Chronic Lyme Disease: An appraisal

Synopsis

“Chronic Lyme disease” is a confusing term that has been used to describe very different patient populations. Studies have shown that most patients diagnosed with “chronic Lyme disease” either have no objective evidence of previous or current infection with B. burgdorferi or are patients that should be classified as having post-Lyme disease syndrome, which is defined as continuing or relapsing non-specific symptoms (such as fatigue, musculoskeletal pain, and cognitive complaints) in a patient previously treated for Lyme disease. Despite extensive study, there is currently no clear evidence that post-Lyme disease syndrome is due to persistent infection with B. burgdorferi. Four randomized placebo-controlled studies have shown that antibiotic therapy offers no sustained benefit to patients with post-Lyme disease syndrome. These studies also showed a substantial placebo effect and a significant risk of treatment-related adverse events. Further research to elucidate the mechanisms underlying persistent symptoms after Lyme disease and controlled trials of new approaches to the treatment and management of these patients are needed.

‘The beginning of wisdom is the definition of terms’. Socrates

Introduction

“Chronic Lyme disease” is probably the most confusing term in the Lyme disease field. The term “chronic Lyme disease” has been used to describe vastly different patient populations, that should not be grouped together. These include patients with objective manifestations of late Lyme disease (for example, arthritis, encephalomyelitis or peripheral neuropathy, addressed in detail in other chapters), patients with post-Lyme disease syndrome, and patients with nonspecific signs and symptoms of unclear cause who receive this diagnosis based on unproven and/or non validated laboratory tests and clinical criteria. In a recent article [1], patients diagnosed with “chronic Lyme disease” were classified in 4 categories (Table 1). This article addresses mainly patients with post-Lyme disease syndrome (category 4) as there have been relatively fewer studies addressing patients in categories 1 and 2; and no studies focusing on patients in category 3.

Table 1.

Categories of “Chronic Lyme Disease”

Category 1	Symptoms of unknown cause, with no evidence of Borrelia burgdorferi infection
Category 2	A well-defined illness unrelated to B. burgdorferi infection
Category 3	Symptoms of unknown cause, with antibodies against B. burgdorferi but no history of objective clinical findings that are consistent with Lyme disease
Category 4	Post-Lyme disease syndrome

From Feder, H.M., Jr., et al., A critical appraisal of “chronic Lyme disease”. N Engl J Med, 2007. 357(14): p. 1422–30

Chronic Lyme disease

Most patients who are labeled as having “chronic Lyme disease” will fall into Categories 1 and 2. Patients in Category 1 are diagnosed with “chronic Lyme disease” based on unexplained symptoms without objective or valid laboratory evidence of infection B. burgdorferi. Patients in Category 2 have other recognized diseases and have been misdiagnosed with Lyme disease. The distribution of patients who fall into these categories can be estimated by the difficulty in accruing patients into the placebo-controlled studies of antibiotic treatment in patients with post-Lyme disease syndrome (Category 4), where only 1 to 10% of the screened individuals were eligible [2–4].

There have been a number of studies addressing the issue of over diagnosis of Lyme disease (Table 2), and while these studies represent the experience of referral centers, they are informative regarding the range of patients seeking further evaluation for suspected Lyme disease. In general, only about one quarter to one third of the patients evaluated were thought to have Lyme disease; in comparison, between 50 to 60% of the patients had no present or past evidence of Lyme disease. A large portion of patients presented with fatigue, myalgias, arthralgias, sleep disturbances, memory complaints and/or depression, and many fulfilled criteria for chronic fatigue syndrome or fibromyalgia [5–10]. Common and related problems contributing to the over diagnosis of Lyme disease included the use of serological testing in clinical situations in which the pre-test probability of Lyme disease was low, misinterpretation of test results, and use of non-validated methods and criteria for interpretation of laboratory results.

Table 2.

