RF and ACPA are established predictors of RA severity. Their presence associates with a more severe disease course characterised by higher rates of radiological damage and extra-articular manifestations. In one observational study of 135 early RA patients, individuals with a persistently positive RF had greater X-ray damage, disease activity and disability and required more aggressive treatment [3]. Similarly in a cohort study of 93 early RA patients identified amongst blood donors, the presence of ACPA prior to and at disease onset associated with worse radiological outcomes [4].

There is some evidence that smoking influences RA severity. In one study of 100 early RA patients observed over 24 months smokers had more swollen and tender joints, when compared to non-smokers [5]. The impact of smoking is, however, difficult to assess as it predispose to ACPA formation, which could mediate its effect on disease severity.

Two observational studies have indicated that alcohol consumption may reduce RA severity. In one study of 873 RA cases, more frequent alcohol intake significantly associated with lower DAS28-CRP, Larsen and HAQ scores [6]. In another large observational study of 2,908 RA patients a J-shaped dose-response effect was seen with occasional/daily drinkers having less X-ray progression compared with non-drinkers/heavy drinkers [7].

Several studies have observed a correlation between social deprivation and poor RA outcomes. In 869 patients from the Early Rheumatoid Arthritis Study (ERAS) higher social deprivation scores associated with higher HAQ scores and joint counts [8]. This effect may be driven by other lifestyle factors that associate with low socioeconomic status such as smoking.

RA may be more severe in women. In a prospective study of 225 women and 67 men with seropositive early RA, women had worse disease progression over 2 years with higher DAS28 scores despite receiving similar therapy to men; men had higher remission rates [9]. In another Swedish study women had significantly higher DAS28 and HAQ scores compared with males at all time-points over 5 years [10].

A number of studies have tried to identify genetic markers that can be used to predict RA severity. Most have focussed on genetic associations with X-ray progression. To date 12 genetic loci that associate with radiological progression have been identified and replicated [11]. In a Dutch study of 426 early RA patients, these genetic variants explained approximately 18 % of the variance in X-ray progression, measured as the change in the Sharp-van der Heijde score over 6 years (Fig. 14.1). Most loci had minor effects on X-ray progression. Additionally, their impact was reduced when adjusting for traditional prognostic factors such as age, gender and ACPA, suggesting that their effects may be partially mediated by these clinical factors.

Research suggests that males probably respond better to methotrexate than females. A systematic review of studies reporting predictors of methotrexate response found that women were up to three times less likely to have a good clinical response when treated with methotrexate compared to men [13].

Data from the British Society for Rheumatology Biologics Register (which collects clinical data on patients receiving biologic drugs in the UK) suggests that women may also respond less well to anti-TNF therapy. In this analysis of nearly 3,000 RA patients female gender associated with a reduced odds of remission with etanercept (OR 0.61; 95 % CI 0.38–0.94) and infliximab (OR 0.60; 95 % CI 0.40–0.89) treatment [14].

Several studies have reported that smoking reduces the efficacy of methotrexate and anti-TNF treatment. One example is the Epidemiological Investigation of Rheumatoid Arthritis (EIRA) study, undertaken in Sweden [15]. In this analysis current smokers were less likely to attain a good response when receiving both methotrexate and anti-TNF therapy. Interestingly, a history of previous smoking did not affect treatment responses.

Although RF and ACPA do not clearly affect responses to DMARD treatment there is good evidence that the requirement for intensive combination treatment differs between RA patients with and without ACPA. In an analysis of 431 patients with early, active RA enrolled to the Combination Anti-Rheumatic Drugs in Early RA (CARDERA) trial, only ACPA-positive patients benefited from intensive combination treatment [16]. No benefit beyond monotherapy with methotrexate was observed in ACPA-negative patients. This is demonstrated in Fig. 14.2, which shows mean changes from baseline in modified Larsen scores (measuring X-ray damage) in ACPA-positive and ACPA-negative patients receiving combination treatment or methotrexate monotherapy over 2-years. Combination treatment significantly reduced X-ray progression beyond methotrexate monotherapy at all time-points in ACPA-positive patients. By contrast ACPA-negative patients had minimal X-ray progression irrespective of the treatment strategy used.

There is some evidence that ACPA status can also predict which patients are most likely to respond to biologic drugs. Rituximab appears more effective in seropositive (RF and/or ACPA-positive) patients, with a meta-analysis of four RCTs demonstrating a significantly greater reduction in DAS28 scores between seropositive and seronegative RA patients at 24 weeks [17]. By contrast TNF-inhibitors seem more efficacious in ACPA-negative disease. In an analysis of the UK Biologics Registry, ACPA-negative patients had a 0.39 (95 % CI 0.07–0.71) greater mean improvement in DAS28 with anti-TNF compared with ACPA-positive patients [18]. Overall the impact of ACPA status on biologic responses is modest, and is of too small an effect size to guide clinical decision making on its own.