Diane Holmgren

Duration of antibiotic therapy for early Lyme disease. A randomized, double-blind, placebo-controlled trial.

Context Optimal antibiotic treatment for patients with early Lyme disease is unclear. Contribution This single-center randomized, double-blind, placebo-controlled trial found that patients with erythema migrans given any of the follwoing regimens had high response rates, defined as resolution of erythema migrans and symptoms at 30 months: 20 days of doxycycline, 83.9%; 10 days of doxycycline, 90.3%; and 10 days of doxycycline plus a single intravenous dose of ceftriaxone, 86.5%. Patients given doxycycline plus ceftriaxone more often had diarrhea than patients given doxycycline alone. Implications Oral doxycycline alone for 10 days is sufficient treatment for patients with early Lyme disease that manifests as erythema migrans. The Editors Although Lyme disease is the most common vector-borne disease in the United States (1), the appropriate duration of treatment for its most common manifestation, erythema migrans, remains unclear. A small open-label prospective study reported in 1983 found that outcome did not improve when tetracycline therapy was extended from 10 days to 20 days (2). However, most recent treatment trials have used antibiotic regimens of approximately 3 weeks (3-5), and some authorities recommend 20 to 30 days of therapy (6). This change in practice has occurred in the absence of additional prospective studies on the duration of treatment and is a source of ongoing controversy. Another uncertain issue in the management of patients with erythema migrans is whether Borrelia burgdorferi, the etiologic agent, has disseminated to the central nervous system at the time of presentation (7). If so, the outcome of therapy might be enhanced by treatment with a parenteral agent such as ceftriaxone, which readily crosses the bloodbrain barrier. To address these concerns, we conducted a placebo-controlled study comparing 10 days of doxycycline with both 10 days of doxycycline plus one dose of ceftriaxone and 20 days of doxycycline. Methods Patients at least 16 years of age who had erythema migrans and satisfied the U.S. Centers for Disease Control and Preventions surveillance definition of Lyme disease (an annular erythematous skin lesion 5 cm in diameter) (8) entered the study between 1992 and 1994. Volunteers were recruited primarily through the walk-in Lyme Disease Diagnostic Center at the Westchester Medical Center, Valhalla, New York. Exclusion criteria included pregnancy or lactation, allergy to a tetracycline or a -lactam antibiotic, receipt of antibiotic treatment for Lyme disease for more than 48 hours before enrollment, meningitis or advanced heart block, or any underlying conditions that might interfere with evaluability or follow-up. All patients gave written informed consent, and the institutional review board at New York Medical College approved the study protocol. Patients were randomly assigned to one of three treatment groups: 1) a single 2-g dose of intravenous ceftriaxone followed by 10 days of oral doxycycline capsules twice daily, then by 10 days of oral placebo capsules [cornstarch], identical in appearance to the doxycycline capsules, twice daily; 2) a placebo injection [5% dextrose] followed by 10 days of oral doxycycline, 100 mg twice daily, and then by 10 days of oral placebo twice daily; or 3) a placebo injection followed by 20 days of oral doxycycline twice daily. Since ceftriaxone is yellow, infusion bags were kept covered to maintain masking. The pharmacy dispensed study medications in accordance with a randomization schedule that was based on a computer-generated random-number code with a permuted block size of 9. Clinical staff involved in recruitment of participants or in assessing clinical outcomes were isolated from the allocation process and blinded to treatment assignment. Randomization was stratified by whether patients were symptomatic (defined as having any systemic symptoms or having multiple erythema migrans lesions) or asymptomatic (defined as having a single erythema migrans lesion and no systemic symptoms). This was done to ensure that patients who may have had dissemination of B. burgdorferi were equally represented in the three treatment groups. Evaluation Trained study personnel interviewed participants and performed physical examination at baseline and 10 days, 20 days, 3 months, 6 months, 12 months, 24 months, and 30 months after initiation of therapy. If appointments were missed, patients were interviewed by telephone; however, this occurred in fewer than 5% of patient interactions. A neurologist performed complete neurologic examination at baseline, 20 days, 3 months, 12 months, 24 months, and 30 months. At 18 months, an evaluation was conducted by telephone. Structured questionnaires using both closed- and open-ended questions were used. Symptom scores were recorded on a visual analogue scale (9). A blood sample was obtained for serologic testing (enzyme-linked immunosorbent assay) at all