We examined multiple cause-of-death mortality for children 1–59 months of age reported via death certificates to NCHS for 2005–2016, excluding 2007, the transitional year after vaccine introduction. Diarrhea-associated deaths were defined using ICD-10 codes listed in any position on the death certificate (Table, Supplemental Digital Content 1, http://links.lww.com/INF/D283).

Using national census data, we calculated overall diarrhea-associated death rates, as well as death rates by sex, age group and race, per 100,000 children <5 years of age. We report rate ratios and 95% CI comparing each of 3 postvaccine periods (2008–2010, 2011–2013 and 2014–2016) to the prevaccine period (2005–2006) using Poisson regression. We calculated the proportion of deaths occurring in the rotavirus season (January–June) as compared with the nonrotavirus season (July–December) by age group, comparing the proportion who died in-season in each post vaccine period to the prevaccine reference period using a χ² test with postvaccine<0.05 considered statistically significant.

During 2005–2016, 4355 diarrhea-related deaths among 1- to 59-month-olds were reported to NCHS. In the prevaccine period, 32 (3%) of diarrheal deaths included specific codes indicating viral-diarrhea, while in the postvaccine periods, this was specified in 30 (2%), 20 (2%) and 34 (3%) deaths, respectively. Males, 1- to 11-month-olds, and black children had higher death rates in all time periods (Table 1). The overall death rate among children <5 years was 2.3/100,000 in the prevaccine period. This remained stable in 2008–2010, but decreased to 1.7/100,000 in both 2011–2013 and 2014–2016 [rate ratio: 0.73 (95% CI: 0.67–0.80) and 0.71 (95% CI: 0.65–0.78), respectively], indicating a 26.7% reduction (95% CI: −32.9 to −19.9) in 2011–2013 and a 28.6% (95% CI: −34.7 to −21.9) reduction in 2014–2016. By 2014–2016, declines in death rates were seen among children 1–23 months of age (Fig., Supplemental Digital Content 2, http://links.lww.com/INF/D284), both sexes, and black and white children, compared with 2005–2006.

In the prevaccine period, a higher proportion of diarrheal deaths occurred in the rotavirus season among all age groups, except for those 1–3 months of age. Among 4- to 11-month-olds, 56% of deaths occurred in-season in the prevaccine period, while 51% occurred in-season in both 2011–2013 and 2014–2016. For 12- to 23-month-olds, 57% of deaths occurred in-season before vaccine introduction, 53% during 2011–2013 and 58% in the last postvaccine period. Finally, among the oldest 24- to 59-month age group, 68% of deaths occurred in-season in 2005–2006, and 61%, 59% and 58% in each successive postvaccine period, respectively. None of these changes were statistically significant. Among 1- to 3-month-olds, 44% of deaths occurred in-season in the prevaccine period, increasing to 50%, 51% and 50% in 2008–2010, 2011–2013 and 2014–2016, respectively (P < 0.05 for each postvaccine period compared with prevaccine).

Compared with the 2 years before rotavirus vaccine implementation, diarrheal death rates among US children <5 years of age old decreased significantly during 2010–2016, concurrent with increase in full-series rotavirus vaccine coverage from 43.9% in 2009 to 74.1% in 2016 in 18- to 35-month-old children.^5,6 This is in line with findings from Mexico and Malawi, where, after introduction of rotavirus vaccine, a 41% and 31% reduction in diarrhea-associated deaths were reported, respectively.^7,8 Additionally, the decline in mortality in our study was greater and achieved faster by 4- to 11-month-olds, who experienced a 50% decline in mortality by 2011–2013; in 2014–2016, the 11- to 23-month-olds also experienced a similar decrease. These age groups bear the greatest burden of rotavirus disease, and collectively these data support that the declines observed may be, in part, due to the effect of vaccination. However, we did not observe greater declines in diarrheal deaths during the rotavirus season months. We also recorded a decrease in diarrheal mortality among 1- to 3-month-olds who were too young to be vaccinated against rotavirus, suggesting other factors may also contribute to the decreasing diarrhea-associated deaths among US children.

In each period analyzed, racial disparities existed. While death rates decreased among both white and black children during the study period, black children had between 3–5.5 times the death rate of white children. While the general decline in death rates is encouraging, further exploration of reasons for the higher death rates among blacks and interventions targeted at further reducing the rate, such as improving rotavirus vaccine coverage, which trails that for whites by 10 percentage points⁵ in the most recent coverage estimates, are warranted.

This analysis was subject to several limitations. First, an error in the ICD coding system before 2005 precluded inclusion of more than 2 years of prevaccine data, thus limiting comparison years. Next, the timing and geospatial pattern of the rotavirus season has shifted in the US since vaccine introduction² and we were unable to adjust for these changes, thereby potentially masking effects of shifting seasons on in-season declines in mortality. Finally, we did not have enough individual level data on rotavirus as the etiology of diarrheal death, thus we cannot definitely attribute the decline in diarrheal deaths in our analysis to the rotavirus vaccine program.

All-cause diarrheal mortality has declined among all <5-year-olds in the later years post-rotavirus vaccine implementation compared with prevaccine years; further research to assess the role of rotavirus vaccine in decreasing pediatric deaths is warranted.