Primary carnitine deficiency (PCD) is an autosomal recessive disorder caused by mutations in the SLC22A5 gene, which encodes the organic cation transporter 2 (OCTN2). These mutations impair carnitine transport and fatty acid metabolism, leading to a wide range of clinical symptoms, from mild fatigue to severe cardiomyopathy. Although over 100 mutations have been identified, the correlation between specific genotypes and pathogenicity remains unclear, necessitating further investigation into their molecular and clinical impacts. This systematic review aims to evaluate the association between SLC22A5 mutations and PCD, with a focus on cardiomyopathy, to characterise the molecular and functional consequences of these variants, and to explore their implications for diagnosis, treatment, and personalised care. A systematic search of PubMed was conducted, focusing on studies reporting genetic, clinical, and biochemical data related to SLC22A5. Variants were classified according to ACMG guidelines and CADD scoring, and the data were synthesised through descriptive analyses. Most reported variants are missense mutations, accounting for approximately 80% of cases, followed by deletions (12.7%), while duplications, insertions, and deletion-insertion variants are rare. According to ACMG classification, the majority of variants are pathogenic or likely pathogenic (≈72%), with VUS comprising about 24%, and benign variants being uncommon. Exons 1 and 8 have the highest frequencies, whereas Exons 5 and 6, despite fewer variants, exhibit higher mean CADD scores, suggesting greater functional impact. The findings revealed that Exons 1 and 8 harboured the most frequent mutations, while Exons 5 and 6 showed the highest pathogenicity. Nonsense variants exhibit the highest pathogenicity and CADD scores, followed by frameshift and missense mutations. Severe truncating mutations were linked to early-onset cardiomyopathy. Intronic variants are infrequent but include sites with high CADD scores, indicating potential regulatory relevance. Overall, these findings demonstrate that variant pathogenicity is influenced more by functional impact than by frequency alone.