Accurate timing of BNT162b2 boosters to prevent breakthrough infections of coronavirus disease 2019 (COVID-19) requires reliable estimates of immune status. We hypothesized that spike IgG levels at 3 months after two doses of the BNT162b2 vaccine might predict subsequent spike IgG levels. Spike IgG levels were tested at 3, 6, and 8 months after the second dose of the BNT162b2 vaccine in 251 Japanese health care workers (median age: 39 years, female: 187). The median level of spike IgG was 2,882 AU/mL at 3 months. This decreased to 875 AU/mL at 6 months and 579 AU/mL at 8 months. There were good correlations of log-transformed spike IgG levels between 3 and 6 months (r = 0.86) and between 3 and 8 months (r = 0.82). The correlation further improved after excluding three subjects who had possible COVID-19 infections (r = 0.91, r = 0.86). Log-transformed spike IgG levels at 6 or 8 months yields the following equation: log spike IgG at 6 (8) months = 0.92 (0.86) X log spike IgG at 3 months- 0.23 (0.18). Predicted spike IgG at 6 months of ≥ 300 or < 300 AU/mL had 98% sensitivity, 47% specificity, and 94% accuracy for discriminating subjects whose actual spike IgG titers at 6 months were above or below 300 AU/mL. Corresponding values of predicted spike IgG at 8 months were 97%, 70%, and 93%, respectively. We conclude that predictive formulae using spike IgG levels at 3 months after two-dose vaccination with BNT162b2 reliably estimate subsequent spike IgG levels up to 8 months and provide useful information in terms of vaccination booster timing.