Serological Follow-Up after Treatment of Patients with

Transcription

1 JOURNL OF CLINICL MICROIOLOGY, June 1994, p /94/$4.+ Copyright (C 1994, merican Society for Microbiology Vol. 32, No. 6 Serological Follow-Up after Treatment of Patients with rythema Migrans and Neuroborreliosis SUSNN HMMRS-RGGRN,I* NN-MTF LCH,2 MTS KRLSSON,' O SVNUNGSSON,3 KLUS HNSN,2 ND GORN STIRNSTDT3 Departments of Infectious Diseases, Danderyd Hospital' and Huddinge University Hospital,3 Karolinska Institut, Stockholm, Sweden, and orrelia Laboratory, Department of Infection Immunology, Division of iotechnology, Statens Seruminstitut, Copenhagen, Denmark2 Received 27 December 1993/Returned for modification 8 February 1994/ccepted 18 March 1994 To investigate the duration and kinetics of immunoglobulin M (IgM) and IgG antibodies against orrelia burgdorferi in serum after treatment of Lyme borreliosis, consecutive serum samples from 3 seropositive patients with erythema migrans and 91 seropositive patients with neuroborreliosis were analyzed with a capture IgM enzyme-linked immunosorbent assay (LIS) and an indirect IgG LIS, both using. burgdorferi flagella as the antigen. ll the patients improved after treatment: 97 patients had a complete clinical recovery, while 24 patients had sequelae. The results showed that patients with erythema migrans and early neuroborreliosis more often initially had highly elevated IgM optical density () values and low IgG values against. burgdorferi, while the opposite was found in patients with late neuroborreliosis. During follow-up, the majority of patients had developed negative or significantly declining IgM s after 1 to 1.5 years but persistently positive IgM s were found up to 17 months after treatment of erythema migrans and 3 years after treatment of neuroborreliosis. IgG antibody levels declined more slowly and remained elevated to a larger extent, but more than half of the patients had developed negative IgG s within 5 years after therapy. However, positive IgG values were found after 9 to 1 years for patients treated for neuroborreliosis as well as erythema migrans. oth IgM and IgG antibodies against. burgdorferi may persist for months to years after successful treatment of Lyme borreliosis. Consequently, a single serum sample with antibodies against. burgdorferi must always be carefully evaluated and correlated to clinical symptoms. rythema migrans (M), neuroborreliosis, arthritis, and acrodermatitis chronica atrophicans are different clinical manifestations of Lyme borreliosis, which may be diagnosed by the medical history and typical clinical findings (2, 3, 16, 18, 23, 24). However, to confirm the diagnosis, serological methods, mainly the enzyme-linked immunosorbent assay (LIS), are available and often necessary (1, 12, 22). Serological studies have shown seropositivity rates of 25 to 5% in patients with M and 5 to 8% in patients with neuroborreliosis (9, 1, 14, 21, 22). The diagnostic sensitivity of LIS has been increased by the introduction of new antigens, e.g., the flagellum antigen (1, 14), and capture LIS instead of indirect LIS for the immunoglobulin M (IgM) analysis (12). n increase in test sensitivity is of importance for early diagnosis. lthough Lyme borreliosis in all stages is a treatable disease, doubts might be raised in clinical practice whether treatment is successful. One reason for persisting symptoms after antibiotic treatment is true treatment failure, but most probably incorrect diagnosis or symptoms due to irreversible tissue damage are more common causes (1, 19). However, this points out the need for a laboratory parameter to help the clinician follow treatment efficacy. One aim of this study was to investigate whether the kinetics of the antibody response against orrelia burgdorferi might be such a parameter for patients with Lyme borreliosis.. burgdorferi may cause subclinical, mild clinical, or incorrectly diagnosed clinical infection, followed by a spontaneous * Corresponding author. Mailing address: Department of Infectious Diseases, Danderyd Hospital, S Danderyd, Sweden. Phone: 46 /8/ Fax: 46 /8/ recovery (7, 15, 17). This may account for the high seropositivity rate among healthy people living in areas where the disease is endemic without a previous known history of Lyme borreliosis (8). If symptoms of the skin, the nervous system or joints (due to an etiological agent other than. burgdorferi) develop in patients that are seropositive against. burgdorferi because of previous exposure, diagnostic difficulties may