It is 1am and Ms Jones has been labouring for 11 hours. She has made slow but semi-steady progress, and now at 9 cm dilation, the fetal heart rate tracing has reverted to a Category II. The fetal head is asynclitic and the estimated fetal weight is 4280 grams. This has been Ms Jones first near-term pregnancy and, together with your midwives, you have cared for her and her husband through her two prior miscarriages. Over this same time frame of pregnancy loss, she gave up her former passion for running marathons and gained nearly 20 kg on to her 180 cm frame. She had an early positive diabetes screen this pregnancy, and by 29 weeks she had failed oral hypoglycaemic therapy and finally achieved euglycaemia on her current dosing of 80 units of neutral protamine hagedron (NPH) and regular insulin. Both she and her husband attended the birthing and breastfeeding classes you give on Thursday nights with your team of midwives, and their birth plan calls for a safe vaginal delivery, tummy time and early immediate breastfeeding.

However, both Ms Jones and her husband have told you that while their goal is a vaginal birth, they have built a trusting relationship with you over these many years and would ask for a recommendation as to when a caesarean section (CS) will be the better option. As you interpret the Category II tracing, reassess the fetal weight, slow descent, persistent asyncliticism and her pelvis, your many years of experience as an obstetrician come to bear. Weighing the risks of ongoing labour and potential for maternal and fetal harm versus the benefit of a CS, you make the decision to now offer Ms Jones a CS. As you discuss the risks, benefits, alternatives and limitations, the immediate benefits relative to the apparently imminent risks of ongoing labour appeal to both Ms Jones and her husband. As you prepare to leave the room and ready for the CS, she asks you one last question, ‘Excuse me Dr Apple, but are there any long-term risks to my baby with a CS?’ With a kind but gentle sigh, you sit back down at her bedside and begin a longer and currently less evidence-based discussion that ends with ever more questions than answers.

For the practising obstetrician and midwife, the above scenario is everyday medicine, where the minute-to-minute decisions must be made in the best interest of mother and baby. To the epidemiologist, this is another CS contributing to a population trend that has been steadily increasing for nearly 50 years. To the neonatologist, this represents an occasion where potential immediate newborn harm will be avoided. To the paediatrician and the lactation consultant, the next focus will be enabling exclusive breastfeeding for the next six months. To the paediatric allergist and immunologist, this is a potential atopic child in the making. To the nutritionist, this is a risk for early childhood obesity. And to the parents? This is their firstborn child for whom they want a healthy life, free of disease and harm. While once all we had to consider was ‘what is best for Ms Jones and her baby’ on the delivery ward, we are now challenged to balance and consider often competing risks with unclear long-term health outcomes.

In this article, we will attempt to clarify and summarise several key studies that have led to the belief (but not yet proven truth) that a CS has the potential for long-term harm for the offspring. We will balance this discussion with evidence reminding us that it may be the underlying indication for the CS or related post-delivery course, rather than the surgery itself, which renders these risks. Finally, we will highlight where we are lacking evidence and clinical trials that may provide meaningful interventions to prevent some of the downstream harm to children of CS deliveries, without compromising immediate maternal or infant health.

While it is understood by the obstetrician and midwife, it is critically important to recognise that there are many pathways that lead to an indication for a CS. Some of these risk factors may necessitate or precipitate CS, or the resultant impact of a CS on the child’s care in the immediate postnatal period, and thus may confound interpretations of longer term CS-attributed risk. Similarly, the immediate infant postnatal and later paediatric interval may share environmental risk factors with the prenatal environment, and if those risks factors are independent predictors of the CS, then mitigation of risk must involve longer-term and precedent interventions. In other words, an infant can only be delivered one of two ways, but the developmental and clinical paths that result in either of these two delivery outcomes are certainly more diverse and complex than the dichotomous nature of delivery suggests.

