Frequent infection with a viral pathogen, parvovirus B19, in rheumatic diseases of childhood




To find further evidence of the association of parvovirus B19 infection with juvenile rheumatic diseases, and to get new insights into the immunopathogenesis of these diseases.


Paired serum and synovial fluid samples from 74 children with rheumatic disease were analyzed with respect to their content of viral DNA and antibodies directed against the B19 viral proteins VP1, VP2, and NS1. Control sera from 124 children with noninflammatory bone diseases or growth retardation were also analyzed. The sequence of the viral DNA, amplified by polymerase chain reaction (PCR), was determined. IgG-complexed virus was isolated from sera and synovial fluid by adsorption to protein A beads. The amount of free virus versus immunocomplexed virus particles was determined by quantification of the viral genomes by quantitative PCR.


Twenty-six of the 74 patients (35%) had detectable amounts of parvovirus B19 DNA in the serum (n = 22 [30%]) and/or the synovial fluid (n = 16 [22%]), whereas only 9 of the 124 control sera (7%) were positive for the viral DNA (P < 0.0001). Forty-six patients (62%) had serum IgG against the structural proteins, indicating past infection with B19. NS1-specific antibodies were detected in sera from 29 patients (39%) and 27 controls (22%) (P < 0.001). In addition, 3 patients (4%) showed VP2-specific IgM. In 15 patients, viral DNA could be repeatedly detected in followup samples of serum and synovial fluid. Sequencing revealed low-degree nucleotide variations that are in the range of genotype 1 of parvovirus B19. Immunocomplexed virus was present in varying amounts, both in the sera and in the synovial fluid samples.


Parvovirus B19 is frequently found in serum or synovial fluid of children with rheumatism. The rate of persistent B19 infection in these patients is significantly higher than in age-matched controls.

Parvovirus B19, a single-stranded DNA virus, has been associated with a wide spectrum of diseases. In addition to acute infection resulting in anemia and erythema infectiosum (fifth disease; a rash illness of childhood), hydrops fetalis in pregnant women and acute symmetric polyarthropathy in adults have been reported as clinical manifestations. Depending on the hematologic status of the host, B19 infection may be associated with hematopoetic disorders such as aplastic crisis, thrombocytopenia, or pancytopenia. Hepatitis, myocarditis, myositis, neurologic disease, vasculitis, and persistent arthropathy may occur occasionally (for review, see refs. 1–3). Persistent parvovirus B19 infection has been reported both in patients with and in patients without underlying immunodeficiencies (4, 5). In children who are immunocompetent, the B19 infection is usually mild and has a short duration.

Productive parvovirus B19 replication takes place in erythroid progenitor cells (6). In addition, B19 genomes have been detected in cells of bone marrow and synovial membranes (7, 8). The infectious particle is composed of the major capsid protein VP2 and the minor capsid protein VP1, which are identical except for the 227 amino acids at the amino-terminal end of VP1, the so-called VP1 unique region (9). Recently, a phospholipase A2–like activity has been linked to the VP1 unique region of several parvoviruses, including human parvovirus B19 (10, 11). Immunoglobulins directed to this domain of the VP1 protein offer life-long protection against reinfection. Viral persistence has been assumed to be due to a qualitatively or quantitatively inadequate humoral immune response against the viral capsid proteins, particularly against the VP1 unique region (12, 13). During infection, a third viral protein, the nonstructural protein NS1, is produced. It has been shown to be involved in the transactivation of the viral promoter and in the induction of apoptotic processes (14–16). NS1-specific antibodies were found preferentially in patients with prolonged B19 viremia and persistent infection (17–21).

In this present study, we wanted to establish the general role and involvement of parvovirus B19 as a causative agent or trigger in arthritis during childhood. After initial parvovirus B19 serologic analysis, paired serum and synovial fluid from 74 children with different rheumatic diseases were tested for parvovirus B19 DNA by quantitative polymerase chain reaction (PCR). The amplified viral DNA was submitted for sequence determination. Furthermore, we analyzed the amount of immunocomplexed virus present in the sera and synovial fluid. The results indicate that the presence of parvovirus B19 may have been previously underestimated and that it is, in fact, found frequently in both sera and synovial fluid of children with acute arthritis. The long-term presence of immunocomplexes may be involved in the pathogenesis of rheumatic diseases.



