Lyme arthritis was first described in 1977 by Steere and colleagues in a cluster of children thought to have juvenile rheumatoid arthritis. They lived in and around Old Lyme, Connecticut. Subsequent studies documented that the disease was caused by the spirochete Borrelia burgdorferi and that arthritis was only one of many possible manifestations of this infection, now known as Lyme borreliosis or Lyme disease .
Clinical case descriptions of various manifestations of this disease date back more than a century. Acrodermatitis chronica atrophicans was described in Germany in 1883. Erythema migrans, the early skin manifestation of Lyme borreliosis, was reported in Sweden in 1909. The first case of neuroborreliosis and its association with a tick bite was observed in France in 1922. Among the cases of lymphocytic meningitis and inflammatory polyneuritis studied by Bannwarth in Germany in 1941, several patients described had “rheumatism,” probably the first report of what is now called Lyme arthritis . Successful treatment with penicillin was described in 1946. Erythema migrans was transferred by skin biopsy to healthy human volunteers in 1955. These observations suggested an infectious cause.
Definition and Classification
Lyme disease is a complex disease with cutaneous, articular, neurological, and other manifestations that result from infection with the spirochete B. burgdorferi transmitted by the bite of a tick of the genus Ixodes . Various components of the disease (e.g., erythema migrans, arthritis, neuroborreliosis) may occur in isolation. The term Lyme borreliosis is often used in Europe; Lyme disease is the most frequent term used in North America. Lyme arthritis is usually responsive to antibiotics. The term antibiotic-refractory Lyme arthritis is used for patients who do not show clearance of Lyme arthritis following two courses of appropriate antibiotics.
Lyme disease has been documented only in the temperate zones of the northern hemisphere. In North America, the disease is recognized most commonly in the northeastern, mid-Atlantic and north-central United States; it occurs less frequently on the West Coast and in southern Canada. Lyme borreliosis is rare or absent in the other parts of the United States and Canada. In Europe, the disease is most common in central Europe but occurs endemically from southern Sweden to the northern Mediterranean and from Portugal to Russia. Although sporadic cases of Lyme disease have been reported in eastern Russia, China, Korea, and Japan, it appears to be much less common in Asia than in the endemic areas of North America or Europe. A recent review explores the drivers and mechanisms for changing geographic ranges of ticks and tick-borne pathogens.
Incidence and Prevalence
The Centers for Disease Control and Prevention (CDC) have reported a rapid increase in the frequency of Lyme disease in the United States since 1982. Between 1992 and 2009, the incidence tripled, and a maximum of 29,959 confirmed cases were reported in 2009. Case numbers were slightly smaller for the years 2010 to 2012, the last year for which a detailed analysis has been published. Recently, the CDC estimated that due to underreporting, the true U.S. incidence might be as high as 300,000 cases annually. However, this figure may be inflated because it was based, in part, on patient self-reporting of a Lyme diagnosis and on laboratory diagnosis, so false-positive results may be included. The highest statewide incidence of reported confirmed cases was 111.2 cases per 100,000 persons of the general population in Delaware in 2009. Throughout this period, the highest local incidence shifted from the island of Nantucket, Massachusetts, to Columbia County, New York, with a peak incidence of 962 per 100,000 during the years 2002 to 2006. There has also been a geographical spread of cases both along the Northeastern seaboard as well as in the Midwestern states.
Data from the Slovenian National Registry document a maximum annual incidence of Lyme disease of 309 cases per 100,000 persons in 2009 and subsequent declines to 244 and 274 for 2010 and 2011, respectively (F. Strle, personal communication, 2013). A study in southern Sweden reported an incidence of 69 cases per 100,000; Lyme arthritis was present in 7% of all cases. In a population-based study in Würzburg, Germany, the incidence was 111 per 100,000, with higher rates among children younger than 16 years old.
In a community-based Connecticut cohort study of 201 consecutive cases in children in whom Lyme disease had been newly diagnosed, 13 (6%) had arthritis, and 5% had facial palsy. In Europe, Lyme arthritis and neuroborreliosis have been reported in similar frequencies, and in the Würzburg study, arthritis was more common. Compared with adults, children more frequently had manifestations other than isolated erythema migrans. Early onset of cutaneous disease and neural involvement are closely related to tick activity in the spring to autumn; there is no seasonal pattern for late manifestations, such as Lyme arthritis, because arthritis may occur a year or longer after tick exposure.