visits. Complete blood count was determined and serum chemistries were performed at baseline, day 10, and day 20. Electrocardiography was done at baseline and was repeated if results were abnormal. Patients were considered unevaluable if they did not adhere to study medication. Nonadherence was defined as taking fewer than 90% of prescribed capsules or not returning pill containers at the 10- or 20-day visit, receiving an intercurrent antibiotic within the first 20 days, not meeting study inclusion criteria, or not attending follow-up visits after the baseline visit. Patients who had an intercurrent episode of erythema migrans due to reinfection were considered unevaluable from that point onward. Outcome was characterized as complete response, partial response, or failure. Early treatment response was assessed at 20 days. At this time, complete response was defined as resolution of erythema migrans and associated symptoms and return to preLyme disease health status. Partial response was defined as resolution of erythema migrans but incomplete resolution or development of subjective symptoms. Failure was defined as the occurrence of any one of the following during the first 20 days: no clinical improvement by day 10; recurrence of erythema migrans; recurrence of fever attributed by the study physician to Lyme disease; development of new objective rheumatologic, cardiac, or neurologic manifestations of Lyme disease that were not present within the first 10 days; or the occurrence of meningitis, advanced heart block, or other objective manifestation of Lyme disease requiring intravenous therapy. Late response was evaluated at 3 months, 12 months, and 30 months. Complete response was defined as no recurrence of erythema migrans or associated symptoms and the continued absence of objective rheumatologic, cardiac, or neurologic manifestations of Lyme disease, with return to preLyme disease health status. Partial response was defined as no recurrence of erythema migrans and the continued absence of objective manifestations of Lyme disease, but incomplete resolution or development of subjective symptoms of uncertain cause. Failure was defined as the occurrence of objective manifestations of Lyme disease. Safety Assessment Drug safety was monitored by recording adverse events (solicited by open-ended semi-structured interview) and results of laboratory tests during treatment. Neurocognitive Evaluation At baseline, 12 months, and 30 months, a psychologist performed neurocognitive testing consisting of the Booklet Categories Test, the California Verbal Learning Test, the Wechsler Memory ScaleRevised, the Block Design Test, the Trail Making Test (Parts A and B), the Boston Naming Test, the Symbol Digit Modalities Test (written and oral), the Beck Depression Inventory, and the Symptom Checklist-90Revised. Healthy volunteers without a history of Lyme disease were recruited to serve as controls for the psychiatric interview and the neuropsychological testing. Spouses were preferred as controls since they were usually similar to the patients in age, socioeconomic level, and place of residence. Controls were 27 persons with a mean age (SD) of 42.6 15.2 years; 22% were men. Statistical Analysis Groups were compared by using one-way analysis of variance for continuous variables and the chi-square test for categorical variables (two-tailed). Data were analyzed both for participants who adhered to the study protocol and on an intention-to-treat basis. Patient blinding was evaluated at the 30-month visit by using the chi-square test and the statistic (10). For the neuropsychological tests, analysis of variance techniques were used to examine differences among treatment and control groups. For analysis of raw test scores, analysis of covariance was used, controlling for the effects of age, sex, and education. Test scores that were normalized to population standards (11-19) were analyzed by using one-way analysis of variance. For all analysis of variance models, residual plots were generated to verify adequate model fit. In the few situations in which test scores were not normally distributed or model fit was questionable, nonparametric procedures (the KruskalWallis test) or conventional transformations were applied. For categorical results, the chi-square test or the Fisher exact test was used. If the treatment variable in the global tests achieved the liberal cut-point of a P value less than 0.10, pairwise comparisons were conducted. In these instances, the Bonferroni adjustment was applied to the P value to control for multiple comparisons. Sample size was based on the estimated frequency of postLyme disease syndrome, defined as an array of subjective symptoms, such as fatigue, arthralgias, or myalgias, that may occur after infection. It was assumed that 30% of patients receiving the least effective treatment would develop postLyme disease syndrome. A more effective treatment was assumed to reduce the rate to 5%. Using an level of 0.05, two-tailed testing, and a Bonferroni multiple compari

