Global Health Action ISSN: (Print) (Online) Journal homepage: www.tandfonline.com/journals/zgha20 Is exclusive breastfeeding for six-months protective against pediatric tuberculosis? Juan A. Flores, Julia Coit, Milagros Mendoza, Segundo R. Leon, Kelika Konda, Leonid Lecca & Molly F. Franke To cite this article: Juan A. Flores, Julia Coit, Milagros Mendoza, Segundo R. Leon, Kelika Konda, Leonid Lecca & Molly F. Franke (2021) Is exclusive breastfeeding for six-months protective against pediatric tuberculosis?, Global Health Action, 14:1, 1861922, DOI: 10.1080/16549716.2020.1861922 To link to this article: https://doi.org/10.1080/16549716.2020.1861922 © 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. View supplementary material Published online: 03 Jan 2021. Submit your article to this journal Article views: 3220 View related articles View Crossmark data Citing articles: 1 View citing articles Full Terms & Conditions of access and use can be found at https://www.tandfonline.com/action/journalInformation?journalCode=zgha20 GLOBAL HEALTH ACTION 2021, VOL. 14, 1861922 https://doi.org/10.1080/16549716.2020.1861922 SHORT COMMUNICATION Is exclusive breastfeeding for six-months protective against pediatric tuberculosis? Juan A. Flores a,b, Julia Coit c, Milagros Mendoza d, Segundo R. Leon a, Kelika Konda e,f, Leonid Lecca d and Molly F. Franke c aEscuela Profesional de Tecnología Médica, Universidad Privada San Juan Bautista, Lima, Peru; bFacultad de Salud Pública, Universidad Peruana Cayetano Heredia, Lima, Peru; cDepartment of Global Health and Social Medicine, Harvard Medical School, Boston, MA, USA; dDirection of Research, Socios En Salud at Partners in Health, Lima, Peru; eDivision of Infectious Diseases, University of California, Los Angeles, CA, USA; fCenter for Interdisciplinary Studies in Sexuality, AIDS and Society, and Laboratory of Sexual Health, Universidad Peruana Cayetano Heredia, Lima, Peru ABSTRACT ARTICLE HISTORY Experts recommend exclusive breastfeeding from birth to six months because it protects against Received 7 October 2020 deadly childhood illness, including respiratory tract infections and diarrhea. We hypothesized that Accepted 6 December 2020 exclusive breastfeeding would decrease the risk of active tuberculosis (TB) in children. We analyzed RESPONSIBLE EDITOR cross-sectional data from 279 children in Lima, Peru aged 6 to 59 months with TB symptoms and Jennifer Stewart Williams, a close adult contact with TB. Mothers self-reported breastfeeding, and children were evaluated for TB Umeå University, Sweden per national guidelines. To quantify the association between exclusive breastfeeding and TB, we estimated prevalence ratios using a generalized linear model with a log link, binomial distribution, and KEYWORDS robust variance. Twenty-two percent of children were diagnosed with TB and 72% were exclusively Mycobacterium tuberculosis; breastfed for six months. We found no evidence that six months of exclusive breastfeeding was child; risk factor; breastfeeding; infant associated with TB disease in either bivariate analyses (prevalence ratio [PR] = 1.5; 95%CI = 0.8–2.5) or multivariable analyses adjusting for sex and socioeconomic status (adjusted PR = 1.6; 95% [CI] = 0.9–2.7). In post hoc analyses among children whose close TB contact was their mother, we found evidence of a weak positive association between breastfeeding and TB (aPR = 2.1; 95% [CI] = 0.9–4.9). This association was not apparent among children whose close contact was not the mother (aPR = 1.2; 95%[CI] = 0.6–2.4). Our results raise the possibility that children who are breastfed by mothers with TB may be at increased risk for TB, given the close contact. Due to the cross-sectional study design, these results should be interpreted with caution. If these findings are confirmed in longitudinal analyses, future interventions could aim to minimize TB transmission from mothers with TB to breastfeeding infants. Background