appear. arlier studies have shown the persistence of IgG against. burgdorferi in serum for years after untreated Lyme borreliosis infection (4, 15) or the persistence of specific IgG for years in asymptomatic untreated persons in areas of endemicity (2). It has been suggested that such findings predict subsequent disease activity (4). It is therefore of interest to investigate whether the antibody response against. burgdorferi also persists after antibiotic treatment in well-defined cases of Lyme borreliosis. To answer these questions, we analyzed consecutive serum samples from patients treated for M or Lyme neuroborreliosis with a capture IgM LIS and an indirect IgG LIS, both using. burgdorferi flagella as the antigen. MTRILS ND MTHS Patients. Patients with M and patients with Lyme neuroborreliosis were included in this retrospective study. The patients were admitted to the Department of Infectious Diseases or Department of Dermatology, Danderyd Hospital, or the Department of Infectious Diseases, Roslagstull Hospital (present Huddinge Hospital), Stockholm, Sweden, between 1973 and Only patients with positive IgM and/or IgG antibodies against. burgdorferi in serum before or within a month after the start of antibiotic therapy were included. Serological testing of individual patients was performed until Downloaded from on November 22, 218 by guest

2 152 HMMRS-RGGRN T L. TL 1. Clinical data for patients with M Characteristic Value for group No. of patients... 3) Male/female (no. of patients)... 14/16 ge (median) in yrs (54) Preceding tick bite (no. of cases)...2) (67%) Clinical manifcstations (no. of cases) Localized M Singlc M + gencral symptoms' Multiple M... 2 Multiple M + general symptoms...3 Duration of symptoms pretrcatmcnt...2 days-7 mo (median 3.5 wk) Treatment 1t)-14 days (no. of patients) Oral pcv, I g q8-12h"...2) Oral doxycycline, 2t)t) mg q24h...7 Oral erythromycin, St()(mg ql2h...3 Retreatment for 14 days within 3 mo with oral pcv-*oral doxycycline (no. of patients)... 2 Follow-up time... 3 mo-9 yr (mcdian 3 yr) Improvement (no. of patients)...3 Complete recovery Sequclac I (arthralgia aftcr 1 yr) General symptoms: fever, myalgia, and headache. " q8-12h, every 8 to 12 h. seronegativity occurred or until the end of this study. dditional serological testing of patients who had become seronegative was also performed if new symptoms compatible with Lyme borreliosis appeared. M. total of 3 patients with M were included in the study. The diagnosis of M was made on the basis of clinical examination. Clinical data from the 3 patients are presented in Table 1. Neuroborreliosis. total of 91 patients with neuroborreliosis were included in the study. The diagnosis of neuroborreliosis was made on the basis of a typical clinical picture (n = 91), the presence of pleocytosis in cerebrospinal fluid (n = 91), and intrathecal antibody production against. burgdorferi (n = 62). In the 29 patients without specific intrathecal antibody production at diagnosis, a majority of whom had a disease duration of less than 1 month, preceding or present M (n = 8), positive IgM antibodies against. burgdorferi in serum (n = 7), or both (n = 14) were required for inclusion. Clinical data from the 91 patients are presented in Table 2. LIS. ll serum samples, stored at -7 C, were analyzed with a capture IgM LIS and an indirect IgG LIS, both using. burgdoiferi flagella as the antigen. The methods have earlier been described in detail (1, 12). In brief, in the capture IgM LIS the solid phase was coated with,u-chain-specific rabbit anti-human IgM, binding to human IgM antibodies in sera diluted 1/2.. burgdorferi-specific IgM antibodies reacted with the test antigen consisting of a. burgdorferi flagellum-biotin-avidin-peroxidase complex. ound flagella were visualized by the addition of O-phenylenediamine substrate. In the indirect IgG LIS, microdilution plates were coated with purified. burgdorferi flagella. Serum samples diluted 1/2 were added, and specific antibodies were bound. The binding was detected by the addition of peroxidaseconjugated rabbit anti-human IgG antibodies and O-phenylenediamine substrate. In both assays the optical density () was read spectrophotometrically and cutoff was defined as the 98th percentile of values obtained among sera from 2 healthy controls and was.5 in the IgM LIS and.18 in the IgG LIS (1, 12). Consecutive serum samples TL 2. Clinical data for patients with neuroborreliosis Characteristic Value for group No. of patients Male/female (no. of patients)...34/57 ge (median) in yrs (49) Preceding tick bite (no. of cases)...17 (19%) Preceding M (no. of cases)...21 (23%) Preceding tick bite and M (no. of cases)...17 (19%) Clinical manifestations (no. of cases) Meningitis Meningoradiculitis, cranial neuritis...66 ncephalomyelitis Pleocytosis (t)t)%) Leukocyte count (1"/liter)...6-1,4) (median 128) Intrathecal antibody production (no. of cases) (68%) Duration of neurological symptoms pretreatment... 