Association of Specific Subtypes of Borrelia burgdorferi with Hematogenous Dissemination in Early Lyme Disease

To investigate whether genetic diversity of Borrelia burgdorferi sensu stricto may affect the occurrence of hematogenous dissemination, 104 untreated adults with erythema migrans from a Lyme disease diagnostic center in Westchester County, New York, were studied. Cultured skin isolates were classified into 3 groups by a polymerase chain reaction amplification and restriction fragment length polymorphism (RFLP) method. A highly significant association between infecting RFLP type in skin and the presence of spirochetemia was found (P<.001). The same association existed for the presence of multiple erythema migrans lesions (P=.045), providing clinical corroboration that hematogenous dissemination is related to the genetic subtype of B. burgdorferi sensu stricto. There were no significant associations between RFLP type and seropositivity or clinical symptoms and signs except for a history of fever and chills (P=.033). These results suggest that specific genetic subtypes of B. burgdorferi sensu stricto influence disease pathogenesis. Infection with different subtypes of B. burgdorferi sensu stricto may help to explain differences in the clinical presentation of patients with Lyme disease.

Long-term follow-up of patients with culture-confirmed lyme disease

PURPOSE To determine the long-term outcome of patients with culture-confirmed Lyme disease. METHODS We analyzed data collected prospectively on adult patients from a highly endemic area in New York State who were diagnosed with early Lyme disease between 1991 and 1994. Patients with culture-confirmed erythema migrans were evaluated at baseline, 7 to 10 days, 21 to 28 days, 3 months, 6 months, 1 year, and annually thereafter. All patients were treated with antibiotics at the time of diagnosis. RESULTS We evaluated 96 cases on 709 separate occasions (median, eight evaluations per case). The erythema migrans rash resolved within 3 weeks in all of the 94 evaluable cases, none of whom developed an objective extracutaneous manifestation of Lyme disease. Of the 81 cases who were followed for >/=1 year, all but 8 (10%) were asymptomatic at their last visit, a mean (+/- SD) of 5.6 +/- 2.6 years into follow-up, and only 3 (4%) were symptomatic at every follow-up visit. Intercurrent tick bites were reported by 45 cases (47%), and 14 (15%) developed a second episode of erythema migrans. Four other cases who were asymptomatic seroconverted between years 2 and 5. CONCLUSION The long-term outcome of patients with erythema migrans after antibiotic therapy was excellent, but patients from a highly endemic area in New York State remained at high risk of re-exposure to ticks and reinfection. Subjective symptoms during follow-up evaluations tended to be mild to moderate, intermittent, and associated with more symptomatic illness at the time of initial diagnosis.

Prospective Clinical Evaluation of Patients from Missouri and New York with Erythema Migrans—Like Skin Lesions

BACKGROUND The most common and most recognizable feature of Borrelia burgdorferi infection (Lyme disease) is the skin lesion erythema migrans (EM). An illness associated with an EM-like skin lesion, but which is not caused by B. burgdorferi, occurs in many southern states in the United States (southern tick-associated rash illness [STARI], also known as Masters disease). METHODS Clinical features of 21 cases of EM-like skin lesions in 21 patients from Missouri were compared in a prospective study with those of 101 cases in 97 patients with EM-like skin lesions from New York. RESULTS Among Missouri cases, the peak incidence of EM-like skin lesions occurred earlier in the year than it did among New York cases (P<.001). Case patients from Missouri were more likely to recall a tick bite than were case patients from New York (85.7% and 19.8%, respectively; P<.001), and the time period from tick bite to onset of the skin lesion was shorter among Missouri case patients (6.1+/-4.2 days and 10.4+/-6.1 days, respectively; P=.011). Missouri case patients were less likely to be symptomatic than were New York case patients (19.0% and 76.2%, respectively; P<.001), and Missouri case patients were less likely to have multiple skin lesions (4.8% and 26.7%, respectively; P=.042). EM-like lesions in Missouri cases were smaller in size than those in New York cases (8.3+/-2.2 cm and 16.4+/-11.5 cm, respectively; P<.001), more circular in shape (P=.004), and more likely to have central clearing (76.2% and 21.6%, respectively; P<.001). After antibiotic treatment, Missouri case patients recovered more rapidly than did New York case patients (P=.037). CONCLUSION Cases of EM-like skin lesion in patients from Missouri and New York have distinct clinical presentations.