Exclusive breastfeeding protects against common pediatric illnesses including diarrhea, gastro-enteritis, otitis Tuberculosis (TB) is an important cause of morbidity and media and respiratory tract infections including pneumo- mortality among children. In 2018, 1.12 million children nia [12–15]. However, it is unknown whether it affords under 15 years old became ill with TB; nearly 25% of protection against TB. Given its role in immune system whom died of the disease [1]. Neonatal TB mortality can maturation, response and memory [16], we hypothesized be as high as 40–60% [2–5] and mothers with active TB that exclusive breastfeeding during the first 6 months of life can transmit the infection to their infants in-utero or post- may confer protection against TB. To explore this hypoth- partum [6]. esis, we quantified the association between exclusive Globally, 41% of infants are exclusively breastfed for the breastfeeding and pediatric TB in children aged 6 to first six month of life, as recommended by the World 59 months of age. Health Organization (WHO) [7]. However, this rate is far lower than the 2030 global target of 70% [8]. Exclusive breastfeeding means that an infant receives Methods only breastmilk and no other food or liquids, including Study design and setting water; the only exceptions are oral-rehydration salts or medications [9]. Evidence supporting this recommenda- We conducted a secondary analysis of cross-sectional tion suggests that breast milk contains bioactive compo- data from a prospective pediatric TB diagnostics study nents and anti-inflammatory properties that promote (the ‘parent study’) in Lima, Peru that enrolled children immune development [10,11]. under 15 years old [17–20] who were undergoing CONTACT Juan A. Flores antonio.flores@upsjb.edu.pe Escuela Profesional de Tecnología Médica, Universidad Privada San Juan Bautista, Av. José Antonio Lavalle N° 302-304, Lima 15067, Peru This article has been republished with minor changes. These changes do not impact the academic content of the article. Supplemental data for this article can be accessed here. © 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 2 J. A. FLORES ET AL. evaluation for TB in accordance with Peruvian guidelines less than six month) to reflect current WHO recommen- [21]. In 2016, Peru just met the 2030 WHO global dations and continuous (in months) to examine a dose breastfeeding target with 70% exclusive breastfeeding response. coverage [8,22]. We estimated prevalence ratios using a generalized linear model with a log link, binomial distribution and robust variance to quantify the association between Study population and data collection exclusive breastfeeding and TB. Given the small dataset, Between May 2015 and February 2018, children were we aimed for a parsimonious multivariable model. We recruited into the parent study from 49 health centers in adjusted for socioeconomic index (SEI) and sex, which Lima. To maximize TB case yield, eligible children were were identified a priori as potential confounders. Then, those with a history of contact with an adult with from a list of other biologically plausible potential con- pulmonary TB within the previous two years and with founders (e. age in continuous [24], whether the contact one or more symptoms compatible with TB, as defined was the child’s mother, shared a bed with the child, by an expert panel [23]: persistent, unremitting, and smoked cigarettes or drank alcohol [25,26], we identi- unexplained cough for >2 weeks, unexplained weight fied those associated with TB at a p-value threshold of loss, unexplained fever for >1 week, or unexplained ≤0.20 and adjusted for them in the final model. The SEI fatigue or lethargy. Study staff implemented standar- variable was constructed using a principal component dized questionnaires to collect clinical, social and demo- analysis (PCA) based on the household characteristics graphic data via a face-to-face