Given the overall success in reducing perinatal HIV infection and the ability to prevent transmission in the United States, each perinatal transmission is a sentinel health event that reflects missed opportunities for prevention by the medical or public health system. It is estimated that in 2006 (the most recent year for which data are available), 8650–8900 HIV-exposed infants were born, and 133,700–137,553 HIV-infected women aged 13–45 years were living in the United States.1 In 2009, the Centers for Disease Control and Prevention (CDC) estimated that 131 children were diagnosed as having perinatal HIV infection.2 However, in 2009, CDC estimated that 21% of 166 diagnoses of HIV/AIDS among children (<13 years) were due to modes other than perinatal transmission, including hemophilia, blood transfusion, and “risk factor not reported or not identified”.2 To achieve the CDC goal of reducing perinatal transmission below 1 per 100,000 live births annually, it is essential to better understand the risk of pediatric HIV transmission incurred by premastication—chewing foods or medicines by a caregiver and then feeding the prechewed food to a child.
In a 2009 case series report, premastication was the suspected mode of transmission for 3 pediatric HIV infections.3 This report, which included an investigation of a pediatric case whose mother was HIV-uninfected, emphasized that HIV transmission via premastication requires the presence of blood in the mouth possibly due to periodontal disease, loose or missing teeth, or cuts or sores in the mouth. The prevalence of premastication among caregivers in contemporary Western cultures has received relatively little attention. However, in a US study reported in 2008, 14% of mothers of healthy singletons premasticated food for their child.4 Given the recently published findings regarding HIV transmission through premastication or prechewing infant foods, CDC conducted a case–control investigation in 2009 to assess the degree of risk for HIV transmission represented by premasticating. A cross-sectional investigation was initiated in concert with the case–control investigation to establish the prevalence of premastication by surveying the caregivers of HIV-exposed children who attended pediatric HIV clinics.5
Six HIV pediatric clinics in the following cities participated in the case–control investigation: Atlanta (GA), Dallas (TX), Houston (TX), Memphis (TN), Miami (FL), and Washington (DC). The 3 potential cases of premastication-related HIV transmission were originally reported from Memphis and Miami.3 Each of the clinics at these sites had collaborated previously with CDC in premastication-related or longitudinal HIV-related epidemiologic studies. At the time of interview, case-patients were being treated in the clinic; controls were being or had previously been treated in the clinic. CDC classified this investigation as a public health disease control activity. The sites obtained institutional review board approval according to their local requirements.
Case-patients were those with “late-diagnosed HIV infection”, defined as children with HIV infection diagnosed between 4 months and 10 years of age, with at least 1 negative HIV test result after 6 weeks of age before their first-positive test result. The HIV diagnosis required 2 positive HIV tests results, including repeatedly positive HIV antibody tests after 18 months of age or positive nucleic acid amplification tests after 4 months of age. The case-control investigation was limited to children older than 6 months but younger than 13 years at the time of the investigation. These restrictions allowed time for exposure to premasticated food at younger ages and limited the likelihood of sexually acquired infection at older ages.
Clinic and state HIV databases were used to identify case-patients. Children with only 1 negative test result before the first-positive test result were considered “probable” case-patients; those with 2 or more prior negative test results (at least 1 after 6 weeks of age) were considered “definitive” case-patients. The investigation was conducted during December 2009 to February 2010. Staff at the sites contacted caregivers to ascertain their willingness to participate in the investigation. Interviews were typically scheduled in conjunction with routine clinic appointments.
Controls were HIV-exposed children (born to an HIV-infected mother) who were not HIV infected. Three controls per case-patient were selected by using patient databases to select the children in the clinic nearest in age to the case. Siblings of case-patients were excluded as controls. Face-to-face or telephone interviews lasting approximately 20 minutes were conducted with caregivers who characterized themselves as the primary caregiver or who stated that they were aware of the feeding practices used when the child was <2 years old. Neither chart reviews nor specimen collections were conducted during the study. The interview encompassed information on demographics, premasticating behaviors, HIV and hepatitis infection history, oral conditions (eg, oral thrush, loose or missing teeth, cut/sores, or bleeding gums), breastfeeding, sexual abuse, exposure to blood or needles, and the feeding practices of a nonprimary caregiver. After the interview, participants were given health information about pathogens potentially transmitted through premastication. More discussion on the risk of premastication was conducted with participants who expressed interest to discuss this topic or had questions about the practice.