Paired serum and synovial fluid samples derived from 74 unselected children and adolescents with rheumatic diseases (29 white males, 45 white females) were collected. The catchment area was the northern part of Germany, including Schleswig-Holstein, Hamburg, parts of Mecklenburg-Vorpommern, and Niedersachsen. The patients' ages varied between 2.2 years and 19.4 years (mean 10.1 years). All were consecutive inpatients of a pediatric rheumatology ward between July 2000 and October 2001. The mean disease duration was 3.9 years (range 6 weeks–14.5 years). A mean of 2.7 joints (range 1–17) was affected, and all diagnoses of rheumatic disease fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) criteria for active rheumatoid arthritis (22). The clinical diagnosis, defined according to the juvenile idiopathic arthritis criteria of the International League of Associations for Rheumatology (23), encompassed a broad spectrum of diseases, including oligoarthritis, extended oligoarthritis, rheumatoid factor (RF)–negative polyarthritis, RF-positive polyarthritis, systemic arthritis, enthesitis-related arthritis, psoriatic arthritis, Lyme borreliosis or other reactive arthritis, and chronic inflammatory bowel disease associated with inflammatory joint disease (Table 1). The study received approval from the hospital ethics committee of the University of Lübeck in Germany.

Table 1. Patient classification according to the International League of Associations for Rheumatology criteria for rheumatic diseases*
Clinical courseNumber of patientsIgG antibodiesIgM antibodiesPresence of viral DNA or protein
VP1/VP2VP1VP2NS1VP1/VP2VP1VP2Serum/synovial fluidSerumSynovial fluid
  • *

    Values are the number (%) of patients. RF = rheumatoid factor.

 Oligoarticular3319 (58)17 (52)15 (45)11 (33)0009 (27)8 (24)4 (12)
 Oligoarticular extended63 (50)2 (33)2 (33)1 (17)0003 (50)2 (33)2 (33)
 Polyarticular RF neg.159 (60)9 (60)6 (40)7 (47)0004 (27)2 (13)4 (27)
 Polyarticular RF pos.11 (100)1 (100)1 (100)01 (100)01 (100)1 (100)1 (100)1 (100)
 Systemic44 (100)4 (100)4 (100)3 (75)1 (25)01 (25)4 (100)4 (100)2 (50)
 Enthesitis related21 (50)1 (50)1 (50)0000000
 Psoriatic75 (71)5 (71)5 (71)4 (57)1 (14)01 (14)2 (29)2 (29)2 (29)
 Lyme borreliosis32 (67)1 (33)1 (33)00001 (33)1 (33)0
 Other reactive arthritis21 (50)1 (50)1 (50)2 (100)0001 (50)1 (50)0
 Chronic inflammatory bowel disease11 (100)1 (100)1 (100)1 (100)0001 (100)1 (100)1 (100)
Total7446 (62)42 (57)37 (50)29 (39)3 (4)03 (4)26 (35)22 (30)16 (22)
Controls12464 (52)61 (49)59 (47)27 (22)7 (6)4 (3)7 (6)9 (7)9 (7)


Sera from 124 children and adolescents (68 male, 56 female; ages 2 months–16 years, mean 10.2 years) with noninflammatory bone diseases (e.g., osteogenesis imperfecta, rickets, or growth retardation) were obtained from 3 different centers located in the northern part of Germany. Any acute disease (e.g., erythema infectiosum, anemia, fever, malaise) that was related to an infection with parvovirus B19 was excluded.

Detection of parvovirus B19–specific antibodies and immunocomplexes.

