the initial samples and on the IgM immunoblots of the follow-up samples; six bands were lost and one band was gained. There was a total of 172 bands on the IgG immunoblots of the initial samples, and on the IgG im-
Downloaded from http://cid.oxfordjournals.org/ at University of Arizona on July 10, 2015
symptoms with fever, arthritis, arthralgias, myalgias, neck pain, and neuropathies. A follow-up serum specimen was also drawn and stored at -20°C. The initial and follow-up serum samples from each patient were then analyzed simultaneously in pairs for the presence of IgM and IgG antibodies to B. burgdorferi by ELISA and immunoblot assay. The ELISA for IgM and IgG antibodies was performed, and positive titers were defined as previously described [13, 14]. B. burgdorferi strain no. 2591 from the Connecticut Agricultural Station (West Haven) was used. The plates were read spectrophotometrically at an absorbance of 414 nm when the optical density of the positive control well (1: 160 dilution) minus the optical density of the nonspecific binding well equaled 1.0 for IgG or 0.5 for IgM. A serum dilution was considered positive if the net absorbance (antigen well minus nonspecific binding well) was ~3 SDs above the mean absorbance of the negative control wells in which pooled sera known to be negative for antibodies to B. burgdorferi were used. Appropriate positive and negative controls were included with each assay. The IgM titer was determined by ELISA and was reported as positive (~l: 160) or negative (< 1:160). Further titers of IgM antibodies were not determined. The IgG titer was reported as positive (~l :320), borderline positive (1: 160), or negative (~ 1:80). If the IgG titer was positive, then the serum specimen was serially diluted to determine end point titers, up to 1:5,120. The immunoblot assay for IgM and IgG antibodies was performed as previously described [14]. Using a modification ofpreviously established criteria [15], we defined seropositivity as the presence of a 4l-kD band and at least one low-molecular-weight band-including 18 kD, 25 kD, 31 kD (Osp A), and 34 kD (Osp B)-for either the IgM or the IgG immunoblot. The data were analyzed using the Student's r-test.
789
790
Feder et al.
CID 1992;15 (November)
Table 1. Clinical manifestations of Lyme disease and results of tests for antibody to Borrelia burgdorferi.
Case no. I 2 3 4
6 7
Serology interval (mo)
IgM ELISA initial*/ follow-up
IgO initial*/ follow-up
IgM immunobIot initial/ follow-up
EM/CNS disorder EM/arthritis EM/arthritis Arthritis
57/F 9/M 37/F 62/M
6 20 18 4
-/-/-/-/-
80 t/80 160/
Persistence of Serum Antibodies to Borrelia burgdorferi in Patients Treated for Lyme Disease Henry M. Feder, Jr., Michael A. Gerber, Stephen W. Luger, and Raymond W. Ryan
From the Departments of Family Medicine, Pediatrics, and Laboratory Medicine, University ofConnecticut Health Center, Farmington; and the Old Lyme Family Practice, Old Lyme, Connecticut
Lyme disease is a multisystem disorder caused by the spirochete Borrelia burgdorferi. The clinical manifestations of this disease were originally divided into three stages but are now described in terms of early localized (erythema migrans), early disseminated (secondary erythema migrans, arthralgias, meningitis, neuritis, or carditis), and late persistent disease (arthritis or encephalomyelitis). At the onset of early localized Lyme disease, serum antibodies to B. burgdorferi are frequently not detectable, while during early disseminated and late persistent infections, antibodies to B. burgdorf eri are almost always identifiable [1-5]. Several methods have been used to measure serum antibodies to B. burgdorjeri, the first being indirect immunofluorescence assay (IFA) [6]. Later, ELISA was developed [2], and most recently an immunoblot assay has been employed [7]. When patients who have Lyme disease are treated with appropriate antibiotics, their signs and symptoms usually resolve [I, 8, 9]. Without treatment, and occasionally even with appropriate treatment, patients with Lyme disease may suffer from persistent or recurrent signs and symptoms [I, 10, II]. Little is known about the persistence of serum antibodies to B. burgdorferi in patients who are treated for Lyme disease or about the relationship between the persistence of these antibodies
Received 24 February 1992; revised 19 May 1992. This work was presented in part at the 30th Interscience Conference on Antimicrobial Agents and Chemotherapy held on 21-24 October 1990 in Atlanta. Reprints or correspondence: Dr. Henry M. Feder, Jr., Department of Family Medicine, University of Connecticut Health Center, Farmington, Connecticut 06032. Clinical Infectious Diseases 1992;15:788-93 © 1992 by The University of Chicago. All rights reserved. 1058-4838/92/1505-0006$02.00
and the persistence or recurrence of signs and symptoms of Lyme disease. We measured serial titers ofserum antibody to B. burgdorferi in a group of patients with Lyme disease who had received appropriate antibiotic therapy to determine if (I) freezing and storage affect these antibody titers, (2) these antibodies persist after treatment, and (3) the persistence of these antibodies correlates with the persistence or recurrence of signs and symptoms of Lyme disease.