3 days-2 yr (median I mo) Treatment 1-14 days (no. of patients) i.v. pcg, 3 g q6-8h" (15t) mg/kg/day for children) i.v. cefuroxime, 3 g q8h...1 i.v. doxycycline, 2 mg q24h...4 Oral doxycycline, 2 mg q24h...2) i.v. penicillin + oral doxycycline (5 days) and oral doxycycline (9 days)...3 Retreatment for 14 days within 8 mo (no. of patients) with: Oral doxycycline--i.v. cefuroxime...1 i.v. pcg--i.v. cefotaxime, 3 g ql2h...2 Follow-up time... 3 mo-1 yr (median 2.5 yr) Improvement (no. of patients)...91 Complete recovery (75%) Sequelac (25%) Impaired sensitivity, pain...6 Impaired hearing...1 Unilateral partial facial palsy...6 ilateral partial facial palsy...i Monoparesis of extremity...2 Paraparesis...3 Reduced intellectual capacity...3 sthenia...1 " q6-8h, every 6 to 8 h. from each patient were analyzed simultaneously on the same microdilution plate. twofold decrease or increase of the in consecutive serum samples was considered significant for both the IgM capture LIS and the indirect IgG LIS. For patients with neuroborreliosis, intrathecal antibody production against. burgdorferi was analyzed in paired serumcerebrospinal fluid samples drawn before or within a month of start of antibiotic treatment. The LIS methods used were either a capture IgM and IgG LIS, using. burgdorferi flagella as the antigen (11), or an indirect IgM and IgG LIS, using whole-cell sonicated. burgdorferi spirochetes as the antigen (22). RSULTS J. CLIN. MICRC)IC)L. M pretreatment to 1 month after start of antibiotic treatment. Nine of the thirty patients had only positive IgM s, 12 patients had both positive IgM and IgG s, and 9 patients were only positive by IgG LIS. For 21 of 3 patients with positive IgM s, the range was.54 to >2.5 (median, 1.16) and for 21 of 3 patients with positive IgG s the range varied from.18 to 1.9 (median,.34) in Downloaded from on November 22, 218 by guest

3 VOL. 32, 1994 SROLOGICL FOLLOW-UP OF LYM ORRLIOSIS 1521 the first serum sample drawn from each patient during the period. M follow-up. (i) IgM values. The kinetics of IgM in serum are demonstrated in Fig. I. The range and median of the values at different time intervals of follow-up are shown in Table 3. In summary, 17 of 21 initially IgM-positive patients (81%) became negative by the IgM LIS during follow-up from 3 weeks to 9 years (median, 3 years), while 4 patients (19%) still had positive s of IgM against. burgdorferi 7, 12, 15, and 17 months after antibiotic therapy. (ii) IgG values. The kinetics of IgG in serum are demonstrated in Fig. l. The range and median of the values are shown in Table 3. In summary, 9 of 21 initially IgG-positive patients (43%) became IgG negative during follow-up from 4 months to 9 years (median, 3 years). The remaining 12 patients (57%) still had positive IgG s after 3, 6, 12, 15, and 17 (n = 2) months and 2.5 (n = 2), 3 (n = 2), and 9 (n = 2) years of follow-up. Retreated M patients. Serum samples from two patients treated by a second antibiotic regimen showed declining IgM levels before retreatment. oth patients had negative IgG s. Comparison of M patients who recovered and those with sequelae. The only patient with persistent symptoms at follow-up (arthralgia 1 year posttreatment) became seronegative. mong the 29 patients with a complete clinical recovery, 1 patient (3%) still had a positive IgM after 7 months, 3 patients (1%) were still positive by both IgM and IgG LIS after 12 to 17 months (median, 15 months), and 9 patients (31%) had positive IgG s at follow-up after 3 months to 9 years (median, 2.5 years). The remaining 16 patients (55%) without residual symptoms became seronegative during a follow-up time which varied from 4 months to 9 years (median, 3 years). Neuroborreliosis. To illustrate the kinetics of the IgM and IgG s for patients with neuroborreliosis graphically (Fig. 2, 3, and 4), the 91 patients were divided into three different groups according to disease duration (<1 month [n = 32]), 1 to 2 months [n = 33], and >2 months [n = 26]). Neuroborreliosis pretreatment to 1 month after start of treatment. Ten patients had only s positive