Variables hypothesized to increase risk of transmission were used to model case status using separate logistic regression models. Unadjusted odds ratio (OR) and 95% confidence intervals (CIs) were calculated by using exact logistic regression in SAS version 9.2 (SAS Institute, Inc, Cary, NC). Fisher exact test and Wilcoxon 2-sample tests were used to compare categorical and continuous variables, respectively. The P values of less than 0.05 were considered statistically significant.
In addition to sites that participated in the case–control investigation, major pediatric HIV hospitals or clinics in Newark (NJ), San Juan, (Puerto Rico), and New Orleans (LA) were included in a cross-sectional investigation (1 site per location). This aspect of the investigation used a convenience sampling frame to collect data from caregivers of children currently being seen in these HIV clinics. A 10-minute self-administered paper survey on child-feeding practices was distributed to primary caregivers in waiting rooms of these clinics. Participants of the case–control investigation were excluded from participation in the cross-sectional investigation. Cross-sectional surveys on premastication were administered in each clinic for a minimum of 1 week or until at least 30 surveys were completed. Those unable to administer a minimum of 30 surveys within 1 week were not required to collect surveys beyond this time frame. One survey per child with an appointment was allowed. Premastication prevalence ratios and 95% CIs were calculated by child and caregiver demographic characteristics.
We interviewed the caregivers of 12 of 19 case-patients with late-diagnosed HIV infection in 6 HIV/AIDS pediatric clinics during December 2009 to February 2010. Seven of the 19 case-patients could not be located or did not consent to be interviewed. We excluded 1 case-patient who was an HIV-infected child of a mother who was not infected because this child may have a different likelihood of HIV infection than other case-patients. Thus, we limited our analysis to the 11 case-patients with HIV-infected mothers. Of these, 9 were definitive case-patients, and 2 were probable case-patients.
We compared the 11 case-patients to 35 controls interviewed at 6 sites. The vast majority of children and caregivers were African American (83%–91%). The median age of the child at time of the interview was statistically similar for case-patients and controls, 75 and 90 months, respectively (P value > 0.05). Case-patients were also similar to controls with respect to other caregiver demographic characteristics (Table 1).
Sixteen (35%) of 46 children were fed premasticated food, 10 (22%) by an HIV-infected caregiver. Twenty-seven percent of case-patients received premasticated food from an HIV-infected caregiver compared with 20% of controls (OR = 1.5; 95% CI = 0.3 to 7.1; Table 2). Eighteen percent of case-patients compared with 6% of controls received premasticated food from an HIV-infected caregiver who also had predisposing oral conditions, such as thrush, missing teeth, or bleeding gums (OR = 3.6; 95% CI = 0.4 to 31.2). A smaller proportion of case-patients (27%) than controls (34%) had oral conditions when receiving prechewed food (OR = 0.7; 95% CI = 0.1 to 3.8). Antiretroviral drug use was common among HIV-infected caregivers who premasticated (2 of 2 case-patients, 5 of 6 controls).
The odds of having received blood or blood products were significantly higher among case-patients (40%) than among controls (3%) (OR = 19.2; 95% CI = 1.6 to > 999.9). According to caregiver reports, none of the children who received blood or blood products had received premasticated food from an HIV-infected caregiver. No case-patients or controls were breastfed. No case-patients and 1 control were reported as victims of sexual abuse.
Of the 203 approached, we surveyed 192 (95%) primary caregivers (11 declined participation) from 9 HIV clinics in the United States and Puerto Rico from December 2009 to February 2010. Given the small likelihood of feeding children solid foods in the first months of life, we limited our analysis to caregivers of children who were 6 months of age or older at time of investigation (155 of 192, or 81%). The vast majority of caregivers were biological mothers (80%) and US born (88%) (Table 3). Sixty-eight percent of the sample was African American and 21% was Hispanic or Latino. The median age was 32 years for primary caregivers and 3 years for children.
Among those who responded to the premastication questions, 96 (65%) of 148 primary caregivers stated they had heard of premastication before the survey. Among primary caregivers, 44 (29%) of 153 reported ever premasticating food for the child. Fourteen (10%) of 140 primary caregivers reported that someone else had fed premasticated food to the child. Overall, 48 (31%) of 154 primary caregivers stated that the child had received premasticated food from either themselves, someone else, or both. Additionally, 40 (31%) of 128 biological mothers either self-reported or were reported to have premasticated food for the child.
