Background <p>Lipoprotein(a) (Lp(a)) is an independent risk factor for fatal cardiovascular disease (CVD) in dialysis patients; however, evidence regarding its association with CVD, particularly nonfatal events, remains inconsistent. This study aimed to evaluate the relationship between Lp(a) levels at dialysis initiation and subsequent CVD events in Japanese patients.</p> Methods <p>We conducted a single-center, retrospective, observational cohort study of patients who initiated dialysis in Japan between January 2015 and December 2019. Patients were classified into low (&lt; 30&#xa0;mg/dL) and high (≥ 30&#xa0;mg/dL) Lp(a) groups on the basis of Lp(a) levels measured at dialysis initiation. The primary outcome was the occurrence of CVD events (death from CVD, hospitalization due to coronary artery disease, arteriosclerosis obliterans, stroke, and congestive heart failure). The patients were monitored for up to 5&#xa0;years. The associations were evaluated using a Cox proportional hazards model.</p> Results <p>Among 181 eligible patients, 69 (38.1%) had high Lp(a) levels and 64 patients (35.4%) experienced CVD events during the follow-up. The 5-year CVD-free survival rate was significantly lower in the high Lp(a) level group than in the low Lp(a) level group (44.1% versus 63.1%; log-rank test, <i>p</i> = 0.008). Elevated Lp(a) was independently associated with increased CVD risk using Cox proportional hazard models adjusted for traditional CVD risk factors (hazard ratio: 1.99; 95% confidence interval: 1.19–3.34; <i>p</i> = 0.012).</p> Conclusions <p>Lp(a) levels ≥ 30&#xa0;mg/dL at dialysis initiation were an independent predictor of cardiovascular events, including nonfatal outcomes.</p>