for IgM, 53 patients were positive by both IgM and IgG LIS, and 28 patients had only s positive for IgG. In 63 of 91 patients with s positive for IgM against. burgdorferi, the range varied from.5 to >2.5 (median, 1.14) and in 81 of 91 patients positive for IgG in serum, the range was.18 to >2.5 (median,.49) in the first serum sample drawn from each patient during the period. One of the 1 patients with only positive IgM was initially IgG negative but had seroconverted at 1.5 months of follow-up (IgG,.33). Neuroborreliosis follow-up. (i) IgM values. The kinetics of IgM in serum are demonstrated in Fig. 2, 3, and 4. The range and median of the values at different time intervals of follow-up are shown in Table 3. In summary, 47 of 63 initially IgM-positive patients (75%) became negative in the IgM LIS during follow-up from I month to 1 years (median, 2 years), while 16 patients (25%) still had positive IgM s when followed for 6 months to 5.5 years (median, 15 months). (ii) IgG values. The kinetics of IgG in serum for the 81 patients initially IgG positive and for one patient IgG positive 1.5 months after therapy are demonstrated in Fig. 2, 3, and 4. The range and median of the values are shown in Table 3. In summary, during follow-up, 55 of 82 IgG-positive patients (67%) became negative by IgG LIS when followed CD _ i >3 3 Co-. CDO 1 > - S - - z O<, I, \ L NN< >~< 222, 2 ONl Vt. on ON ON,,Ji111 S; ON O_N ON. -J O - O. ON _ ONON _ O cno - --N N V X t.#j - ON ON ON'o " " t-j" "C un 4S C > - j " 5 C ) - ON- - _Jl c ONIONONON 5J N tononon Uo ON Uo - L 'IC 'c s fo Cr wo ooon C) = - vl xc tltq t'j- c-con IO,:~ ON 4co w f.)# ON4 1 w ON ON tj en ONo en = L o O c t-q NO - -k " Q(w l.:~ U.) 'IC= -_ -_ oc 4p c U.) oc x x ON ON xn 4NO, ON 4 C~ l:)-j o ON.N.N.N ONONON = = - K. i 4NO NON -_ : o oc c - v,c -J " IC " - uo C~ c. &. _:.: HZ C~ &. 3 : -4 C C. _: z x -V ) : _. rni z C 5- _.I _r. L 1- c: ) D r zz 2 - _ Ct _ O ~ - -t CD CD. C p,. 3: C 3: CD C _ : f _~.. C Ct Downloaded from on November 22, 218 by guest

4 1522 HMMRS-RGGRN T L. > 2,5 2,- 2 1,5-,, 1,,5, 2,5-2 2,- 1,5-1, -,5, r- 1 mo 6mo 1 yr 5 yrs 1yrs 1 mo 6mo 1 yr 5 yrs 1yrs FIG. 1. s of IgM () and IgG () against. burgdorferi in serum before treatment and during follow-up for 3 patients with M. The horizontal dashed lines mark the cutoff. Mo, months; yrs, years. for 1 month to 1 years (median, 2 years). Twenty-seven patients (33%) still had positive IgG s when followed for 6 months to 1 years (median, 3 years) posttreatment, including one patient (III [see additional results below]), who was initially IgG positive and turned negative after 9 months but became IgG positive again after 1.5 years of follow-up and was still positive for IgG after 6 years of follow-up. Retreated patients. Serum samples from all three patients treated by a second antibiotic regimen for neuroborreliosis showed significantly declining IgM and/or IgG levels before retreatment. Patients with rising antibody levels and/or new symptoms of Lyme borreliosis during follow-up. One patient (I [Fig. 3]) with meningoradiculitis and facial palsy treated with intravenous (i.v.) penicillin G (pcg), improved completely within 2 weeks after therapy. He was continuously followed because of his seropositivity and after an initial decline showed an increasing IgM from.82 to 1.27 between 6 months and 1.5 years after treatment but had not noticed any new symptoms. The IgG decreased during the same period. second decline (but persistently positive IgM ) was measured after 5.5 years (,.62). IgG s for two patients (II and III, [Fig. 3]) showed a significant rise at 1 and 1.5 years of follow-up, respectively, before a second decline. One of the patients (II), who at --w J. CLIN. MICROIOL. diagnosis had meningoradiculitis and facial palsy treated with oral doxycycline, had a negative IgM LIS. He showed a gradually decreasing IgG level at 1 months of follow-up. t that time, the patient had recovered completely. significant increase of the IgG from.2 to.73 was seen between 1 months and 1 year of follow-up without clinical symptoms. The IgM LIS remained negative. new serum sample was drawn 2.5 years after treatment, and a second decline of the IgG level was found (,.26 [Fig. 3]). The other patient (III) with meningoradiculitis showed a gradually decreasing IgG level, which became negative 9 months after treatment with i.v. pcg. One year posttreatment, she developed M, which was treated with oral penicillin V (pcv). serum sample drawn 6 months later showed a significant rise