Seroprevalence of Cytomegalovirus among Children 1 to 5 Years of Age in the United States from the National Health and Nutrition Examination Survey of 2011 to 2012 Tatiana M. Lanzieri,a Deanna Kruszon-Moran,b Minal M. Amin,a Stephanie R. Bialek,a Michael J. Cannon,c Margaret D. Carroll,b Sheila C. Dollarda
Cytomegalovirus (CMV) seroprevalence among U.S. children 1 to 5 years old was assessed in the National Health and Nutrition Examination Survey of 2011 to 2012. The overall seroprevalence (95% confidence interval) of IgG was 20.7% (14.4 to 28.2%), that of IgM was 1.1% (0.4 to 2.4%), and that of low IgG avidity was 3.6% (1.7 to 6.6%), corresponding to a 17.3% (10.1 to 26.7%) prevalence of recent infection among IgG-positive children.
C
ongenital cytomegalovirus (CMV) infection is a major cause of sensorineural hearing loss and developmental disabilities in children in the United States (1). An estimated 0.7% of U.S. infants are born with congenital CMV infection (1). Early in life, CMV infection generally results from mother-to-infant transmission through exposure to the virus in genital secretions during birth or postnatally via breast milk (2). CMV seroprevalence continues to rise throughout early childhood, as infants and young children acquire CMV infection via exposure to body fluids from other infected individuals, especially from close contact with young children in household or day care settings (3, 4). Young CMV-seropositive children, who may shed virus in body fluids for months after primary CMV infection, are thought to be an important source of CMV transmission to adults (5). Population-based estimates of CMV seroprevalence among young children in the United States have not been previously reported. In this study, we describe CMV seroprevalence among U.S. children 1 to 5 years old, who were sampled in the National Health and Nutrition Examination Survey of 2011 to 2012 (NHANES 2011-2012). The methods of the NHANES, a nationally representative, cross-sectional survey of the civilian noninstitutionalized U.S. population, have been published elsewhere (6). NHANES 20112012 oversampled non-Hispanic Asians, in addition to Hispanics (including Mexican Americans), non-Hispanic blacks, and lowincome persons who were non-Hispanic whites or designated as “other” to increase sample size and obtain more statistically reliable estimates of these subgroups (6). National estimates of CMV seroprevalence for CMV IgG, IgM, and low IgG avidity were calculated using the exam weights developed for the NHANES to represent the total civilian noninstitutionalized U.S. population and to account for oversampling and nonresponse to the household interview and physical examination (7). Seroprevalence of CMV IgG was examined by age (in years), sex, race/Hispanic origin, and poverty index ratio. Information on race/Hispanic origin was collected by proxy, and individuals were categorized as non-Hispanic white, non-Hispanic black, non-Hispanic Asian, Mexican American, other Hispanic, and other race (which includes multiracial) (8). In our analysis, we combined Mexican American and other Hispanic to form the Hispanic group. The “other” race group (n ⫽ 39) is included in overall
February 2015 Volume 22 Number 2
estimates but is not shown separately due to small sample sizes and heterogeneity of participants in this group. The poverty index ratio was calculated by dividing family income by a poverty threshold specific for family size using the U.S. Department of Health and Human Services’ poverty guidelines and categorized as either below the poverty line (⬍1) or at or above the poverty line (ⱖ1) (9). Standard error estimates were calculated using Taylor series linearization to incorporate the complex sampling design. Estimates were considered unstable if (i) the relative standard error around the proportion of seropositive or seronegative participants was ⬎30%, (ii) the estimate was based on ⬍10 seropositive or seronegative persons, or (iii) the variance estimates were based on ⬍12 degrees of freedom (7). For estimates of standard errors based on ⬍12 degrees of freedom, additional standard error estimates were calculated using a model-based “average design effect” method (10). Confidence intervals and t statistics using the design-based method (Taylor series linearization) and the modelbased average design effect method were compared. Because both methods yielded similar results, we report only those obtained with the design-based method. The exact binomial method was used to calculate 95% confidence intervals (CIs) (11). Pairwise differences between seroprevalence estimates and tests for trends were evaluated using a t statistic from an orthogonal linear contrast procedure, and adjusted odds ratios and independent predictors of positivity were estimated from a logistic regression model,
Received 6 November 2014 Returned for modification 8 December 2014 Accepted 12 December 2014 Accepted manuscript posted online 17 December 2014 Citation Lanzieri TM, Kruszon-Moran D, Amin MM, Bialek SR, Cannon MJ, Carroll MD, Dollard SC. 2015. Seroprevalence of cytomegalovirus among children 1 to 5 years of age in the United States from the National Health and Nutrition Examination Survey of 2011 to 2012. Clin Vaccine Immunol 22:245–247. doi:10.1128/CVI.00697-14. Editor: R. L. Hodinka Address correspondence to Tatiana M. Lanzieri, [email protected]. Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/CVI.00697-14
Clinical and Vaccine Immunology
cvi.asm.org
245
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, USAa; National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Maryland, USAb; National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, Atlanta, Georgia, USAc
Lanzieri et al.