Experience of referral centers with patients suspected of Lyme disease

Reference	Patients	Results
[5]	100 patients referred to the Lyme Disease Center at Robert Wood Johnson Medical School, New Brunswick, NJ.	37 patients had Lyme disease. 25 patients fulfilled criteria for fibromyalgia (15 had a history compatible with previous Lyme disease, 3 were thought to have fibromyalgia coincidently with Lyme disease). Other diagnoses were made in 22 patients, while in 14 patients no specific diagnosis was reached. The authors considered that approximately half of the 91 courses of antibiotic therapy given to these patients were unnecessary.
[75]	65 patients referred to the Borrelia Referral Clinic at University Hospital in Vancouver, Canada.	Only 2 patients were judged to have probable Lyme disease. Definite alternative diagnoses were made for 50 patients(77%). Chronic fatigue syndrome and fibromyalgia were diagnosed in 11 patients (17%).
[6]	788 patients referred to the Lyme disease clinic at the New England Medical Center, Boston, MA.	180 (23%) had active Lyme disease, usually arthritis, encephalopathy, or polyneuropathy. 156 patients (20%) had previous Lyme disease and another current illness, with 84 presenting mainly with musculoskeletal pain or fatigue. 452 patients (57%) did not have Lyme disease. The majority of these patients had chronic fatigue syndrome (142) or fibromyalgia (84); the others were diagnosed with rheumatic(143), neurological (41) or other diseases (17).
[7]	227 children referred to the pediatric Lyme disease clinic at the Alfred I. duPont Institute, Wilmington, DE.	138 children did not have Lyme disease and were divided in 4 groups: predominantly subjective symptoms (54 children), alternative diagnosis (52 children), previous Lyme disease (8 children), 20 children were referred because of tick bites, and 4 children because of a family member with Lyme disease. The majority of the children with subjective symptoms had chronic fatigue, most had received previous antibiotic therapy, and 6 children had received prolonged intravenous antibiotic therapy (range 3–36 weeks).
[8]	146 pediatric patients referred with possible Lyme disease to the University of Connecticut Health Center, Farmington, CT.	56 (38%) were considered overdiagnosed, 12 (8%) were underdiagnosed, and 75 (51%) were correctly diagnosed with Lyme disease.
[9]	209 patients referred to the Yale University Lyme Disease Clinic, New Haven, CT.	44 (21%) met criteria for active Lyme disease, 40 (19%) had previous but not active Lyme disease, and 125 (60%) had no evidence of current or previous infection. Patients with previous Lyme disease and patients with no evidence of Lyme disease had a longer median duration of symptoms, and about1/3 had received antibiotic therapy for more than 100 days. At follow up about 4 months later, 71% of the patients with previous Lyme disease and 82% of patients with no evidence of Lyme disease reported persistent symptoms and about 50% disagreed with the diagnosis provided at the Yale clinic. 31% of patients with previous Lyme disease and 20% of patients with no evidence of Lyme disease had sought further evaluation for Lyme disease, and 21% and 11%, respectively, received additional antibiotic therapy.
[10]	216 children referred for Lyme disease to the pediatric infectious diseases clinic at State University of New York at Stony Brook, Stony Brook, NY.	68 (31%) children had active Lyme disease. 39 (18%) children had a prior history of Lyme disease, with 23 having an intercurrent illness or lower school grades; and 16 referred due to confusion in the interpretation of immunoblot results. 109 (50%) children had no past or current evidence of Lyme disease, yet 86 (79%) had been started on therapy before referral.
[76]	86 patients referred to the Rheumatology Unit at the Medical University Policlinic in Bonn, Germany.	Only 8 patients had ongoing or recent Lyme disease. The most common diagnoses were degenerative disorders of the spine (29%), arthropathies related to psoriasis or rheumatoid arthritis (17%), and spondiloarthopathy.

