Rheumatoid arthritis (RA) is a chronic inflammatory disease associated with cardiovascular disease.1 There are conflicting data as to whether or not this increased risk of cardiovascular disease is already present before the clinical onset of RA.2 In active RA an unfavourable lipid profile is present and is associated with inflammation.3 Patients with arthralgia positive for rheumatoid factor (RF) and/or anticyclic citrullinated peptide antibodies (aCCP) (seropositive) are at risk of developing RA and can be considered as patients with symptoms and systemic autoimmunity associated with RA without clinical arthritis.4 ,5 Considering the close relationship between inflammation and serum lipid levels, we investigated whether an unfavourable lipid profile was associated with the development of RA in patients with seropositive arthralgia.

Total cholesterol (TC), high-density lipoprotein cholesterol (HDLc), low-density lipoprotein cholesterol (LDLc), triglycerides (TG), apolipoprotein (apo) A1 and apoB were determined at baseline in the serum of 348 prospectively followed patients with seropositive arthralgia.4 Age, sex, smoking, alcohol use, C-reactive protein, RF, aCCP, shared epitope status, statin use and cardiovascular comorbidities (myocardial infarction, coronary or peripheral arterial bypass surgery, transient ischaemic attack or cerebral vascular accident) were recorded at baseline and analysed for confounding. The development of arthritis was followed over a median (IQR) of 24 (14–49) months. Differences in lipid profiles between patient categories were analysed with the Student t test or Mann–Whitney U test as appropriate. Predictive properties of lipid profiles for the development of arthritis were analysed by Cox proportional hazard analyses.

One hundred and sixteen patients (33%) developed arthritis after a median (IQR) follow-up of 12 (6–23) months, of whom 91% were diagnosed with RA according to the 2010 ACR/EULAR criteria.6 Their baseline characteristics are shown in table 1. Lower mean levels of TC, HDLc, LDLc, ApoA1 and ApoB and higher mean levels of TG were observed in patients who developed arthritis than in patients who did not, although these differences did not reach statistical significance for every lipid (table 2). ApoA1 showed a linear association with the risk of arthritis development (HR 0.44, 95% CI 0.26 to 0.79). HDLc showed a non-linear association with arthritis development and was therefore analysed as a categorical variable (HR lower tertile vs second and third tertiles 0.64, 95% CI 0.43 to 0.94). The other lipids were not significantly associated with arthritis development. aCCP status was a confounder for all lipids. No other confounders were found. aCCP-positive patients had significantly lower TC levels. For the other lipids, trends were seen towards lower lipid levels in aCCP-positive patients. After adjustment for aCCP status, only ApoA1 was significantly predictive of arthritis development (HR 0.52, 95% CI 0.29 to 0.92) whereas, for HDLc, an obvious trend was observed (HR first vs second and third tertiles 0.70, 95% CI 0.47 to 1.02).

Our data show that patients with arthralgia (particularly those with aCCP positivity) who developed arthritis had a different lipid profile from those who did not. After adjustment for aCCP status, a lower ApoA1 level was predictive for the development of arthritis. Lower lipid levels were present prior to arthritis, which might indicate subclinical inflammation. These data are partly in line with previous reports which also showed differences in lipid profiles prior to RA diagnosis.7 ,8 We detected lower HDLc levels, but also lower ApoA1 levels. Unlike van Halm et al, we did not find significant differences in ApoB or TG levels, but there was a significant association between aCCP and TC as well as a trend towards a higher TC in patients who later developed RA. These different findings could be due to a difference in patient selection. Nevertheless, we have shown for the first time that seropositivity is linked with differences in lipid profile. Whether inflammation is the cause or consequence of lipid abnormalities in pre-RA remains to be investigated.