TABLE 1 CMV IgG seroprevalence in U.S. children 1 to 5 years of age by selected demographic factors, using data from NHANES 2011-2012 CMV IgG weighteda seroprevalence
No. (%) of children Demographic factor
Examined
Tested for CMV IgG
CMV IgG positive
% (95% CI)
Total
1,135
699 (62)
192
20.7 (14.4–28.2)
Age (yr) 1 2 3 4 5
219 302 215 220 179
114 (52) 176 (58) 138 (64) 145 (66) 126 (70)
20 46 39 40 47
12.3 (6.8–19.8) 20.0 (13.6–27.9) 19.9 (8.8–36.0)b 17.3 (10.7–25.9) 31.1 (19.0–45.3)c
Ref. 1.7 (0.8–3.8) 2.0 (1.0–3.9) 1.6 (0.7–3.6) 3.6 (1.7–7.8)d
Sex Male Female
561 574
360 (64) 339 (59)
94 98
19.9 (13.2–28.0) 21.7 (14.7–30.2)
Ref. 1.2 (0.7–1.8)
Race/Hispanic origine Non-Hispanic white Non-Hispanic black All Hispanic Non-Hispanic Asian
209 345 383 132
119 (57) 230 (67) 254 (66) 57 (43)
14 57 78 34
10.6 (3.3–23.8)f 24.5 (19.9–29.6)c 31.0 (26.3–36.0)c 59.2 (39.4–77.0)c,g
Ref. 2.5 (0.8–7.5) 3.0 (1.1–8.4)d 14.9 (4.7–47.3)d
Index ratio of family income to poverty line Below poverty line At or above poverty line
423 621
286 (68) 368 (59)
92 87
31.1 (24.4–38.6)c 14.9 (8.6–23.4)
2.1 (1.2–3.6)d Ref.
aOR (95% CI)
Estimated seroprevalence based on original exam weights from NHANES 2011-2012. aOR, adjusted odds ratio from logistic regression model; Ref., reference group. The estimate was considered unstable because the relative standard error was 32.4%. c P ⬍ 0.05 from the t statistic comparing the subgroup to the reference group within demographic characteristic by univariate analysis. d P ⬍ 0.05 from the Satterthwaite-adjusted F statistic comparing the subgroup to the reference group, adjusting for other demographic characteristics (gender, age, race/Hispanic origin, and index ratio of family income to poverty line) by using a logistic regression model. e Children of “other” races are included in total estimates but are not shown separately. f The estimate was considered unstable because the relative standard error of the estimate was 44.3% and the degrees of freedom for the estimate of the standard error was ⬍12. g The estimate considered unstable because the number of degrees of freedom for the estimate of the standard error was ⬍12 and the sample size for the estimate was based on ⬍100 persons. b
both in SUDAAN version 9.0 (Research Triangle Institute, Research Triangle Park, NC); P values of ⬍0.05 were considered significant. Serologic testing of plasma specimens from the NHANES was conducted at the CDC CMV diagnostic laboratory. CMV IgG, IgM, and IgG avidity were all measured by VIDAS (bioMérieux) (12). For measuring IgG avidity, a modified cutoff value of 0.7, which has been shown to provide improved sensitivity for detection of recent infection, was used (12). A positive CMV IgG result indicates past or recent CMV infection, whereas the presence of CMV IgM is transient and can indicate a recent primary infection, reactivation, or reinfection. Low IgG avidity is typically present for 3 to 4 months after primary infection, after which high-avidity IgG develops as a result of maturation of the humoral immune response (13). Among the 1,463 children 1 to 5 years old who were sampled in NHANES 2011-2012, 1,203 (82.2% of those sampled) completed the home interview, 1,135 (94.3% of those interviewed) completed the health examination component, 741 (65.3% of those examined) had a blood sample successfully collected, and 699 (61.6% of those examined) were tested for CMV antibody. Those tested for CMV differed from those who were examined but not tested for CMV by age, poverty status, and race/Hispanic origin (Table 1). Because the percentage of sample persons with sufficient blood available for CMV testing out of those examined was ⬍60% in many subgroups and differed by age, poverty status, and
246
cvi.asm.org