Post-Lyme Disease Syndrome

Many studies have shown that Lyme disease is treated successfully with antibiotics in the majority of cases, and patients with objective evidence of treatment failure are rare with currently recommended regimens [11–14]. Patients with late manifestations can have a slower response to therapy, sometimes taking weeks or months to recover [15–23]. Some patients may have incomplete resolution due to irreversible damage, as can occur in facial nerve palsy with residual facial weakness. A few patients may develop antibiotic-refractory Lyme arthritis, when synovitis persists for months to years after antibiotic therapy, it is most likely due to autoimmunity triggered by the infection [24].

A minority of patients treated for Lyme disease will have persistent or relapsing non-specific symptoms (such as fatigue, musculoskeletal pain, and cognitive complaints) after receiving an adequate course of antibiotic therapy. In the absence of another condition that would explain these non-specific symptoms, such patients are classified as having post–Lyme disease syndrome (Table 3). The best estimates of the prevalence of post-Lyme disease syndrome come from studies of patients with erythema migrans who received appropriate antibiotic treatment. From 10–20% of such patients have persistent or intermittent subjective symptoms of mild to moderate intensity 12 months after completion of therapy (Table 4). The most common post-Lyme disease symptoms are fatigue, arthralgias, myalgias, headache, neck stiffness, paresthesias, sleeplessness, irritability, and difficulty with memory, word finding, and concentration [12, 13, 25–28]. The appearance of post-Lyme disease symptoms seems to correlate with disseminated disease, a greater severity of illness at presentation, and delayed antibiotic therapy [12, 29–33]; but not with the duration of the initial antibiotic therapy [13, 23]. Children appear to be less likely to develop post-Lyme disease symptoms [34–42].

Table 3.

Proposed definition of post–Lyme disease syndrome

Inclusion criteria	Exclusion criteria
An adult or child with a documented episode of early or late Lyme disease fulfilling the case definition of the Centers for Disease	An active, untreated, well-documented coinfection, such as babesiosis.
Control and Prevention. If based on erythema migrans, the diagnosis must be made and documented by an experienced health care practitioner.	The presence of objective abnormalities on physical examination or on neuropsychological testing that may explain the patient’s complaints.
After treatment of the episode of Lyme disease with a generally accepted treatment regimen, there is resolution or stabilization of the objective manifestation(s) of Lyme disease.	A diagnosis of fibromyalgia or chronic fatigue syndrome before the onset of Lyme disease.
Onset of any of the following subjective symptoms within 6 months of the diagnosis of Lyme disease and persistence of continuous or relapsing symptoms for at least a 6-month period after completion of antibiotic therapy:
Fatigue	A prolonged history of undiagnosed or unexplained somatic complaints, such as musculoskeletal pains or fatigue, before the onset of Lyme disease.
Widespread musculoskeletal pain	A diagnosis of an underlying disease or condition that might explain the patient’s symptoms
Complaints of cognitive difficulties	Laboratory or imaging abnormalities that might suggest an undiagnosed process distinct from post–Lyme disease syndrome.
Subjective symptoms are of such severity that, when present, they result in substantial reduction in previous levels of occupational, educational, social, or personal activities.	Although testing by either culture or PCR for evidence of Borrelia burgdorferi infection is not required, should such testing be done by reliable methods, a positive result would be an exclusion.

From Wormser GP, Dattwyler RJ, Shapiro ED, et al., The clinical assessment, treatment, and prevention of Lyme disease, human granulocytic anaplasmosis, and babesiosis: clinical practice guidelines by the Infectious Diseases Society of America. Clin Infect Dis, 2006. 43(9): p. 1089–134

Table 4.