Initially, all patient samples were analyzed for IgG and IgM antibodies against the structural and nonstructural proteins of parvovirus B19, using Western blot assays (RecomBlot; Mikrogen, Munich, Germany). In addition, IgM and IgG antibodies against VP2 capsids were determined using an enzyme-linked immunosorbent assay (ELISA) (Biotrin, Dublin, Ireland). For detection of immunocomplexes, 150 μl of serum or synovial fluid was diluted with 300 μl of buffer (0.025M Tris HCl, pH 7.5, 0.075 NaCl, 0.5% Triton X-100), after which 200 μl was incubated for 120 minutes at 37°C with a 20-μl suspension of protein A–Sepharose beads (Sigma-Aldrich, Bornem, Belgium) that had been previously washed in the buffer described above. The beads were then pelleted and the supernatant was subjected to DNA purification via QIAquick spin columns (Qiagen, Hilden, Germany). Viral genomes from the supernatant's DNA preparation, representing free virus particles, were quantified by TaqMan PCR. The beads with bound immunocomplexes were washed several times with buffer, the DNA was purified, and the number of viral genomes was determined as described above.

Detection of B19 genomes.

Sera and synovial fluid samples were individually diluted 1:1 with water, and incubated at 95°C for 10 minutes. After this heat treatment, aggregated proteins were removed by centrifugation at 14,000 revolutions per minute for 15 minutes. Five microliters of these mixtures was directly submitted for DNA amplification (ABI PRISM 7700 sequence detection system; Perkin-Elmer/Applied Biosystems, Weiterstadt, Germany). Parvovirus B19 DNA was detected with real-time TaqMan PCR assay, amplifying a conserved genomic region encoding the capsid protein VP2, as previously reported (24). In each TaqMan PCR assay, 10-fold serial dilutions of plasmid pJB (generously provided by Dr. John Clewley, Virus Reference Division, Central Public Health Laboratory, London, UK) were analyzed. A standard curve was generated and the number of B19 genomes was calculated for each clinical sample. All reactions were carried out in duplicate. In a control reaction, distinct quantities of a bacterial plasmid containing an unrelated DNA segment (Volvox carteri) was mixed with 5 μl of serum or synovial fluid. Samples that displayed inhibitory effects on the amplification of Volvox DNA were subjected to DNA purification (QIAmp blood kit; Qiagen) prior to amplification of viral DNA.

DNA sequence analysis.

PCR fragments were purified via QIAquick spin columns (Qiagen). Nucleotide sequences for the region encoding the carboxy-terminal part of the VP1 unique region and the amino-terminal part of the VP2 protein were obtained with a 373A Sequencer (Applied Biosystems) by the cycle-sequencing method, using primers as described previously (25). The nucleotide sequence data from this study have been deposited in the GenBank database (AY211615–AY211622).

Statistical analysis.

For statistical analysis, the chi-square test was applied to compare proportions.


Prevalence of parvovirus B19–specific antibodies in patients and controls.

Seventy-four children and adolescents who had acute arthritis in association with various rheumatic diseases were included in this study. All of these patients were in a pediatric rheumatology clinic over a period of at least 16 months. With respect to their clinical diagnosis, the children were divided into 10 groups (Table 1): oligoarthritis (n = 33), RF-negative polyarthritis (n = 15), psoriatic arthritis (n = 7), extended oligoarthritis (n = 6), systemic arthritis (n = 4), Lyme borreliosis (n = 3), enthesitis-related arthritis (n = 2), reactive arthritis (n = 2), RF-positive polyarthritis (n = 1), and chronic inflammatory bowel disease associated with inflammatory joint disease (n = 1). As controls, 124 children with noninflammatory bone diseases or growth retardation were investigated.