interview with caregivers. and family possessions. Because we expected the SEI to Figure 1 shows the flowchart for inclusion in this predict TB risk in a linear fashion, we modelled it as analysis. a continuous linear variable (supplementary material 1). When the binomial model failed to converge, we used a Poisson distribution. Data analysis We conducted post hoc stratified analyses to exam- Children classified as having TB were those who had ine potential effect modification by age and whether the a positive culture result or, in its absence, a clinical diag- child’s close contact with TB was his/her mother. With nosis made on the basis of chest x-ray, symptomatology regard to age, we hypothesized that any protective effect and tuberculin skin testing, by a pediatric pulmonologist. of breastfeeding on TB might be most apparent in Exclusive breastfeeding was determined based on mater- children less than 24 months of age; the immune system nal report and analyzed as dichotomous (6 months vs. is weakest at this time [27] and exposure to Figure 1. Flowchart of enrollment and exclusions. GLOBAL HEALTH ACTION 3 breastfeeding is likely to be more proximal to TB expo- (prevalence ratio [PR] = 1.5; 95%[CI] = 0.8–2.5). sure in younger children than in older children. None of the potential confounders identified Whether the mother was the child’s close contact with a priori were associated with TB at the defined TB is relevant because a protective effect of breastfeed- threshold of ≤0.20; therefore, multivariable analyses ing could be attenuated by increased exposure to were adjusted for sex and SEI. In multivariate ana- Mycobacterium tuberculosis during breastfeeding. Data lyses, children who were exclusively breastfed for six analyses were performed using Stata v16.0 (StataCorp, months had a 60% higher prevalence of active TB College Station, TX). compared to children who were not exclusively breastfed; however, the confidence interval was wide (adjusted PR = 1.6; 95%[CI] = 0.9–2.7; p = 0.111). Results There was no association when we evaluated exclu- Of the 628 children enrolled in the parent study, 298 sive breastfeeding as a continuous variable in months (47%) were between 6 and 59 months of age and (adjusted PR = 1.0; 95%[CI] = 0.9–1.1, p = 0.983). therefore eligible for inclusion in the present analysis. Of these, 19 (6%) lacked data on breastfeeding and Effect modification by age and maternal TB were excluded (Figure 1). Of the 279 children history included, 61 (22%) were diagnosed with TB, of whom 12 (20%) had a positive culture. Among all Stratifying by age, we found that children between 6 children included, 72% (200/279) were exclusively and 23 months of age who were exclusively breastfed breasted for six months. Among those who were for six months had twice the prevalence of TB as exclusively breastfed for less than six months or not compared to those that were not exclusively breastfed at all (n = 79), the median duration of exclusive for 6 months. This association was greatly attenuated breastfeeding was 4 months (IQR = 2–5). Patient among older children; however confidence intervals characteristics are described in Table 1. were wide and significance testing suggested that chance was a likely explanation for these differences (aPR = 2.2; 95%[CI] = 0.8–5.7 in children aged Bivariate and multivariate analyses 6–23 months old versus aPR = 1.2; 95% We found no association between exclusive breast- [CI] = 0.6–2.5 in children over 23 months; p-value feeding and active TB in the bivariate analysis for interaction = 0.390). Similarly, among children whose close contact was their mother, we found Table 1. Characteristics of children exposed to exclusive that six months of exclusive breastfeeding was asso- breastfeeding (N = 279)*. ciated with twice the prevalence of TB, and this was Exclusive breastfeeding for six not observed among children whose contact was months someone other than their mother (aPR = 2.1;95% Total Yes (N = 200) No (N = 79) [CI] = 0.9–4.9 versus aPR = 1.2;95%[CI] = 0.6–2.4, Variables n (%) n (%) n (%) respectively; p-value for interaction = 0.380) (Table Age (months) < 24 112 (40) 80 (40) 32 (41) 2). Once again, confidence intervals were wide, and 24 to 59 167 (60) 120 (60) 47 (59) p-values for interaction suggested that chance was Sex Male 149 (53) 103 (52) 46 (58) one likely explanation for these differences. Female 130 (47) 97 (48) 33 (42) Index case is the mother (N = 275) Discussion Yes 96 (35) 63 (32) 33 (42) No 179 (65) 133 (68) 46 (58) We found no evidence of a protective association Index case shares bed with children (N = 275) between exclusive breastfeeding and TB in children Yes 118 (43) 80 (41) 38 (48) between 6 and 59 months of age. Among children No 157 (57) 116 (59) 41 (52) Index case drinks aged 6 to 23 months and children whose TB contact (N = 274) was their mother, we found weak evidence that 6 Yes 171 (62) 126 (64) 45 (58) No 103 (38) 70 (36) 33 (42) months of exclusive breastfeeding was positively asso- Index case smokes ciated with TB. This raises the possibility that expo- (N = 273) Yes 110 (40) 84 (43) 26 (33) sure to TB during breastfeeding plays a more No 163 (60) 111 (57) 52 (67) determinant role in the child’s risk of disease than SEI (N = 278) the theoretical protection against TB that six months Median (IQR) −0.05 (2.8) −0.05 (2.3) −0.13 (2.4) TB diagnosed of exclusive breastfeeding may grant. Yes 61 (22) 48 (24) 13 (16) Breastfeeding is a cheap and healthy way to feed No 218 (78) 152 (76) 66 (84) a baby. The WHO recommends that breastfeeding *Unless otherwise noted. Abbreviation: IQR = interquartile range, TB = tuberculosis, continue irrespective of the TB status of the mother SEI = Socioeconomic Index. when she is on appropriate anti-TB treatment for 4 J. A. FLORES ET AL. Table 2. Associations between exclusive breastfeeding for six months and TB in children aged 6 to 59 months. Models‡ N PR CI 95% p value aPR CI 95% p value All children 279 1.5 0.8–2.5 0.183 1.6** 0.9–2.7 0.111 Stratified by age 6 to 23 months of age 112 2.2 0.8–5.9 0.118 2.2** 0.8–5.7 0.109 24 to 59 months of age 167 1.1 0.6–2.2 0.722 1.2 0.6–2.5 0.540 Stratified by contact Contact was mother 96 2.0 0.8–4.9 0.132 2.1 0.9–4.9 0.097 Contact was not mother 179 1.2 0.6–2.4 0.672 1.2** 0.6–2.4 0.640 ‡Analyses were conducted using generalized linear binomial regression models, which adjusted for sex and socioeconomic index. **These variables had missing data for one child. Abbreviation: PR = Prevalence ratio; aPR = Adjusted prevalence ratio; CI = Confidence interval. active disease [28]. Breastfeeding cessation is interpretation of results, writing, editing and data manage- recommended if a woman has active tubercular ment. MM: study implementation in the field. SRL: Study breast lesions or tubercular mastitis [6]. In cases implementation. KK: review of manuscript. LL: study design. where a mother has active TB and is breastfeeding MF: conceptualized and designed the study and contributed to writing and editing. All authors critically reviewed the manu- her infant, general protective measures, such as script. All authors read and approved the final manuscript. masks, could be used, particularly in the early phases of treatment [6,28]. Counseling breastfeed- ing women with TB on the recommendations and Disclosure statement risks and ensuring timely and appropriate TB- No potential conflict of interest was reported by the treatment will contribute to TB prevention in author(s). infants [29]. Future longitudinal research on TB in pregnant and breastfeeding women will be impor- Ethics and consent tant to more fully understand the relationship between exclusive breastfeeding and TB and poten- The study was approved by the Ethics Committee at Peru’s tial interventions that will maximize health and National Institute of Health and the Office of Human well-being for them and their