Background Despite the pivotal role hydroxychloroquine (HCQ) plays in treating SLE, less than 50% of patients take HCQ as prescribed. Non-adherence versus lack of effect to HCQ are difficult to distinguish, underscoring the importance of measuring HCQ blood levels to assess adherence. Despite this, information and consensus on the clinical impact of incorporating routine testing of HCQ blood levels is lacking. Therefore, we systematically reviewed publications examining the correlation between 1) HCQ levels and adherence, and 2) HCQ levels and SLEDAI scores, in SLE patients. We hypothesized that low HCQ levels would correlate with non-adherence and higher SLEDAI scores.

Methods A comprehensive search was performed using MeSH heading and keywords in Medline, Embase, CINHL and Web of Science databases. We selected observational and interventional studies that measured HCQ levels and assessed adherence and/or SLEDAI scores in adults with SLE. Newcastle Ottawa Scale and Cochrane Collaboration Risk Assessment tools were used to rate the quality of observational and intervention studies, respectively. We used Forest plots to compare pooled estimates (95% CI) of correlations between HCQ levels and patient or physician reported nonadherence and SLEDAI scores. Heterogeneity was assessed using I2.

Results From 306 manually reviewed studies, four studies analyzing correlation between HCQ levels and adherence, and five studies examining the correlation between SLEDAI and HCQ blood levels, met inclusion criteria. The odds of nonadherence measured by physician or reported by the patient was 3 times higher in patients with below threshold HCQ levels, compared to those with higher HCQ levels (OR 2.95, 95% CI 1.63, 5.35, p<0.001, I2 49%) (figure 1, panel A). The mean SLEDAI score was 3.33 points higher in groups with HCQ levels below threshold, but this trend was not statistically significant (3.33, 95% CI −0.60, 7.26, p=0.097, I2 99%) (figure 1, panel B). Risk of bias assessment revealed three poor quality studies which were excluded in sensitivity analysis which did not change results. Limitations of our analysis include study heterogeneity and lack of consensus on HCQ level interpretations.

• Download figure
• Open in new tab
• Download powerpoint

Abstract 221 Figure 1

Forest plot showing correlation of HCQ blood levels with A) Assessed Adherence B) SLEDAI scores

Conclusions We found good correlation between HCQ levels and non-adherence. Low HCQ levels showed higher mean SLEDAI scores but due to the limited number of studies and heterogeneity we did not find significant correlation. Findings support the clinical utility of measuring HCQ levels in SLE patients. Future larger studies should investigate standards for HCQ level interpretation and target levels needed to minimize SLEDAI scores.

Funding Source(s): n/a