race/Hispanic origin (Table 1), the analysis was repeated using new sample weights that accounted for the missing CMV testing data. Individual exam weights for those tested were inflated to account for those examined but not tested within all race/Hispanic origin and age strata (14). Results from the reweighted analysis did not vary substantially from the original analysis, so only the results obtained with the original exam weights are reported. The overall seroprevalence of CMV IgG among children 1 to 5 years old in the United States was 20.7%. IgG seroprevalence was lowest among 1-year-olds (12.3%) and highest among 5-year-olds (31.1%) (P ⬍ 0.01 linear test for trend) (Table 1). There was no significant difference in IgG seroprevalence between males and females. IgG seroprevalences were significantly higher among non-Hispanic Asians, Hispanics, and non-Hispanic blacks than among non-Hispanic whites (P ⬍ 0.05). However, the estimates for non-Hispanic whites and non-Hispanic Asians were unstable. IgG seroprevalence was significantly higher among those living below the poverty line than among those living at or above the poverty line (P ⬍ 0.001). In the logistic regression model, age (5 years compared to 1 year), race/Hispanic origin (non-Hispanic Asian and Hispanic compared to non-Hispanic white persons), and index ratio of family income (below versus at or above the poverty line) were independently associated with IgG seroprevalence (Table 1). The overall weighted seroprevalence of CMV IgM was 1.1% (95% CI, 0.4 to 2.4%). This estimate was based on 11 seropositive children with a relative standard error of ⬎40%, so
Clinical and Vaccine Immunology
February 2015 Volume 22 Number 2
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
a
CMV Seroprevalence among U.S. Children
ACKNOWLEDGMENTS No external funding was used for this study. We have no financial relationships relevant to this article to disclose.
February 2015 Volume 22 Number 2
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
REFERENCES 1. Dollard SC, Grosse SD, Ross DS. 2007. New estimates of the prevalence of neurological and sensory sequelae and mortality associated with congenital cytomegalovirus infection. Rev Med Virol 17:355–363. http://dx .doi.org/10.1002/rmv.544. 2. Stagno S, Reynolds DW, Pass RF, Alford CA. 1980. Breast milk and the risk of cytomegalovirus infection. N Engl J Med 302:1073–1076. http://dx .doi.org/10.1056/NEJM198005083021908. 3. Pass RF, Hutto SC, Reynolds DW, Polhill RB. 1984. Increased frequency of cytomegalovirus infection in children in group day care. Pediatrics 74:121–126. 4. Staras SA, Flanders WD, Dollard SC, Pass RF, McGowan JE, Jr, Cannon MJ. 2008. Cytomegalovirus seroprevalence and childhood sources of infection: A population-based study among pre-adolescents in the United States. J Clin Virol 43:266 –271. http://dx.doi.org/10.1016/j.jcv.2008.07.012. 5. Hyde TB, Schmid DS, Cannon MJ. 2010. Cytomegalovirus seroconversion rates and risk factors: implications for congenital CMV. Rev Med Virol 20:311–326. http://dx.doi.org/10.1002/rmv.659. 6. Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. 2014. National Health and Nutrition Examination Survey: sample design, 2011-2014. National Center for Health Statistics. Vital Health Stat 2:162. 7. Centers for Disease Control and Prevention. Accessed 12 December 2014. National Center for Health Statistics. Division of Health and Nutrition Examination Surveys. National Health and Nutrition Examination Survey: analytic guidelines, 2011-2012. Centers for Disease Control and Prevention, Atlanta, GA. http://www.cdc.gov/nchs/data/nhanes/analytic _guidelines_11_12.pdf. 8. Centers for Disease Control and Prevention. Accessed 12 December 2014. National Center for Health Statistics. Division of Health and Nutrition Examination Surveys. National Health and Nutrition Examination Survey, 2011-2012: data documentation, codebook, and frequencies. Centers for Disease Control and Prevention, Atlanta, GA. http://wwwn.cdc.gov/nchs/nhanes/2011-2012/DEMO_G.htm. 9. U.S. Department of Health and Human Services. Accessed 12 December 2014. Poverty guidelines, research, and measurement. U.S. Department of Health and Human Services, Washington, DC. http://aspe.hhs.gov /POVERTY/index.shtml. 10. Kovar MG, Johnson C. 1986. Design effects from the Mexican American portion of the Hispanic Health and Nutrition Examination Survey: a strategy for analysis, p 396 –399. Proceedings of the American Statistical Association, Section on Survey Research Methods. American Statistical Association, Alexandria, VA. 11. Korn EL, Graubard BI. 1998. Confidence intervals for proportions with small expected number of positive counts estimated from survey data. Survey Methodol 24:98 –201. 