infants. Research Administration at Harvard Medical School. Caregivers provided written informed consent for the child’s Among the limitations of this study are the cross- participation in the TB diagnostic study. We also obtained sectional design and lack of information on the timing of approval from the Ethics Committee of the Peruvian breastfeeding relative to exposure to a close contact with University Cayetano Heredia for secondary data analysis. TB and limited data on other potential confounders. Although we found no significant association between Funding information TB and exclusive breastfeeding in children aged 6 to 59 months, limitations in our study design preclude us This work was supported by the National Institutes of from discarding the possibility that exclusive breastfeed- Health (NIH) under the Center of Excellence in ing may afford protection against TB. The benefits of Translational Research (CETR) grant U19 AI109755. Juan A. Flores is a doctoral student studying Epidemiological breastfeeding in children are indisputable [10–15]. In Research at Universidad Peruana Cayetano Heredia under young children whose close contact was their mother FONDECYT/CIENCIACTIVA scholarship EF033-235- and in children between 6 and 23 months of age, we 2015. found weak evidence that exclusive breastfeeding was associated with an elevated risk of TB, underscoring the critical importance of rapid diagnosis and treatment of Paper context pregnant women and breastfeeding mothers with TB. Despite not having found any association between breastfeed- ing and TB in children aged 6 to 59 months overall, we found weak evidence that exclusive breastfeeding was associated Acknowledgments with an elevated risk of TB among children whose close contact was their mother. These associations should be eval- The authors are grateful to children and guardians who uated in longitudinal studies. Timely diagnosis and treatment participated in this study as well as the staff at Socios En of women with TB, along with counseling for those who plan Salud Sucursal Perú in Lima, where the study was mana- to breastfeed, will contribute to TB prevention in infants. ged. Moreover, the first author gives special thanks to the Department of Global Health and Social Medicine at Harvard University for providing him with a productive ORCID training and mentorship by Dr. Franke. Juan A. Flores http://orcid.org/0000-0002-5162-0782 Julia Coit http://orcid.org/0000-0002-9837-8444 Author’s contributions Milagros Mendoza http://orcid.org/0000-0002-6424- 5459 JAF: Study implementation, led the data analysis and interpre- Segundo R. Leon http://orcid.org/0000-0002-5630-5714 tation and wrote the manuscript. JC: study implementation, Kelika Konda http://orcid.org/0000-0003-2836-0174 GLOBAL HEALTH ACTION 5 Leonid Lecca http://orcid.org/0000-0003-1363-3985 [16] Kelly D, Coutts AG. Early nutrition and the develop- Molly F. Franke http://orcid.org/0000-0002-4890-5728 ment of immune function in the neonate. Proc Nutr Soc. 2000;59:177–185. [17] Flores JA, Calderon R, Mesman AW, et al. Detection References of Mycobacterium tuberculosis DNA in Buccal Swab samples from children in Lima, Peru. Pediatr Infect [1] WHO. Global tuberculosis report 2019. WHO; 2019 Dis J. 2020;39:e376–e80. [cited 2020 Nov 18]. Available from: https://www. [18] Mesman AW, Soto M, Coit J, et al. Detection of who.int/tb/publications/global_report/en/ Mycobacterium tuberculosis in pediatric stool samples [2] Loto OM, Awowole I. Tuberculosis in pregnancy: a using TruTip technology. BMC Infect Dis. 2019;19:563. review. J Pregnancy. 2012;2012:379271. [19] Tafur KT, Coit J, Leon SR, et al. Feasibility of the [3] Peng W, Yang J, Liu E. Analysis of 170 cases of string test for tuberculosis diagnosis in children congenital TB reported in the literature between between 4 and 14 years old. BMC Infect Dis. 1946 and 2009. Pediatr Pulmonol. 2011;46:1215–1224. 