We are pleased to see the great interest raised by our results, which showed for the first time a dramatic effect of rituximab therapy on endothelial function in patients with severe RA refractory to tumor necrosis factor α blockers. Sandoo et al emphasize the significant and sustained effect of up to 6 months of flow-mediated endothelium-dependent vasodilatation (FMD%) compared with baseline values of FMD% obtained prior to rituximab therapy. However, although this maintained positive effect on endothelial function at 6 months is clear, they are concerned about the short-term effects on FMD% and the lipid profile at 2 weeks following the first administration of rituximab.
In our rituximab-treated patients, premedication with corticosteroids was given following the recommendations indicated by the manufacturer. They suggested prescribing either hydrocortisone (100 mg intravenously) or dexamethasone (8 mg intravenously), 30–60 minutes before rituximab administration. Certainly, it is clear that corticosteroids by themselves may yield improvement of FMD%. We agree on this point with Sandoo et al. However, the effect of a single intravenous administration of corticosteroids prior to rituximab infusion cannot offer a solid explanation for the persistent improvement of endothelial function observed at 6 months after rituximab therapy. Moreover, the effect observed at 2 weeks, immediately before the last infusion of rituximab, was associated with a dramatic fall of peripheral CD20+ B cell levels. This result may support a role of rituximab in the improvement of FMD% found at week 2. On the other hand, the potential long-term effect of corticosteroids on endothelial function in patients with vasculitides, such as giant cell arteritis (GCA), cannot be extrapolated to that observed in RA, a chronic inflammatory disease associated with endothelial dysfunction in long-term corticosteroid-treated patients (1). In this regard, the chronicity and persistence of the inflammatory response may explain the presence (2) and progression (3) of subclinical atherosclerosis as well as the increased incidence of cardiovascular events and cardiovascular mortality (4) in patients with RA who received a long-term, high cumulative prednisone dosage.
In accordance with the observation above, we found that subclinical atherosclerosis, determined by the ultrasonographic assessment of the carotid artery intima-media thickness in 40 biopsy-proven GCA patients, who at the time of the ultrasonographic study had ended steroid therapy (mean treatment duration 32.1 months, mean cumulative prednisone dose 8,775 mg), was not greater than that found in matched controls (5). Regrettably, this was not the case for 47 long-term corticosteroid-treated patients with RA (mean prednisone cumulative dose 15,900 mg) (6).
Sandoo and colleagues also believe that corticosteroids used as premedication before rituximab infusion may also influence the changes in the lipid profile observed in our patients. With respect to this, we cannot exclude that the increase of total cholesterol and LDL cholesterol observed at week 2 after the commencement of rituximab might not be due to the specific effect of rituximab therapy, but represents a predictable response to attenuation of inflammation. In this regard and in keeping with findings observed in other conditions associated with inflammation or infection, untreated patients with RA have reductions in high-density lipoprotein (HDL) cholesterol, LDL cholesterol, and total cholesterol. Moreover, treatment with a broad range of drugs, including biologic therapies, leads to increases not only of HDL cholesterol, but also of total and LDL cholesterol in patients with RA. However, unlike the effect attributed to corticosteroids by Boers et al (who found a significant increase in total cholesterol levels 28 days after corticosteroid treatment; levels returned to baseline after corticosteroid treatment was stopped) (7), and besides a persistent significant increase of HDL cholesterol at 6 months after rituximab therapy (compared with the levels observed before the onset of this drug), a nonsignificant increase in the median values of total cholesterol and LDL cholesterol was still evident at 6 months in our series, compared with values obtained prior to rituximab therapy. In any case, we realize that the small number of patients included in our study precludes further conclusions on these results.
Finally, we are in complete agreement with Sandoo et al on the fact that the persistent and continuing improvement of FMD% observed in this series at 6 months coincides with the timing of rituximab maximum efficacy (8). Based on our promising results and the importance of FMD% as a marker of subclinical atherosclerosis and predictor of cardiovascular events in patients with RA (9), we encourage other investigators to reproduce our results with a larger number of RA patients with severe disease that is refractory to other biologic therapies.