Associations between statins and all-cause mortality and cardiovascular events among peritoneal dialysis patients: A multi-center large-scale cohort study

To the Editor: The high morbidity and mortality related to cardiovascular disease (CVD) in the peritoneal dialysis (PD) population has given rise to the urgent need to explore interventional strategies. As hyperlipidemia is closely related to CVD, lipid management is essential. However, hyperlipidemia is a nutritional marker in dialysis patients and has been shown to have a reverse epidemiological phenomenon in the prediction of mortality. These paradoxical findings raise the perennial question of whether statins are truly beneficial for PD patients. The latest guidelines recommend that dialysis patients continue but not initiate statins; however, some small sample observational studies showed the beneficial role of statins in PD patients. Thus, we conducted this multi-center large-scale cohort study to investigate the association between statins and prognosis in the PD population. This study was affiliated with the PD Telemedicine-assisted Platform (PDTAP) Cohort Study, a nationwide large-scale PD cohort study in China. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of Peking University First Hospital (No. 2018-100). The inclusion criteria were: (1) aged >14 years; (2) diagnosed with end-stage kidney disease; (3) receiving PD >3 months. The exclusion criteria were: (1) receiving temporary PD because of acute kidney injury; (2) being treated with non-statin lipid-lowering drugs. The included parameters were demography, comorbidity, dialysis-related data, and drug information, which were recorded at baseline by the PD staff; and laboratory data that were exported directly from the laboratory information management system of hospitals, or input at baseline by the PD staff. Information on statins included whether they were taken and their types, which were collected at baseline and at each month's visit. The time ratio of statins was defined as [duration of statins (months) / follow-up time (months)] and only included patients with a follow-up time of over 6 months. The primary outcome was all-cause mortality, and the secondary outcomes were major adverse cardiovascular events (MACE) and modified MACE+. Data were expressed as mean ± standard deviation (SD) or median (Q1, Q3). The t-test, non-parametric test, or chi-squared test were used to compare baseline characteristics. Cox regression and Kaplan–Meier analysis were performed to analyze the association between variables and outcome events by HR (95% confidence intervals [CIs]). Propensity score matching (PSM, 1:1) was used to match the patients in the statins and non-statins group, with the caliper value of 0.02. A linear mixed model was used to evaluate the effect of statins on lipid trends. Statistical analysis was performed using SPSS 27.0 (IBM, Armonk, NY, USA) and R 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria). All probabilities were two-tailed, and the significance level was 0.05. There were 7429 PD patients enrolled from 27 hospitals, and they were followed up for 29.00 (14.00, 45.00) months. Among them, 1380 (18.6%) patients died, and the the leading cause of mortality was CVD (843 [4.71/100 person-year]) [Figure 1A]. MACE and MACE+ occurred in 1059 (6.05/100) and 1857 (10.95/100) patients, respectively. A total of 2211 (29.8%) patients had used statins, with the median time of medication usage as 16.73 (6.70, 31.00) months, and the leading drug was atorvastatin (81.4%).Figure 1: Association between the use of lipid-lowing drugs and outcomes in the PD population. (A) Flowchart of this study. (B) Cox regression analysis. Normal font and circle: univariable analysis; bold font and triangle: multivariable analysis (variables included in multivariable analysis: age, annual income, health insurance, hospital grade, Eastern Cooperative Oncology Group index, Charlson comorbidity index, hemoglobin, albumin, creatinine, calcium, sodium, potassium, inflammation status, renal, and peritoneal Ccr). MACE: Major adverse cardiovascular events; PSM: Propensity score matching; PD: Peritoneal dialysis.Statins users were characterized as older, fatter, having higher education and annual income, living in an urban area with medical insurance for urban residents, and having higher Charlson comorbidity indexes (P <0.001, [Supplementary Table 1, https://links.lww.com/CM9/C569]). They had better nutritional status as indicated by hemoglobin, albumin, calcium, potassium, triglycerides, and total cholesterol levels (P <0.001 or 0.010) and had better renal Kt/Vure (P = 0.010). We divided the statins users who were followed up for more than 6 months into long-term and short-term users by the median value of time ratio of taking statins, and the long-term users showed similar characteristics