Symptoms after antibiotic therapy in patients with erythema migrans

Year	Country	Study Design	Patients and Treatment	Results	Post-Lyme Disease Symptoms	Reference
1992	US	Randomized investigator- blinded multicenter study	63 patients were randomized to cefuroxime 500 mg orally twice a day for 20 days and 60 patients to doxycycline PO 100 mg three times a day for 20 days.	Satisfactory outcome was seen in 51(93%) of patients who received cefuroxime and 45(88%) of patients who received doxycycline at 1 month. Ten patients were considered to not have a satisfactory outcome. In 9 patients, the EM had resolved but they had arthralgias, myalgias, paresthesias, fatigue and headache.	At 1 year, 43(90%) in the cefuroxime group and 35 (92%) in the doxycycline had satisfactory outcomes, while 8 patients were considered failures as they had arthralgias, myalgias, headache and fatigue. Patients who were assessed as clinical improvements at1 month post- treatment were more likely to become clinical failures at 1 year follow up.	[54]
1993	Slovenia	Randomized open label single-center study	55 patients received azithromycin 500 mg orally twice a day for the1st day followed by 500 mg once a day for 4 days and 52 patients received doxycycline100 mg orally twice a day for 14 days	There were 3 definite and 4 probable treatment failures in the doxycycline group and 1 probable treatment failure in the azithromycin group.	At 1 year, 15 patients on doxycycline and10 on azithromycin had minor symptoms(arthralgias, myalgias, fatigue, headache and concentration disturbances). Most minor symptoms appeared in the first 6 months after therapy.	[26]
1995	US	Randomized investigator- blinded multicenter study	119 patients receive cefuroxime axetil 500 mg orally twice a day and 113 patients received doxycycline100 mg orally three times a day for 20 days	Satisfactory clinical response was seen in 90% of patients on cefuroxime and 95% of patients on doxycycline. Presenting with paresthesias, arthralgia and irritability at the initial visit was associated with failure at 1 month.	Of the 118 patients evaluated at 1 year, satisfactory outcomes were seen in 95% of the patients in the cefuroxime group and 100% in the doxycycline group.	[77]
1996	US	Observational cohort multicenter study	Prospective evaluation of 201 children with Lyme disease in Connecticut. 132 (66%) presented with single EM, 56 (28%) presented with early disseminated disease and 13 (6%) with Lyme arthritis. All were treated with 2–4 weeks of antibiotics, (94% for 3–4 weeks) and 96% were treated orally. 137 received amoxicillin and 51 received doxycycline.	All children responded promptly and 94% were asymptomatic by 4 weeks, with 5% having arthralgia, myalgia and fatigue, and 1% had residual facial palsy. All patients recovered completely at 6 months but for one patient with mild recurrent arthralgia.	At 2-year follow- up, only 1 child had mild recurrent arthralgia.	[35]
1996	US	Randomized double- blinded multicenter study	111 patients received azithromycin 500 mg orally once a day (with placebo orally twice a day) for 7 days (followed by placebo three times a day for 13 days) versus 106 patients to amoxicillin500 mg orally three times a day for 20 days.	At day 20, 84 (76%) had a complete response in the azithromycin group vs. 93 (88%) in the amoxicillin group. Partial response was seen in 24 (22%) in the azithromycin group vs.13 (12%) in the amoxicillin group. There were 3 failures in the azithromycin group.	At 180 days, 17 patients in the azithromycin group vs. 4 patients in the amoxicillin group were considered relapses. A partial response at day20 was predictive of relapse	[53]
1997	US	Randomized open-label multicenter study	68 patients received ceftriaxone IV 2 gr once a day for 14 days and72 patients received doxycycline 100 mg orally twice a day for 21 days.	At 3 months, 55(92%) of ceftriaxone and 63 (94%) in the doxycycline had recovered completely.	At 9 months, 56(97%) in the ceftriaxone group and 58 (94%) in the doxycycline group were considered cured. At the last follow- up visit, there were persistent symptoms in 18 patients treated with ceftriaxone and 10 patients treated with doxycycline. Most symptoms were considered mild.	[27]
2000	Croatia	Randomized open-label multicenter study	48 patients received azithromycin 500 mg orally twice a day for the first day followed by 500 mg daily for 4 days and40 patients received doxycycline 100 mg orally twice a day for 14 days.	There was one clear treatment failure in the azithromycin group.	At 1 year, minor symptoms occurred in 2/47 of patients who received azithromycin and3/35 of patients who received doxycycline.	[78]
2002	US	Observational cohort multicenter study	Follow up of 118 patients participating in a vaccine study who had EM with positive PCR and/or culture. Most patients were treated with oral doxycycline or amoxicillin for 14–30 days.	Most patients had resolution of all symptoms by 3 weeks. At 30 days after therapy, 13(11%) still had symptoms, and 5(4%) had symptoms for more than 60 days (3 with fatigue, headache, arthralgia; and 2 with residual facial numbness or weakness).	One patient had myalgias at the end of the study.	[11]
2002	Slovenia	Randomized open label single-center study	42 children received azithromycin 20 mg/kg/day (maximum1000 mg/day) for the first day followed by 10 mg/kg/day (maximum500 mg/day) for 4 days and 42 children received phenoxymethylpenicillin100,000 IU/kg/day(maximum 3 million IU/day) divided in 3 daily doses for 14 days	Appearance of minor manifestations(17.5% versus24.4%) and major manifestations of Lyme (one patient in each group) was not different between the groups.	At 1 year, all patients were asymptomatic.	[40]
2003	US	Observational cohort single- center study	From 99 patients with101 episodes of EM who were culture positive, there were 96 evaluable cases. 87 cases (91%) received a first-line oral antimicrobial regimen, such as doxycycline, amoxicillin, or cefuroxime axetil, or received intravenous ceftriaxone for 10 to 21 days. Nine cases received a 7-day course of azithromycin.	After 3 months, 84% to 92% of cases were asymptomatic.	Only 8 (10%) of the 81 cases followed for ≥ 1 year were symptomatic at their last visit, a mean of 5.6 ± 2.6 years of follow-up. Their symptoms tended to be intermittent and mild, with only 3 patients (4%) consistently symptomatic at each follow-up visit. Presenting with symptoms during follow up was associated with presenting with more symptoms and of greater severity, and presenting with multiple EM at the first visit.	[12]
2003	US	Randomized double- blinded single- center study	60 patients received a single 2gr dose of IV ceftriaxone followed by doxycycline 100 mg orally twice a day for 10 days, followed by placebo orally twice a day for 10 days. 61 patients received a single dose of IV placebo, followed by doxycycline 100 mg orally twice a day for 10 days, followed by placebo twice a day for10 days. 59 patients received a single placebo injection followed by doxycycline100 mg orally twice a day for 20 days.	The complete response rate was similar in the 3 groups at all time points. At20 days, 97 patients had a complete response, 47 had a partial response, and1 was a failure. The only failure was a patient treated with doxycycline for 10 days who developed meningitis at 18 days and was treated with ceftriaxone.	At 12 months, 103 patients had acomplete response and 24 had a partial response. At 30 months, 86 patients had a complete response and 12 had a partial response.	[13]