Initial testing of all sera for the presence of parvovirus B19–specific humoral immune reactions revealed that 46 of 74 patients (62%) and 64 of 124 controls (52%) demonstrated IgG against the structural proteins VP1 and/or VP2 of parvovirus B19, indicating previous B19 infection. IgG antibodies against the structural proteins were detected in these patients irrespective of the manifestation of the rheumatic disease (Table 1). This antibody prevalence in both the patient and the control group is in a range that would be expected in children and adolescents, since parvovirus B19 infection is known as one of the classic diseases of childhood (26). However, the frequency of NS1-specific IgG antibodies differed significantly, being present in 29 patients (39%) and 27 controls (22%) (P < 0.001); the frequency obtained in the control group is in accordance with the values described for the detection of NS1-specific IgG in individuals with past B19 infection (18, 19, 27). Western blot analysis of the synovial fluid samples derived from the arthritis patients revealed an antibody pattern that was qualitatively identical to that obtained from the sera (data not shown).

IgM antibodies against the structural protein VP2 were detectable by Western blot and ELISA in serum and/or synovial fluid samples from 3 patients (4%) (patients 17, 18, and 22 in Table 2). In these patients, subsequent evaluation of consecutive samples of serum or synovial fluid persistently revealed IgM, in addition to the presence of VP1- and VP2-specific IgG, over a period of more than 12 months. Therefore, the possibility that the positive IgM values were associated with acute or recent B19 infections could be excluded, although they did indicate that these patients' samples had continuously produced VP2-specific IgM. The possibility that unspecific IgM reactivity was due to the presence of RF could be excluded in almost all patients, since only 1 patient was RF positive. IgM antibodies against NS1 or VP1 proteins were not detectable in the patients. In the control group, 7 of 124 children (6%) showed VP2-specific IgM antibodies (Table 1).

Table 2. Individual data on the patients who tested positive for parvovirus B19 DNA in serum and/or synovial fluid
PatientRheumatic diseaseParvovirus B19 serologyAmplification of viral DNA
Clinical courseDuration, months before studyInitial time point, months before study*Status at initial time pointSeraSynovial fluidConsecutive samples
AvailablePositiveObservation period, months
  • *

    A time point of 0 months indicates that the respective patient was tested for the first time in this study.

  • Samples were obtained over an interval of at least 4 weeks.

  • Interval during which consecutive samples were obtained.

10Oligoarticular extended6752Acute++No
11Oligoarticular extended410Acute+YesNo2
12Oligoarticular extended20Past+YesYes6
13Polyarticular, RF neg.5050Acute++YesYes7
14Polyarticular, RF neg.773Past+No
15Polyarticular, RF neg.330Acute++YesYes6
16Polyarticular, RF neg.2018Past+No
17Polyarticular, RF pos.1610Past++YesYes20
25Reactive arthritis119119Acute++YesYes20
26Chronic bowel disease6549Acute++No

Detection of parvovirus B19 genomes in patient samples.

Viral genomes were detectable in 26 patients, in either the serum or synovial fluid or in both samples (Tables 1 and 2). This indicates that 35% of all patients (P < 0.0001) and 57% of those with positive serologic detection of B19 were persistently infected with parvovirus B19. Twenty-two (30%) of the patients' serum samples could be shown to be viremic, since viral DNA could be amplified from these serum samples. Synovial fluid samples from 16 patients (22%) contained detectable amounts of parvovirus B19 DNA, of which 12 had viral genomes in both the serum and the synovial fluid.

After the start of the study, viral DNA could be detected in consecutive serum and/or synovial fluid samples, obtained during time periods of up to 29 months from the time of flares, from 15 arthritis patients (patients 2, 3, 5, 6, 7, 12, 13, 15, 17, 18, 19, 20, 22, 23, and 25), indicating a chronically active form of B19 infection with ongoing viral replication or repeated reactivation of latent virus (Table 2). Quantitative PCR revealed that between 5 × 102 and 3.5 × 106 viral genome equivalents could be detected per ml of serum or synovial fluid. In comparison, only 9 children (7%) of the control group could be shown to be viremic, with the viral load in a range similar to that observed in the arthritis patients (P < 0.0001). Fourteen (48%) of the 29 patients who were positive for anti–NS1-IgG contained viral genomes in either the serum or synovial fluid, whereas only 3 (11%) of the 27 anti–NS1-IgG–positive controls displayed detectable amounts of DNA. Control sera were obtained from children exhibiting symptoms (noninflammatory bone diseases, growth retardation) that are not correlated with B19 infection. Since followup serum samples were not available from the controls, we assume that these children with positive values for B19-specific IgM and DNA were experiencing acute, subclinical B19 infection at the times when the sera were obtained.