12. Dollard SC, Staras SA, Amin MM, Schmid DS, Cannon MJ. 2011. National prevalence estimates for cytomegalovirus IgM and IgG avidity and association between high IgM antibody titer and low IgG avidity. Clin Vaccine Immunol 18:1895–1899. http://dx.doi.org/10.1128/CVI.05228-11. 13. Lazzarotto T, Guerra B, Lanari M, Gabrielli L, Landini MP. 2008. New advances in the diagnosis of congenital cytomegalovirus infection. J Clin Virol 41:192–197. http://dx.doi.org/10.1016/j.jcv.2007.10.015. 14. Lohr S. 1999. Sampling design and analysis. Duxbury Press, Albany, NY. 15. Bate SL, Dollard SC, Cannon MJ. 2010. Cytomegalovirus seroprevalence in the United States: the national health and nutrition examination surveys, 1988-2004. Clin Infect Dis 50:1439 –1447. http://dx.doi.org/10.1086 /652438.
Clinical and Vaccine Immunology
cvi.asm.org
247
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
results were considered unstable. The overall weighted prevalence of low IgG avidity was 3.6% (95% CI, 1.7 to 6.6%), corresponding to a prevalence of recent infection among IgG-positive children of 17.3% (95% CI, 10.1 to 26.7%). Our study is the first to provide an estimate of CMV seroprevalence (20.7%) in a nationally representative sample of U.S. children 1 to 5 years old. The only other study in healthy U.S. children that we are aware of reported 17% seroprevalence among children 1 to 5 years old recruited from a single pediatric practice in Alabama in the 1980s (3). Our study did not include children that were ⬍1 year old; thus, the detection of CMV IgG was not likely due to the presence of residual maternal antibody. We found that CMV IgG seroprevalence was higher in older children and in nonwhite and poorer children, consistent with trends reported for older children and adults in the United States using previous NHANES data (15). During 1999 to 2004, CMV seroprevalence among children 6 to 11 years old was 38% overall and was higher in non-Hispanic black (⬃45%) and Mexican American (⬃ 60%) children than in non-Hispanic white (⬃30%) children. Throughout adolescence and adulthood, the prevalence differences across racial and ethnic subgroups increased with age, particularly among women (15). Wide variations in maternal seroprevalence, breastfeeding rates, and group child care practices among racial and ethnic subgroups likely contribute to the differences in CMV seroprevalence observed across subgroups in children 1 to 5 years old. We could not examine some other factors reported to be associated with CMV IgG seroprevalence in older children (4) because they either were not collected in NHANES 2011-2012 (e.g., child care attendance and maternal CMV serostatus) or were not yet available for analysis (family size, crowding index, household education level, and history of breastfeeding). We could not assess the extent to which higher seroprevalences among non-Hispanic Asian and Hispanic children reflected birth outside the United States, because the number of foreign-born children was not large enough to provide reliable estimates and parental country of origin was not collected. Our study lacked the power to provide stable national seroprevalence estimates of IgG, low IgG avidity, or IgM for all subgroups. The seroprevalence data described in our study will be useful for modeling the potential impact of interventions. A better understanding of immune correlates of protection against CMV in young children and adults will be critical for the development of an effective CMV vaccine. Young children have been identified as a potential target population for CMV vaccination with the ultimate goal of prevention of congenital CMV infection. Because young children often shed CMV in urine and saliva for many months, a vaccine that provides protection against CMV infection in infants and toddlers could prevent congenital CMV infections by reducing CMV transmission from young children to women during pregnancy.