As birth attendants readily appreciate, in most cases, CS is performed for a specific obstetric, fetal or maternal indication aimed at reducing the near risk of morbidity or mortality for either the mother or the fetus. Some of these indications are sporadic, and may have relatively limited antecedent risk. For example, in instances of complete placenta praevia or vasa praevia, vaginal birth would be lethal to the mother and/or fetus. However, other relative risks have antecedent factors that are likely more insidious and parlay as postnatal risks. In Ms Jones case, her recently acquired preconception obesity rendered a platform for her probable type II diabetes. Together, these served as a likely strong contributor leading to the development of a borderline macrosomic infant, which in turn was ascynclitic and did not descend readily into the pelvis. Moreover, these same maternal metabolic disturbances may be associated with a reduced likelihood for initiating lactation, and similarly have a lower chance of being able to maintain exclusive breastmilk feedings over the ensuing six months. Add into this scenario such typical confounders as familial subpar nutrition and decreased activity, and this becomes an only too often scenario that will culminate in risk for childhood obesity and atopic disease. When examining risks in population-wide studies, it becomes challenging to discern whether it was it the CS or the company it kept that rendered these longer-term risks across the lifetime of the progeny.

CS is the most commonly performed major surgical procedure in the US, accounting for over one million operations annually.¹ Whereas one-in-20 births was via CS four decades ago, now one-in-three to -four births occur by this route. The appropriate rate of CS necessary to assure healthy delivery is not easily determined, since it is dependent on not only multiple maternal, fetal and obstetrical comorbidities and risk factors, but also on the incidence of primary CS. Both the prevalence and relative safety of CS in competent hands and suitable settings has resulted in less apprehension on the part of patients and physicians alike, and the attributable risk of significant morbidity and mortality (at least with primary CS) is extremely low. However, in recent years, tremendous scrutiny has been placed on CS, largely in part to its real or perceived risk of future disease or harm.²

While there is limited maternal risk with primary CS, repeated surgeries in subsequent pregnancies may be associated with bowel and bladder trauma, surgical adhesions and scarring, uterine scar separation, and morbidly adherent placentation (such as, placenta accreta or percreta with indicated need for caesarean hysterectomy and high risk of accompanying massive blood transfusions and surgical complications). Recognising these real hazards (albeit relatively low occurring but highly morbid), there have been multi-pronged approaches undertaken to reduce the CS rate, including efforts specifically aimed at fewer primary surgeries. The completion of the Human Microbiome Project in 2011, which sought to describe and catalogue the ‘healthy’ adult repertoire of commensal bacteria, has since galvanised investigators to closely examine how our microbiota contribute to both health and disease.³ One prominent focus of the field has been to determine when and how newborns begin to acquire commensal microbiota, and how these early patterns of colonisation may influence normal developmental processes. Literature to date indicates that microbiota play a key role in energy homeostasis, as well as patterning the immune and neurologic systems in early life; ergo, acquiring the right microbes at the right time is thought to be essential.⁴

As such, the potential impact of a CS on the early colonisation of the infant microbiome has garnered significant attention in both the scientific and lay press, in part because of the worldwide increase of CS over the past four decades and the reported association of CS with obesity and atopic disease. In particular, there is much debate on whether CS increases the infant’s risk of several diseases later in life, and whether any of these diseases result from a lack of exposure to the mother’s vaginal microbiota during birth.⁵ Recent evidence has indicated a potential association between CS and increased rates of atopic disease and IgE-mediated sensitisation to food allergens, as well as metabolic syndrome and obesity later in life. For instance, a recent large prospective study with 22 068 participants tracked over 16 years found a mere 13 per cent adjusted increased risk of obesity later in life if that individual was delivered by CS.⁶ Additional large and robust epidemiological studies have been performed in an effort to explore the influence of birth mode on disease risk. Such studies are typically in agreement that the late-life risk of either metabolic disease obesity or atopy is increased in infants delivered by CS. For instance, in an epidemiological analysis of 2917 children aged eight years, CS was found to increase the risk of allergy by 1.5 fold (increasing to 4.5 fold if at least one parent was allergic).⁷ A similar trend was reported in asthma risk from a large meta-analysis of 1 206 679 infants (23 studies), where CS was reported to increase the risk of allergy by more than 20 per cent.⁸ Of note, these are relatively small effect size estimates, despite their statistical significance.

Despite the robustness in terms of numbers of subjects analysed and statistical methodology, given the small and modest effect size measures it is still unclear how much of this attributed risk is actually due to the CS procedure itself, rather than the risk factors that led to the CS in the first place (or the postnatal factors that followed).^{6 9 10 11} Investigators studying the human microbiome have been quick to attribute this perceived risk of CS to a lack of exposure to the mother’s vaginal microbiota during delivery.⁶ So, what do we know from the available literature about the association of CS on the gut microbiota and associated disease risk?

