Intended for healthcare professionals

Clinical Review State of the Art Review

Lyme borreliosis: diagnosis and management

BMJ 2020; 369 doi: (Published 26 May 2020) Cite this as: BMJ 2020;369:m1041
  1. Bart Jan Kullberg, professor of internal medicine and infectious diseases1,
  2. Hedwig D Vrijmoeth, Lyme disease project leader and, consultant1,
  3. Freek van de Schoor, Lyme disease project leader and consultant1,
  4. Joppe W Hovius, professor in vector-borne infections and head2
  1. 1Department of Medicine and Radboudumc Center for Infectious Diseases, Radboud University Medical Center, Nijmegen, Netherlands
  2. 2Amsterdam University Medical Centers, location AMC, Department of Medicine, Division of Infectious Diseases, and Amsterdam Multidisciplinary Lyme borreliosis Center, Amsterdam, Netherlands
  1. Correspondence to BJ Kullberg BJ.Kullberg{at}


Lyme borreliosis is the most common vectorborne disease in the northern hemisphere. It usually begins with erythema migrans; early disseminated infection particularly causes multiple erythema migrans or neurologic disease, and late manifestations predominantly include arthritis in North America, and acrodermatitis chronica atrophicans (ACA) in Europe. Diagnosis of Lyme borreliosis is based on characteristic clinical signs and symptoms, complemented by serological confirmation of infection once an antibody response has been mounted. Manifestations usually respond to appropriate antibiotic regimens, but the disease can be followed by sequelae, such as immune arthritis or residual damage to affected tissues. A subset of individuals reports persistent symptoms, including fatigue, pain, arthralgia, and neurocognitive symptoms, which in some people are severe enough to fulfil the criteria for post-treatment Lyme disease syndrome. The reported prevalence of such persistent symptoms following antimicrobial treatment varies considerably, and its pathophysiology is unclear. Persistent active infection in humans has not been identified as a cause of this syndrome, and randomized treatment trials have invariably failed to show any benefit of prolonged antibiotic treatment. For prevention of Lyme borreliosis, post-exposure prophylaxis may be indicated in specific cases, and novel vaccine strategies are under development.


Lyme borreliosis, or Lyme disease, is an emerging tickborne disease, primarily caused by the bacterium Borrelia burgdorferi sensu stricto (ss) in North America and predominantly Borrelia afzelii or Borrelia garinii in Europe. Reported incidence has been increasing, and the clinical manifestations of Lyme borreliosis are diverse. Establishing the diagnosis may be complex, particularly for early manifestations, before a serological response has developed, and where there is disseminated infection in the joints, the heart, or the central nervous system. Another limitation of serological testing is that antibodies can remain for years after infection, and serology can therefore not be used to assess the efficacy of antibiotic therapy. Thus, the diagnosis and evaluation of Lyme borreliosis mainly depends on clinical evaluation, as is discussed in this review. Most people respond well to antibiotic therapy as recommended by treatment guidelines. However, some report post-infectious signs or symptoms, despite recommended antibiotic therapy and putative clearance of infection. These symptoms, which include fatigue, pain, and neurocognitive symptoms, may be persistent and highly disabling. Common uncertainties among patients and physicians include the reliability of diagnostic tests for Lyme borreliosis, and the pathogenesis and therapy of persistent symptoms. In this review, we assess the diagnostic and therapeutic approach to Lyme borreliosis, and the evidence related to pathogenesis and management of sequelae and post-infectious symptoms attributed to Lyme borreliosis.

Sources and selection criteria

We searched PubMed for publications between January 2009 and January 2019, using the search terms “Borrelia Infections”, “Borrelia”, “borrelia”, “b. afzelii”, “b. burgdorferi”, “b. garinii”, “Borrelia afzelii,,” “Borrelia burgdorferi”, “Borrelia garinii”, “borreliosis”, “Erythema Chronicum Migrans”, “Erythema Migrans”, “lyme”, “Lymes”, “Lyme’s”, “Neuroborreliosis”

We categorized human, animal, and in vitro studies based on title and abstract into the topics of this review, and favored randomized trials, systematic reviews, and guidelines in English. Observational patient studies, in vitro, and animal studies with adequate study design and statistical methods were also reviewed. Because of the clinical focus of this review, we also included case series.

We also reviewed clinical trials of limited quality and correspondence to describe actual controversies. Studies published before 2009 were included if they were referred to by selected papers, guidelines, or reviews. We included articles of sufficient quality published after January 2019 (during the writing process of this review) if relevant to the scope of this review.

Incidence and epidemiology

Borrelia burgdorferi sensu lato and their vectors

Lyme borreliosis is caused by spirochetes belonging to the Borrelia burgdorferi sensu lato (sl) complex, which consists of ~20 genospecies with a complex genomic structure.1 Not all B burgdorferi sl species are pathogenic, and in North America, Borrelia burgdorferi ss is the dominant genospecies associated with Lyme borreliosis,2 although a novel genospecies, Borrelia mayonii, was recently identified.3 In Eurasia, B afzelii and B garinii are the most common B burgdorferi sl genospecies in ticks and humans.45B burgdorferi sl genospecies are genetically distinct from other species within the genus Borrelia—ie, those causing relapsing fever. Recently, it was suggested to divide the genus Borrelia into two, with the new genus name for the Lyme borreliosis group of spirochetes being Borreliella,6 although it is debatable whether such a split is justified.78

B burgdorferi sl spirochetes are transmitted through the bite of tick species belonging to the genus Ixodes, which are largely confined to temperate climate zones of the northern hemisphere.2 In North America, the Ixodes species that transmits the causative agent of Lyme borreliosis is primarily Ixodes scapularis; however, Ixodes pacificus also acts as a vector in the western coastal regions.9 In Europe, Ixodes ricinus is the tick species primarily responsible for transmitting B burgdorferi sl, whereas Ixodes persulcatus is predominant in large parts of Russia and Asia.2 On average, in Europe 12% of nymphal and 15% of adult I ricinus ticks are infected with B burgdorferi sl,5 and 2-3% of humans develop Lyme borreliosis after a tick bite,101112 which is similar to the incidence in the US.13