Parvovirus B19 genomes could be detected in association with almost all disease manifestations (Tables 1 and 2): oligoarticular (9 of 33 [27%]), extended oligoarticular (3 of 6 [50%]), polyarticular RF negative (4 of 15 [27%]), polyarticular RF positive (1 of 1 [100%]), systemic arthritis (4 of 4 [100%]), psoriatic arthritis (2 of 7 [29%]), reactive arthritis (1 of 2 [50%]), Lyme borreliosis (1 of 3 [33%]), and inflammatory bowel disease associated with arthritis (1 of 1 [100%]). The duration of the rheumatic disease (from 0 to 181 months prior to the start of this study) as well as the time points and results of the initially performed parvovirus serology were deduced from the individual records of these 26 viral DNA–positive patients (Table 2). These data allowed further insights into the different influences that are exerted by the viral infection.

For 5 patients (patient 13, who had polyarticular disease, patients 18, 19, and 20, all of whom had systemic arthritis, and patient 25, who had other reactive arthritis), the beginning of the rheumatic disease was directly correlated with the diagnosis of acute parvovirus B19 infection, as indicated by positive B19-specific IgM. IgM could not be detected in the followup samples from patients 13, 19, 20, and 25; it was replaced by IgG. In patient 18, a girl who developed systemic arthritis as a consequence of the acute B19 infection, low level IgM antibodies could be detected over a time period of 12 months, in combination with IgG. After the onset of the disease, all of these patients developed persistent viremia, and in patients 13, 19, 20, and 25, virus could also be detected in the synovial fluid (Table 2).

Six patients (patients 1, 6, 10, 11, 15, and 26) had different rheumatic diseases prior to the viral infection. In all these cases, seroconversion indicating acute or recent B19 infection coincided with a long-lasting flare. Two of these children (patients 6 and 15) developed persistent parvovirus B19 infection, since viral DNA was subsequently detected in followup samples. Two children (patients 1 and 11) tested negative for viral DNA in consecutive samples, indicating virus elimination. Followup samples from patients 10 and 26 were not available.

In the other patients, the exact time point of acute B19 infection could not be determined. All patients, except patients 17 and 22, displayed the serologic parameters of past B19 infection (anti–VP1/VP2-IgG positive, anti–VP2-IgM negative) at the time point when parvovirus B19 serology was performed for the first time. In patient 17, who had RF-positive polyarticular arthritis, and in patient 22, who had psoriatic arthritis, VP2-specific IgM antibodies, in addition to VP1/VP2-specific IgG, have been detectable since onset of the disease (Tables 1 and 2). In both patients, viral DNA could be amplified from serum and synovial fluid during periods of several months.

For DNA sequence analysis, we selected the region encoding the carboxy-terminal domain of the VP1 unique region and the amino-terminal domain of the VP2 protein as the part of the viral genome with high sequence conservation. Using PCR-amplified DNA, sequence determination was successful for 8 viral isolates (in patients 3, 7, 8, 11, 14, 17, 18, and 22). With respect to the other patients, the material was either quantitatively or qualitatively not sufficient to allow sequencing. Comparison with the 3 standard B19 genotypes (28), Pvbaua (genotype 1, accession no. M13178), Lali (genotype 2, accession no. AY044266), and V9 (genotype 3, accession no. AX003421), revealed the highest sequence rates of homology (97–98%) to the sequence of genotype 1 in all cases. Individual nucleotide alterations in the sequence of the viral genomes from patients 3, 7, 8, 11, 17, 18, and 22 resulted in a rather limited number of amino acid exchanges as compared with the VP1-protein region (amino acids 173–326) encoded by genotype 1 Pvbaua (patient 3 D175→A; patient 7 V248→A, S259→T; patient 8 T304→N; patient 11 V248→A, S259→T, T304→N; patient 17 G211→A, Y225→T; patient 18 V248→A, S259→T; patient 22 V248→A, S259→T).