Cytomegalovirus (CMV) seroprevalence among U.S. children 1 to 5 years old was assessed in the National Health and Nutrition Examination Survey of 2011 to 2012. The overall seroprevalence (95% confidence interval) of IgG was 20.7% (14.4 to 28.2%), that of IgM was 1.1% (0.4 to 2.4%), and that of low IgG avidity was 3.6% (1.7 to 6.6%), corresponding to a 17.3% (10.1 to 26.7%) prevalence of recent infection among IgG-positive children.
C
ongenital cytomegalovirus (CMV) infection is a major cause of sensorineural hearing loss and developmental disabilities in children in the United States (1). An estimated 0.7% of U.S. infants are born with congenital CMV infection (1). Early in life, CMV infection generally results from mother-to-infant transmission through exposure to the virus in genital secretions during birth or postnatally via breast milk (2). CMV seroprevalence continues to rise throughout early childhood, as infants and young children acquire CMV infection via exposure to body fluids from other infected individuals, especially from close contact with young children in household or day care settings (3, 4). Young CMV-seropositive children, who may shed virus in body fluids for months after primary CMV infection, are thought to be an important source of CMV transmission to adults (5). Population-based estimates of CMV seroprevalence among young children in the United States have not been previously reported. In this study, we describe CMV seroprevalence among U.S. children 1 to 5 years old, who were sampled in the National Health and Nutrition Examination Survey of 2011 to 2012 (NHANES 2011-2012). The methods of the NHANES, a nationally representative, cross-sectional survey of the civilian noninstitutionalized U.S. population, have been published elsewhere (6). NHANES 20112012 oversampled non-Hispanic Asians, in addition to Hispanics (including Mexican Americans), non-Hispanic blacks, and lowincome persons who were non-Hispanic whites or designated as “other” to increase sample size and obtain more statistically reliable estimates of these subgroups (6). National estimates of CMV seroprevalence for CMV IgG, IgM, and low IgG avidity were calculated using the exam weights developed for the NHANES to represent the total civilian noninstitutionalized U.S. population and to account for oversampling and nonresponse to the household interview and physical examination (7). Seroprevalence of CMV IgG was examined by age (in years), sex, race/Hispanic origin, and poverty index ratio. Information on race/Hispanic origin was collected by proxy, and individuals were categorized as non-Hispanic white, non-Hispanic black, non-Hispanic Asian, Mexican American, other Hispanic, and other race (which includes multiracial) (8). In our analysis, we combined Mexican American and other Hispanic to form the Hispanic group. The “other” race group (n ⫽ 39) is included in overall
February 2015 Volume 22 Number 2
estimates but is not shown separately due to small sample sizes and heterogeneity of participants in this group. The poverty index ratio was calculated by dividing family income by a poverty threshold specific for family size using the U.S. Department of Health and Human Services’ poverty guidelines and categorized as either below the poverty line (⬍1) or at or above the poverty line (ⱖ1) (9). Standard error estimates were calculated using Taylor series linearization to incorporate the complex sampling design. Estimates were considered unstable if (i) the relative standard error around the proportion of seropositive or seronegative participants was ⬎30%, (ii) the estimate was based on ⬍10 seropositive or seronegative persons, or (iii) the variance estimates were based on ⬍12 degrees of freedom (7). For estimates of standard errors based on ⬍12 degrees of freedom, additional standard error estimates were calculated using a model-based “average design effect” method (10). Confidence intervals and t statistics using the design-based method (Taylor series linearization) and the modelbased average design effect method were compared. Because both methods yielded similar results, we report only those obtained with the design-based method. The exact binomial method was used to calculate 95% confidence intervals (CIs) (11). Pairwise differences between seroprevalence estimates and tests for trends were evaluated using a t statistic from an orthogonal linear contrast procedure, and adjusted odds ratios and independent predictors of positivity were estimated from a logistic regression model,
Received 6 November 2014 Returned for modification 8 December 2014 Accepted 12 December 2014 Accepted manuscript posted online 17 December 2014 Citation Lanzieri TM, Kruszon-Moran D, Amin MM, Bialek SR, Cannon MJ, Carroll MD, Dollard SC. 2015. Seroprevalence of cytomegalovirus among children 1 to 5 years of age in the United States from the National Health and Nutrition Examination Survey of 2011 to 2012. Clin Vaccine Immunol 22:245–247. doi:10.1128/CVI.00697-14. Editor: R. L. Hodinka Address correspondence to Tatiana M. Lanzieri, [email protected]. Copyright © 2015, American Society for Microbiology. All Rights Reserved. doi:10.1128/CVI.00697-14
Clinical and Vaccine Immunology
cvi.asm.org
245
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia, USAa; National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Maryland, USAb; National Center on Birth Defects and Developmental Disabilities, Centers for Disease Control and Prevention, Atlanta, Georgia, USAc
Lanzieri et al.