Sex Ratio and Age at Onset
Both sexes are affected equally. Cases have been reported among all age groups, with peaks occurring in school-age children and people between 40 and 74 years old.
Although Lyme disease may affect several members of the same family, genetic factors appear to have a limited influence on its occurrence. Nonetheless, host factors influence the course of the disease. In American patients, the development of chronic Lyme arthritis and an antibiotic-refractory course have been associated with the presence of human leukocyte antigen (HLA)-DR4. HLA-DR2 is an additional risk factor, especially in patients who are HLA-DR4 negative. These results have not been confirmed in European patients. A DR4-positive mouse model has been established.
Etiology and Pathogenesis
Lyme disease is the most common vector-borne infection in North America and Europe and is transmitted by hard-bodied ticks of the genus Ixodes ( Fig. 42-1 ). Transmission by other ticks or flying hematophagous insects has been suggested but has not been proven. Ticks of the genus Ixodes include Ixodes ricinus in central Europe, Ixodes persulcatus in Eastern Europe and Asia, Ixodes scapularis in the northeastern and north-central United States and Ontario, Canada, and Ixodes pacificus in the western United States.
To become active, ticks require a warm and humid environment and are affected by climatic variability. Infection is acquired in tick habitats, including forests, shaded valleys, gardens, lawns, and inner-city parks. After gaining access to unprotected skin, ticks crawl to the preferred feeding locations in the popliteal region, thighs, groins, breasts, axillae, neck, or head. Ixodes ticks feed only once during each of the three stages of their life cycle. Most human infections occur after the painless bite of nymphs ( Fig. 42-2 ). Ixodes ticks also transmit tick-borne encephalitis virus, Ehrlichia, Anaplasma phagocytophila, and Babesia organisms. Coinfection of these organisms with B. burgdorferi has been reported in rare cases. Infection with Anaplasma results in higher fever and more severe illness than in patients with early Lyme disease. Recently, it was shown in Russia and in the United States that Ixodes ticks also transmit some relapsing fever Borreliae such as B. miyamotoi; clinical cases have been described in Russia, the United States, and Central Europe.
Lyme disease is caused by infection with one of several species of B. burgdorferi sensu lato. These spirochetes have a protoplasmic cylinder surrounded by a cell membrane, a periplasmic flagellum, and an outer membrane. They are microaerophilic and grow best at 33°C in a special liquid medium. They grow slowly, with doubling times between 12 and 24 hours. B. burgdorferi sensu lato has been subdivided into several species, of which only B. burgdorferi sensu stricto has been found to cause human disease in North America; Borrelia garinii and Borrelia afzelii also have been identified regularly in patients in Europe. Two newer species, Borrelia spielmanii and B. bavariensis, have been isolated in Europe and associated with cases of erythema migrans. In rare instances, these species may also cause Lyme arthritis. Borrelia lusitaniae was isolated from a 13-year-old girl with a vasculitis-like syndrome in Portugal. The pathological potential of this species remains to be defined. In general, diversity in B. burgdorferi organisms has been greater in Europe and Asia than in North America. Concurrent infection with more than one species of B. burgdorferi was described in a patient with acrodermatitis and erythema migrans, as was culture-confirmed reinfection in patients with several episodes of erythema migrans. In all patients who had another erythema migrans following antibiotic treatment for the first episode, reinfection with a genetically different B. burgdorferi strain, and not relapse of disease by the strain causing the first episode, could be demonstrated. These results demonstrate no evidence for persistent infection as a cause of human Lyme borreliosis after appropriate antibiotic treatment.
B. burgdorferi species differ genomically. Even within a species, different strains express proteins of different molecular weights as identified on gel electrophoresis. The major proteins identified in sonicates of B. burgdorferi are the 41-kD flagellar antigen; the 60-kD GroEL heat-shock protein; the three major outer surface proteins (Osp) OspA (30 to 32 kD), OspB (34 to 36 kD), and OspC (21 to 25 kD); the vlsE lipoprotein, the 39-kD BmpA protein; and the 83- to 100-kD antigen. Two membrane glycolipids of B. burgdorferi have recently been shown to lead to strong IgG responses in patients with Lyme arthritis. The linear chromosome and plasmids of B. burgdorferi sensu stricto strain B31 as well as of 2 B. afzelii and 2 B. garinii strains and one strain each of B. spielmanii, B. valaisiana, and B. bissettii have been sequenced.