of the IgG level to 1.2. She initially tested negative by the IgM LIS, and serum at 18 months was still negative. serum sample drawn 6 years after the first treatment showed a significant decline, but was still positive for IgG (,.34) (Fig. 3). One patient (IV [Fig. 4]) with meningoradiculitis and facial palsy treated with i.v. pcg initially had an IgG of 1.14 and a negative IgM. t 3 years of follow-up, the IgG had declined to.44. t that time, she had persistent partial facial palsy but showed no signs of active disease. n insignificant rise to.64 by the IgG LIS was seen after 4 years, when she was treated with oral pcv for M. serum sample drawn 2 years later showed a slight decline to.52. The IgM value remained negative (Fig. 4). One other patient with meningoradiculitis and treated with i.v. pcg recovered completely. In the following summer, 9 months posttreatment, he developed M and was treated with oral pcv. No simultaneous rise of the IgM or IgG levels was seen for this patient. Comparison of neuroborreliosis patients who recovered and those with sequelae. Of the 23 patients with sequelae, one patient (4%) with extremity monoparesis still had an level positive for IgM against. burgdorferi after 15 months. ight patients (35%) with residual symptoms consisting of neuralgic pain and/or impaired sensitivity (n = 2), partial facial palsy (n = 3), reduced intellectual capacity (n = 2), and paraparesis (n = 1) were still positive by IgG LIS when followed for 9 months to 1 years (median, 3 years, 9 months). The remaining 14 patients (61%) with sequelae became seronegative during follow-up of 3 months to 8 years (median, 2.5 years). Of the 68 patients with a complete clinical recovery, 9 patients (13%) had positive IgM s when followed for 8 months to 3 years (median, 17 months). Six patients (9%) had s positive for both IgM and IgG against. burgdorferi at 6 months to 5.5 years of follow-up (median, 15 months). The patient (I [Fig. 3]) with a positive IgM level 5.5 years after treatment had an increasing IgM level at 1.5 years of follow-up without clinical symptoms. Thirteen patients (19%) were still positive by IgG LIS when followed for 9 months to 1 years (median, 4 years). This included the patient (III [Fig. 3]) who was seronegative at 9 months of follow-up but showed a significant rise in IgG against. burgdorferi after 18 months due to M. The remaining 4 of 68 (59%) patients with a complete clinical recovery became seronegative during follow-up between 4 months and 1 years (median, 2 years). DISCUSSION few previous studies have investigated the duration of seropositivity after orrelia infection, and the results indicate that the IgG antibody response against. burgdorferi may remain positive for years (4, 6, 8, 15, 2, 22). Most of the earlier Downloaded from on November 22, 218 by guest

5 VOL. 32, 1994 SROLOGICL FOLLOW-UP OF LYM ORRLIOSIS 1523 > 2,5k > 2, z 2, a C1,5 2 2,5- m 1,5 aimo 6mo 1 yr 5 yrs 1yrs qo 1;, Imo i6mo 1 yr 5 yrs Ioyrs FIG. 2. s of IgM () and IgG () against. burgdo,feri in serum before treatment and during follow-up for 32 patients with neuroborreliosis with a symptom duration of less than 1 month before therapy. The horizontal dashed lines mark the cutoff. Mo, months; yrs, years. publications on this subject have concerned untreated patients with a spontaneous recovery (4, 15) or asymptomatic persons in epidemiological screening surveys (8, 2). One publication has demonstrated specific IgM and IgG against. burgdorferi for 2 to 3 years after antibiotic treatment of different Lyme disease manifestations (6), and another study has shown persistent IgG against. burgdorferi in serum for several years after therapy of neuroborreliosis (22). The present study also demonstrates that the serological response against. burgdorferi may persist for years even after antibiotic treatment of M and neuroborreliosis, although the specific antibody response decreases or vanishes over time. No earlier publication has demonstrated the kinetics of the antibody response after treatment of Lyme borreliosis for as long as 1 years or for such a large selection of patients. s has been shown earlier, IgM and IgG antibodies against. burgdorferi develop gradually and rather slowly during the disease before treatment (1, 12, 14, 22). This is also demonstrated in the present study by Fig. 1 to 4 showing that patients with M and early neuroborreliosis more often have initially high IgM and low IgG s, while the opposite results are found for patients with a longer disease duration. In