The possible causes of post-Lyme disease symptoms

The mechanisms underlying post-Lyme disease symptoms are not known and are likely to be multifactorial. Possible explanations include persistent infection with B. burgdorferi, other tick-borne infections, part of the expected resolution of symptoms after treatment, post-infective fatigue syndrome, autoimmune mechanisms, and intercurrent conditions.

In many patients, these symptoms probably represent the natural evolution of response after therapy, as the percentage of patients reporting symptoms after antibiotic treatment decreases over time. In one study of patients treated for erythema migrans, 34% had symptoms at 3 weeks, 24% at 3 months, and 17% at 12 months [13]. In other patients, a post-infective fatigue syndrome may be triggered by Lyme disease, as has been shown to occur with other infections. Prolonged fatigue after infections is relatively common, and it can be disabling and persistent. A recent study showed that post-infective fatigue syndrome could be predicted by the severity of the acute illness, and its incidence was similar after the different infections [43]. In this cohort, the case rate for provisional post-infective fatigue syndrome was 35% (87/250) at six weeks, 27% (67/250) at three months, and 9% (22/250) at 12 months [43]; rates similar to those reported in patients treated for erythema migrans (see above) [13]. The mechanisms that are triggered during the acute illness and that sustain the persistent symptoms in post-infective fatigue syndrome are currently unknown.