While the fetus does not develop in a sterile intrauterine environment,^{12 13 14 15 16 17 18 19}, there is an additional influx of viable microbes approximating the time of delivery. Early studies concluded that because maternal skin and vaginal microbiomes are distinct, with dominance of Staphylococcus sp. or Lactobacillus sp. respectively, that the microbiome of the infant at various sites (skin, oral, respiratory and gut) immediately following birth would closely resemble the corresponding maternal site per birth mode.15 In this landmark work, Dominguez-Bellow and colleagues very eloquently described differences among Mestizo and Amerindian women in Venezuela. Their cohort was comprised of nine women and 10 neonates; four gravidae and their four neonates made up the vaginal cohort, and five gravidae and six neonates were included in the CS cohort. Of these five gravidae delivered by CS, one surgery was performed for a set of male twins. While, except for the twins, the exact weight of each neonate was not given, the methods section of the manuscript states, ‘All mothers had healthy pregnancies and all babies were born at term, without complications. Babies weighed between two and 5.2 kg (the smallest baby was the twin in second order of birth, after his 3 kg brother.’ However, this description does not meet standard definitions of ‘healthy and uncomplicated’. First, the presumptive dizygotic twins (chorionicity was not provided, and both were male, so mono- versus di-zygosity is unclear) were 33 per cent discordant in growth with reported weights of 2 kg and 3 kg (discordance=[birthweight of larger twin-birthweight of smaller twin/larger twin birth weight] x 100). Second, at least one neonate was impressively macrosomic. While we do not know whether this 5.2 kg infant was delivered vaginally or via CS, in the USA, CS is typically offered to diabetic mothers with a fetus estimated at >4.5 kg and a non-diabetic mother with a fetus estimated at more than 5 kg. A 5.2 kg fetus would be 0.8 to 1 kg larger than >98th percentile of the birth population using published WHO growth standards for male and female newborns, respectively.

While this small, but highly cited, landmark study, unfortunately, did not provide the underlying indication for the CS,²⁰ it is not alone in its lack of reporting antecedent risk. In fact, a detailed examination of additional often cited and recent microbiome studies reporting on a presumptively positive association between CS and a perturbed gut microbiome in the offspring, there is a  notable absence of ability to account for primary maternal or fetal comorbidities known to render increased risk of CS, such as maternal prepregnancy BMI, nulliparity, gestational or type II diabetes, maternal weight gain, or fetal ^{21 22 23 24 25 26},Of interest, Yassour et al22 demonstrated that 20 per cent of vaginally delivered infants manifest a microbial pattern hallmarked by a lack of Bacteroides over the first 12 months of life, after which the mode of birth had no effect on microbiome profiles. This lack of significant difference between CS or vaginally delivered infants after one year of life is aligned with data from well-designed studies of the infant microbiome in early childhood.^{22 27 28} When considered collectively and carefully, it may be argued that since low or absent Bacteroides may also occur among vaginally delivered infants,²⁷ the summary conclusions of other studies that are either small in number, low in CS rate, or limited due to treatment of CS as an independent categorical variable and failing to account for antecedent maternal risk factors, ought be taken with a note of caution.

Born by caesarean – but what about the baby’s biome?  © jomphong Shutterstock 239084974

It is not our intent to be critical of the landmark work of our respected colleagues and the importance of their observations, but rather to seek to identify true causal links and not spurious or potentially misclassified risks. Nevertheless, if the medical indication or antecedent risk for the CS, (rather than the surgical procedure itself) puts the child at risk for later-in-life disease, earlier interventions that target the maternal health issue will not only reduce the prevalence of CS, but would be both necessary and sufficient for preventing those later health risks in the child. For example, if it is maternal diet overly rich in calories that both disturbs the establishment of the infant’s microbiome and renders a large baby that cannot readily or safely fit through the birth canal, then any truly helpful interventions would be aimed at reducing the excess calories in her diet. Conversely, if it is a postnatal event that happens to co-associate with CS, then the mode of delivery becomes irrelevant and effective modifiers would be needed after birth. Such postnatal factors would, for instance, include breastmilk versus formula feeding; obese women, who coincidentally have higher rates of CS, are more likely to formula feed. Hence, it is vitally important to understand whether it is the surgery, its indication, or postnatal factors that may be behind reported links between CS and infant atopic diseases or obesity risks later in life.