only 3 of 65 patients with M or neuroborreliosis with IgM and IgG co 2 2,5-2, - 1,5-1,,5, 1 mo 6io 1 yr 5 yrs 1Oyrs FIG. 3. s of IgM () and IgG () against. burgdorferi in serum before treatment and during follow-up for 33 patients with neuroborreliosis with a symptom duration of I to 2 months before therapy. The horizontal dashed lines mark the cutoff. Mo, months; yrs, years., patient I;, patient II; *, patient III. s both positive, the IgG antibodies developed after the IgM response and after the start of antibiotic therapy but within the first month, while the remaining 62 patients had both IgM and IgG antibodies against. burgdorferi in their first serum analysis (data not shown). On the other hand, as only 1 of 19 patients with M or neuroborreliosis with only positive IgM s before therapy became IgG positive 1.5 months posttreatment, it seems that antibiotic treatment usually prevents the development of further antibody response. Our results also show that the percentage of patients that become positive by the IgG LIS increases with the duration of disease before therapy to 1% in patients with a disease duration of neuroborreliosis for more than 2 months. Fewer than half of the patients in the same group tested positive for IgM against. burgdorferi, compared with 8% of the patients with neurological symptoms for 2 months or less. This indicates that IgM positivity against. burgdorferi may disappear spontaneously although disease activity persists. In this study, the intervals between consecutive serum analyses differed among the 121 patients. In some patients who were found to be negative several years after the last positive IgM-IgG serum sample, the time to seroconversion is not known. However, surprisingly, we found significantly declining Downloaded from on November 22, 218 by guest

6 1524 HMMRS-RGGRN T L. 2 2,5-2, - 1,5 1,,5, 1,,,5 1 mo 6 mo 1 yr 5 yrs 1 yrs 2 o 6o1y r r FIG. 4. s of IgM () and IgG () against. burgdorferi in serum before treatment and during follow-up for 26 patients with neuroborreliosis with a symptom duration of more than 2 months before therapy. The horizontal dashed lines mark the cutoff. Mo, months; yrs, years., patient IV. but still positive s of IgM against. burgdorferi 17 months after treatment of M with general symptoms and 3 years (5.5 years in one patient with an unexplained IgM level rise after 1.5 years) after therapy of neuroborreliosis without residual symptoms. These results show that not only specific IgG but also IgM positivity against. burgdorferi may persist for years without disease activity. The higher sensitivity of the IgM capture LIS than that of the indirect IgM LIS used in previous studies (12, 22) is probably one important factor contributing to this finding. When, as in this study, analyzing the IgG response against. burgdorferi in consecutive serum samples after therapy of M or neuroborreliosis, a gradual decrease of the IgG for each patient was the most common finding. However, IgG positivity was demonstrated several years after successful antibiotic treatment of both M and neuroborreliosis. s the duration of seropositivity after treatment of clinical infections may persist for years, it seems likely that the same results might also be found after subclinical infection with. burgdorferi. This is probably the explanation for the high percentage of seropositive persons without a history of present or previous Lyme borreliosis living in areas where. burgdorferi is endemic (8). s seropositivity may persist for a long time, even after J. CLIN. MICROIOL. adequate antibiotic treatment and clinical recovery, it seems unlikely that in healthy persons a single positive IgG value indicates live spirochetes or latent infection. s has been pointed out recently, overdiagnosis of Lyme borreliosis occurs for patients with symptoms affecting the skin, the nervous system, or the joints, together with a positive antibody reaction against. burgdorfeni (19). Since this study shows persistent seropositivity for years after therapy of M and neuroborreliosis, it is of great importance to emphasize that a single serum sample with a positive antibody response to. burgdorferi must always be carefully evaluated and correlated to the clinical picture of the patient. The precise time to clinical recovery is not possible to determine in this retrospective study, although the clinical outcome at the last follow-up serum analysis is known for each patient. However, an earlier prospective study concerning treatment of neuroborreliosis