Incidence of Lyme borreliosis in the northern hemisphere

In 2006 roughly 85 000 cases of Lyme borreliosis were reported annually in Europe,14 and more recently, this was estimated to be approximately 230 000 in Western Europe, although this is thought to be an underestimate.15 Incidences in some countries peak as high as 350 per 100 000 population and have increased in the past two decades.151617 Lyme borreliosis is also highly prevalent in North America; the number of reported cases has gradually increased over time in the US,18 and the Centers for Disease Control and Prevention (CDC) has estimated that there are more than 300 000 new cases each year.19 The causes for this increase include greater abundance of wildlife hosts on which ticks feed and propagate, and climatic changes, which result in expansion of the latitude, altitude, and seasonality at which ticks are found.2021 This increase has led to a substantial disease burden and economic costs,22 and has brought societal and political concerns in both the US and Europe.2324

Clinical manifestations and diagnosis

Below we describe the most common clinical disease manifestations in Europe and North America in children and adults. The clinical spectrum is more diverse,25 but rare disease manifestations are beyond the scope of this review. The diagnostic strategies for Lyme borreliosis are reviewed in box 1 and table 1.

Box 1

Laboratory support for diagnosis of Lyme borreliosis

Lyme borreliosis has a diverse clinical presentation, and, with the exception of erythema migrans, laboratory support for evidence of infection should be sought


  • Most readily available form of laboratory support. Should be interpreted in combination with the clinical symptoms and signs

  • Two tier testing is typically performed, where equivocal or positive ELISA based screening test results are technically confirmed using a second test, which can be another ELISA, western blot, or immunoblot26

  • Clinical signs often precede an antibody response. Therefore, in early phases of the disease, antibody responses may be absent, but sensitivity increases over time*

  • Absence of antibody response has been reported in cases confirmed by polymerase chain reaction (PCR) or culture27

  • Antibiotic therapy may abort serologic response or prevent seroconversion28

  • Background seroprevalence exists, from 5% in the general population in endemic regions to 50% in hunters2930

  • Serum Borrelia IgG may persist for decades. Therefore, serology cannot be used to monitor disease activity or eradication31

Polymerase chain reaction

  • Reasonable sensitivity on skin and synovial samples, low sensitivity on cerebrospinal fluid*

  • In other materials, such as blood or urine, PCR has no or limited diagnostic value3233

  • Limited utility in monitoring treatment response, as Borrelia DNA may be detected after successful antibiotic treatment3435

Cerebrospinal fluid (CSF) analysis

  • Similar to PCR, CSF cultures for Borrelia spp have a low yield*

  • Diagnosis relies on indirect measures of meningeal inflammation: pleocytosis, intrathecal antibody production3637

  • Intrathecal Borrelia antibody is measured by calculating the CSF:serum antibody index and has been shown to persist for years after successful treatment, and thus cannot be used to monitor treatment38

  • Chemokine C-X-C motif ligand 13 (CXCL13) is an early biomarker and its concentration falls rapidly after initiation of antibiotic therapy. Elevated CXCL13 concentrations in CSF may also be detected in other disorders, particularly neurosyphilis and central nervous system lymphoma39

Cellular immune response tests and other non-recommended tests

  • Based on assessing T cell mediated immune responses following in vitro stimulation by a specific pathogen. Interferon gamma based cellular tests are well established for tuberculosis, but experimental for Lyme borreliosis40

  • In Lyme borreliosis, results are inconsistent because of small patient cohorts, no clear case definitions, no or poorly defined control groups, and lack of independent academic validation41

  • In a small cohort study, the sensitivity of a cellular assay measuring IFN-γ release was suggested to exceed that of serology during early infection42

  • Validation of four cellular assays in larger populations with early Lyme borreliosis is ongoing43

  • A variety of commercially available testing methods, including urine antigen testing,44 quantitative CD57 assay,45 or dark field microscopy on blood,46 lack a solid scientific basis as well as independent, reproducible validation and should be avoided for clinical use47

  • *Sensitivity and specificity are reported in table 1.

Table 1

Diagnostic tests for Lyme borreliosis

View this table:

Cutaneous manifestations

Erythema migrans

Erythema migrans, the most common manifestation of Lyme borreliosis,77 is characterized by a red or bluish-red macular skin lesion expanding over the course of days to weeks (fig 1).78 In contrast, a tick bite rash may develop within hours to days and generally fades after several days. Only half of people diagnosed with erythema migrans recall a tick bite.78 Historically, a typical case of erythema migrans has been characterized by a bright red outer border with central clearing. However, B burgdorferi ss and B garinii usually lead to homogeneous erythema migrans, as opposed to B afzelii erythema migrans, which is characterized by central clearing in 60% of cases.7980 In 2-18% of cases, erythema migrans is multiple.777880 Untreated erythema migrans may persist for several weeks and occasionally months.81

Fig 1
Fig 1

Divergent characteristics of erythema migrans. Patients with proven erythema migrans, presenting with classic lesions with central clearing (panels 1, 2), homogeneous lesions (panels 3, 4), large lesion with sharply defined borders (panels 5, 6), or patchy skin discolorations (panels 7, 8).
B burgdorferi skin infection was proven by positive culture and PCR (panels 1, 2, 4, 5, 7) or positive PCR only (panels 3, 6, 8)

Borrelial lymphocytoma

Borrelial lymphocytoma is a rare skin manifestation characterized by a painless bluish-red nodule, which is predominantly reported in children.82 Typically, borrelial lymphocytoma is localized on the ear lobe, nipple, or scrotum. Often, there is a preceding or concomitant erythema migrans.5477 With antibiotics, borrelial lymphocytoma is usually cleared within several weeks.

