Trends in intussusception-associated hospitalizations and deaths among US infants

Context. The newly licensed tetravalent rhesus-human reassortant rotavirus vaccine has been withdrawn following reports of intussusception among vacci nated infants. Objective. To describe the epidemiology of intussusception-associated hospi talizations and deaths among US infants. Design. This retrospective cohort study examined hospital discharge data fr om the National Hospital Discharge Survey for 1988-1997, Indian Health Serv ice (IHS) for 1980-1997, California for 1990-1997, Indiana for 1994-1998, G eorgia for 1997-1998, and MarketScan for 1993-1996, and mortality data from the national multiple cause-of-death data for 1979-1997 and linked birth/i nfant death data for 1995-1997. Patients. Infants (< 1 year old) with an International Classification of Di seases, Ninth Revision, Clinical Modification code for intussusception (560 .0) listed on their hospital discharge or mortality record, respectively. Results. During 1994-1996, annual rates for intussusception-associated infa nt hospitalization varied among the data sets, being lowest for the IHS (18 per 100 000; 95% confidence interval [CI] = 9-35 per 100 000) and greatest for the National Hospital Discharge Survey (56 per 100 000; 95% CI = 33-79 per 100 000) data sets. Rates among IHS infants declined from 87 per 100 0 00 during 1980-1982 to 12 per 100 000 during 1995-1997 (relative risk = 7.6 , 95% CI = 3.2-18.2). Intussusception-associated hospitalizations were unco mmon in the first 2 months of life, peaked from 5 to 7 months old, and show ed no consistent seasonality. Intussusception-associated infant mortality r ates declined from 6.4 per 1 000 000 live births during 1979-1981 to 2.3 pe r 1 000 000 live births during 1995-1997 (relative risk = 2.8, 95% CI = 1.8 -4.3). Infants whose mothers were <20 years old, nonwhite, unmarried, and h ad an education level below grade 12 years were at an increased risk for in tussusception-associated death. Conclusions. Intussusception-associated hospitalization rates varied among the data sets and decreased substantially over time in the IHS data. Althou gh intussusception-associated infant deaths in the United States have decli ned substantially over the past 2 decades, some deaths seem to be related t o reduced access to, or delays in seeking, health care and are potentially preventable.