ObjectivesPatients with juvenile‐onset systemic lupus erythematosus (JSLE) have increased atherosclerosis risk. This study investigated novel atherosclerosis progression biomarkers in the APPLE trial, the largest investigator‐led randomised control trial of atorvastatin versus placebo for atherosclerosis progression in JSLE, using carotid intima‐media thickness (CIMT) as primary outcome.MethodsUnsupervised clustering of baseline CIMT and CIMT‐progression over 36 months was used to stratify JSLE patients. Disease characteristics, cardio‐vascular risk scores and baseline serum metabolome were investigated in CIMT‐stratified patients. Machine learning techniques were used to identify/validate a serum metabolomic signature of CIMT progression.ResultsBaseline CIMT stratified JSLE patients (N=151) into three groups with distinct high, intermediate, and low CIMT trajectories irrespective of treatment allocation, despite most patients having low cardiovascular disease‐risk based on recommended assessment criteria. In the placebo group (N=60), patients with high vs, low CIMT‐progression had higher total (P=0.001) and low‐density lipoprotein (LDL) (P=0.002) cholesterol levels, although within the normal range. Furthermore, a robust baseline metabolomic signature predictive of high CIMT‐progression was identified in the placebo arm (area under the curve‐AUC 80.7%). Patients treated with atorvastatin (N=61) had reduced LDL cholesterol levels after 36 months as expected, however, despite this, 36% still had high atherosclerosis progression, which was not predicted by metabolomic biomarkers, suggesting non‐lipid drivers of atherosclerosis in JSLE with management implications for this subset of patients.ConclusionSignificant baseline heterogeneity and distinct subclinical atherosclerosis progression trajectories exist in JSLE. Metabolomic signatures can predict atherosclerosis progression in some JSLE patients with relevance for clinical trial stratification.image