Detection of parvovirus B19 complexed with immunoglobulins in patient samples.

Immunocomplexed virus particles were isolated, by adsorption to protein A beads, from serum and/or synovial fluid samples obtained from patients 2, 3, 15, 17, 19, 22, 23, and 25. The amounts of B19 genomes as part of the adsorbed immunocomplexes, and of uncomplexed virus particles remaining in the supernatant, were determined by quantitative PCR (Table 3). With the exception of patients 17 and 22, samples obtained at consecutive times from 2 months up to 24 months or from different joints were analyzed. Over the various time periods, either decreasing (patients 3 and 15), fluctuating (patients 23 and 25), or unchanged (patients 2 and 19) amounts of viral genomes were detected. In patients 2, 3, 17, 19, 22, 23, and 25, a major percentage of the viral genomes were found as part of the immunocomplexed fraction of viral particles, both in the serum and synovial fluid samples. A different situation was observed in patient 15. This patient had recently been infected with parvovirus B19 and displayed a relatively high virus load, with the majority of the B19-virus particles uncomplexed with IgG.

Table 3. Immunocomplexed versus free parvovirus B19 in sera and/or synovial fluid samples from patients with rheumatic disease
Patient, date obtained*Viral DNA genome equivalents/mlParvovirus B19, % of pos. samples
SerumSynovial fluidFree particleIgG complex
  • *

    The date obtained is the month and year at which the respective samples were obtained.

 January 20011,3000100
 March 20011,0000100
 February 200170,0002575
 July 200119,000991
 September 20011,0000100
 May 2001900,0008119
 May 200160,000964
 October 20011,3007624
 October 200050,0000100
 May 20011,0000100
 May 20011,2000100
 August 20001,4000100
 January 20002,0002080
 December 200018,0005743
 May 20013,5004555
 October 20018000100
 December 20012,3000100
 October 20003,5002080
 March 20018,0000100
 August 20019004060


The present data show an impact of parvovirus B19 infection in the pathogenesis of juvenile rheumatic diseases. B19 viremia usually reaches a peak at days 7–9 after infection and is resolved by the development of an IgM- and IgG-antibody response starting a few days later. Despite the development and presence of B19-specific immune reactions, ∼20% of all B19 infections show a prolonged state of viremia or viral persistence restricted to the synovial fluid, and viral genomes are detected in bone marrow or other organs, e.g., synovial tissue, liver, or myocardium, for several years after infection (4, 8, 12, 13, 19, 27, 29–33). Compared with our age-matched control group and compared with healthy adult blood donors, who have been shown to contain B19 DNA in 7% and in 0.1% up to 0.6% of cases, respectively (34–36), viral genomes were detected in 35% of arthritis patients in the present study. In those cases in which DNA sequence analysis after PCR amplification was successful, parvovirus B19 genotype 1 genome sequences with minor nucleotide and amino acid changes were detected.

An elevated prevalence of B19-specific IgG had been previously reported in Japanese patients with polyarticular juvenile rheumatoid arthritis (37). In addition, Nocton et al reported joint symptoms associated with recent B19 infection in 22 children (38), of whom 6 developed persistent oligo- or polyarticular arthritis for 2–13 months. We can confirm these observations in European children with arthritis, as compared with age-matched controls. This points to a significant worldwide role of parvovirus B19 infection in the generation of childhood rheumatism.