It also important to recognize that there is a substantial background prevalence of similar symptoms in the general population. Musculoskeletal pain is a very common complaint. For example, in a random survey of 3664 persons aged 25 years and over, stratified by age and gender, 44.4% of the individuals reported musculoskeletal pain lasting longer than 3 months, with lower back, shoulder, neck and knee being the most frequently affected sites; and 15.6% reporting chronic pain involving 2 to 3 sites. The prevalence of chronic widespread pain was 5.2% [44]. In another population-based cross-sectional survey that included 2299 subjects, 15% reported chronic widespread pain, and 8% reported chronic fatigue [45]. Insomnia is also common, and can be associated with anxiety, depression and pain [46]. Musculoskeletal pain, fatigue and sleep disturbance are often reported together [47].

Recent studies showed little evidence of a substantial role of other tick-borne infections in the majority of patients with post-Lyme disease syndrome [4, 48–50]. There has been little research in the role of autoimmunity in post-Lyme disease syndrome, but one study showed no association between a class II allele or genotype [51].

A major concern has been that the symptoms of post-Lyme disease syndrome may represent persistent infection with B. burgdorferi. A review of the earliest studies of patients with Lyme disease demonstrate the uncertainty that surrounded the disease and explain in part some of the confusion regarding “chronic Lyme disease”. During those initial years, nonspecific symptoms were classified as part of “minor” late manifestations or complications of Lyme disease, to differentiate from the “major” manifestations, which included arthritis, meningoencephalitis and carditis [25, 29–31]. In some cases, facial palsy and brief episodes of arthritis were grouped together with nonspecific symptoms as part of minor manifestations of late Lyme disease [29, 30], and, in some studies, all patients were grouped together [29, 31]. While arthritis, meningoencephalitis, carditis and other objective manifestations of Lyme disease are clear evidence of treatment failure and require antibiotic therapy [14], there was uncertainty about whether nonspecific “minor” symptoms could also represent treatment failures and that longer courses of antibiotics or different antibiotic regimens may be needed in some of the patients [30, 31, 52, 53].

As the studies progressed and antibiotic therapy for Lyme disease evolved, it became rare for patients with erythema migrans treated with currently recommended antibiotic regimens to develop an objective manifestation of Lyme disease [13]. Physicians also gained more experience following patients who were treated with antibiotics, and, with longer periods of observation, it became apparent that these nonspecific symptoms frequently resolved without further antibiotic treatment, and that antibiotic therapy did not hasten their resolution [33, 54]. Further studies also showed that symptomatic patients were not more likely to be seropositive than patients without symptoms and that patients did not develop objective manifestations of late Lyme disease [12, 18]. While earlier, smaller studies showed a higher prevalence of recurrent arthralgias, symptoms of memory impairment, and other symptoms in persons with a history of Lyme disease compared with controls [32, 33], larger cohort studies showed no differences on physical examination and neurocognitive testing [55], and no difference in the frequency of symptoms between patients with Lyme disease and age-matched controls [39].

Objective evidence of Borrelia infection in patients with post-Lyme disease syndrome has not been found using PCR [4, 49] or culture [4, 49]. It should be noted however, that B. burgdorferi culture and PCR have low sensitivity in most body fluids from patients with Lyme disease [56, 57]. The initial report claiming frequent isolation of B. burgdorferi from patients with post-Lyme disease syndrome using MPM media [58] has not been reproduced by other researchers [49, 59, 60]. One study reported a high percentage of B. burgdorferi PCR in urine samples of patients diagnosed with “chronic Lyme disease” [61], but these results have not been validated. Other tests that have not been helpful to evaluate patients with post-Lyme disease syndrome include changes in C6 antibody levels [62], and antibodies in immune complexes [63].