Alternatively, what evidence exists for other potential factors that independently affect the microbiome and are also associated with increased occurrence of CS? Obesity and a maternal high-fat diet are perhaps the most well-studied confounders that may alternatively explain the observed impact of CS on the microbiome. It is widely accepted that dietary components and obesity status are major drivers of the gut microbiota composition, with well-established examples of causality in murine models. Transference of gut microbiota from obese to germ-free mice (devoid of any microbiota) promotes increased adiposity, while consumption of a high-fat diet similarly promotes adiposity by rapidly altering the gut microbiome.

We have similarly used a primate model  to demonstrate that maternal diet during pregnancy and lactation infers significant and persistent alterations to the juvenile microbiome, even when juveniles are cohoused and switched back to a healthy diet after weaning.²⁹ More recently, we have recapitulated many of these findings in a large human cohort of mothers and their infants, demonstrating that a maternal diet high in fat (>40%), is associated with changes to the infant microbiome at birth and up to at least four to six weeks after birth.³⁰ Notably, one of the major microbiota negatively associated with a maternal high-fat diet, Bacteroides, has been reported by others to be in low abundance due to CS,^{20 21 22 23 24 25 26}, and thus may alternatively account for the perceived impact of CS on the infant microbiome. However, as noted above, maternal diet during pregnancy and other similar potential risk factors for CS have been largely underreported and unaccounted for in many studies.

So, can anything potentially beneficial be done at the time of CS? To answer this question, Dominguez-Bello and colleagues took an innovative and provocative approach to ‘restore’ the infant microbiome of CS infants with maternal vaginal flora.³¹ By incubating a gauze in the mother’s vagina for one hour before surgery and subsequently ‘wiping’ the neonate’s mouth, face and body at the time of delivery, the team described that they were able to ‘partially restore’ bacterial members from the vaginal flora into the neonate microbiome when measured over 30 days. Here again, the details of the study may be important in the interpretation. While only four of 11 CS neonates were exposed to vaginal wiping, in CS babies exposed to maternal vaginal fluids, the skin and oral sites were most comparable to vaginally delivered infants, whereas the anal site remained more comparable to CS infants who did not undergo vaginal wiping. They conclude that CS babies ‘lacked the vaginal bacteria that were restored by swabbing infants with gauze or that were present in vaginally delivered infants, particularly anal and skin Lactobacillus early in life’. However, there are a number of interesting considerations and possible confounders. First, the only neonates and infants that were exclusively breastfed were born vaginally and all CS infants received at least some formula. Second, the supplemental methods reveal inconsistent and limited sampling among subjects at all of the time points and between body sites. For example, the authors comment, ‘In anal samples from exposed infants and vaginally delivered infants, there was an early enrichment of Lactobaccillus followed by a bloom of Bacteroides from week two, which was not observed in newborns that were not exposed to vaginal fluids.’ However, this conclusion is based on a single CS infant sampled at day 7 and 14 and the referenced relative abundance plots do not appear to display ³² Nonetheless, this proof-of-principle study suggests one innovative potential method of partially restoring microbiota in neonates.

Coming back around to Ms Jones, what do we tell her and her husband about the longer-term risks of CS for their baby? While that is ultimately up to the confidence and interpretation of the provider, one option is to simply state that there are potential early signs from studies in other corners of the world that warn of increased risk for asthma, allergy and obesity. However, the proportion of those risks that are actually due to the CS surgery itself, how much can be restored with exclusive breastfeeding and healthy habits throughout childhood, and how much will vary by factors we don’t yet understand.