The natural reservoirs of B. burgdorferi are mice and voles, although hedgehogs and birds may also serve this function. The life cycle of Ixodes ticks lasts 2 years (see Fig. 42-2 ). The eggs hatch and larvae develop in the spring of the first year. The larvae feed once that summer on their preferred host (i.e., mice and voles) and so become infected with B. burgdorferi . The next spring, the larvae molt into nymphs, which feed on the preferred host before becoming mature ticks, at which time larger animals (e.g., deer) act as hosts. Humans are accidental hosts. Mating occurs while the female tick feeds. The female then detaches and lays her eggs on the ground.
Lyme arthritis provides a fascinating model for other arthritides because the causative organism and the clinical picture are well known. However, knowledge of the pathogenesis of this disease remains fragmented. B. burgdorferi excreted through tick salivary glands spread locally in the skin and can frequently be found at the advancing edge of erythema migrans. They attach to human cells by binding to various integrins, such as the fibronectin and vitronectin receptors. Binding of the organism to platelets may play a role in its hematogenous spread. Some of the molecular mechanisms involved in vascular interactions of B. burgdorferi have been described in a living mouse model. B. burgdorferi organisms are presumed to reach the synovium through the bloodstream. It is probable that their presence in synovium is required at the onset of arthritis.
Survival of B. burgdorferi for decades in the lesions of acrodermatitis chronica atrophicans indicates that the spirochetes are able to evade the host immune response. There is also evidence that B. burgdorferi may survive intracellularly in endothelial cells, fibroblasts, and synovial cells. B. burgdorferi use the mechanism of sequential variation of their outer surface proteins to attempt to evade the host immune response. This contributes to the survival of the organisms in spite of a strong antibody response. Arthritis appears to be largely due to the host inflammatory response. Neutrophil-activating protein A is elevated in synovial fluid from patients with Lyme arthritis recruiting inflammatory cells into the joint cavity. B. burgdorferi have stimulatory effects on B cells, and a dominant T-helper cell 1 (Th1) response has been found in synovial fluid of patients with Lyme arthritis. A Toll-like receptor 1 polymorphism has been associated with heightened Th1 inflammatory responses and antibiotic-refractory Lyme arthritis. Higher numbers of regulatory T cells were implicated in contributing to antibiotic-responsive Lyme arthritis. Recent work from the same group found higher expression of activation coreceptors and less effective inhibition of proinflammatory cytokines as well as more activated natural killer cells in patients with antibiotic-refractory Lyme arthritis. A B. burgdorferi –specific CD8 + cytotoxic T-cell response has also been reported in patients with Lyme arthritis. These cells were found only after the disappearance of arthritis. B. burgdorferi also stimulate synovial γ/δ T cells from patients with Lyme arthritis, leading to high and prolonged expression of Fas ligand associated with cytolytic activity. A number of cytokines are induced, including interleukin (IL)-1 and IL-6, tumor necrosis factor, CXCR3, CCR5, CXCL9, and interferon (IFN)-γ, and IL-10. The chemokine CXCL13 appears to be a key regulator of B cell recruitment in acute neuroborreliosis.
Molecular mimicry may also play a role in the pathogenesis of some of the manifestations of Lyme disease. Sequence homologies have been identified between B. burgdorferi flagellin and human myelin basic protein, as well as cross-reactivity between flagellin and a human axonal protein. Antibody reactivity to OspA and OspB occurred late in the course of infection in American patients with chronic Lyme arthritis. Tick-specific borrelial antigens such as OspA and OspD were upregulated in American but not European patients with Lyme arthritis. Th cells from patients with antibiotic-refractory Lyme arthritis demonstrated dominant recognition of an OspA peptide of B. burgdorferi , and high levels of CXCL9 and IFN-γ were found in synovial fluid and tissue. Human homologues of the borrelial T-cell epitope were found, but reactivity with the self-peptides was lower, implying that molecular mimicry is unlikely to be the critical mechanism. The human leukocyte function-associated antigen-1 (LFA-1) was implicated as a candidate autoantigen in treatment-resistant Lyme arthritis. However, later work cast doubt that LFA-1 was a relevant autoantigen. Recently, endothelial cell growth factor, a novel human autoantigen, was implicated in the pathogenesis of antibiotic-refractory Lyme arthritis. IL-17 and Th17 lymphocytes have an additional role in Lyme arthritis.