Acrodermatitis chronica atrophicans

Acrodermatitis chronica atrophicans (ACA) is reported in 1-3% of Lyme borreliosis cases in Europe, and is predominantly caused by B afzelii.7783 ACA manifests as a chronic, slowly progressive red or bluish skin lesion, which eventually may become atrophic.5784 ACA is a late manifestation of Lyme borreliosis, and may present several months to years after an untreated erythema migrans.578485 It has been postulated that ACA does not resolve spontaneously, in contrast to most other manifestations of Lyme borreliosis.57 Indeed, even lesions present for 10 years may reveal active infection by culture or PCR positivity, and respond to antimicrobial therapy.8687 Concurrent peripheral neuropathy is common, and local joint involvement may occur.25

Nervous system manifestations

Dissemination of Borrelia spp primarily involves the skin, nervous system, joints, or, more rarely, the heart. Neurologic manifestations, henceforth called Lyme neuroborreliosis, are reported in ~10% of all cases of Lyme borreliosis.77 Early Lyme neuroborreliosis usually presents days to weeks after a tick bite, as a lymphocytic meningitis, cranial neuritis, or radiculoneuritis.3688 Cranial neuritis most commonly affects the facial nerve, which may be bilateral in up to 25% of individuals.8889 Involvement of other cranial nerves may occur, leading to diplopia, pain, hearing loss, or vertigo. Lyme neuroborreliosis is more frequent in children, commonly manifesting as facial nerve palsy,7790 in contrast to adults who typically present with radiculoneuritis and lymphocytic CSF pleocytosis.37 Lyme radiculitis may present with signs resembling disc herniation. Pain is neuropathic, and dermatomal in distribution, while sensory defects and paresis may occur.3637 Rarely, Borrelia spp may cause a wide variety of peripheral nerve disorders, characterized by a mononeuropathia multiplex.9192 Parenchymal brain involvement is extremely rare. Only sporadic cases of chronic encephalitis or encephalomyelitis owing to Lyme borreliosis have been reported.3793 Additionally, cerebrovascular events resulting from CNS vasculitis have been associated with Lyme borreliosis, based on brain biopsy or intrathecal synthesis of anti-Borrelia antibodies, responding to antibiotic therapy.94 Most of the individuals who presented with these conditions lived in an endemic area and had a history of erythema migrans, cranial neuritis, or radiculoneuritis.95

Lyme arthritis

Lyme arthritis typically presents as an oligo- or monoarthritis, often involving the knee joint, three to six months after infection.5881 In contrast to other causes of septic arthritis, Lyme arthritis usually is less painful and not accompanied by fever. Most patients do not report a preceding tick bite or erythema migrans. If untreated, symptoms include intermittent or persistent joint swelling and pain during a period of months to several years. With appropriate antimicrobial therapy, the arthritis is cured and Borrelia is eradicated in most patients, but in some cases proliferative synovitis may persist for months or several years, requiring anti-inflammatory therapy.5896

Lyme carditis

During early disseminated infection, acute cardiac involvement may occur, characterized by atrioventricular conduction defects in varying degrees.97 Less common cardiac manifestations include acute myopericarditis or, rarely, cardiomyopathy. Lyme carditis usually is self-limited, but is potentially fatal if untreated.98 Most case series on the relation between carditis and Lyme borreliosis lack causality; however, selected studies have identified B burgdorferi sl in endomyocardial biopsy samples from patients with dilated cardiomyopathy.99100

Differences between clinical manifestations in North America and Europe

The differences in Borrelia genospecies between the continents101 result in differences in clinical presentation.102 In North America, central clearing of erythema migrans is uncommon: up to 18% of erythema migrans cases are multiple, and Lyme borreliosis is more often associated with constitutional symptoms, such as fever and malaise.80103B burgdorferi ss in North America is more arthritogenic, and Lyme arthritis is more frequently encountered in North America (28% of Lyme borreliosis cases)104 than in Europe (3-7%).77105 Lyme neuroborreliosis in North America mostly presents as subacute meningitis with or without cranial neuropathy (usually facial palsy), and less frequently as painful radiculoneuritis.88 In Europe, B garinii is particularly neurotropic, and typically associated with painful radiculoneuritis and lymphocytic meningitis (originally described as Bannwarth syndrome).36B afzelii primarily causes skin infections, including ACA and borrelial lymphocytoma, both of which are virtually absent in North America.52101105106

Pregnancy and congenital infection

Adverse outcomes of infection on offspring during pregnancy have been known for other spirochetal diseases, such as syphilis.107 During spirochetemia in the acute phase of infection, B burgdorferi sl may spread transplacentally, and evidence for congenital infection has indeed been reported in a few cases where Borrelia species were cultured from the newborn post mortem.108109110111 A meta-analysis of nine studies suggested fewer adverse birth outcomes in women who were treated for gestational Lyme borreliosis compared with untreated cases, suggesting indirect evidence for adverse birth outcomes.112 Conversely, untreated ACA during pregnancy, an active chronic infection, was not associated with adverse outcomes in a retrospective survey,113 and in a systematic review, eight epidemiological studies reporting on potential associations between Borrelia sl exposure and adverse birth outcomes did not suggest any relation.112


Early localized/disseminated disease

For treatment of erythema migrans, doxycycline, amoxicillin, and oral cephalosporins were equally effective in randomized clinical trials, with complete response rates >90%.115 Azithromycin was as effective as doxycycline or amoxicillin in European open label trials,116117118119 but not in a randomized double blind controlled study in 246 patients in North America.120 As a result, doxycycline generally is regarded first choice therapy for erythema migrans.115 Randomized trials have shown that doxycycline for a duration of 10 days is as effective as 15 or 21 days.121122 Persistent symptoms after treatment were no more frequent in patients treated for ≤10 days as compared with longer courses in a retrospective cohort study with a mean follow-up duration of 2.9 years.123 For multiple erythema migrans, oral doxycycline for 14 days has been shown to be as effective as intravenous ceftriaxone in an open label alternate treatment trial among 200 patients.106

Lyme neuroborreliosis

Lyme neuroborreliosis is typically treated with intravenous ceftriaxone for at least 14 days.124 After meningoradiculitis, clinical recovery often is slow, and neurologic sequelae or subjective symptoms may persist in up to 40-50% of patients after 30 months.36125 While no studies have assessed the optimal duration of ceftriaxone therapy, the slow clinical resolution and long term neurologic sequelae have led some clinicians to extend the duration of treatment to up to 28 days.95 In a prospective, double blind study in Europe of 102 adults with early Lyme neuroborreliosis, oral doxycycline was found to be as effective as ceftriaxone.126 Long term outcomes (neurologic sequelae, quality of life, fatigue, cognition) were similar after either treatment.125 Based on this study, various treatment guidelines now consider doxycycline as a reasonable choice for Lyme neuroborreliosis.