The fact that in a subgroup of 15 patients, parvovirus DNA was repeatedly detected in followup samples indicates that these patients have developed persistent B19 infection and are incapable of eliminating the virus. This may be due to an inadequate immune reaction against the viral capsid proteins. Both the VP1 and VP2 proteins have been shown to contain epitopes that elicit the production of virus-neutralizing antibodies. Following patient testing for humoral immune reactions against viral capsid proteins, no defects in the elicitation of either VP1- or VP2-specific IgG were observed. In addition, no differences could be observed between the patients with persistent and past B19 infection with respect to antibody affinity to VP1 and VP2 proteins (data not shown). It cannot be excluded that in some patients, the capacity of the immune system to eliminate the virus may be restricted by the immunosuppressive treatment with steroid drugs. However, since 32 of 33 children with oligoarticular arthritis have received exclusively monotherapy with nonsteroidal antirheumatic drugs, and all children with enthesitis-related arthritis (2), Lyme borreliosis (3), or other reactive arthritis (2), and 3 of 7 children with psoriatic arthritis have exclusively been treated with methotrexate and/or nonsteroidal antirheumatic drugs, steroid-related immunosuppression cannot be used as a general explanation of the establishment of persistent B19 infection.

In general, our data allow the following conclusions. Parvovirus B19 infections may be associated with the beginning of a rheumatic disease. In patients with ongoing active arthritis, the clinical status worsens coincidentally with B19 infection. With respect to the different groups of patients, it may be important that viral DNA could be detected in all of the children whose rheumatic disease was systemic. Similar data have been obtained in a previous study that also demonstrated an elevated incidence of persistent B19 infection in children with long-lasting arthropathies (21). Our data from children affected by rheumatic disease are clearly different from those obtained from adults (8). Among 73 adult patients with different forms of arthritis, only 1 synovial fluid sample (1.4%) contained parvovirus B19 DNA (8), whereas in our group, 16 children (22%) could be shown to be viral DNA positive. This may be explained by the different time points of B19 infection that usually occur during childhood and adolescence. However, this finding also indicates that parvovirus B19 infection is more frequently associated with rheumatic disease in childhood than in adults.

Since viral DNA has been shown to persist in the synovial tissue after clearance from the peripheral blood and/or synovial fluid (7, 39–42), it has to be assumed that virus reactivation may occur in the synovial tissue sporadically or in association with inflammatory processes due to rheumatic disease and/or other infectious agents such as borrelia or streptococci. Similar interactions between inflammation, cell differentiation, and reactivation of latent or persistent virus are well known from studies with herpes and papillomavirus systems. In patients with rheumatic disease, reactivation results in the production of B19 particles that are released into the synovial fluid and form complexes with the B19-specific immunoglobulins. These immunocomplexes have been detected in different amounts in all samples tested so far (Table 3). They indicate the production and presence of virus particles and may elicit further immune reactions and trigger inflammation and arthritis by continuous stimulation of immune reactions via activation of the classical complement cascade and cytokine secretion by neutrophils.

In addition, a phospholipase A2–like activity has been shown to be associated with the VP1 unique region (11) and may be involved in the continuous production of leukotrienes and prostaglandins due to the persisting virus infection in the joints. Moreover, Ray and coworkers have recently shown that parvovirus B19 is capable of inducing a cartilage-invasive phenotype in normal human synovial fibroblast cultures (43). Invasive pannus is one of the main complications leading to joint destruction in arthritis. It has to be assumed that parvovirus B19 is directly involved in this serious complication. At present, data about the inherited susceptibility of individuals to persistent B19 infection are not available. Kerr et al recently identified an association between acute symptomatic B19 infection and HLA–DRB*01, *04, and *07 alleles (44). Respective data on persistent parvovirus infection are not yet available. Therefore, future research has to be done on the genetic basis of chronic B19 infection to further strengthen the assumed causative or modifying role of parvovirus B19 in juvenile rheumatism.


The authors thank Mikrogen (Munich, Germany) for the generous donations of RecomBlot tests, and Karin Beckenlehner and Alexia Herrmann for their excellent technical assistance. In addition, we thank Sean Doyle (National University of Ireland, Maynooth) for critically reading the manuscript.