Many persons infected with B. burgdorferi are asymptomatic, and the risk of developing Lyme disease following a tick bite is low, even if the tick was infected by B. burgdorferi . Very often, a tick bite is not recalled. In symptomatic patients, the cutaneous, nervous, and musculoskeletal systems are most frequently involved. Symptoms of Lyme disease can be divided into early and late manifestations ( Table 42-1 ). Early signs of infection become evident within days to weeks of the tick bite, whereas late organ involvement begins several months or even years later. Early symptoms are usually self-limiting, whereas late manifestations may become chronic and rarely lead to irreversible damage of involved organs. In most patients with the disease, only one organ system is affected.
|ORGAN SYSTEM||EARLY LYME DISEASE||LATE LYME DISEASE|
|Skin||Erythema migrans |
Borrelial lymphocytoma *
|Acrodermatitis chronica atrophicans *|
|Nervous system||Cranial nerve palsy||Chronic encephalomyelitis *|
The earliest and most common skin manifestation, erythema migrans, typically occurs days to several weeks after infection as an enlarging, warm, but usually painless, erythematous rash at the site of the bite, and it lasts for days or weeks ( Fig. 42-3 ). In its classic form, this lesion begins as a red macule or papule and expands peripherally with partial central clearing occurring later. Sometimes, the rash does not clear; in other cases, the clearing is so complete that erythema migrans becomes a mere curved red streak. In children, the neck and head are the most frequently affected sites, and the erythema may look more cellulitic. It is also common in the groin, axilla, or thigh and may expand up to 30 cm. Secondary lesions can occur at sites distant from the tick bite. Erythema migrans may be accompanied by flulike symptoms of fever, chills, arthralgia, musculoskeletal pain, headaches, malaise, and fatigue. Lyme disease may also begin as a flulike illness in the absence of erythema migrans.
Weeks to months after infection, borrelial lymphocytoma , also known as lymphadenosis cutis benigna (e.g., a purple swelling most commonly at an earlobe, the scrotum, or a nipple), is occasionally reported in European patients. Recently, Lyme chondritis presenting as ear erythema has been distinguished from borrelial lymphocytoma. Acrodermatitis chronica atrophicans , a late skin manifestation, rarely affects European children and then only years after infection. In the early phase of acrodermatitis, the affected limb develops inflammatory changes with a red or bluish discoloration. Later, cutaneous atrophy becomes apparent. A peripheral neuropathy can accompany the skin lesion. Lymphocytoma and acrodermatitis are extremely uncommon in North America.
Nervous System Disease: Neuroborreliosis
Early neuroborreliosis most frequently presents as lymphocytic meningitis or a cranial nerve palsy weeks to months after infection. It may be accompanied by fever, headache, nausea, vomiting, radicular paresthesias, or pain. In endemic areas Lyme meningitis may be the most common form of meningitis in childhood. Unilateral or bilateral facial nerve palsy is the most common focal neurological manifestation, but cranial nerves III, IV, VI, and VIII may also be affected. Also, the optic nerves may be involved, but mostly secondary to increased intracranial pressure. Signs of meningitis may be mild or absent in spite of increased protein and lymphocytes in the cerebrospinal fluid. A painful meningoradiculoneuritis is the most common neurological manifestation in European adults but is relatively uncommon in children. Months to years after infection, a small number of patients develop late neuroborreliosis, progressive encephalomyelitis, or an encephalopathy. Other rare neurological manifestations include Guillain–Barré syndrome, pseudotumor cerebri, cerebral vasculitis, and neurogenic bladder.
After erythema migrans, arthritis is the most common manifestation of Lyme borreliosis in many series of pediatric patients and is more common in North America than in Europe. There is little evidence that the musculoskeletal symptoms of Lyme borreliosis otherwise differ between Europe and North America. Myalgia, myositis, and enthesitis are uncommon, although B. burgdorferi have been identified in a few patients with enthesitis or nodular fasciitis. Arthralgia and myalgia develop as early as days to weeks after infection, sometimes concurrent with erythema migrans or flulike symptoms. However, arthritis appears typically months to years after infection.
