Strongest evidence: L-Carnitine has high-strength evidence for reducing fasting blood glucose (7 of 7 studies beneficial, small effect, 2000–3000 mg/day) and C-reactive protein (6 of 7 studies beneficial, moderate effect, 2–3 g/day). High-strength evidence also supports LDL cholesterol reduction (4 of 4 studies, moderate effect, 50 mg/kg/day to 3000 mg/day). Moderate-strength evidence shows consistent benefits for triglycerides (5 of 8 studies, small effect, 1000–3000 mg/day), hemoglobin A1c (4 of 4 studies, mixed effect, 1–2 g/day), HOMA-IR (4 of 4 studies, small effect, ≥2 g/day), body weight (4 of 4 studies, moderate effect, ~3000 mg/day), and serum albumin in critically ill patients (3 of 4 studies, small effect, 3 g/day).
Mixed or weaker evidence: For LDL cholesterol, moderate-strength evidence from 7 studies shows only 4 reporting beneficial effects (small), with 3 neutral, particularly in hemodialysis patients. Body mass index reduction (4 of 6 studies beneficial, small effect, 1–4 g/day) and total cholesterol reduction (4 of 6 studies beneficial, mixed effect, 1–3 g/day) show moderate but inconsistent results, with some neutral findings possibly due to dose differences (e.g., 1 g/day vs. 3 g/day). No outcomes in the database had low or very low evidence strength.
Effective dose patterns: Across multiple outcomes, effective doses converge on 2000–3000 mg/day for glycemic and inflammatory markers, with 1–2 g/day commonly used for lipid improvements. For body weight and albumin, 3 g/day was the most-studied dose. One consistent threshold is that doses ≥2 g/day appear more reliably effective for insulin resistance (HOMA-IR).
Population insights: Benefits are most consistently observed in clinical populations with metabolic or inflammatory conditions — including type 2 diabetes, PCOS, overweight/obesity, and hemodialysis patients. Effects in healthy individuals are not well established, as most studies recruited people with existing metabolic or inflammatory conditions. Two outcomes (serum albumin, CRP) drew almost exclusively from acute hospital settings (ICU, sepsis), limiting generalizability.
Notable caveats: A major caveat across most syntheses is the possibility of publication bias — null-result studies may be less likely to be published. Many studies had short durations (e.g., median 7 days for CRP), small sample sizes, and did not specify the form of L-carnitine used (e.g., L-carnitine tartrate vs. acetyl-L-carnitine), limiting form-specific conclusions. Effect sizes were predominantly small, and heterogeneity was high in several meta-analyses (e.g., I²=93% for BMI).