Lyme encephalomyelitis, involving the brain parenchyma as evidenced by focal neurologic defects or magnetic resonance imaging findings, is extremely rare, and is typically treated with two to four weeks of intravenous ceftriaxone.127

Lyme arthritis

After treatment for Lyme arthritis with antibiotics, clinical resolution may take six to 12 months in some patients.96128 A 30 day course of either oral doxycycline or amoxicillin led to resolution of arthritis in ~90% of 38 patients in a randomized trial,128 while 10-14 days of intravenous ceftriaxone led to lower success rates (19/40; 48%) in a small randomized, placebo controlled, double blind trial.129 Patients with Lyme arthritis typically are treated with 30 days of doxycycline, while those who continue to have symptoms of arthritis after oral antibiotics subsequently are re-treated with either doxycycline or intravenous ceftriaxone.128130 Between 10% and 20% of patients may develop “antibiotic refractory” Lyme arthritis, a proliferative synovitis that no longer responds to antimicrobial therapy and requires therapy with intra-articular corticosteroids, non-steroidal anti-inflammatory drugs (NSAIDs), hydroxychloroquine, methotrexate, biologic response modifiers, or even synovectomy.130

Acrodermatitis chronica atrophicans

For ACA, observational studies have shown disappointing results of shorter term therapies—eg, intravenous ceftriaxone for two weeks or doxycycline for 20 days, while treatment success was 85% to 100% with doxycycline for up to four weeks. Hence, a treatment duration of four weeks is recommended.131132 Resolution may take many months after antibiotic therapy, while skin atrophy and neuropathy often are irreversible.132

Chronic symptoms attributed to Lyme borreliosis

Most people with Lyme borreliosis respond well to antimicrobial treatment. Despite antibiotic therapy, some patients with Lyme borreliosis develop disabling persistent symptoms, including fatigue, pain, and neurocognitive disturbances.2133 Their exact incidence, pathogenesis, and prognosis are not well known and are an ongoing source of debate.134 Longlasting signs and symptoms attributed to Lyme borreliosis often are referred to as “chronic Lyme.” It is essential to discriminate antibiotic therapy failure, which results in progressive infection or persistent signs at the primary sites of infection, from new subjective symptoms developing after resolution of the initial disease manifestations. Other patients seek medical attention for longlasting symptoms which are usually medically unexplained, questioning whether these may be attributable to an unnoticed episode of Lyme borreliosis, even when there is little or no evidence of previous B burgdorferi sl infection (fig 2).

Fig 2
Fig 2

Potential courses of disease after a tick bite. PTLDS=post-treatment Lyme disease syndrome; EM=erythema migrans; ACA=acrodermatitis chronica atrophicans. *Based on Strnad et al 20175; †based on Wilhelmsson et al 201611; and Hofhuis et al 201710

Persistent infection

In patients with ACA, diagnosis is often delayed, and duration of signs and symptoms for many years have been reported.8687 Even skin biopsy samples from untreated lesions present for ≥10 years reveal positive PCR and culture results,368687 indicating an active infection that can persist for years. After antibiotic treatment, culture results become negative, and skin lesions and accompanying signs may resolve completely, whereas local atrophy may persist in others.2557 Whereas a primate model has suggested that the persistence of bacteria after antibiotic therapy may drive other Lyme borreliosis manifestations,135136 ACA is the only manifestation of Lyme borreliosis in humans where chronic infection has been unequivocally demonstrated. Reported antimicrobial treatment failure rates, defined as development of disseminated Lyme borreliosis after treatment of early Lyme borreliosis, are low (≤1%).121123 In most studies where B burgdorferi sl was cultured from skin biopsy specimens of erythema migrans lesions before antibiotic therapy, post-treatment cultures yielded negative results.51122137 Recurrences of erythema migrans are not uncommon in endemic areas, and molecular typing showed that repeat episodes of erythema migrans in 17 appropriately treated patients were caused by re-infection and not relapse.138

Immune response

Ongoing infection has not been shown in patients who have been treated for Lyme arthritis. However, persistent or recurrent synovial inflammation is observed in up to one third of patients after first antibiotic treatment, and in 10-17% after repeated courses, the latter being referred to as antibiotic refractory Lyme arthritis.96128 In these cases, there is no clinical benefit from prolonged or recurrent antibiotic courses, but most patients do respond to immunosuppressive therapy.58 Hypotheses on the underlying mechanisms include persistence of non-viable spirochetal components or debris in the joint resulting in recurrent synovitis,61114 and potential immunological mechanisms, such as auto-antibodies, autoinflammatory processes, and dysregulated T cell responses.139140141 Likewise, association with HLA-DR alleles,142 upregulated expression of specific microRNA,143 and a Toll-like receptor 1 polymorphism144 have been suggested. Together with the observation that synovial PCR and culture results are negative or show non-viable spirochetes after antibiotic treatment, antibiotic refractory Lyme arthritis is likely based on immunological mechanisms, and persistence of B burgdorferi sl infection as a cause is highly unlikely.34130

Residual damage

In patients with Lyme neuroborreliosis, longlasting signs may be attributable to irreversible neurological injury caused by the infection. Persistent symptoms have been reported by 12-48% of 77 and 85 patients in two prospective follow-up studies.36145 Specific neurologic findings have been observed in 30% of patients included in a randomized treatment trial at four months,126 and 25-28% at two to five years in prospective and retrospective cohorts, including radiculopathy, paresis, hyposensibility, or hearing loss, without evidence for microbiological persistence.146147148 In a prospective case-control study including 50 patients with Lyme neuroborreliosis and matched controls, patients had a statistically significantly lower quality of life after 30 months, measured by the Medical Outcomes Study 36-Item Short-Form General Health Survey (SF-36) physical component summary.125 Delayed start of treatment, symptoms before treatment, and non-complete recovery after four months were possible predictors of a poorer quality of life and severe fatigue.146147 The association of unfavorable outcomes with longer duration or severity of symptoms before the start of treatment supports the concept of irreversible neurological damage as the underlying mechanism.36146148 Similar long term neurological sequelae after microbial eradication have been observed in bacterial meningitis caused by other bacteria.149

Functional disability and neurocognitive symptoms

The reported prevalence in observational studies of persistent symptoms, such as musculoskeletal pain, fatigue, and cognitive complaints varies considerably, between 0% and 48%.145150151152 Case definition, Lyme borreliosis manifestation, follow-up, geographic location, and use of self-reported symptoms rather than validated measures might explain the divergence in prevalence. In a prospective study, 26/71 (36%) patients had self-reported ongoing symptoms at six months after treatment of erythema migrans.151 In 11%, subjective symptoms were associated with functional disability as measured with the SF-36 questionnaire, while none had microbiological or clinical evidence for ongoing infection. Another prospective study among 128 US patients with culture confirmed erythema migrans followed for >10 years found a 10.9% incidence of self-reported ongoing symptoms, predominantly memory or concentration difficulties, fatigue, and joint pain.152 In most patients, one or more symptoms persisted for >10 years.