2018;18:574. [4] Skevaki CL, Kafetzis DA. Tuberculosis in neonates [20] Coit J, Mendoza M, Pinedo C, et al. Performance of and infants: epidemiology, pathogenesis, clinical man- a household tuberculosis exposure survey among chil- ifestations, diagnosis, and management issues. dren in a Latin American setting. Int J Tuberc Lung Paediatr Drugs. 2005;7:219–234. Dis. 2019;23:1223–1227. [5] Whittaker E, Kampmann B. Perinatal tuberculosis: [21] MINSA. Norma técnica de salud para la atención integral new challenges in the diagnosis and treatment of de las personas afectadas por tuberculosis - Ministerio de tuberculosis in infants and the newborn. Early Hum Salud del Perú Lima: MINSA. 2013 [cited 2020 Nov 18]. Dev. 2008;84:795–799. Available from: http://www.tuberculosis.minsa.gob.pe/por [6] Di Comite A, Esposito S, Villani A, et al. Italian taldpctb/recursos/20180308083418.pdf pediatric TBSG. How to manage neonatal [22] INEI. Encuesta Demográfica y de Salud Familiar 2016. tuberculosis. J Perinatol. 2016;36:80–85. Lima; 2016. [cited 2020 Nov 18]. Available from: [7] WHO/UNICEF. Global breastfeeding scorecard, 2018. https://www.inei.gob.pe/media/MenuRecursivo/publi 2018 [cited 2020 Nov 18]. Available from: https:// caciones_digitales/Est/Lib1433/index.html www.who.int/nutrition/publications/infantfeeding/glo [23] Cuevas LE, Browning R, Bossuyt P, et al. Evaluation of bal-bf-scorecard-2018.pdf tuberculosis diagnostics in children: 2. Methodological [8] WHO/UNICEF. The extension of the 2025 Maternal, issues for conducting and reporting research evalua- Infant and Young Child nutrition targets to 2030. tions of tuberculosis diagnostics for intrathoracic WHO/UNICEF; 2018 [cited 2020 Nov 18]. Available tuberculosis in children. Consensus from an expert from: https://www.who.int/nutrition/global-target panel. J Infect Dis. 2012;205:S209–15. -2025/discussion-paper-extension-targets-2030.pdf [24] Marais BJ, Gie RP, Schaaf HS, et al. Childhood pul- [9] WHO. The World Health Organization’s infant feed- monary tuberculosis: old wisdom and new challenges. ing recommendation. 2001 [cited 2020 Nov 18]. Am J Respir Crit Care Med. 2006;173:1078–1090. Available from: http://www.who.int/nutrition/topics/ [25] Patra J, Bhatia M, Suraweera W, et al. Exposure to infantfeeding_recommendation/en/ second-hand smoke and the risk of tuberculosis in [10] Hanson LA. Session 1: feeding and infant develop- children and adults: a systematic review and ment breast-feeding and immune function. Proc meta-analysis of 18 observational studies. PLoS Med. Nutr Soc. 2007;66:384–396. 2015;12:e1001835; discussion e. [11] Palmeira P, Carneiro-Sampaio M. Immunology of [26] Fiske CT, Hamilton CD, Stout JE. Alcohol use and breast milk. Rev Assoc Med Bras (1992). clinical manifestations of tuberculosis. J Infect. 2016;62:584–593. 2009;58:395–401. [12] Chantry CJ, Howard CR, Auinger P. Full breastfeed- [27] Simon AK, Hollander GA, McMichael A. Evolution of ing duration and associated decrease in respiratory the immune system in humans from infancy to old tract infection in US children. Pediatrics. age. Proc Biol Sci. 2015;282:20143085. 2006;117:425–432. [28] WHO. Guidance for national tuberculosis programmes on [13] Duijts L, Jaddoe VW, Hofman A, et al. Prolonged and the management of tuberculosis in children Ginebra. exclusive breastfeeding reduces the risk of infectious Switzerland: WHO; 2014 [cited 2020 Nov 18]. Available diseases in infancy. Pediatrics. 2010;126:e18–25. from: https://apps.who.int/iris/bitstream/handle/10665/ [14] Howie PW, Forsyth JS, Ogston SA, et al. Protective 112360/9789241548748_eng.pdf?sequence=1 effect of breast feeding against infection. BMJ. [29] WHO. Breastfeeding and maternal tuberculosis. 1990;300:11–16. WHO; 1998 [cited 2020 Nov 18]. Available from: [15] Popkin BM, Adair L, Akin JS, et al. Breast-feeding and http://www.who.int/maternal_child_adolescent/docu diarrheal morbidity. Pediatrics. 1990;86:874–882. ments/pdfs/breastfeeding_and_maternal_tb.pdf