The two largest series of pediatric patients with Lyme arthritis include 90 children from Connecticut and 109 from Germany, 62 of whom have been described in detail. Monoarthritis of a knee occurred in approximately two thirds of all children. Both knees or other large joints may also be affected. Polyarticular involvement of small joints was rare. At onset, the arthritis was usually episodic, with relatively painless swelling lasting only a few days and disappearing without the need for therapy. Recurrent episodes of arthritis may become prolonged, and chronic arthritis (duration of more than 3 months) has been reported in up to 18% of patients. Among 109 German children with Lyme arthritis, 70 had monoarthritis, 32 oligoarthritis, and 7 polyarthritis. The pattern of oligoarticular involvement differed from that found in patients with early-onset oligoarticular juvenile idiopathic arthritis or juvenile spondyloarthritis. However, ocular involvement with keratitis and anterior and intermediate uveitis may occur in children with Lyme arthritis. Lyme arthritis may occasionally be confused with septic arthritis, especially when there is isolated hip involvement. In several studies, fever and complete refusal to bear weight on the affected joint were negative predictors for Lyme arthritis versus septic arthritis, whereas knee involvement, sedimentation rate less than 40, and absolute blood neutrophil count less than 10,000 were positive predictors. Children are more likely than adults to show acute disease.
Lyme arthritis was reported in a woman after autologous chondrocyte transplantation. The investigators hypothesize that B. burgdorferi was asymptomatically present in the patient’s joint before the chondrocyte transplant procedure. Myositis has only rarely been described in adult patients, and only myalgia has been reported in children. A dermatomyositis-like picture has also been described in a single adult patient.
Involvement of other organ systems is much more uncommon. Carditis is rare in children and most commonly manifests as a reversible atrioventricular block. A recent report documented three adult deaths from undetected lymphoplasmacytic pancarditis ( Fig. 42-4 ). Ocular involvement, including conjunctivitis, keratitis, iridocyclitis, intermediate uveitis, choroiditis, or optic neuritis, has been described in children with Lyme disease. Even more uncommonly, patients may develop hepatitis. There have been anecdotal reports of transplacental transmission of B. burgdorferi , but this has not been confirmed in controlled studies. The offspring of 5 of 19 pregnant women who had Lyme disease during pregnancy had one or more of the following abnormal outcomes: prematurity, syndactyly, rash, cortical blindness, developmental delay, or intrauterine fetal death. Whether any of these complications is attributable to infection with B. burgdorferi or with other spirochetes is not certain. There is no evidence that maternal infection presents a significant risk to the fetus.
The synovitis of Lyme arthritis resembles that of juvenile idiopathic arthritis or juvenile rheumatoid arthritis, with villous hypertrophy, synovial cell hyperplasia, and infiltration of lymphocytes and plasma cells. Lymphoid follicles may also be present. Endarteritis is a characteristic finding in patients with Lyme synovitis. In one study, spirochetes were detected in 2 of 17 synovia, mainly in a perivascular distribution. Other studies using special silver stains have also identified B. burgdorferi in synovium or synovial fluid. The organism has been recovered from the margins of erythema migrans lesions and cardiac tissue. Although cardiomyopathy may result from the initial myocarditis, valvular endocarditis does not develop. Myositis may in part account for the myalgia and fatigue that occur in this disease, and DNA from B. burgdorferi has been identified in the muscle of such patients.
The erythrocyte sedimentation rate is elevated in half of the patients, especially during the early phase of Lyme arthritis. Meningoencephalitis causes mild cerebrospinal fluid lymphocytic pleocytosis, with a median cerebrospinal fluid (CSF) cell count of 160 cells/mm 3 , mostly lymphocytes. The mean synovial fluid white blood cell count ranges from less than 5000/mm 3 to greater than 100,000/mm 3 , with a predominance of neutrophils in samples with high cell counts. Children have higher synovial cell counts than adults.
Confirmation of Infection With Borrelia burgdorferi
Laboratory methods to document infection with B. burgdorferi include direct tests, such as culture or the polymerase chain reaction (PCR) to detect borrelial sequences, and indirect tests, such as serology ( Table 42-2 ). The latter tests are most frequently used and universally available. In spite of the standardization of the laboratory evaluation in North America, the approach suggested by the American College of Physicians has not been widely adopted for European patients. Common problems with standardization of test procedures for the diagnosis of Lyme disease have been reviewed.