Microbiologic Evaluation of Patients from Missouri with Erythema Migrans

BACKGROUND Borrelia lonestari infects Amblyomma americanum, the tick species that is the most common cause of tick bites in southeast and south-central United States, and this spirochete has been detected in an erythema migrans (EM)-like skin rash in 1 patient. Therefore, B. lonestari is considered to be a leading candidate for the etiologic agent of EM in this region. METHODS Skin biopsy specimens obtained from patients from the Cape Girardeau area of Missouri who had EM-like lesions were cultured in Barbour-Stoenner-Kelly medium and evaluated by polymerase chain reaction (PCR) targeting multiple genes. Serum specimens were tested by enzyme-linked immunosorbent assay for antibodies against sonicated whole-cell Borrelia burgdorferi. Results were compared with those obtained over the same period for patients from New York State who had EM. RESULTS B. lonestari was not detected by PCR in any of 31 skin biopsy specimens collected from 30 Missouri patients. None of 19 cultures of Missouri skin samples that were suitable for evaluation were positive for B. burgdorferi, compared with 89 (63%) of 142 cultures of samples collected from New York State patients (P<.001). None of the 25 evaluable Missouri patients were seropositive for antibodies against B. burgdorferi, compared with 107 (75%) of 143 New York State patients (P<.001). CONCLUSIONS Neither B. lonestari nor B. burgdorferi is likely to be the cause of EM-like skin lesions in patients from the Cape Girardeau area of Missouri. The etiology of this condition remains unknown.

Differentiation of Reinfection from Relapse in Recurrent Lyme Disease

BACKGROUND Erythema migrans is the most common manifestation of Lyme disease. Recurrences are not uncommon, and although they are usually attributed to reinfection rather than relapse of the original infection, this remains somewhat controversial. We used molecular typing of Borrelia burgdorferi isolates obtained from patients with culture-confirmed episodes of erythema migrans to distinguish between relapse and reinfection. METHODS We determined the genotype of the gene encoding outer-surface protein C (ospC) of B. burgdorferi strains detected in cultures of skin or blood specimens obtained from patients with consecutive episodes of erythema migrans. After polymerase-chain-reaction amplification, ospC genotyping was performed by means of reverse line-blot analysis or DNA sequencing of the nearly full-length gene. Most strains were further analyzed by determining the genotype according to the 16S-23S ribosomal RNA intergenic spacer type, multilocus sequence typing, or both. Patients received standard courses of antibiotics for erythema migrans. RESULTS B. burgdorferi isolates obtained from 17 patients who received a diagnosis of erythema migrans between 1991 and 2011 and who had 22 paired episodes of this lesion (initial and second episodes) were available for testing. The ospC genotype was found to be different at each initial and second episode. Apparently identical genotypes were identified on more than one occasion in only one patient, at the first and third episodes, 5 years apart, but different genotypes were identified at the second and fourth episodes. CONCLUSIONS None of the 22 paired consecutive episodes of erythema migrans were associated with the same strain of B. burgdorferi on culture. Our data show that repeat episodes of erythema migrans in appropriately treated patients were due to reinfection and not relapse. (Funded by the National Institutes of Health and the William and Sylvia Silberstein Foundation.).