Materials and Methods Sixty serum samples collected between 1986 and 1989 from patients with Lyme disease who had been seen at the Old Lyme Family Practice (a four-physician private practice in Old Lyme, CT) had been stored at -20°C (in a non-frostfree freezer). All ofthese samples had been positive for either IgM or IgG antibodies to B. burgdorferi when examined by ELISA (positive IgM titer, ~ I: 160; positive IgG titer, ~ I:320) at the Clinical Microbiology Laboratory, University of Connecticut Health Center (Farmington), prior to freezing and storage. In addition, VORL tests of all these specimens were nonreactive. An attempt was then made to contact the patients from whom these 60 serum samples had been drawn. After informed consent was obtained, each patient's chart was reviewed. Patients who fulfilled the Lyme disease national surveillance case definition of the Centers for Disease Control (CDC; Atlanta) [12] and who were treated with appropriate antibiotics [I, II] were asked to return to the Old Lyme Family Practice. An interim history was then obtained that focused on the persistence or recurrence of the initial signs and symptoms of Lyme disease and on the appearance of any new signs or symptoms of the disease. Specifically, patients were asked about rash, flulike
Downloaded from http://cid.oxfordjournals.org/ at University of Arizona on July 10, 2015
To determine if antibodies to Borrelia burgdorferi persist after antibiotic treatment, werecalled 32 patients with Lyme diseasefrom a primary care practice a mean of 16 months after treatment and analyzed initial and follow-up serum samplesbyELISA and immunoblot assays.Of the eight patients whose initial serum specimens were positive for IgM antibody by ELISA, three had positive titers of IgM antibody at follow-up; of the 23 patients whose initial serum specimens were positive for IgG antibody by ELISA, 19 had positive titers of IgG at follow-up. Of the five patients whose initial serum specimenswere positive for IgM antibody by immunoblot, two had positive titers of IgM antibody at follow-up; of the 30 patients whose initial serum specimens were positive for IgG antibody by immunoblot, 29 had positive titers of IgG antibody at followup. The bands on the IgG immunoblot remained remarkably constant during the period from analysis of the initial specimen to that of the follow-up specimen. Nine of the 32 patients had persistent or recurrent symptoms, and ELISA and immunoblot were not helpful for identifying these nine patients.
CID 1992; 15 (November)
Persistence of Antibodies to B. burgdorferi
Results Of the 60 patients, 50 were successfully contacted and 41 agreed to participate in the study. Nine of the 41 patients were eliminated as they did not fulfill the CDC's Lyme disease national surveillance case definition [12]. These nine patients experienced only nonspecific symptoms without ever having had erythema migrans, arthritis, objective neurological signs, or heart block. The remaining 32 patients, who are the subjects of this investigation, all lived in Old Lyme, Connecticut, or in one of the adjacent towns. Their ages ranged from 10 to 84 years (mean, 54 years), and 18 were female. The interval from the initial specimen to the followup serum specimen ranged from 2 to 36 months (mean, 15.7 months) (table 1). The predominant clinical findings for these 32 patients are summarized in table 1. Patients were treated with appropriate antibiotic regimens, including ceftriaxone (1 g twice a day for 14-28 days), doxycycline (100 mg twice a day for
14-28 days), or amoxicillin (500 mg three times a day for 14-28 days). The results of the ELISA for IgM did not change with respect to positive vs. negative following freezing and storage for 2 to 36 months; however, specific IgM titers were not determined and, therefore, could not be compared. The results of the ELISA for IgG changed little with freezing and storage. The titers measured before and after freezing showed a significant (two dilution) decrease for only one patient (3%). Five patients had antibody titers that decreased from borderline positive (l: 160) to negative (~l :80) (table 1). The results of the ELISAs performed on the initial and follow-up serum specimens are presented in table 1. The antibody titers for the initial samples are those obtained after freezing and storage. Of the eight patients for whom results of IgM ELISA of initial specimens were positive, three had positive results at