|Culture of Borrelia burgdorferi||Requires weeks; rarely successful|
|Histochemistry using silver stain or monoclonal antibodies||Rarely successful in synovial tissue|
|Polymerase chain reaction for borrelial DNA||Efficiency varies widely: Urine, 5% to 30%; synovium, 6% to 90% (higher in membrane than in fluid)|
|Enzyme immunoassay or immunofluorescence assay using serum||High sensitivity, low specificity, rarely false negative, >10% false positive|
|Immunoblot using serum||Confirmatory test with high specificity; healthy blood donors <3% positive|
|Lymphocyte-proliferation assay with borrelial antigens||Sensitivity and specificity <80% |
Direct Methods to Detect Infection
Culture of B. burgdorferi usually takes 2 weeks to a few months, requires immediate suspension of the test material in special medium, and has high rates of recovery only from skin biopsies of patients with dermatological manifestations of the disease. Culture of the organism from blood and especially from synovial fluid has been relatively unsuccessful. The possibility of obtaining positive cultures is better from the CSF of patients with early neuroborreliosis. Methods such as silver staining of spirochetes in tissue specimens or staining with monoclonal antibodies are not routinely performed and are prone to artifacts.
The PCR can demonstrate DNA of B. burgdorferi in tissues or bodily fluids, including synovial fluid. In a large North American study of synovial fluid from patients with Lyme arthritis, PCR results were positive for 96% of patients not previously treated with antibiotics and 37% of those who had been treated. In a later study from the same group, borrelial sequences were undetectable in synovial specimens from patients with chronic Lyme arthritis after appropriate antibiotic therapy.
Other groups have found a smaller percentage of positive PCR results in synovial fluid of patients with Lyme arthritis. The precise role of PCR in routine diagnosis remains unclear. False-positive or false-negative results may occur. Optimization of PCR includes using more than one primer pair, targeting genes situated on the bacterial chromosome and the plasmids, performing nested PCR, and analyzing synovial fluid and urine. There is evidence that PCR of synovial tissue may have a higher positivity rate than in synovial fluid and may remain positive in patients with ongoing arthritis whose synovial fluid is negative by PCR after antibiotic treatment. B. burgdorferi in the skin lesions of patients with erythema migrans were active and viable, whereas the spirochetes in synovial fluid or tissue from Lyme arthritis patients were moribund or dead. PCR results for urine may be positive in healthy humans whose sera contain B. burgdorferi –specific antibodies; urine testing for Lyme antigens is not an approved assay.
Indirect Methods to Detect Infection
Specific antibodies can be demonstrated after B. burgdorferi infection by a variety of tests including enzyme immunoassay (EIA), immunofluorescence, hemagglutination, and Western blotting ( Fig. 42-5 ). It is recommended that a sensitive EIA be used as a screening test and that all results in the indeterminate or positive ranges be confirmed by Western blotting; this has been called two-tier testing.
Typical IgM and IgG responses of patients with Lyme arthritis and the North American criteria for a positive IgG blot are shown in Fig. 42-5 . In early Lyme borreliosis, IgM blots are considered positive if at least two of the following three bands are present: 21-kD OspC, 39-kD BmpA, and 41-kD flagellin. Antigens from different strains of B. burgdorferi have different molecular weights, and, for this reason, North American criteria cannot easily be applied in Europe or Asia, where there is greater strain diversity. Even so, a two-test approach is the best available method for diagnosis of Lyme disease in European children. In a German pediatric Lyme arthritis study, at least six specific bands were required for a positive IgG Western blot, similar to the American criteria. Commonly, patients with Lyme arthritis have 10 or more IgG bands (see Fig. 42-5 ), including the ones mentioned earlier. Specific blot-positivity criteria have been established for each of the three pathogenic species in Europe, with different positivity criteria for each strain. North American studies found that kinetic EIAs detecting IgG responses to recombinant B. burgdorferi antigen vlsE1 or to a conserved internal sequence of vlsE1 were equally sensitive and specific in assessing patients with Lyme arthritis and more sensitive in assessing patients with early skin or neurological manifestations of Lyme borreliosis. There was also a cost-reduction with that strategy. European studies have also shown the usefulness of EIAs or Western blots using recombinant vlsE1 peptide antigen in children with Lyme arthritis or adults with neuroborreliosis. In North American studies comparing two-tier testing with enzyme-linked immunosorbent assay (ELISA) and Western blot with a vlsE C6 peptide ELISA, both methods had similar sensitivity, but two-tier testing had slightly better specificity. Using an IgG Western blot with a vlsE IgG band as the second-tier test improved sensitivity while maintaining the specificity of standard IgM and IgG two-tier testing. Two recent studies have examined the use of American serologic assays for Lyme borreliosis acquired in Europe and found significantly less sensitivity. European serologic tests performed well for Lyme disease acquired in the United States.