to the statins users [Supplementary Table 2, https://links.lww.com/CM9/C569]. After being followed up for 29.00 (14.00, 45.00) months, the incidence of all-cause mortality was lower in the statins group (427 [6.85/100 person-year] vs. 953 [8.17/100], P = 0.004], which was mainly due to a lower incidence of CVD-related death (252 [4.04/100] vs. 591 (5.06/100), P = 0.007]. The incidence of MACE (388 [6.49/100] vs. 671 [5.83/100], P = 0.051) and MACE+ (599 [10.22] vs. 1258 [11.34], P = 0.090) were comparable. See Supplementary Table 3, https://links.lww.com/CM9/C569. The association of statins with a lower risk of all-cause mortality (HR 0.489, 95% CI 0.393-0.608, P <0.001) was proven using Cox multivariable regression analysis after adjustment for other significant predictors. Compared with simvastatin, atorvastatin (HR 0.663, 95% CI 0.487–0.903, P = 0.009) and pitavastatin (HR 0.474, 95% CI 0.245–0.918, P = 0.027) users were associated with lower mortality upon univariate regression analysis, but not with multivariate analysis. The time ratio of statins was not significant with multivariable analysis (HR 1.364, 95% CI 0.927–2.077, P = 0.115). Seen in Figure 1B and Supplementary Table 4, https://links.lww.com/CM9/C569. The usage of statins was associated with a lower level of modified MACE+ after Cox regression analysis (HR 0.691, 95% CI 0.577–0.827, P <0.001), but it was not associated with MACE. Compared with simvastatin, rosuvastatin (HR 3.588, 95% CI 1.111–11.588, P = 0.033) was associated with a higher level of MACE+ upon multivariate regression analysis. Similar to mortality, the time ratio of statins was associated with a higher risk of MACE and MACE+ upon univariate analysis (HR and 95% CI >1.000, P ≤ 0.001), but became insignificant with multivariate analysis. Seen in Figure 1B and Supplementary Table 4, https://links.lww.com/CM9/C569. After PSM [Supplementary Table 5 https://links.lww.com/CM9/C569], there was still an association between statins and a lower risk of all-cause mortality (χ2 = 23.89, P <0.001) and MACE+ (χ2 = 10.37, P = 0.001) after Kaplan–Meier analysis. All the lipid parameters were relatively stable over the first year in the whole PSM cohort (P >0.050) [Supplementary Table 6 https://links.lww.com/CM9/C569]. In the PDTAP cohort, we found that only one-third of patients took statins, and there was an association between statins and a reduced risk of all-cause mortality and MACE+. After being adjusted for potential confounders with multivariate analysis and with PSM for individual-level factors, there was still a beneficial effect on outcome events of the statins users. The possible mechanisms between statins and better outcomes are as follows: First, statins inhibit inflammation activation. Cohort studies have demonstrated that statins can effectively reduce the levels of inflammatory cytokines.[1] Second, statins improve endothelial cell function by increasing endothelial nitric oxide synthase activity and promoting vasodilation, which helps to reduce the incidence of cardiovascular events.[2] In our cohort, the changes in lipids levels in statins and non-statins users showed no difference during the observation period using a linear mixed model. We speculated that this was due to the follow-up time being insufficient. The median duration of statins was 16.83 months in our cohort, vs. 4.9 years in the SHARP trial in which the lipids levels were significantly increased after statins treatment.[3] We also found that compared with simvastatin, rosuvastatin was associated with a higher incidence of MACE+. A meta-analysis including 11,000 patients from 13 trials reported similar results of lipophilic statins being superior to hydrophilic rosuvastatin in mortality.[4] Lipophilic statins (atorvastatin, simvastatin, and pitavastatin) are non-hepatoselective and penetrate more easily through cell membranes, which may explain the better prognosis. Due to insufficient evidence in the dialysis population of our data, there is a need to explore via randomized trials which statins have advantages in improving the clinical prognosis of PD patients. This is a national-level study to investigate the association of statins and outcomes in the PD population. The results were robust with multivariable and PSM analysis adjusted for many covariates, and the relationships of PD outcomes with several different statins were compared. However, as an observational study, it could not provide causality findings and the presence of residual biases cannot be ignored. The doses and possible mechanisms of the statins were not included, which we will have to improve on in our future study. Through this multi-center PD cohort, the association of statins with a lower risk of all-cause mortality and MACE+ has been proven. However, further investigation into statins would be needed in the PD population.