To the Editor:
We read with great interest the report by Mazzantini, et al1. That retrospective analysis of a large cohort of patients with polymyalgia rheumatica (PMR) demonstrated that duration or cumulative dose of longterm low-dose glucocorticoid (GC) therapy was significantly associated with higher risk of arterial hypertension and acute myocardial infarction. However, in a multivariate analysis adjusted for traditional cardiovascular (CV) risk factors, arterial hypertension was confirmed as the only adverse effect significantly associated with treatment duration.
It is thought that longterm GC treatment, especially at high dose, may indirectly increase the risk of CV disease through its well-recognized effect on traditional CV risk factors, including arterial hypertension, dyslipidemia, hyperglycemia, and obesity2. As summarized in Table 1, GC have been associated with higher incidence of subclinical atherosclerosis, and their detrimental effects on the CV system seem to occur at a preclinical phase. Although results are not uniform, duration of treatment exposure and higher cumulative dose are significantly correlated with higher risk of subclinical endothelial damage, the cumulative dose of prednisone being the only risk factor associated with progression of carotid intima-media thickness9. But whether and to what extent the subclinical endothelial dysfunction caused by GC affects the risk of overt CV events remains uncertain and unpredictable. Indeed, GC, especially at low dose, may exert a paradoxical cardioprotective effect through their antiinflammatory activity on the vessel wall. Interpretation of data is also hampered because some rheumatic immune-mediated inflammatory disorders, such as rheumatoid arthritis (RA) and systemic lupus erythematosus, represent per se an independent risk factor for atherosclerosis and CV disease10,11,12, making it difficult to distinguish between effects of GC and those of the underlying inflammatory and potentially proatherogenic disease.
In this context, data from studies investigating the harmful effect of GC on clinically manifested CV events are not conclusive, because there are no large randomized trials powered to detect differences in CV mortality and morbidity. Nevertheless, results from retrospective and case-control studies involving large cohorts of patients with chronic rheumatic and non-rheumatic inflammatory disorders (RA, PMR, connective tissue diseases, chronic obstructive pulmonary and inflammatory bowel disease) highlight the intriguing considerations (Table 2).
Specifically, after adjustment for confounders and traditional CV risk factors, a higher risk for CV events due to GC treatment appears to be shared by subjects with one of these heterogeneous conditions. Moreover, the risk appears more strictly associated with occurrence of heart failure than every other CV event, including myocardial infarction, stroke, transient ischemic attack, and peripheral artery disease. Continuous intake of GC, oral GC administration, current exposure, and higher cumulative dose may represent variables significantly associated with higher risk of CV events. Finally, the adverse effects of GC on the CV system may differ according to the underlying disease for which they are prescribed. Notably, and similar to data from Mazzantini and colleagues, a retrospective analysis of a wide cohort of patients with PMR found no increased CV event risk associated with GC treatment, but a trend toward a protective effect of GC on the combined CV endpoints was observed21. Of interest, substantial risk reduction was demonstrated in patients exposed to GC for at least 1 year prior to the event compared to those never exposed21. It is conceivable that different pathogenic mechanisms underlying PMR in comparison to other inflammatory/autoimmune diseases may partially explain such discordant results. Indeed, it is well known that PMR is characterized by a high burden of inflammation that may reasonably predispose to functional and structural vascular endothelial alterations. This hypothesis is supported by our findings showing that aortic stiffness of patients with untreated PMR at disease onset was greater than that of matched healthy subjects, with a strong direct correlation between pulse-wave velocity and inflammatory markers22. The imbalance between endothelial fragmentation and repair, as assessed by increased endothelial microparticle (EMP) formation associated with reduced availability of endothelial progenitor cells (EPC) in our cohort, may be suggested as one of the pathogenic mechanisms involved in endothelial dysfunction9. Interestingly, a short treatment with low-dose GC was associated with consistent reduction of both arterial stiffness and the EMP/EPC ratio, suggesting beneficial effects on vascular endothelium homeostatis22,23.
Although a negative effect of longterm GC therapy on vascular tone and endothelial dynamic structure cannot be ruled out, rapid and profound inhibition of systemic inflammation induced by low-dose GC may quickly attenuate CV risk in systemic disorders characterized by high degree of inflammation. We believe that deeper knowledge of the effects of GC on the complex network of mechanisms leading to CV damage will be valuable in preventing CV events in patients with systemic inflammatory diseases.