The appreciation of long term outcomes of Lyme borreliosis is hampered by the lack of proper control groups in most studies. In a controlled prospective study from Europe in 285 patients with erythema migrans, the prevalence of symptoms was highest at baseline (33.3%), decreasing to 4.6% at six months and 2.2% at 12 months. In 259 matched controls without Lyme borreliosis, these rates were similar (3.0% at 12 months). Microbiological failure requiring retreatment was documented in two cases only.150 This particular study did not include patients with multiple erythema migrans or manifestations such as Lyme neuroborreliosis, who might have a greater likelihood of developing persistent symptoms. Also, B burgdorferi ss infected patients in the US are more likely to be symptomatic at baseline, and therefore might have a greater likelihood of developing persistent symptoms.152 While many patients report cognitive problems such as memory loss, word-finding difficulties, and lack of concentration, subjective memory complaints were not associated with impaired objective test performances in a prospective study on 279 patients with persistent symptoms attributed to Lyme borreliosis.153 Only 3% of patients included in that study were classified as having clinically impaired cognitive performance compared with normative data.154

It has been hypothesized that chronic pain and fatigue syndromes may be part of a central sensitization syndrome that follows non-infectious or infectious diseases.155 Central sensitization is thought to involve activation of central neurons, leading to synaptic and neurotransmitter changes, and an increased sensitization to pain signals that may last for months or years.155 After Lyme borreliosis, such a mechanism might be elicited by past infection, persistence of remnant bacterial proinflammatory triggers,61156 or other physiological or behavioral mechanisms.155

Post-treatment Lyme disease syndrome

The complex of persistent or recurrent symptoms despite antibiotic treatment is often referred to as post-treatment Lyme disease syndrome (PTLDS), for which a case definition was proposed by the Infectious Diseases Society of America (IDSA) (box 2).157 This definition includes a clinical syndrome of patient reported symptoms in the context of a previously treated, physician diagnosed case of Lyme borreliosis meeting the CDC surveillance criteria.151157 Additional criteria exclude patients with untreated Lyme borreliosis or other tickborne infections, and those with objective persistent signs of the initial episode of Lyme borreliosis, such as recurrent arthritis, ACA, or neurologic sequelae of Lyme meningoradiculitis. Others have proposed to better define the “functional impact” component of the case definition, using an SF-36 questionnaire threshold.151 By definition, patients with PTLDS have no compelling clinical or laboratory support for the diagnosis of ongoing B burgdorferi sl infection, and neither have they signs suggesting immunological phenomena, such as recurrent synovitis, or irreversible nerve damage after Lyme neuroborreliosis. Several hypotheses on the causes of PTLDS exist. First, microbiological mechanisms have been considered, including tickborne co-infections. Whereas ticks that transmit B burgdorferi sl are known vectors of other human pathogens, none of these pathogens are known to cause chronic symptoms.159 Likewise, round morphologic forms of B burgdorferi sl have been suggested to cause persistent disease. While morphologic variants have been reported in human tissue specimens in a small number of cases, none of the patients had symptoms resembling those of PTLDS, and conversely, morphologic variants have never been identified in patients with chronic Lyme attributed symptoms.160 Second, immunogenetic mechanisms have been proposed, but prospective observational studies on dysregulated immune responses161162163 and case-control studies on autoimmune processes164165 have been inconclusive. Third, associations of PTDLS have been described with demographic, clinical, and epidemiological patient characteristics, such as advanced age, female sex,166167 comorbidity, and duration of pre-treatment symptoms.168169 Finally, cognitive behavioral characteristics, including depression, anxiety, negative affect, and catastrophizing have been associated with the risk of developing persistent symptoms.170

Box 2

Proposed definitions of post-treatment Lyme disease syndrome (PTLDS)

IDSA guidelines criteria157

Inclusion criteria

  1. Documented episode of early or late Lyme disease fulfilling the case definition of the CDC.158 If based on erythema migrans, the diagnosis must be made and documented by an experienced healthcare practitioner

  2. Resolution or stabilization of the objective manifestation(s) of Lyme disease after treatment of the episode of Lyme disease with a generally accepted treatment regimen

  3. Onset of any of the following subjective symptoms within six months of the diagnosis of Lyme disease and persistence for at least a six month period after completion of antibiotic therapy:

    • fatigue

    • widespread musculoskeletal pain

    • complaints of cognitive difficulties

  4. Subjective symptoms of such severity that they result in substantial reduction in previous levels of occupational, educational, social, or personal activities

Exclusion criteria

  1. Active Borrelia burgdorferi infection diagnosed by a reliable method based on either culture or PCR

  2. Active, untreated, well documented co-infection, such as babesiosis

  3. Presence of objective abnormalities on physical examination or on neuropsychologic testing that may explain the patient’s complaints. For example, a patient with antibiotic refractory Lyme arthritis or with late neuroborreliosis associated with objective cognitive dysfunction would be excluded

  4. Diagnosis of fibromyalgia or chronic fatigue syndrome or a prolonged history of undiagnosed or unexplained somatic complaints, such as musculoskeletal pains or fatigue, before the onset of Lyme disease

  5. Underlying disease or condition that might explain the patient’s symptoms, or laboratory or imaging abnormalities that might suggest an undiagnosed process distinct from post-Lyme disease syndrome


Trials of prolonged antimicrobial therapy

Empirical antibiotic treatment studies have targeted the possibility of concealed infection in patients with persistent symptoms, despite the weight of evidence against persistent infection as the explanation for PTLDS. Initial open cohort studies have claimed successful antimicrobial therapy in patients with “chronic Lyme.”171172 Oral tetracycline for a median of four months was reportedly associated with a 90% success rate in a case series of 277 patients. Likewise, the combination of clarithromycin and hydroxychloroquine reportedly was as effective as prolonged tetracycline in another series of 235 cases. Most patients improved within two weeks after initiation of therapy, and all patients had improved after three months.172 However, in both studies, inclusion criteria were not clearly defined, and serologic reactivity against B burgdorferi sl, but no documented Lyme borreliosis, was required. The studies were non-randomized and uncontrolled, and did not use standardized questionnaire outcomes.

