<b><i>Background:</i></b> Total serum bilirubin (TSB) is used in managing neonates with jaundice, but clear evidence on its association with major outcomes is lacking. <b><i>Objectives:</i></b> We evaluated the association between TSB and kernicterus spectrum disorder (KSD). <b><i>Methods:</i></b> We searched PubMed, EMBASE, and CENTRAL till July 2021. Two authors independently selected relevant cohort studies, extracted data (CHARMS checklist), assessed risk of bias (RoB) (QUIPS tool), and rated certainty-of-evidence (Grades of Recommendation, Assessment, Development, and Evaluation). We pooled adjusted odds ratio (aOR) (random-effect) via generic inverse variance methods. <b><i>Results:</i></b> From 2,826 records retrieved, we included 37 studies (<i>n</i> = 648,979). Fifteen studies had low, 16 moderate, and 6 high RoB, with majority having concerns on confounder adjustment and statistical analysis. Twenty-two studies contributed meta-analysis data, and 15 were summarized narratively. TSB appears associated with KSD in infants with certain risk factors (aOR 1.10, 95% CI: 1.07–1.13; 5 studies [<i>n</i> = 4,484]). However, TSB (aOR 1.10, 95% CI: 0.98–1.23; 1 study [<i>n</i> = 34,533]) or hyperbilirubinemia (aOR 1.00, 95% CI: 0.51–1.95; 2 studies [<i>n</i> = 56,578]) have no clear association with kernicterus or neurological diagnosis in overall neonatal population (moderate-certainty-evidence). One study shows that infants with hyperbilirubinemia appear likelier to develop attention-deficit disorder (aOR 1.90, 95% CI: 1.10–3.28) and autistic spectrum disorder (aOR 1.60, 95% CI: 1.03–2.49, <i>n</i> = 56,019) (low-certainty-evidence). Certain clinical factors appear associated with KSD, although very few studies contributed to the analyses. <b><i>Conclusions:</i></b> Despite the importance of this question, there is insufficient high-quality evidence on the independent prognostic value of TSB for adverse neurodevelopmental outcomes in most neonatal populations. Future studies should incorporate all known risk factors alongside TSB in a multivariable analysis to improve certainty-of-evidence.