Background: Subclinical hypothyroidism (SCH) has been reported to be associated with adverse pregnancy outcomes, however universal screening and treatment is controversial. Objectives: Our objectives were to determine population-specific pregnancy reference values (R1) for serum thyroid-stimulating hormone (TSH) and free thyroxine (FT<sub>4</sub>) at 14 weeks' gestation, along with the prevalence of SCH and thyroid peroxidase antibody (TPOAb). Methods: This was a prospective hospital-based cohort study. 1,402 subjects were recruited. Blood samples were obtained from 769 singleton pregnancies due to default between recruitment and scheduled blood draw. The prevalence of SCH was determined using R1, the laboratory non-pregnant reference values (R2) and previously recommended pregnancy reference values (R3). Results: R1 for TSH and FT<sub>4</sub> was 0.03-3.17 mU/l (mean ± SD, 1.1 ± 0.76) and 8.85-17.02 pmol/l (mean ± SD, 11.96 ± 2.06), respectively. The prevalence of SCH using reference values R1, R2 and R3 was 1.4% (11/769), 0.5% (4/769) and 1.9% (15/769). Prevalence was significantly greater using R3 when compared to R2 (p = 0.011). TPOAb prevalence was 2.6%. A significantly greater prevalence of TPOAb was found in subclinical hypothyroid subjects using all three reference values than in euthyroid subjects (∼25 vs. 2%, p < 0.05). Conclusions: These reference values are the first to be reported for an Afro-Caribbean population. Our findings support the use of pregnancy-specific reference values in our population.