Five randomized controlled clinical trials have been performed in patients with persistent symptoms attributed to Lyme borreliosis (table 2).173174175176177 One trial did not find any beneficial effects on quality of life in 115 patients randomized to prolonged therapy compared with the matching placebo group.173 While the study found improvement in self-reported cognitive functioning in both randomization arms, no improvement in objective neurocognitive functioning was found.173182

Table 2

Overview of randomized controlled treatment trials in patients with symptoms attributed to previously documented Lyme borreliosis

View this table:

In a small study, 37 patients with memory impairment were randomized to ceftriaxone versus placebo.175 At 12 weeks, the ceftriaxone group showed a statistically significantly greater effect on objective neurocognitive functioning, but this was not sustained to week 24, whereas the effect on self-reported fatigue and physical functioning was only sustained among a subgroup more severely affected at baseline.175 How this short term cognitive improvement relates to other patients with PTLDS is uncertain, as participants were selected from a cohort of >3000 patients, and were required to have objectified memory impairment on neuropsychological testing, which is rare among patients with PTLDS.153182

Another trial randomized 55 patients with severe fatigue to four weeks of ceftriaxone versus placebo.174 The primary endpoint failed to show improvement in neuropsychological performance, but self-reported fatigue improved in the group taking ceftriaxone. Despite the statistically significant effect reported, the authors themselves conclude that the study does not support the use of additional antibiotic therapy, because fatigue, a non-specific symptom, was the only outcome that improved.174

A trial that randomized 86 patients with “a recurrence of Lyme borreliosis symptoms” to oral amoxicillin for three months or placebo was hampered by several shortcomings. It was prematurely terminated because of slow recruitment, the publication did not provide details on the intention-to-treat population, and a large proportion of patients were excluded from the analysis “because of persistent symptoms.”

In another trial, the PLEASE trial, 281 patients were randomized to receive 14 days of ceftriaxone, followed by 12 weeks of either doxycycline, clarithromycin plus hydroxychloroquine, or placebo.177 Neither regimen of 12 weeks of therapy yielded benefit over placebo with respect to serial mental and physical health related quality of life measures during follow-up until 52 months. This study has built upon lessons learnt from earlier studies. Choice and duration of the treatment regimens were based on uncontrolled studies which reported that almost 75% of patients improved within one month and 92% to 100% within three months of treatment, which suggested that a three month regimen should be optimal to assess whether those observations are sustained when placebo controlled.171172 Whereas earlier randomized trials might have been influenced by baseline differences, the PLEASE trial analysis was corrected for baseline health related quality of life. Finally, the endpoint was based on a minimal clinically relevant treatment effect on the SF-36 summary score specific for patients with PTLDS, as prospectively assessed in a pilot study.183 Three aspects of the study design have been subject of debate. First, patients received two weeks of open label antibiotics preceding the randomized treatment phase, to standardize and synchronize pretreatment, while previous trials allowed for a wide variance of antibiotic pretreatments. Consequently, the study was designed as a randomized, placebo controlled trial to compare longer term with standardized shorter term therapy. Sensitivity analyses showed that excluding patients without prior oral pretreatment did not affect the outcomes.177 Second, all three study groups slightly improved during the 52 weeks of follow-up, irrespective of randomization arm, and some have attributed this to the standardized pretreatment with ceftriaxone, rather than placebo effects or regression to the mean. However, regression of symptoms, including fatigue severity, has consistently been reported in the placebo arm of other studies, and was of similar magnitude (table 2).174175177 This is in agreement with the finding that positive pretreatment expectancies and higher self-efficacy were the major predictors of outcome, regardless of randomization arm.184 Third, it has been argued by some that 14 weeks of treatment was insufficient to show a beneficial effect on alleviation of symptoms, in contrast to the earlier uncontrolled studies.171172 While prolonged antimicrobial treatment is not uncommon for various infectious diseases, such as tuberculosis, in the case of TB it is aimed at preventing microbiological relapse, and not at a delayed onset of clinical alleviation, as no infectious diseases have been described in which the initial effect on signs, symptoms, and laboratory findings is delayed beyond the first three months of effective therapy.

In summary, five randomized clinical trials have provided little support for prolonged antibiotic treatment in patients with persistent symptoms attributed to Lyme borreliosis. While smaller studies reported limited, or non-sustained effects on selected outcomes,174175 the two largest trials did not find beneficial effects of prolonged treatment on any of the domains studied (table 2).154173177182

Management of patients with persistent Lyme attributed symptoms

Attributing a cause to medically unexplained symptoms is challenging, and the lack of abnormal test results or objective findings can be frustrating for both patients and physicians. A stepwise approach can help assess whether persistent symptoms attributed to Lyme borreliosis are indeed related to previous B burgdorferi sl infection, and may be caused by active infection (thus, potentially amenable to antimicrobial treatment), or caused by immune mediated mechanisms or residual damage (fig 3, box 3). A temporal relation of persistent constitutional symptoms with primary Lyme borreliosis may suggest PTLDS (box 2). For a large group of patients who have little or no evidence of previous B burgdorferi sl infection, who understandably seek explanation for chronic fatigue, pain, and other incapacitating symptoms, Lyme borreliosis has become a common consideration. Anchoring bias to ascribe symptoms to Lyme borreliosis or even tick bites is a potential pitfall. Alternative explanations, such as chronic fatigue syndrome or fibromyalgia, may be considered but may be less acceptable for some patients. As described, the weight of evidence is strongly against persistent infection as the explanation for persistent symptoms in such patients, and treatment trials have consistently provided evidence against prolonged antibiotic treatment.