TABLE 1 CMV IgG seroprevalence in U.S. children 1 to 5 years of age by selected demographic factors, using data from NHANES 2011-2012 CMV IgG weighteda seroprevalence
No. (%) of children Demographic factor
Examined
Tested for CMV IgG
CMV IgG positive
% (95% CI)
Total
1,135
699 (62)
192
20.7 (14.4–28.2)
Age (yr) 1 2 3 4 5
219 302 215 220 179
114 (52) 176 (58) 138 (64) 145 (66) 126 (70)
20 46 39 40 47
12.3 (6.8–19.8) 20.0 (13.6–27.9) 19.9 (8.8–36.0)b 17.3 (10.7–25.9) 31.1 (19.0–45.3)c
Ref. 1.7 (0.8–3.8) 2.0 (1.0–3.9) 1.6 (0.7–3.6) 3.6 (1.7–7.8)d
Sex Male Female
561 574
360 (64) 339 (59)
94 98
19.9 (13.2–28.0) 21.7 (14.7–30.2)
Ref. 1.2 (0.7–1.8)
Race/Hispanic origine Non-Hispanic white Non-Hispanic black All Hispanic Non-Hispanic Asian
209 345 383 132
119 (57) 230 (67) 254 (66) 57 (43)
14 57 78 34
10.6 (3.3–23.8)f 24.5 (19.9–29.6)c 31.0 (26.3–36.0)c 59.2 (39.4–77.0)c,g
Ref. 2.5 (0.8–7.5) 3.0 (1.1–8.4)d 14.9 (4.7–47.3)d
Index ratio of family income to poverty line Below poverty line At or above poverty line
423 621
286 (68) 368 (59)
92 87
31.1 (24.4–38.6)c 14.9 (8.6–23.4)
2.1 (1.2–3.6)d Ref.
aOR (95% CI)
Estimated seroprevalence based on original exam weights from NHANES 2011-2012. aOR, adjusted odds ratio from logistic regression model; Ref., reference group. The estimate was considered unstable because the relative standard error was 32.4%. c P ⬍ 0.05 from the t statistic comparing the subgroup to the reference group within demographic characteristic by univariate analysis. d P ⬍ 0.05 from the Satterthwaite-adjusted F statistic comparing the subgroup to the reference group, adjusting for other demographic characteristics (gender, age, race/Hispanic origin, and index ratio of family income to poverty line) by using a logistic regression model. e Children of “other” races are included in total estimates but are not shown separately. f The estimate was considered unstable because the relative standard error of the estimate was 44.3% and the degrees of freedom for the estimate of the standard error was ⬍12. g The estimate considered unstable because the number of degrees of freedom for the estimate of the standard error was ⬍12 and the sample size for the estimate was based on ⬍100 persons. b
both in SUDAAN version 9.0 (Research Triangle Institute, Research Triangle Park, NC); P values of ⬍0.05 were considered significant. Serologic testing of plasma specimens from the NHANES was conducted at the CDC CMV diagnostic laboratory. CMV IgG, IgM, and IgG avidity were all measured by VIDAS (bioMérieux) (12). For measuring IgG avidity, a modified cutoff value of 0.7, which has been shown to provide improved sensitivity for detection of recent infection, was used (12). A positive CMV IgG result indicates past or recent CMV infection, whereas the presence of CMV IgM is transient and can indicate a recent primary infection, reactivation, or reinfection. Low IgG avidity is typically present for 3 to 4 months after primary infection, after which high-avidity IgG develops as a result of maturation of the humoral immune response (13). Among the 1,463 children 1 to 5 years old who were sampled in NHANES 2011-2012, 1,203 (82.2% of those sampled) completed the home interview, 1,135 (94.3% of those interviewed) completed the health examination component, 741 (65.3% of those examined) had a blood sample successfully collected, and 699 (61.6% of those examined) were tested for CMV antibody. Those tested for CMV differed from those who were examined but not tested for CMV by age, poverty status, and race/Hispanic origin (Table 1). Because the percentage of sample persons with sufficient blood available for CMV testing out of those examined was ⬍60% in many subgroups and differed by age, poverty status, and
246
cvi.asm.org