demonstrated that the majority of 54 patients treated for neuroborreliosis had a complete recovery within 3 to 6 months (13). few patients had persistent sequelae after 12 months of follow-up but no signs of disease activity (13). Concerning M, previous studies have shown recovery during therapy or within weeks to a few months posttreatment for the majority of patients (3). s this study has demonstrated that positivity for IgM against. burgdorferi persists for several months and even years after therapy of M or neuroborreliosis in patients with a complete clinical recovery, it is doubtful that analysis of a single posttreatment serum sample for IgM can be a laboratory parameter of treatment efficacy. However, as shown by the IgM graphs (Fig. 1, 2, 3, and 4), decreasing IgM levels were demonstrated for almost all patients after 1 month posttreatment. Consequently, since all the patients in the present study improved, it is possible that the decline in the IgM response level in consecutive serum samples after therapy might be an effect of antibiotic treatment, although the same development could have been a spontaneous event. comparison of the serological responses after antibiotic therapy of patients whose treatment failed as a result of persistent infection or false diagnosis would have been valuable but was not possible in this study. lso, the IgG response against. burgdorferi in consecutive serum samples showed a gradual decrease of IgG, starting within the first months after the start of treatment for most of the patients. n earlier study has demonstrated increasing IgG titers with longer disease durations (22), while the decline of the IgG s after the start of therapy in the majority of patients was probably influenced by antibiotic treatment. Since the duration of both IgM and IgG positivity against. burgdorferi after therapy may persist for a long time, other authors have suggested the measurement of cytokines to determine the effect of treatment (5). This suggestion is interesting and should be investigated further. s demonstrated by the graphs, for a few patients, the IgM or IgG s seemed to rise slowly or remain stationary for months or even years after therapy, despite clinical recovery. Since the intervals between serum analyses of the patients differed in the present study, a long interval between pretreatment and posttreatment samples may give the false impression of a stable value even though there might have been an increase of the in between. nother explanation is of course a subclinical or an undiagnosed clinical reinfection causing a rise in the IgG. s seen with a few patients in the present study (Fig. 3 and 4), clinical reinfections of patients living in areas endemic to. burgdorferi seem to be a reality. Three patients in the present study developed M months to years after treatment of neuroborreliosis. For two of these Downloaded from on November 22, 218 by guest

7 VOL. 32, 1994 three patients, an increase of the IgG was measured. Since one patient developed a rise in IgG without symptoms, subclinical reinfections also seem to occur. One patient (I [Fig. 3]) showed an increasing IgM 1.5 years after treatment. He initially had a complete clinical recovery and had developed neither any new symptoms of infection with. burgdorferi nor symptoms of any other infection when the IgM rise was measured. No retreatment was given. The IgM kinetics for this patient are difficult to explain, but in the absence of additional symptoms treatment failure seems unlikely. On the other hand, as declining levels of IgM and/or IgG antibodies against. burgdorferi were seen before retreatment of the five patients who received a second antibiotic course, it is doubtful that any of these cases should be judged as true treatment failures. In conclusion, both IgM and IgG antibodies against. burgdorferi may persist for months to years without signs of disease activity after treatment of Lyme borreliosis. The decrease of IgM and IgG s often seems to be slower than the clinical recovery, but it is likely that the declining tendency is influenced by antibiotic therapy. positive measurement of antibody to. burgdorferi in a single serum analysis for healthy persons or patients with dermatologic, neurologic, or joint diseases must always be carefully evaluated. lso, reinfection may maintain seropositivity. CKNOWLDGMNT We thank Dorthe Seborg Pedersen for excellent technical assistance. RFRNCS 1. ckermann, R.,. Rehse-Kupper,. Gollmer, and R. Schmidt Chronic neurologic manifestations of erythema migrans borreliosis. nn. N. Y. cad. Sci. 539: sbrink,.,. Hovmark, and I. Olsson Clinical manifestations of acrodermatitis chronica atrophicans in 5 Swedish patients. Zentralbl. akteriol. Mikrobiol. Hyg. 263: sbrink,., and I. Olsson Clinical manifestations of erythema chronicum migrans fzelius in 161 patients. comparison with Lyme disease. cta Dermato-Venereol. 