Fig 3
Fig 3

Presentation of persisting symptoms attributed to Lyme borreliosis. PTLDS=post-treatment Lyme disease syndrome; EM=erythema migrans; ACA=acrodermatitis chronica atrophicans

Box 3

Stepwise approach to management of patients with persisting symptoms attributed to Lyme borreliosis

  • Goal: to assess whether persisting symptoms are related to previous B burgdorferi infection, and may be due to active infection and, thus, potentially amenable to antimicrobial treatment

  • A careful medical history should show whether there may have been a B burgdorferi infection and clinical signs of localized or disseminated Lyme borreliosis

  • Physical examination should focus on persistent signs at the primary sites of infection or new, localized signs of disseminated infection that may indicate persistent infection. If persisting infection is deemed unlikely, immune mediated arthritis or neuroborreliosis induced residual damage should be ruled out

  • A temporal relation of subjective symptoms and primary Lyme borreliosis may suggest PTLDS

  • Consider other underlying diseases or conditions that may explain the patient’s symptoms, or laboratory or imaging abnormalities that might suggest an undiagnosed process

  • In patients with constitutional symptoms and little or no evidence of previous B burgdorferi infection, the evidence is against persistent B burgdorferi infection as the explanation for persistent symptoms, and treatment trials have provided evidence against prolonged antibiotic therapy

  • Listen to the patient’s concerns and acknowledge their suffering, and discuss the pitfalls and misconceptions surrounding the diagnosis and management of Lyme borreliosis


Clinicians caring for patients with poorly understood syndromes of fatigue, pain, and impaired mental acuity often have the difficult task of explaining that there is no straightforward diagnosis or treatment, such as an active infection to be cured with antibiotics. Understandably, this often appears to be removing hope for seriously ill patients, who are indisputably suffering, regardless of the cause of their symptoms. Good medical care for this group includes listening to patients, understanding their concerns and suffering, and discussing the pitfalls and misconceptions surrounding the diagnosis and management of Lyme borreliosis, while remaining aware of potential alternative diagnoses. In close consultation with the patient, the optimal management strategy should be determined, without reverting to unsubstantiated, irrational, or even potentially harmful therapies.


Other microorganisms present in ticks

Numerous bacteria, parasites, and viruses have been detected in Ixodes ticks178185 and Ixodes ticks are well known vectors for other human diseases, including anaplasmosis,186 babesiosis,187Borrelia miyamotoi disease,188 ehrlichiosis,186 rickettsioses, and disease caused by tickborne flaviviruses.189190 In the US, the number of reported cases of these Ixodes tickborne diseases has increased over time.191192 Multiple new tickborne pathogens have been described in Ixodes species in Europe, the US, and Asia; for example, Ehrlichia muris eauclairensis,193Neoehrlichia mikurensis,194195Borrelia miyamotoi,196Borrelia mayonii,3 and Alongshan virus.197 Notably, the mere presence of a microbe—let alone its DNA—does not render it a tickborne pathogen. Indeed, ticks have many commensal bacteria and endosymbionts,198 and can acquire bacteria, parasites, and viruses from an infected host through a blood meal, which they are unable to transmit, or which, if transmitted, do not lead to disease.185

Lyme borreliosis: evidence for role of co-infections

Co-infection of Ixodes ticks with multiple tickborne pathogens is well established,178185 although even in regions where multiple tickborne diseases are endemic, human co-infections appear to be relatively rare.199200 In the US, B burgdorferi ss-Anaplasma phagocytophilum and B burgdorferi ss-Babesia microti co-infections are relatively common.178 Experimental evidence suggests that Anaplasma phagocytophilum may alter the course of acute Lyme borreliosis201 but clinical data are inconclusive.157 In contrast, human co-infection with Babesia microti can increase the duration and severity of symptoms caused by acute Lyme borreliosis.202 Of note, patients with acute Lyme borreliosis accompanied by persisting high grade fever in the US, any fever in Europe, or abnormal blood counts (anemia, thrombocytopenia, or leukopenia) should raise a suspicion of co-infection.

In contrast, there is little evidence for the notion that “chronic Lyme” can be attributed to Lyme borreliosis and a wide range of co-infections. These include the pathogens mentioned above, but also other microorganisms, such as Chlamydia, Brucella, or Mycoplasma species, Toxoplasma gondii, Epstein-Barr virus, cytomegalovirus, or human herpes virus-6. However, these patients frequently lack clinical symptoms compatible with such infections.203 A systematic review of patients diagnosed with “chronic Lyme” did not find evidence for chronic anaplasmosis or babesiosis in humans, tick transmission of Bartonella species, or B burgdorferi sl-Bartonella co-infections.159 In Europe, a large prospective clinical study is currently ongoing, assessing the role of several known Ixodes Ricinus-borne pathogens in the development of longlasting symptoms.134

Post-exposure prophylaxis

A meta-analysis of four placebo controlled clinical trials (totaling 1082 patients) in the US showed a risk of developing Lyme borreliosis of 2.2% (95% confidence interval 1.2% to 3.9%) in the placebo group, compared with 0.2% (95% confidence interval 0.0% to 1.0%) in the prophylaxis group,13 indicating that prophylaxis to ~50 individuals prevents one case of Lyme borreliosis.204 A trial with topical azithromycin was stopped prematurely because of a lack of effect.205 In current guidelines from the US and Europe, watchful waiting is primarily recommended, while a single dose of doxycycline (200 mg) within 72 hours after a tick bite204 may be offered in highly endemic settings, when the tick has been attached for longer periods.157206207


The OspA vaccine

In the late 1990s, two vaccines, LYMErix208 and ImuLyme,209 were assessed in large phase III clinical, double blind, randomized, placebo controlled trials. Both vaccines were based on recombinant OspA of B burgdorferi ss. The vaccines were well tolerated, efficacy ranged from 76% to 92% after three immunizations, and they were shown to be cost effective.208209210211 LYMErix was commercially launched in the US in 1998, and in 2002, manufacturer GSK voluntarily withdrew the vaccine, citing poor sales on lack of demand.212 However, the reasons were multifactorial and extensively discussed previously,213214 cumulating in class action lawsuits and final withdrawal of the vaccine. Several OspA based veterinary vaccines are still available,215 but a commercial vaccine to prevent Lyme borreliosis in humans does not exist.

