Thyroid peroxidase antibodies (TPOAb) in pregnancy are a marker for postpartum (PPTD) and long-term thyroid dysfunction, with variable sensitivity and specificity in PPTD prediction. To test its utility in prediction, we recruited 308 TPOAb-positive (147 developed PPTD (PPTD group) and 161 remained euthyroid [PPTE group]) and 102 TPOAb-negative women (none developed PPTD), in early pregnancy (median, 18; range, 9-19 weeks' gestation). TPOAb levels were higher in the PPTD group (median) (125.2 kIU/L; p < 0.001), and in its hypothyroid (162.4 kIU.; p < 0.0001), hyperthyroid (114.2 kIU/L; p < 0.007), and biphasic (105.1 kIU/L; p < 0.02) variants, compared to the PPTE group (66.7 kIU/L) The incidence of PPTD was significantly higher with TPOAb levels above 58.2 kIU/L (early pregnancy versus postpartum; relative risk, 1.37 [95% confidence interval [CI] 1.17-1.61] versus 0.78 [95% CI 0.5-1.2]) compared to levels below. The integrated postpartum TPOAb response was higher in the PPTD group (median) (159 kIU/L per week) and its variants (hypothyroid; 199 kIU/L per week; biphasic, 180 kIU/L per week; hyperthyroid, 120 kIU/L per week), compared to the PPTE group (86 kIU/L per week p < 0.004). Median early pregnancy TPOAb levels in the PPTD and PPTE groups correlated well with the postpartum antibody response (r = 0.58, p < 0.001). The sensitivity of TPOAb in PPTD prediction was 100% (early pregnancy and postpartum), specificity 62% (early pregnancy) versus 41% (postpartum) and positive predictive value 48% (early pregnancy and postpartum). The timing of TPOAb testing, the sensitive assay used and the absence of PPTD in TPOAb-negative subjects contributed to this high sensitivity. We recommend TPOAb in early pregnancy as a useful predictor of PPTD, particularly in populations where PPTD does not occur in TPOAb-negative women.