65: Craft, J.., R. L. Grodzicki, and. C. Steere ntibody response in Lyme disease: evaluation of diagnostic tests. J. Infect. Dis. 149: Fawcett, P. T., C. D. Rose, R. Proujansky, K. M. Gibney, D. M. Molloy, and R.. Doughty Serial measurement of soluble interleukin 2 receptor levels: an early indicator of treatment response for Lyme disease. J. Rheumatol. 2: Feder, H. M., M.. Gerber, S. W. Luger, and R. W. Ryan Persistence of serum antibodies to orrelia burgdorferi in patients treated for Lyme disease. Clin. Infect. Dis. 15: Gustafson, R.,. Svenungsson, M. Forsgren,. Gardulf, and M. Granstrom Two year survey of the incidence of Lyme borreliosis and tick-borne encephalitis in a high risk population in Sweden. ur. J. Clin. Microbiol. Infect. Dis. 11: Gustafson, R.,. Svenungsson,. Gardulf, G. Stiernstedt, and M. Forsgren Prevalence of tick-borne encephalitis and Lyme borreliosis in a defined Swedish population. Scand. J. Infect. Dis. 22: Hansen, K., and. sbrink Serodiagnosis of erythema SROLOGICL FOLLOW-UP OF LYM ORRLIOSIS 1525 migians and acrodermatitis chronica atrophicans by the orrelia burgdorferi flagellum enzyme-linked immunosorbent assay. J. Clin. Microbiol. 27: Hansen, K, P. Hindersson, and N. Strandberg-Pedersen Measurement of antibodies to orrelia burgdorferi flagellum improves serodiagnosis in Lyme disease. J. Clin. Microbiol. 26: Hansen, K., and.-m. Lebech Lyme neuroborreliosis: a new sensitive diagnostic assay for intrathecal synthesis of orrelia burgdorferi specific immunoglobulin G, and M. nn. Neurol. 3: Hansen, K., K. Pil, and.-m. Lebech Improved immunoglobulin M serodiagnosis in Lyme borreliosis by using a,u-capture enzyme-linked immunosorbent assay with biotinylated orrelia burgdorferi flagella. J. Clin. Microbiol. 29: Karlsson, M., S. Hammers-erggren, L. Lindqvist, G. Stiernstedt, and. Svenungsson. Comparison of intravenous penicillin G and oral doxycycline for treatment of Lyme neuroborreliosis. Neurology, in press. 14. Karlsson, M., G. Stiernstedt, M. Granstrom,. sbrink, and. Wretlind Comparison of flagellum and sonicate antigens for serological diagnosis of Lyme borreliosis. ur. J. Clin. Microbiol. Infect. Dis. 9: Kruger, H., K. Reuss, M. Pulz,. Rohrbach, K. W. Pflughaupt, R. Martin, and H. G. Mertens Meningoradiculitis and encephalomyelitis due to orrelia burgdorferi: a follow-up study of 72 patients over 27 years. J. Neurol. 236: Pachner,. R., and. C. Steere The triad of neurologic manifestations of Lyme disease: meningitis, cranial neuritis, and radiculoneuritis. Neurology 35: Schmutzhard,., G. Stanek, M. Pletschette,. M. Hirschl,. Pallua, R. Schmitzberger, and R. Schlogl Infections following tick bites. Tick-borne encephalitis and Lyme borreliosis: a prospective epidemiological study from Tyrol. Infection 16: Steere,. C., S.. Malawista, J.. Hardin, S. Ruddy, P. W. skenase, and W.. ndiman rythema chronicum migrans and Lyme arthritis. The enlarging clinical spectrum. nn. Intern. Med. 86: Steere,. C.,. Taylor, G. L. McHugh, and. L. Logigian The overdiagnosis of Lyme disease. JM 269: t). Steere,. C.,. Taylor, M. L. Wilson, J. F. Levine, and. Spielman Longitudinal assessment of the clinical and epidemiological features of Lyme disease in a defined population. J. Infect. Dis. 154: Stiernstedt, G., G. riksson, W. nfors, H. Jorbeck,. Svenungsson,. Skoldenberg, and M. Granstrom rythema chronicum migrans in Sweden: clinical manifestations and antibodies to Ixodes ricinus spirochete measured by indirect immunofluorescence and enzyme-linked immunosorbent assay. Scand. J. Infect. Dis. 18: Stiernstedt, G., M. Granstrom,. Hederstedt, and. Skoldenberg Diagnosis of spirochetal meningitis by enzyme-linked immunosorbent assay and indirect immunofluorescence assay in serum and cerebrospinal fluid. J. Clin. Microbiol. 21: Stiernstedt, G., R. Gustavsson, M. Karlsson,. Svenungsson, and. Skoldenberg Clinical manifestations and diagnosis of neuroborreliosis. nn. N. Y. cad. Sci. 539: Stiernstedt, G. T.,. R. Skoldenberg,. Vandvik,. Hederstedt,. Garde, G. Kolmodin, H. Jorbeck, and. Svenungsson Chronic meningitis and Lyme disease in Sweden. Yale J. iol. Med. 57: Downloaded from on November 22, 218 by guest