Premastication was less common among older caregivers than among younger caregivers (Table 4). Caregivers aged 40 years and older were significantly less likely to report ever premasticating for the child in the clinic (13%) than those aged ≤19 years (44%), those aged 20–29 years old (38%), and those aged 30–39 years (36%). Although only marginally significant, a higher prevalence of premastication was found among African American caregivers when compared with caregivers of other races or ethnicity. We found no significant differences in the prevalence of premastication on the basis of the child's gender or the primary caregiver's place of birth, education level, or income.
The median age of children at the time of the survey at the beginning and the end of receiving premasticated food was similar for primary and other caregivers. Primary caregivers began feeding the child premasticated food when the child was 7 months old and stopped when the child was 13 months old; other caregivers began when the child was 8 months old and stopped when the child was 13 months old. When considering the range of responses, primary caregivers reported premastication as early as 1 month and as late as 36 months of age; however, none of the 31 caregivers of children younger than 6 months at the time of the interview reported premastication (information on age of 6 children was unavailable). When asked about the frequency of premastication in a typical week, 39% (15 of 38) of primary caregivers reported 1 to 3 days in a week; 37% stated that they prechewed food ≥4 days in a typical week. Meats and fish were the most commonly reported food types that were premasticated for the child by primary caregivers (80%, 37 of 46), followed by fruits (39%) and vegetables (37%), and then candies or sweets (30%).
Caregivers were asked to choose reasons for premastication from a list. Among 45 caregivers, the most common responses were “Child wanted some of the caregiver's food” (64%) and “caregiver did not want the child to choke” (62%). “Prechewing is done in my family” was also reported by 31% of caregivers. Reasons that may reflect inability to provide baby food were less commonly reported as follows: “did not have store-bought baby food” (4%) and “did not make baby food” (2%).
Premastication was a common practice among caregivers of HIV-exposed infants in the cross-sectional investigation: almost one-third of the children received premasticated food. We consider this finding particularly important because most of the caregivers were biological mothers and were HIV infected; therefore, premastication by these women presents a risk of HIV transmission to children in their care. Premastication was more commonly reported by younger caregivers; prevalence decreased in each successive age group. Younger caregivers may be less adept at the various alternative methods available for child feeding and may thus benefit from targeted prevention messages.
Knowledge and practice of premastication was commonly reported by African Americans. This finding is supported by a 1987 study that examined a nonrandom sample of African American caregivers as follows: 65% reported premastication, and 90% reported knowledge of the practice.6 Given that racial disparities in HIV infection, including perinatal transmission, are most prominent among African Americans7,8 and that premastication may be a common practice in African American households, these findings suggest a compounded risk of HIV transmission to African American children.
Premastication-related transmission has been reported for a variety of pathogens, including hepatitis B virus,9 group A streptococci,10 and Treponema pallidum.11 Furthermore, premastication has been associated with increased risk of infection by Helicobacter pylori,12 Streptococcus mutans,13 human herpesvirus 8,14 and Epstein-Barr virus.15 Conversely, only 1 study has demonstrated that premastication may be protective, in this case, against respiratory syncytial virus infection in Alaskan Native children <6 months.16 Despite such findings, it has been argued that premastication has played a key role in human evolution and survival and that it is a valuable practice because it provides a nutritional and possibly immunologic benefit to a child.17 For centuries, premastication has enabled caregivers to provide nourishment to children in addition to, or replacement of, breast milk. In 1 cross-cultural study, premastication was practiced in one-third of the 119 cultures surveyed.17
In our investigation, the reasons reported for premastication suggest that in the United States, the practice is mostly situational—a response to the practical logistics of feeding the child. Fewer of the caregivers chose reasons suggesting that they were unable to provide baby food or formula or that they were culturally committed to premastication. Hence, culturally sensitive messages targeting HIV-infected caregivers who are more likely to premasticate may be highly effective among this population, particularly among caregivers with situational reasons for premastication.
