MANCHESTER, U.K. – In a general practice setting, people who developed systemic lupus erythematosus were more than twice as likely to consult a primary care physician in the 5-year run-up to their diagnosis, according to the results of a case-control study.

The median annual consultation rates were 9.2 times for systemic lupus erythematosus (SLE) patients versus 3.8 times for controls (P < .001).

Lupus patients also started to exhibit telltale symptoms during this time, providing an opportunity for making a diagnosis sooner through better recognition of these early symptoms.

The aim of the study was “to find out whether people with lupus came to see their GP [general practitioner] more frequently and what they were going to see their GP for,” Dr. Frances Rees of Nottingham (England) University Hospitals NHS Trust explained in a poster session on connective tissue disease at the British Society for Rheumatology annual conference.

“After that, we developed a risk-prediction model which was a multivariate analysis taking some of the variables that we decided in advance would be most significant,” she added. These variables included age, gender, arthritis or arthralgia, rash, fatigue, alopecia, serositis, sicca, Raynaud’s phenomenon, and the consultation rate.

“We found that the model performed well in both the model development data set and in the model validation data set,” Dr. Rees said.

The rationale for the study was that SLE can be difficult to diagnose early, and that delays of more than 7 years have been found between symptom onset and diagnosis. Perhaps, if SLE is diagnosed earlier, outcomes could be improved, Dr. Rees and her coauthors reasoned.

Using the UK Clinical Practice Research Datalink, an anonymized database of general practice records for approximately 12 million people, all incident cases of SLE occurring between 1999 and 2012 were identified and matched to four individuals without lupus who were seen at the same primary care practices.

A total of 1,739 patients with lupus and 6,952 controls were identified and separated into two cohorts: approximately two-thirds formed the development cohort and the remainder a validation cohort.

Individual odds ratios (OR) were calculated for age, gender, consultation rates, 27 clinical features, and five previous diagnoses of rheumatic disease in the 5 years preceding SLE diagnosis. Some of the clinical features considered were arthritis or arthralgia, rash, fatigue, alopecia, serositis, and sicca; the associated diseases were another connective tissue disease, rheumatoid arthritis (RA), fibromyalgia, chronic fatigue syndrome (CFS), and Epstein-Barr virus (EBV) infection.

Results showed that cases were more than five times more likely than controls to be female (84% vs. 51%; OR, 5.23; P < .001).

Early symptoms suggestive of SLE occurred more often in cases than in controls, with arthritis or arthralgia in 36% vs. 10% (OR, 4.77; P < .001), rash in 43% vs. 10% (OR, 6.84; P < .001), fatigue in 17% vs. 7% (OR, 2.76; P < .001), alopecia in 4% vs. 1% (OR, 4.77; P < .001), serositis in 4% vs. 1% (OR, 5.05; P < .001), and sicca in 6% vs 1% (OR, 4.97; P < .001). Raynaud’s phenomenon was seen in 4% vs. 0%, with an OR of 15.29, which was significant (P < .001), but the confidence interval (CI) was wide (CI, 9.41-24.85).

Patients with SLE also were more than 20 times more likely to have a prior diagnosis of another connective tissue disease (OR, 21.68; P < .001) and 10 times more likely to have been diagnosed with RA (OR, 10.77; P < .001). Fibromyalgia (OR, 7.89; P < .001) and CFS (OR, 6.39; P < .001) also were more frequent among those who developed lupus than in those who did not.

“This suggests that, although there probably is a delay in diagnosis, patients are getting other diagnoses first,” Dr. Rees noted. “So, we would propose that, in the future, our risk-prediction model may be able to reduce that delay in conjunction with educational programs to say if you have a combination of these symptoms think about testing for [antinuclear antibodies] and/or referring to rheumatology.”

The team’s risk-prediction model now needs further external and economic validation, but the hope is that perhaps it could be incorporated into electronic systems used in primary care to flag potential lupus cases.

“The clinical message is to keep SLE in mind if someone is presenting multiple times with these kind of symptoms,” Dr. Rees said in an interview.

“Lupus is very rare,” she acknowledged, nothing that some primary care physicians may only see one or two cases of lupus in their career. So the risk-prediction model would aim to be an aid to help them consider SLE as a possibility in patients with several symptoms that, on their own, might not be suggestive of lupus.

The work was supported by a research grant from Lupus UK. Dr. Rees had no conflicts of interest.

BSR: Flagging early symptoms could diagnose lupus sooner