Clinical and Laboratory Spectrum of Culture-Proven Human Granulocytic Ehrlichiosis: Comparison with Culture-Negative Cases

We describe the clinical and laboratory manifestations of human granulocytic ehrlichiosis (HGE) in eight patients for whom cultures were positive for the HGE agent and compare them with 15 patients for whom cultures were negative but who fulfilled a modified New York State Surveillance definition for HGE. Polymerase chain reaction analysis was positive in 8 (100%) of 8 culture-positive cases vs. 3 (20%) of 15 culture-negative cases (P < .001), morulae were detected in 7 (100%) of 7 culture-positive cases in which tests were performed vs. 0 of 15 culture-negative cases (P < .001), and a fourfold change in antibody titer was demonstrated in 6 (75%) of 8 culture-positive cases vs. 9 (69%) of 13 culture-negative cases (P = not significant). Patients for whom cultures were positive had higher mean oral temperatures +/- SD at presentation than did patients for whom cultures were negative (38.6 degrees C +/- 0.7 degree C vs. 37.2 degrees C +/- 0.8 degree C, respectively; P = .002). Other symptoms and signs were not significantly different between the two groups. Multivariate analysis revealed that the lymphocyte count at presentation was significantly lower in culture-positive cases than in culture-negative cases. Clinical response to treatment was similar in the two groups. Culture confirmation of HGE is the gold standard for defining the sensitivity and specificity of other diagnostic tests presently being developed.

Lyme Disease and Human Granulocytic Anaplasmosis Coinfection: Impact of Case Definition on Coinfection Rates and Illness Severity

BACKGROUND Lyme disease is transmitted by the bite of the Ixodes scapularis tick, which can also transmit Anaplasma phagocytophilum, the cause of human granulocytic anaplasmosis (HGA). Conflicting data exist on the frequency of coinfection and on whether Lyme-HGA coinfected patients have more symptoms than patients with Lyme disease alone. METHODS Blood culture and serology were used to detect HGA infection in patients with early Lyme disease who presented with erythema migrans. The rate of coinfection was determined using different definitions. The clinical and laboratory features of Lyme-HGA coinfection were compared with that of the individual infections. RESULTS Among 311 patients with erythema migrans, the frequency of coinfection with HGA varied from 2.3% to 10.0%, depending on the definition used (P < .001). Only 1 of 4 groups with presumed coinfection had significantly more symptoms than patients with Lyme disease alone P < .05. High fever and cytopenia were less common in Lyme-HGA coinfection than in patients with HGA alone. CONCLUSION The results of this study indicate that how HGA is defined in patients with early Lyme disease has an impact on the apparent rate of coinfection and the severity of illness. The findings also suggest that HGA may be less severe than is usually believed, suggesting the existence of referral bias in testing patients preferentially who present with high fever or cytopenia.

Comparison of five diagnostic modalities for direct detection of Borrelia burgdorferi in patients with early Lyme disease.

Lyme disease, the most commonly reported tick-borne infection in North America, is caused by infection with the spirochete Borrelia burgdorferi. Although an accurate clinical diagnosis can often be made based on the presence of erythema migrans, in research studies microbiologic or molecular microbiologic confirmation of the diagnosis may be required. In this study, we evaluated the sensitivity of 5 direct diagnostic methods (culture and nested polymerase chain reaction [PCR] of a 2-mm skin biopsy specimen, nested PCR and quantitative PCR (qPCR) performed on the same 1-mL aliquot of plasma and a novel qPCR-blood culture method) in 66 untreated adult patients with erythema migrans. Results of one or more of these tests were positive in 93.9% of the patients. Culture was more sensitive than PCR for both skin and blood, but the difference was only statistically significant for blood samples (P<0.005). Blood culture was significantly more likely to be positive in patients with multiple erythema migrans skin lesions compared to those with a single lesion (P=0.001). Positive test results among the 48 patients for whom all 5 assays were performed invariably included either a positive blood or a skin culture. The results of this study demonstrate that direct detection methods such as PCR and culture are highly sensitive in untreated adult patients with erythema migrans. This enabled microbiologic or molecular microbiologic confirmation of the diagnosis of B. burgdorferi infection in all but 4 (6.1%) of the 66 patients evaluated.

Blood Cultures for Patients with Extracutaneous Manifestations of Lyme Disease in the United States

Spirochetemia in US patients with extracutaneous manifestations of Lyme disease is not well documented. In this study, blood culture results were positive for 5 (19.2%; 95% confidence interval, 6.6%-39.4%) of 26 untreated adult patients with extracutaneous manifestations but only for patients with clinical evidence for a short duration of infection.