Premastication was also common among the caregivers of the participants in the case–control study. The point estimates of the odds of HIV infection were greater than 1 among those who received premasticated food from an HIV-infected caregiver and those who received premasticated food from an HIV-infected caregiver who reported having predisposing oral conditions although premasticating; however, these increases were not statistically significant. The lack of statistical significance may have been due to the small sample size and the fact that most caregivers of both cases and controls were receiving HIV antiretroviral therapy, thus potentially decreasing the likelihood of HIV transmission through premastication.
The prevalence of oral condition in children, such as teething, although receiving premasticated food was not more prevalent among case-patients. Oral conditions of the caregiver who prechews food may be more important than the oral conditions of the child. However, it will be difficult to assess the role of the child's oral condition given that most children, whether or not they are HIV infected, experience disruptions in their oral mucosa due to teething at the ages when most children in our study received premasticated food.
Receiving blood or blood products was strongly associated with late-diagnosed HIV infection. Although all case-patients who received blood products received it before their diagnosis, it could not be confirmed whether this receipt was the source or the consequence of HIV infection (eg, blood transfusion because of illness related to HIV infection). However, because of the testing of blood and blood products and the screening of donors since 1985, HIV transmission is presently unlikely to occur from blood transfusion or tissue transplantation in the United States.18 At the time of diagnosis, 3 of these transmissions were considered perinatal; however, due to the review taken in our investigation, these cases were determined not to be perinatal because these children had 2 negative diagnostic test results before their first-positive test result. Further investigation by the health department into the source of the children's infection was limited by incomplete health records for these children. We found no overlap between receiving blood or blood products and receiving premasticated food.
Neither case-patients nor controls were breastfed. This finding may be a result of effective prevention messages for this population. CDC advised HIV-infected women against breastfeeding in 1985,19 and it is likely that the caregivers in our investigation had followed this long-standing recommendation.
To our knowledge, this is the first epidemiologic investigation to assess the risk of HIV infection from receiving premasticated food or to characterize the details of premastication (demographics, frequency, timing, and reasons) in HIV-affected families across several pediatric HIV clinical care sites in the United States, including Puerto Rico. Also, conducting the study in multiple states allowed us to more broadly assess this child-feeding practice in diverse geographical and cultural settings. Moreover, participants were asked to participate in a survey about child-feeding practices, not premastication specifically; therefore, selection bias due to fear that premastication may be viewed negatively by clinic staff was minimal. Studies in South Africa20 and South America21 on premastication have been undertaken as a result of the new information provided from this and previous premastication investigations.
Among the limitations of this study, the small number of case-patients restricted our ability to detect small differences between case-patients and controls and to conduct stratified analysis. Also, our results may have been affected by recall bias. The caregivers of case-patients may have been more likely than the caregivers of controls to remember potential exposures. However, we hypothesize that premastication is a practice that a caregiver is likely to remember, regardless of HIV infection. When considering no caregivers reported breastfeeding, social desirability bias may be an alternative explanation to the interpretation that caregivers adhered to public health recommendations. Furthermore, we interviewed the current caregiver; so, information on the practices of a former caregiver may have not been gathered. Finally, the caregivers' current annual income and education may differ from their income and education at the time of premastication.
Regarding the cross-sectional investigation, given that surveys were completed in a setting where caregivers were accompanied by their children and other family members who may have been unaware of their caregiver's HIV status, HIV status information of caregivers could not be gathered. However, given that all participants were interviewed in pediatric HIV clinics, and 81% of caregivers were biological mothers, it is likely that the majority of the children were born to HIV-infected mothers. Additionally, premastication among caregivers may have taken place many years before survey administration; therefore, recollection of premastication details may be subject to recall bias, especially for caregivers of older children. Furthermore, caregivers may have underreported premastication if they perceived that it would be viewed negatively by clinic staff. Thus, our prevalence estimates may have underestimated premastication in this population. Finally, this cross-sectional investigation included a convenience sample of caregivers of children seen in HIV clinics and may not be generalizable to all HIV-infected caregivers.
Although research on the risk of HIV transmission via premastication is limited and the risk versus benefit of this practice is debated, CDC recommends that to prevent possible transmission, HIV-infected caregivers not premasticate food for children. This recommendation is made carefully, given the implications of such a recommendation on developing or resource-limited countries where food is often scarce and premastication may be necessary to nourish a child. Discouraging premastication among HIV-infected caregivers in the United States is advisable because of the availability of preprepared infant food, including those made more accessible through government-sponsored food programs. More research is needed to determine the role of premastication in postnatal HIV transmission in the United States and in developing countries and to compare the risk of HIV transmission versus the benefit of nutritional supplementation in developing countries where food may be scarce. It will also be important to determine the relative contributions of postnatal HIV transmission risks associated with premastication versus breastfeeding and other modes of transmission. Public health officials should continue to educate caregivers and health care providers about the risk of disease transmission, including HIV infection, via premastication.