race/Hispanic origin (Table 1), the analysis was repeated using new sample weights that accounted for the missing CMV testing data. Individual exam weights for those tested were inflated to account for those examined but not tested within all race/Hispanic origin and age strata (14). Results from the reweighted analysis did not vary substantially from the original analysis, so only the results obtained with the original exam weights are reported. The overall seroprevalence of CMV IgG among children 1 to 5 years old in the United States was 20.7%. IgG seroprevalence was lowest among 1-year-olds (12.3%) and highest among 5-year-olds (31.1%) (P ⬍ 0.01 linear test for trend) (Table 1). There was no significant difference in IgG seroprevalence between males and females. IgG seroprevalences were significantly higher among non-Hispanic Asians, Hispanics, and non-Hispanic blacks than among non-Hispanic whites (P ⬍ 0.05). However, the estimates for non-Hispanic whites and non-Hispanic Asians were unstable. IgG seroprevalence was significantly higher among those living below the poverty line than among those living at or above the poverty line (P ⬍ 0.001). In the logistic regression model, age (5 years compared to 1 year), race/Hispanic origin (non-Hispanic Asian and Hispanic compared to non-Hispanic white persons), and index ratio of family income (below versus at or above the poverty line) were independently associated with IgG seroprevalence (Table 1). The overall weighted seroprevalence of CMV IgM was 1.1% (95% CI, 0.4 to 2.4%). This estimate was based on 11 seropositive children with a relative standard error of ⬎40%, so
Clinical and Vaccine Immunology
February 2015 Volume 22 Number 2
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
a
CMV Seroprevalence among U.S. Children
ACKNOWLEDGMENTS No external funding was used for this study. We have no financial relationships relevant to this article to disclose.
February 2015 Volume 22 Number 2
The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.
REFERENCES 1. Dollard SC, Grosse SD, Ross DS. 2007. New estimates of the prevalence of neurological and sensory sequelae and mortality associated with congenital cytomegalovirus infection. Rev Med Virol 17:355–363. http://dx .doi.org/10.1002/rmv.544. 2. Stagno S, Reynolds DW, Pass RF, Alford CA. 1980. Breast milk and the risk of cytomegalovirus infection. N Engl J Med 302:1073–1076. http://dx .doi.org/10.1056/NEJM198005083021908. 3. Pass RF, Hutto SC, Reynolds DW, Polhill RB. 1984. Increased frequency of cytomegalovirus infection in children in group day care. Pediatrics 74:121–126. 4. Staras SA, Flanders WD, Dollard SC, Pass RF, McGowan JE, Jr, Cannon MJ. 2008. Cytomegalovirus seroprevalence and childhood sources of infection: A population-based study among pre-adolescents in the United States. J Clin Virol 43:266 –271. http://dx.doi.org/10.1016/j.jcv.2008.07.012. 5. Hyde TB, Schmid DS, Cannon MJ. 2010. Cytomegalovirus seroconversion rates and risk factors: implications for congenital CMV. Rev Med Virol 20:311–326. http://dx.doi.org/10.1002/rmv.659. 6. Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. 2014. National Health and Nutrition Examination Survey: sample design, 2011-2014. National Center for Health Statistics. Vital Health Stat 2:162. 7. Centers for Disease Control and Prevention. Accessed 12 December 2014. National Center for Health Statistics. Division of Health and Nutrition Examination Surveys. National Health and Nutrition Examination Survey: analytic guidelines, 2011-2012. Centers for Disease Control and Prevention, Atlanta, GA. http://www.cdc.gov/nchs/data/nhanes/analytic _guidelines_11_12.pdf. 8. Centers for Disease Control and Prevention. Accessed 12 December 2014. National Center for Health Statistics. Division of Health and Nutrition Examination Surveys. National Health and Nutrition Examination Survey, 2011-2012: data documentation, codebook, and frequencies. Centers for Disease Control and Prevention, Atlanta, GA. http://wwwn.cdc.gov/nchs/nhanes/2011-2012/DEMO_G.htm. 9. U.S. Department of Health and Human Services. Accessed 12 December 2014. Poverty guidelines, research, and measurement. U.S. Department of Health and Human Services, Washington, DC. http://aspe.hhs.gov /POVERTY/index.shtml. 10. Kovar MG, Johnson C. 1986. Design effects from the Mexican American portion of the Hispanic Health and Nutrition Examination Survey: a strategy for analysis, p 396 –399. Proceedings of the American Statistical Association, Section on Survey Research Methods. American Statistical Association, Alexandria, VA. 11. Korn EL, Graubard BI. 1998. Confidence intervals for proportions with small expected number of positive counts estimated from survey data. Survey Methodol 24:98 –201. 