Within the first weeks after infection, all serological results may be negative because specific IgM antibodies usually do not appear until 3 to 4 weeks after infection, and IgG antibodies cannot be detected until 4 to 8 weeks after infection. Frequently, an EIA may give false-positive results related to cross-reactive antibodies such as rheumatoid factors or after infection with Epstein–Barr virus or other spirochetes, including Treponema pallidum, Treponema denticola, Borrelia hermsii, and leptospira. Antibodies to B. burgdorferi occasionally are found in children with juvenile rheumatoid arthritis, systemic lupus erythematosus, and other illnesses (e.g., bacterial endocarditis, mumps, Rocky Mountain spotted fever, and other rickettsial diseases).
Serological tests cannot distinguish patients with active infection from those with a previous infection who have responded to therapy. In particular, about 10% of patients with late manifestations of Lyme disease continue to demonstrate an IgM response in addition to the IgG response. Because IgG titers tend to remain elevated for years, serology cannot be used to monitor treatment success or failure. Contrary to claims that the rate of decline of antibodies to the conserved internal sequence of vlsE1 could be useful to monitor treatment efficacy, this was not true in German children with Lyme arthritis. In an endemic area, serological tests should be performed only in patients with clinical signs suggestive of the disease; a positive result in a patient with a low pretest probability of Lyme disease is much more likely to represent a false-positive rather than a true-positive result. Because of the rampant overtesting for Lyme disease, the American College of Rheumatology has recommended not testing for Lyme disease as a cause of musculoskeletal symptoms without an exposure history and appropriate examination findings. Given these restrictions, serology is useful diagnostically in patients with suspected Lyme arthritis because virtually all patients are clearly IgG seropositive. The diagnosis of Lyme arthritis must not be based on inappropriate Western blot testing of synovial fluid alone. Box 42-1 provides a simplified overview of the laboratory evaluation.
Patient living in or having visited an endemic area
Presence of arthritis documented
No other obvious cause of arthritis
Serology positive (enzyme immunoassay [EIA] and Western blot for IgG antibodies to B. burgdorferi ):
No further laboratory test needed
Serology negative (EIA and Western blot):
Rule out other diagnosis
Refer for specialized evaluation
Clinical characteristics of patients with arthritis that suggest a diagnosis of Lyme arthritis include residence in or travel to an endemic area, a previous tick bite, episodic oligoarthritis involving the knee joint, the absence of arthralgias preceding the onset of arthritis, and being of adolescent age at onset. Specific criteria have been combined to form a clinical diagnostic score that would confirm or exclude Lyme arthritis in two thirds of children with arthritis ( Table 42-3 ). For a diagnosis of Lyme arthritis, arthritis (i.e., swelling and effusion or painful limitation of motion in the absence of trauma) must be observed by a physician. Arthralgias alone or reports by the patient or the parents that the joint was swollen are not objective signs of arthritis in this context.
|Arthralgia before onset of arthritis||−3|
|Age at onset of arthritis||+0.3 × age (in years)|
|Initial arthritis in knee joint||+2|
|History of tick bite||+2|
|Number of joints involved||−0.4 × number of large joints affected|
|Scoring example : A 10-year-old boy without prior arthralgia or history of a tick bite developed arthritis in a knee. The arthritis resolved after 10 days but recurred after a 3-month interval. †|
|Episodic arthritis present||+4|
|No initial arthralgias||+0|
|Age at onset × 0.3||+3|
|Initial arthritis in knee||+2|
|No history of tick bite||+0|
|1 large joint affected||−0.4|
* If a criterion is recognized, its indicated value is added to or subtracted from the total score for the patient. If it is not identified, the item is scored as 0. Values of 6 or greater indicate the presence of Lyme arthritis, and values of 2.5 or less exclude the diagnosis.
† The patient’s serum was later shown to contain IgG antibodies to Borrelia burgdorferi by enzyme immunoassay and immunoblot. He was treated with ceftriaxone for 2 weeks. Arthritis disappeared during therapy and did not recur in the subsequent 2 years of follow-up.
Frequently, however, the clinical presentation of Lyme arthritis may be indistinguishable from that of other rheumatic diseases of childhood with arthritis as a principal manifestation, making laboratory tests mandatory in all patients with arthritis living in or having traveled to an endemic area. A recent review addressed a number of common misconceptions about the diagnosis and treatment of Lyme disease.
