Close association of herpes zoster reactivation and systemic lupus erythematosus (SLE) diagnosis: case-control study of patients with SLE or noninflammatory musculoskeletal disorders.

OBJECTIVE

To investigate the prevalence of infections, particularly the frequency of shingles and the timing of varicella zoster virus (VZV) reactivation, and antibiotic use, vaccinations, and joint trauma prior to and at diagnosis of systemic lupus erythematosus (SLE).

METHODS

We sent questionnaires to patients with SLE (n = 93) and controls with noninflammatory musculoskeletal disorders (MSK; n = 353) including osteoarthritis, fibromyalgia, and tendonitis. We matched SLE patients to controls for sex (up to 1:3).

RESULTS

The response rate in SLE was 66% and in controls 69% (p < 0.53). Four of 61 SLE patients and 12 of 173 controls were men. The mean disease duration in the SLE group was 8 +/- 1 years compared to 10 +/- 1 years in controls (p < 0.23). SLE patients were significantly younger than controls (mean age of SLE patients 49 +/- 2 vs 57 +/- 1 years for controls; p < 0.0004), and results were adjusted for age. A significantly higher proportion of SLE participants had a history of VZV (shingles) (19% vs 7%, respectively; OR 2.98, p < 0.003), whereas rubella was reported less in SLE (23% vs 42%; OR 0.43, p < 0.03). VZV infections were clustered just prior to or after diagnosis in SLE but were more widely spaced temporally in the controls (1 +/- 4.5 years after the diagnosis of SLE vs -14.7 +/- 4 years before the diagnosis of noninflammatory MSK disorder; p < 0.003). Diagnosis of shingles was observed in 6 of 11 SLE patients within +/- 2 years of SLE diagnosis, whereas only 2 of 15 controls had shingles within +/- 2 years of diagnosis (OR 7.2, p < 0.03). Only 2 patients with SLE were taking immunosuppressive drugs or steroids at time of shingles, so immunosuppressive therapy was not usually concomitant at time of VZV reactivation. Common infections (respiratory, urinary tract, ear, and eye) in the SLE group exceeded controls, but not significantly (23% vs 9%; OR 2.98, p < 0.06) and SLE patients were more likely to have been vaccinated since 18 years of age with any type of vaccine (69% vs 51%; OR 2.21, p < 0.04). SLE patients were less likely than controls to report joint trauma within one year prior to their diagnosis (25% vs 40%; OR 0.49, p < 0.04). There were no differences with respect to streptococcal throat infection (p < 0.96), diarrhea/vomiting (p < 0.84), rash with fever (p < 0.07), parvovirus infection (p < 0.16), infection after surgery (p < 0.58), respiratory tract infection (p < 0.71), or ear (p < 0.09) and eye infection (p < 0.68) one year prior to diagnosis. A higher proportion of SLE patients had a history of urinary tract infections (46% vs 25%), but this was not significant (p < 0.17), nor was it significant one year prior to diagnosis (p < 0.63). Overall, the likelihood of having any infection one year prior to diagnosis was not significantly higher in the SLE group (p < 0.56). There were no differences one year prior to diagnosis in travel history (p < 0.69), hospitalizations (p < 0.47), use of antibiotics (p < 0.54), history of rheumatic fever, positive TB skin test, or hepatitis A, B or C infection.

CONCLUSION

Varicella reactivation as shingles is increased in patients with SLE and clusters around diagnosis. Vaccinations are increased in those with SLE compared to controls. Common infections are not significantly increased in SLE patients prior to onset of symptoms. We cannot determine if VZV infections are causally associated with SLE in some people, are from an abnormal immune system response due to the lupus itself or from the use of steroids or other immunosuppressive drugs to control the disease, or are spurious.