A dynamic model of pneumococcal infection in the United States: implications for prevention through vaccination

Universal infant vaccination with the 7-valent pneumococcal conjugate vaccine (PCV7) has nearly eliminated PCV7-serotype invasive pneumococcal disease (IPD) in young U.S. children, but has been accompanied by increases in the incidence of serotype 19A IPD. Because antibiotic-non-susceptible 19A has increased more than antibiotic-susceptible 19A, antibiotic selection pressure could be contributing to this trend. We developed a dynamic compartmental transmission model of pneumococcus to better understand the causes of this rise and to estimate the impact of vaccines or changes in antibiotic use on future IPD incidence in the U.S. in <2 year-olds. The model predicted that with current practices, serotype 19A IPD incidence will plateau at about the 2007 level over the next few years. The model suggests that antibiotic usage played a major role in the rise in antibiotic-non-susceptible 19A IPD, with a lesser contribution from PCV7 vaccination. However, hypothetical large decreases in antibiotic use starting in 2008 are predicted to yield only gradual decreases in antibiotic-non-susceptible 19A IPD. On the other hand, vaccines with modest (20%) effectiveness against 19A (or 6A or PCV7-serotypes) carriage are predicted to substantially (by 80%) decrease the incidence of IPD caused by those serotypes within 10 years of implementation. Our findings highlight that vaccine effects on colonization are key to their overall benefits. In addition, serotype changes following vaccine introduction may have multifactorial origins, with antibiotic use an important factor for resistant strains such as 19A.