IJCS | Volume 31, Nº3, May/ June 2018

303 Figure 2 - Meta-analysis of the studies investigating the association between increased cystatin C levels and the development of cardiovascular death or non-fatal infarction by the comparison of the fourth and the first quartiles of cystatin C. Valor de Z 5,094 4,939 4,182 8,179 Valor de p 0,000 0,000 0,000 0,000 Martucheli et al. Cystatin C and acute coronary syndromes International Journal of Cardiovascular Sciences. 2018;31(3)290-307 Review Article increased cystatin C levels and the risk for cardiovascular death or non-fatal myocardial infarction is independent of patient’s kidney function. Analysis of the studies that classified patients according to cystatin C tertiles or quartiles showed that patients with higher cystatin C levels were also older, which results from a progressive, physiological decrease in GFR associated with aging. 34 However, 58.8% (n = 10) of the studies included age in the multivariate analysis, including the two studies included in the meta-analysis, indicating that the association between increased cystatin C levels and worse cardiovascular prognosis is independent of age. All studies assessed patients’ kidney function, and most of them (82.4%, n = 14) (including the two studies included in the meta-analysis) used GFR, which is a better marker of kidney function than serum creatinine. 35 Serum creatinine levels may be affected by several factors like muscle mass, age, sex, and hence, it is not specific for assessment of kidney function. 32 Besides, increases in serum creatinine occur only when there is a decrease greater than 50% in glomerular ultrafiltration, and thereby is not considered a sensitive marker for assessment of kidney function. 36 Determination of GFR by calculation of creatinine clearance or equations based in serum creatinine levels may mitigate or eliminate these limitations. 36 The most common equations used to estimate GFR are the Cockcroft & Gault, MDRD and CKD-EPI equations, which include clinical and demographic variables in place of physiological factors known to affect creatinine serum concentrations. 37 Classification of patients according to cystatin C levels was heterogeneous in the studies. A considerable number of these studies (58.8%) classified patients in quartiles or tertiles, which may have influenced the results. It is easier to obtain a correlation of increased cystatin C levels with poor prognosis when patients in the fourth quartile or third tertile (who have higher levels of cystatin C) are compared with patients in the first quartile or first tertile (whose cystatin C levels are decreased) than in comparison between patients with cystatin C levels above and below reference/median values. Nevertheless, classification of cystatin C levels in quartiles and tertiles is of greater clinical value, since it may be used in the determination of cutoff points above which the risk of cardiovascular events and mortality is significantly greater. Therefore, only studies in which patients were classified by cystatin C quartiles, and higher quartiles were compared with lower quartiles were included in the meta-analysis. Immunonephelometry and immunoturbidimetry are the most used methods of cystatin C determination, 38 which has been confirmed in this systematic review, since 84.2% (n = 14) of the studies used these methods for cystatin C measurement, and only 3 studies used other methods or did not mention the method used. Immunonephelometry and immunoturbidimetry are the methods of choice for determination of cystatin C levels in body fluids due to their high accuracy, convenience, automation, in addition to being simple and fast for daily routine. 38 Besides, immunonephelometry has been suggested as a better method than immunoturbidimetry