Parvovirus B19 has been associated with a growing number of diseases. Besides the frequent manifestations such as erythema infectiosum, aplastic crisis in persons with underlying haemolytic anaemia, hydrops fetalis in pregnant women and acute or chronic arthritis a range of rather rare diseases have been described in recent reports.1 Among them are case reports on persistent parvovirus B19 infection in immune incompetent people, encephalitis, myocarditis, systemic lupus erythematosus (reviewed by Anderson and Young1) and rheumatoid arthritis.2Furthermore, parvovirus B19 has been suspected to play a part in the aetiology of polymyalgia rheumatica (PMR).3 ,4 Because of the acute onset of PMR and its systemic symptoms an infectious cause may be a relevant factor. Additionally, autoimmune processes have been demonstrated in both, PMR and parvovirus B19 infection.5 ,6 As the receptor for parvovirus B19, the P-blood group antigen (globoside), is also present on endothelial cells an interrelation between parvovirus B19 and giant cell arteritis or PMR may be possible.

Parvovirus B19 can only replicate in erythroid precursor cells in human bone marrow, but it is known that infection of cells non-permissive for viral replication leads to an excess production of the viral non-structural protein (NS1) without production of capsid proteins.7 As the NS1 protein is cytotoxic and able to induce apoptosis, it probably plays a part in the pathogenetic process of the parvovirus B19 induced tissue damage. This is confirmed by the fact that antibodies against NS1 of parvovirus B19 are preferentially produced during chronic or persistent parvovirus B19 infections, for example in parvovirus B19 associated chronic arthritis.8 ,9

To test the hypothesis whether PMR is associated with acute parvovirus B19 infection, we tested the seroprevalence of IgG antibodies against the two structural proteins VP1 and VP2 and against the non-structural protein NS1 in 110 PMR patients (patients with giant cell arteritis excluded; mean (SEM) age 67.0 (0.8) years, range: 48—77) and, for comparison, in 135 healthy controls of different ages. At the time point of blood sampling (median disease duration at the time point of blood sampling: 0.6 years, range: 0—7.3, mean (SEM): 1.4 (0.2) years), 35 patients had no corticosteroids and 75 patients received on an average 15.2 (1.8) mg prednisolone/day. Furthermore, we investigated the interrelation between age, symptoms10 or laboratory parameters and the presence of NS1 specific antibodies in healthy controls and patients with PMR. Non-parametric Kruskal-Wallis one way analysis was used to compare means of different subgroups. The significance level was p<0.05.

Subjects in the control group had various ages between 18 to 75 years. Overall seroprevalence of IgG against the capsid proteins VP1 and VP2 was 78% (fig 1). Overall IgG seroprevalence against VP1 and VP2 was 88% in patients with PMR (not significantly different versus the age matched control group). With respect to the NS1 IgG antibody, overall seroprevalence in the control group was 22% (fig 1) and in patients with PMR 20% (p=0.057 versus the age matched control group: 10%; fig1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Seroprevalence of IgG antibodies against VP1/VP2 (black bars) and NS1 (open bars) in control subjects and patients with polymyalgia rheumatica (PMR). The number of healthy controls in each subgroup of normal subjects was at least 20. *p=0.057 v seroprevalence of IgG NS1 in age matched healthy controls (groups 56–65 and 66–75 with 42 controls v 110 patients with PMR).

Furthermore, we investigated the association between the presence of NS1 IgG antibodies and PMR related symptoms10 or laboratory parameters (patients with NS1 as compared with patients without NS1 IgG antibodies were not different in age, sex, and medication). The symptoms were assessed using standard record forms from the medical histories (at the time serum was collected).10 We asked the patients for muscular pain in the left/right shoulder, left/right upper arms, left/right neck, left/right gluteal muscle, and left/right thigh. If one muscle group was painful, the corresponding item was scored with one point (the sum of the item points was the overall muscle score).10 In PMR patients with NS1 IgG as compared with patients without NS1 IgG, arthralgia was more frequent (with versus without: 73%v 40%, p=0.024). However, the overall muscle score was lower in NS1 positive than in NS1 negative patients (0.5 (0.2) SEM v 1.6 (0.3) SEM score points; p=0.021). With respect to other PMR related symptoms, no significant differences were found. In patients with a positive NS1 IgG antibody, interleukin 6 (4.6 (0.9) SEM v 11.3 (2.2) SEM; p=0.037) and soluble ELAM (48.2 (4.8) SEMv 71.4 (5.2) SEM; p=0.024) were significantly lower as compared with patients without NS1 IgG. No significant differences were found with respect to erythrocyte sedimentation rate, C reactive protein, tumour necrosis factor, interleukin 2, and interleukin 1β.

In view of these data, there was a positive association between NS1 and arthralgia. However, other symptoms and parameters of inflammation such as erythrocyte sedimentation rate were not associated with the presence of NS1 IgG. Furthermore, disease related immune mediators such as interleukin 6 or soluble ELAM were lower in patients with as compared with patients without NS1 IgG. As a positive NS1 IgG titre indicates an active infection, an acute parvovirus B19 infection does not seem to be a pathogenetic factor in our patients with PMR.