1. Whitmore S, Zhang X, Taylor A. Estimated number of births to HIV-positive women, U.S., 2006. Presented at: Conference on Retrovirus and Opportunistic Infections (CROI); February 10, 2009; Montreal, Canada.
2. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2009. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2011.
3. Gaur AH, Dominguez KL, Kalish ML, et al.. Practice of feeding premasticated food to infants: a potential risk factor for HIV transmission. Pediatrics. 2009;124:658–666.
4. Fein SB, Labiner-Wolfe J, Scanlon KS, et al.. Selected complementary feeding practices and their association with maternal education. Pediatrics. 2008;122(suppl 2):S91–S97.
5. Centers for Disease Control and Prevention. Premastication of food by caregivers of HIV-exposed children—nine U.S. Sites, 2009–2010. MMWR Morb Mortal Wkly Rep. 2011;60:273–275.
6. Walburn JN, Pergam JM, Perry SH, et al.. Black child care practices in the Midwest. Pediatrics. 1988;82:789–790.
7. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2008. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2010.
8. Centers for Disease Control and Prevention. Racial/ethnic disparities among children with diagnoses of perinatal HIV infection—34 States, 2004–2007. MMWR Morb Mortal Wkly Rep. 2010;59:97–101.
9. Huang MJ. An epidemiological study on prevalence and risk factors of hepatitis B virus (HBV) infection in preschool children [in Chinese]. Zhonghua Liu Xing Bing Xue Za Zhi. 1990;11:129–132.
10. Steinkuller JS, Chan K, Rinehouse SE. Prechewing of food by adults and streptococcal pharyngitis in infants. J Pediatr. 1992;120(4 pt 1):563–564.
11. Zhou P, Qian Y, Lu H, et al.. Nonvenereal transmission of syphilis in infancy by mouth-to-mouth transfer of prechewed food. Sex Transm Dis. 2009;36:216–217.
12. Taylor DN, Blaser MJ. The epidemiology of Helicobacter pylori infection. Epidemiol Rev. 1991;13:42–59.
13. Harrison R, Benton T, Everson-Stewart S, et al.. Effect of motivational interviewing on rates of early childhood caries: a randomized trial. Pediatr Dent. 2007;29:16–22.
14. Mbulaiteye SM, Pfeiffer RM, Whitby D, et al.. Human herpesvirus 8 infection within families in rural Tanzania. J Infect Dis. 2003;187:1780–1785.
15. Mbulaiteye SM, Walters M, Engels EA, et al.. High levels of Epstein-Barr virus DNA in saliva and peripheral blood from Ugandan mother-child pairs. J Infect Dis. 2006;193:422–426.
16. Bulkow LR, Singleton RJ, Karron RA, et al.. Risk factors for severe respiratory syncytial virus infection among Alaska native children. Pediatrics. 2002;109:210–216.
17. Pelto GH, Zhang Y, Habicht JP. Premastication: the second arm of infant and young child feeding for health and survival? Matern Child Nutr. 2010;6:4–18.
18. Zou S, Dorsey KA, Notari EP, et al.. Prevalence, incidence, and residual risk of human immunodeficiency virus and hepatitis C virus infections among United States blood donors since the introduction of nucleic acid testing. Transfusion. 2010;50:1495–1504.
19. Centers for Disease Control and Prevention. Current trends recommendations for assisting in the prevention of perinatal transmission of human T-lymphotropic virus type III/lymphadenopathy-associated virus and acquired immunodeficiency syndrome. MMWR Morb Mortal Wkly Rep. 1985;34:721–726, 731–732.
20. Maritz ER, Kidd M, Cotton MF. Premasticating food for weaning African infants: a possible vehicle for transmission of HIV. Pediatrics. 2011;28:e579–e590.
21. Gaur AH, Freimanis-Hance L, Dominguez K, et al.. Knowledge and practice of prechewing/prewarming food by HIV-infected women. Pediatrics. 2011;127:E1206–E1211.