12. Dollard SC, Staras SA, Amin MM, Schmid DS, Cannon MJ. 2011. National prevalence estimates for cytomegalovirus IgM and IgG avidity and association between high IgM antibody titer and low IgG avidity. Clin Vaccine Immunol 18:1895–1899. http://dx.doi.org/10.1128/CVI.05228-11. 13. Lazzarotto T, Guerra B, Lanari M, Gabrielli L, Landini MP. 2008. New advances in the diagnosis of congenital cytomegalovirus infection. J Clin Virol 41:192–197. http://dx.doi.org/10.1016/j.jcv.2007.10.015. 14. Lohr S. 1999. Sampling design and analysis. Duxbury Press, Albany, NY. 15. Bate SL, Dollard SC, Cannon MJ. 2010. Cytomegalovirus seroprevalence in the United States: the national health and nutrition examination surveys, 1988-2004. Clin Infect Dis 50:1439 –1447. http://dx.doi.org/10.1086 /652438.
Clinical and Vaccine Immunology
cvi.asm.org
247
Downloaded from http://cvi.asm.org/ on May 23, 2015 by Univ of North Dakota
results were considered unstable. The overall weighted prevalence of low IgG avidity was 3.6% (95% CI, 1.7 to 6.6%), corresponding to a prevalence of recent infection among IgG-positive children of 17.3% (95% CI, 10.1 to 26.7%). Our study is the first to provide an estimate of CMV seroprevalence (20.7%) in a nationally representative sample of U.S. children 1 to 5 years old. The only other study in healthy U.S. children that we are aware of reported 17% seroprevalence among children 1 to 5 years old recruited from a single pediatric practice in Alabama in the 1980s (3). Our study did not include children that were ⬍1 year old; thus, the detection of CMV IgG was not likely due to the presence of residual maternal antibody. We found that CMV IgG seroprevalence was higher in older children and in nonwhite and poorer children, consistent with trends reported for older children and adults in the United States using previous NHANES data (15). During 1999 to 2004, CMV seroprevalence among children 6 to 11 years old was 38% overall and was higher in non-Hispanic black (⬃45%) and Mexican American (⬃ 60%) children than in non-Hispanic white (⬃30%) children. Throughout adolescence and adulthood, the prevalence differences across racial and ethnic subgroups increased with age, particularly among women (15). Wide variations in maternal seroprevalence, breastfeeding rates, and group child care practices among racial and ethnic subgroups likely contribute to the differences in CMV seroprevalence observed across subgroups in children 1 to 5 years old. We could not examine some other factors reported to be associated with CMV IgG seroprevalence in older children (4) because they either were not collected in NHANES 2011-2012 (e.g., child care attendance and maternal CMV serostatus) or were not yet available for analysis (family size, crowding index, household education level, and history of breastfeeding). We could not assess the extent to which higher seroprevalences among non-Hispanic Asian and Hispanic children reflected birth outside the United States, because the number of foreign-born children was not large enough to provide reliable estimates and parental country of origin was not collected. Our study lacked the power to provide stable national seroprevalence estimates of IgG, low IgG avidity, or IgM for all subgroups. The seroprevalence data described in our study will be useful for modeling the potential impact of interventions. A better understanding of immune correlates of protection against CMV in young children and adults will be critical for the development of an effective CMV vaccine. Young children have been identified as a potential target population for CMV vaccination with the ultimate goal of prevention of congenital CMV infection. Because young children often shed CMV in urine and saliva for many months, a vaccine that provides protection against CMV infection in infants and toddlers could prevent congenital CMV infections by reducing CMV transmission from young children to women during pregnancy.