follow-up. For two of the three patients, results were still positive after short intervals of 2 and 4 months, respectively. For the third patient (who presented with Bell's palsy), the results of the IgM ELISA remained positive for a period of at least 26 months, despite the fact that the Bell's palsy had resolved within 2 weeks of treatment, and he had remained asymptomatic. Of the 23 patients who had positive or borderline positive results of IgG ELISA of their initial specimens, 19 had persistently positive or borderline positive results at follow-up. The results of the immunoblots performed on the initial and follow-up serum specimens are presented in table 1. The initial immunoblot results are those obtained after freezing and storage. Immunoblots had not been performed before freezing and storage. Of the five patients for whom results of IgM immunoblots oftheir initial specimen were positive, two had persistently positive results (one after an interval of 28 months and the other after an interval of 10 months). Of the 30 patients whose results ofIgG immunoblots of their initial specimen were positive, 29 had persistently positive results. The results of IgG immunoblots converted to negative (by losing one band) for only one patient. Of the four patients whose results ofIgG ELISA converted from positive (or borderline positive) to negative, all four had positive results of IgG immunoblots of their initial specimens, which remained positive at follow-up. For example, an IgG titer determined by ELISA decreased from 1:2,560 to 1:160 for one patient, while the results of IgG immunoblots remained unchanged. The bands on the immunoblots (particularly those identified with IgG antibodies) remained remarkably constant during the interval from the initial serum specimen to the follow-up serum specimen. There was a total of 48 bands on the IgM immunoblots of the initial samples and on the IgM immunoblots of the follow-up samples; six bands were lost and one band was gained. There was a total of 172 bands on the IgG immunoblots of the initial samples, and on the IgG im-
Downloaded from http://cid.oxfordjournals.org/ at University of Arizona on July 10, 2015
symptoms with fever, arthritis, arthralgias, myalgias, neck pain, and neuropathies. A follow-up serum specimen was also drawn and stored at -20°C. The initial and follow-up serum samples from each patient were then analyzed simultaneously in pairs for the presence of IgM and IgG antibodies to B. burgdorferi by ELISA and immunoblot assay. The ELISA for IgM and IgG antibodies was performed, and positive titers were defined as previously described [13, 14]. B. burgdorferi strain no. 2591 from the Connecticut Agricultural Station (West Haven) was used. The plates were read spectrophotometrically at an absorbance of 414 nm when the optical density of the positive control well (1: 160 dilution) minus the optical density of the nonspecific binding well equaled 1.0 for IgG or 0.5 for IgM. A serum dilution was considered positive if the net absorbance (antigen well minus nonspecific binding well) was ~3 SDs above the mean absorbance of the negative control wells in which pooled sera known to be negative for antibodies to B. burgdorferi were used. Appropriate positive and negative controls were included with each assay. The IgM titer was determined by ELISA and was reported as positive (~l: 160) or negative (< 1:160). Further titers of IgM antibodies were not determined. The IgG titer was reported as positive (~l :320), borderline positive (1: 160), or negative (~ 1:80). If the IgG titer was positive, then the serum specimen was serially diluted to determine end point titers, up to 1:5,120. The immunoblot assay for IgM and IgG antibodies was performed as previously described [14]. Using a modification ofpreviously established criteria [15], we defined seropositivity as the presence of a 4l-kD band and at least one low-molecular-weight band-including 18 kD, 25 kD, 31 kD (Osp A), and 34 kD (Osp B)-for either the IgM or the IgG immunoblot. The data were analyzed using the Student's r-test.
789
790
Feder et al.
CID 1992;15 (November)
Table 1. Clinical manifestations of Lyme disease and results of tests for antibody to Borrelia burgdorferi.
Case no. I 2 3 4
6 7
Serology interval (mo)
IgM ELISA initial*/ follow-up
IgO initial*/ follow-up
IgM immunobIot initial/ follow-up
EM/CNS disorder EM/arthritis EM/arthritis Arthritis
57/F 9/M 37/F 62/M
6 20 18 4
-/-/-/-/-
80 t/80 160/