The CDC established the following criteria for the diagnosis of Lyme disease : the presence of erythema migrans larger than 5 cm in diameter, or at least one clinical sign (i.e., arthritis, meningitis, radiculoneuritis, mononeuritis, or carditis) and the presence of specific antibodies to B. burgdorferi.
These criteria were developed for epidemiological research and may not always be applicable in the clinical setting. For example, borrelial lymphocytoma may initially occur in the absence of specific antibodies. Moreover, the mere combination of an objective sign with specific antibodies may include chance associations between two frequent events: arthritis of some kind may affect 1 in 1000 children, and in endemic areas, the prevalence of antibodies to B. burgdorferi is high; 3% or more of healthy blood donors may be positive for specific antibodies to B. burgdorferi. In a population-based survey among 5-year-old children in Sweden the seroprevalence of Borrelia IgG antibodies was 3.2%, and in a nationwide German study among 1- to 17-year-old children and adolescents it was 4.8%. Some high-risk populations, such as forestry workers, have a much higher incidence of seropositivity—up to 45%—without clinical evidence of Lyme disease. Serology can provide evidence of current or prior infection but cannot absolutely confirm a pathogenic link between the infection and the clinical manifestation. The child may have been infected with B. burgdorferi as documented by serology; however, arthritis may result from other known or unknown causes. Overdiagnosis of Lyme disease in children has also been reported.
Synovial fluid analysis is of little help in establishing a diagnosis of Lyme arthritis because the white blood cell count and type of cells vary greatly. However, synovial fluid analysis can exclude septic arthritis and other infection-associated arthritides, yields material for testing by PCR, and confirms the presence of inflammation. Although a positive PCR result from an experienced laboratory may indicate a persistent infection, a negative result does not exclude the diagnosis. Synovial tissue may be more suitable than synovial fluid for PCR testing, especially after antibiotic treatment has failed to produce a remission of symptoms.
A lumbar puncture should be performed in patients with suspected neuroborreliosis. Lymphocytic pleocytosis and elevated CSF protein levels are characteristic. In the presence of neurological symptoms and CSF lymphocytic pleocytosis the serological standard in diagnosing neuroborreliosis in European adult patients remains the detection of intrathecal antibody production. However, especially in children, specific antibody production frequently occurs only after several weeks to months of infection and has been less commonly demonstrable in American patients. Therefore intrathecal antibody production is not required for diagnosing early neuroborreliosis in children. Recent studies showed that the detection of the chemokine CXCL13 improves the diagnostic performance in children with early neuroborreliosis. A clinical prediction rule of fewer than 7 days of headaches, fewer than 70% CSF mononuclear cells, and the absence of cranial nerve palsy has been used in American children to distinguish aseptic meningitis due to other causes from Lyme meningitis.
Recommendations by the Infectious Diseases Society of America (IDSA) for the treatment of Lyme disease ( Table 42-4 ) have been published and vary according to disease manifestations. In the treatment of patients with erythema migrans or neuroborreliosis, amoxicillin is as effective as doxycycline. Cephalosporins were only marginally better than penicillin G. Cefuroxime axetil was equally efficacious as doxycycline in adults or as amoxicillin in children with erythema migrans. Although doxycycline and ceftriaxone have been equally effective, most macrolide antibiotics are inferior to these antibiotics. However, clarithromycin was as effective as amoxicillin in children with erythema migrans. It is not known whether these results are applicable to patients with Lyme arthritis, for whom a variety of antibiotics have been recommended, including parenteral penicillin G, oral penicillins, amoxicillin with or without probenecid, ceftriaxone, cefotaxime, cefuroxime, erythromycin, roxithromycin (plus cotrimoxazole), azithromycin, tetracycline, doxycycline, and others. For Lyme arthritis, oral antibiotic therapy for 4 weeks with amoxicillin in young children or with doxycycline in adolescents has been recommended by the IDSA; however, controlled trials of oral versus intravenous antibiotics have not been performed. Oral treatment may be more convenient for the patient and more cost-effective but involves administration of up to 84 doses of amoxicillin. Intravenous treatment with ceftriaxone once daily for 14 days can be given as an outpatient and may be more convenient for some patients. Before the physician can conclude that antibiotic therapy has failed, at least two courses of sufficient duration and well-documented compliance are required. When confronted with failure of a treatment program with appropriate antibiotics, the correctness of a diagnosis of Lyme arthritis should be questioned and reconfirmed.