There have been interesting reports of B. burgdorferi being present after antibiotic therapy in dogs and mice as assessed by PCR, but not by culture [64, 65, 66]. More detailed studies suggested that these organism were attenuated, non infectious spirochetes [66]. The significance of these findings is, at present, unclear. A recent study reported that B. burgdorferi was found by culture in a few mice treated with anti-TNF antibody either simultaneously or 4 weeks after ceftriaxone therapy [67]. However, the number of mice treated in this study was small and the findings need further verification.

Studies of antibiotic treatment in post-Lyme disease syndrome

There are now 4 randomized, placebo-controlled, double-blinded studies of antibiotic therapy in patients with post-Lyme disease syndrome and all showed that prolonged antibiotic therapy offers no sustained benefit and has potential serious adverse effects (Table 5). The first 2 studies, one for patients who were IgG seropositive for B. burgdorferi at enrollment, and the other for seronegative patients, were published together [49]. All patients had well-documented Lyme disease and had previously received antibiotic therapy. These studies enrolled 78 seropositive patients and 51 seronegative patients. Patients were randomized to receive intravenous ceftriaxone, 2 g daily for 30 days, followed by oral doxycycline, 200 mg daily for 60 days, or matching intravenous and oral placebos. The primary outcome was improvement in the Medical Outcomes Study 36-item Short-Form General Health Survey (SF-36) score on day 180 of the study. Patients had previously received an average of three courses of antibiotic therapy and had had symptoms for a median of 4.6 years. Most patients complained of pain, fatigue and cognitive changes. The studies were stopped early because a planned interim analysis showed that there was little chance of demonstrating a difference between treatment groups. Intention-to-treat analyses showed no significant differences between patients in the antibiotic groups and those in the placebo groups in the seropositive study, the seronegative study, or both studies combined. About one-third of the patients improved, one-third of the patients remained unchanged, and one-third of the patients worsened at each time point. There were 2 serious adverse events related to treatment.

Table 5.

Placebo-controlled, double-blinded randomized treatment studies in post-Lyme disease syndrome

Reference	Patients	Regimen and Primary Endpoints	Results	Serious Adverse Events
[49]	78 seropositive and 51 patients seronegative for IgG antibodies to B. burgdorferi at the time of enrollment.	IV ceftriaxone, 2 g daily for30 days, followed by oral doxycycline, 200 mg daily for60 days (64 patients), or matching IV and oral placebos (65 patients). The primary outcome was improvement on SF-36 score at day 180 of the study.	Intention-to-treat analyses at 30, 90, and 180 days showed no significant differences between the antibiotic group and the placebo group in the seropositive study, the seronegative study, or both studies combined. During the 6-month evaluation period, about a third of the patients improved, a third worsened and a third were unchanged by SF-36.	2 patients had serious adverse events associated with treatment that required hospitalization.
[3]	55 patients with persistent severe fatigue post- Lyme disease	IV ceftriaxone 2 g/day (28 patients) or IV placebo (24 patients) for 28 days. Primary clinical outcomes were improvement in fatigue score and cognitive function at 6 months. Follow up at 6 months was completed by26 patients in the ceftriaxone group and 22 patients in the placebo group.	Patients who received ceftriaxone showed improvement on fatigue but there was no benefit in cognitive function. Exploratory analyses showed that patients with positive western blot, no prior IV therapy and less pain had a significant treatment effect.	4 patients had serious adverse events associated with treatment that required hospitalization.
[4]	37 seropositive patients with objective memory impairment and at least 3 weeks of previous IV antibiotic therapy.	Patients were assigned in a2:1 randomization schedule to receive 10 weeks of IV ceftriaxone 2 g/day (23 patients) or IV placebo (14 patients). The primary outcome was improvement at 12 weeks. Durability of benefit was evaluated at 24 weeks. Twenty patients in the ceftriaxone group and 12 patients in the placebo group completed follow up.	There was a slightly greater cognitive improvement in the antibiotic group at week12, but there was no difference at week 24.	8 patients withdrew from therapy, 7 due to adverse events associated with treatment. One patient on ceftriaxone underwent cholecystectomy at week 16.

