Background: Low admission low-density lipoprotein cholesterol (LDL-C) is paradoxically associated with worse outcomes after acute coronary syndrome, but this may reflect confounding rather than causation. We examined the paradox in statin-naïve patients. Methods: We studied 388 statin-naïve patients with a first ST-segment elevation myocardial infarction (STEMI) treated with primary percutaneous coronary intervention (PCI) and followed for up to five years. Admission LDL-C was analyzed continuously and as three categories (130 mg/dL), with all-cause mortality assessed using Kaplan–Meier, Cox regression, restricted cubic splines, and landmark sensitivity analyses. Results: Crude mortality was highest in the lowest LDL-C group (20.0% vs. 8.3% vs. 10.7%; p = 0.014), and LDL-C < 100 mg/dL predicted higher mortality (hazard ratio 2.03, 95% CI 1.02–4.03). After adjustment, this remained non-significant across the parsimonious and fully adjusted models (adjusted HR 1.27–1.43, all 95% CIs including 1); older age, lower ejection fraction, and diabetes were independent predictors of death. The lowest stratum also had lower albumin and higher CONUT scores, consistent with a frailty phenotype. Conclusions: In statin-naïve STEMI patients undergoing primary PCI, the lipid paradox reflected age- and frailty-related confounding rather than protection; low admission LDL-C marks a higher-risk phenotype and should not discourage guideline-directed lipid-lowering therapy.