Modified OspA vaccines

As multiple B burgdorferi sl genospecies can cause Lyme borreliosis, second generation OspA vaccines targeting multiple B burgdorferi sl serotypes are being developed. Baxter BioScience has developed a chimeric recombinant vaccine that contained six OspA serotypes,216 lacking the alleged “auto-reactive” B burgdorferi ss epitope. Phase I/II vaccine trials have shown safety and immunogenicity in naive and previously B burgdorferi sl exposed individuals.217218 Another novel multivalent OspA vaccine, VLA15, consists of three heterodimers linking C-termini of two OspA serotypes and covering six clinically relevant B burgdorferi sl serotypes,219 while the alleged “auto-reactive” epitope was replaced with the corresponding sequence of B afzelii. A phase I trial indicated seroconversion of 71.4% to 96.4% for multiple OspA serotypes after three doses of VLA15, which were well tolerated.220 VLA15 is currently assessed in phase II clinical trials with higher dosages and alternative vaccination schedules (NCT03769194/NCT03970733).

New vaccination horizons

Other spirochetal recombinant protein based vaccines, including a vaccine that targets a B burgdorferi sl protein of unknown function, BB0405,221 and novel delivery strategies are under development.222223 Alternatively, vaccination against tick proteins is considered,222 based on the phenomenon known as “tick immunity”: guinea pigs, rabbits, and possibly humans, develop immune responses against tick proteins after repeated tick infestation, resulting in impaired tick feeding and protection against B burgdorferi ss infection.223 Regardless of the approach, a future human vaccine would need to be sufficiently safe, efficacious, and cost effective, to achieve acceptance from the medical community and public and to prevent a repetition of the past.223224


Guidelines on management of Lyme borreliosis are available in the US 157225226 and Europe.66206207227228 For the antibiotic treatment of Lyme borreliosis, the recommended agents, doses, and durations are highly consistent through different guidelines, and are predominantly based on studies described in table 3. Alternative recommendations, provided in a position paper by the International Lyme and Associated Diseases Society (ILADS), have not provided any credible clinical or scientific evidence to support prolonged antibiotic therapy. Their designation as “evidence based guidelines” belies their anecdotal nature and lack of coherent and evidence based guidance.226

Table 3

Overview of treatment guidelines for Lyme borreliosis*

View this table:


The manifestations of Lyme borreliosis are diverse, the diagnosis is not always straightforward, diagnostic tests may have limitations, and their results should be interpreted in the context of the clinical symptoms. However, the disease generally responds well to antibiotic treatment. Despite antibiotic therapy, patients may develop disabling persistent symptoms, sometimes referred to as “chronic Lyme.” For proper patient management, it is of critical importance to discriminate antibiotic therapy failure from immune driven post-infectious phenomena such as relapsing Lyme arthritis, residual tissue damage, such as post-neuroborreliosis neuropathy, or new subjective symptoms developing after resolution of the initial disease manifestations. The latter may be classified as PTLDS, of which the pathogenesis has not fully been elucidated.134 Finally, “chronic Lyme” remains a popular consideration for patients, often with little or no evidence of previous B burgdorferi infection, who understandably seek explanation for fatigue, pain, and other incapacitating symptoms. These patients, who are indisputably suffering regardless of the cause of their symptoms, deserve a thorough analysis whether there may have been an undiagnosed B burgdorferi infection, or other post-infectious sequelae. Often, the lack of objective findings and abnormal test results are frustrating for physicians as well as for patients. Good medical care for these patients includes listening, understanding, and discussing personalized treatment options, including non-pharmacological options such as rehabilitation. Future research may reveal additional underlying causal mechanisms, indicating how to best diagnose and treat these patients.

Research questions

  • What are the main factors responsible for the rise in the abundance of Ixodes ticks that vector the causative agents of Lyme borreliosis; can we find effective and interdisciplinary countermeasures to halt this upsurge?

  • Can we discover and develop novel diagnostics markers or tests with increased sensitivity for localized disease and early disseminated Lyme borreliosis, and tests that can differentiate between an active B burgdorferi sl infection and past infection, in particular for patients with unexplained signs and symptoms after completing antibiotic treatment?

  • What are effective strategies for the prevention of antibiotic refractory Lyme arthritis, and can we develop evidence based guidelines for its treatment?

  • What are the underlying causal mechanisms in patients with longlasting symptoms after antibiotic treatment for Lyme borreliosis? Can we identify microbiological determinants, co-infections, immunological or genetic mechanisms, epidemiological determinants, or cognitive behavioral factors that are associated with developing persistent post-treatment symptoms?

  • What is the long term outcome in patients with PTLDS? Which factors may influence their prognosis, and can we define optimal treatment strategies for these patients?

  • Can we find a safe, efficacious, and cost effective vaccine that is able to protect both children and adults against Lyme borreliosis in Europe and in North America? What is needed for the general public to accept such a vaccine?

Patient involvement

No patients were involved in the creation of this article.


  • Acknowledgments: HV and FvdS are supported by the Netherlands Organization for Health Research and Development ZonMw; JWH is supported by the European Union INTERREG program, as part of the NorthTick project.

  • Contributors: HDV and FvdS performed the primary literature review for this manuscript. BJK wrote the sections on therapy, trials of prolonged therapy, and management of patients with persistent Lyme attributed symptoms. JWH wrote the sections on incidence and epidemiology, co-infections, prevention, and vaccination. FvdS wrote the section on clinical manifestations. FvdS and JWH wrote the section on laboratory support. HDV wrote the sections on chronic symptoms attributed to Lyme borreliosis and guidelines. BJK guided the writing of the full manuscript and assumed primary responsibility. All authors reviewed all sections of the manuscript, providing suggestions for included content and references, and approved the published version.

  • Competing interests We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.

  • Further details of The BMJ policy on financial interests are here:

  • Provenance and peer review: commissioned; externally peer reviewed.


View Abstract