IV, intravenous; SF-36, Medical Outcomes Study 36-item Short-Form General Health Survey.

The third study enrolled 55 patients with post-Lyme disease syndrome who had significant fatigue [3]. These patients were randomized to ceftriaxone 2 g (28 patients) or placebo (24 patients) intravenously daily for 28 days. The primary clinical endpoints were improvement in the fatigue and mental speed at 6 months. Eighteen patients (64%) in the ceftriaxone group and 19 patients (70.4%) in the placebo group were ELISA and western blot seropositive at enrollment, while 12 (43%) in the ceftriaxone group and 14 (52%) in the placebo group had received at least 2 weeks of intravenous ceftriaxone before the study. The intent to treat analysis showed modest improvement of fatigue with ceftriaxone therapy, with similar results for patients who received therapy and completed follow up. There was no improvement in mental speed or other neurocognitive measures. Three patients in each group discontinued therapy due to side effects, and 4 had to be hospitalized. In this study, significant more patients who received ceftriaxone were able to correctly guess their assignment comparing with placebo recipients.

The fourth study enrolled patients with post-Lyme disease syndrome who were seropositive by IgG western blot, had objective memory impairment and had received at least 3 weeks of intravenous antibiotic therapy [4]. There were only 37 patients enrolled, and they were randomized 2:1 to receive 10 weeks of intravenous ceftriaxone (23 patients) or intravenous placebo (14 patients). The primary outcome was improvement in memory performance at 12 weeks. Patients were evaluated at 24 weeks for durability of benefit. Twenty patients in the ceftriaxone group and 12 patients in the placebo group completed the follow up. In comparisons using a model with an aggregate of the six domains of neurocognitive performance measured in the study, the ceftriaxone group showed a slightly greater improvement at 12 weeks. At 24 weeks, both groups had improved similarly from baseline. Exploratory analysis suggested a greater improvement in physical functioning and pain among patients with greater baseline impairment treated with ceftriaxone. There were 9 patients who discontinued therapy due to side effects, and in 7 patients these side effects were related to the treatment.

Three of these randomized trials have been criticized as offering “too little, too late” [68–70], based on retrospective, open-label case-series that suggested a possible role of prolonged antibiotic therapy in patients diagnosed with “chronic Lyme disease” [71, 72]. In general, case-series studies are fraught with potential for biases. For example, both patients and physicians’ choices will affect the decision to prescribe a drug to a particular patient. The lack of blinding can affect outcomes, especially for subjective measures. Without a comparison group, it is not possible to know if an outcome is related to an intervention, or to a placebo effect, time, or chance. Case-series and case reports are classified at the lowest level of strength in the hierarchy of evidence based medicine [73]. They are best used for hypothesis generation to be investigated by stronger study designs.

Conclusion

At this point, the overwhelming evidence shows that prolonged antibiotic therapy, as tested in the clinical trials, does not offer lasting or substantive benefit in treating patients with post-Lyme disease syndrome. Therefore, it is time to move forward to test other approaches that may help these patients. Unfortunately, no prospective studies of other treatment modalities for patients with post-Lyme disease syndrome have been performed to date. Due to the significant placebo effect and the variation in symptoms intensity seem in these patients, interventional studies should have a randomized controlled design, with clearly defined target patient populations. For the health care provider taking care of these patients, as always, they should review carefully the evidence for the diagnosis of Lyme disease and not loose sight that these patients can develop other unrelated conditions. It is important that patients be offered the best advice based on current, evidence-based information [74]. Most importantly, there should be a collaborative approach to the treatment process with the patient. Hopefully, further research to understand “chronic Lyme disease” and the reasons underlying persistent symptoms after Lyme disease will lead to the development of beneficial therapies.