1. Boyle A, Reddy UM, Landy HJ, et al. Primary cesarean delivery in the United States. Obstet Gynecol. 2013;122(1):33-40. Available from: www.ncbi.nlm.nih.gov/pubmed/23743454
2. Sevelsted A, Stokholm J, Bønnelykke K, Bisgaard H. Cesarean Section and Chronic Immune Disorders. Pediatrics. 2015;135(1):e92-8.
3. Human Microbiome Project Consortium. Structure, function and diversity of the healthy human microbiome. Nature. 2012;486(7402):207-14. Available from: www.nature.com/doifinder/10.1038/nature11234.
4. Tamburini S, Shen N, Wu HC, Clemente JC. The microbiome in early life: implications for health outcomes. Nat Publ Gr. 2016;22(7):713-22. Available from: http://dx.doi.org/10.1038/nm.4142.
5. Neu J, Rushing J. Cesarean versus vaginal delivery: long-term infant outcomes and the hygiene hypothesis. Clin Perinatol.2011;38(2):321-31. Available from: www.ncbi.nlm.nih.gov/pubmed/21645799.
6. Yuan C, Gaskins AJ, Blaine AI, et al. Association Between Cesarean Birth and Risk of Obesity in Offspring in Childhood, Adolescence, and Early Adulthood. JAMA Pediatr. 2016;64(1):1-65.
7. Roduit C, Scholtens S, de Jongste JC, et al. Asthma at 8 years of age in children born by caesarean section. 2009;64(2):107-13. Available from: www.ncbi.nlm.nih.gov/ pubmed/19052046.
8. Thavagnanam S, Fleming J, Bromley A, et al. A meta-analysis of the association between Caesarean section and childhood asthma. Clin Exp Allergy. 2008;38(4):629-33. Available from: www.ncbi.nlm.nih.gov/ pubmed/18352976.
9. Roduit C, Scholtens S, de Jongste JC, et al. Asthma at 8 years of age in children born by caesarean section. 2009;64(2):107-13.Available from: www.ncbi.nlm.nih.gov/ pubmed/19052046.
10. Thavagnanam S, Fleming J, Bromley A, et al. A meta-analysis of the association between Caesarean section and childhood asthma. Clin Exp Allergy. 2008;38(4):629-33.Available from: www.ncbi.nlm.nih.gov/ pubmed/18352976.
11. Salminen S, Gibson GR, McCartney AL, Isolauri E. Influence of mode of delivery on gut microbiota composition in seven year old children. Gut. 2004;53(9):1388- 9. Available from: http://gut.bmj.com/ content/53/9/1388.2.full.
12. Jones HE, Harris K, Azizia M, et al. Differing prevalence and diversity of bacterial species in fetal membranes from very preterm and term labor. PLoS One. 2009;4(12):e8205. Available from: www.pubmedcentral.nih. gov/articlerender.fcgi?artid=2785424&tool =pmcentrez&render type=abstract.
13. Satokari R, Grönroos T, Laitinen K, et al. Bifidobacterium and Lactobacillus DNA in the human placenta. Lett Appl Microbiol. 2009;48(1):8-12. 
14. Rautava S, Collado MC, Salminen S, Isolauri E. Probiotics modulate host-microbe interaction in the placenta and fetal gut: A randomized, double-blind, placebo-controlled trial. Neonatology. 2012;102(3):178-84.
15. tout MJ, Conlon B, Landeau M, et alIdentification of intracellular bacteria in the basal plate of the human placenta in term and preterm gestations. Am J Obstet Gynecol. 2013;208(3):226.e1-226.e7. Available from: http://dx.doi.org/10.1016/j. ajog.2013.01.018..
16. Dong XD, Li XR, Luan JJ, et al. Bacterial communities in neonatal feces are similar to mothers’ placentae. Can J Infect Dis Med Microbiol. 2015;26(2):90-4.
17. Aagaard K, Ma J, Antony KM, et al. The placenta harbors a unique microbiome. Sci Transl Med. 2014;6(237):237ra65. Available from: www.ncbi.nlm.nih.gov/ pubmed/24848255.
18. Collado MC, Rautava S, Aakko J, et al. Human gut colonisation may be initiated in utero by distinct microbial communities in the placenta and amniotic fluid. Sci Rep. 2016;6(October 2015):23129. Available from: www.nature.com/articles/srep23129.
19. Bassols J, Serino M, Carreras-Badosa G, et al. Gestational Diabetes is associated with changes in Placental Microbiota and Microbiome. Pediatr Res. 2016. Available from: www.nature.com/doifinder/10.1038/ pr.2016.155.
20. Dominguez-Bello MG, Costello EK, Contreras M, et al. Delivery mode shapes the acquisition and structure of the initial microbiota across multiple body habitats in newborns. Proc Natl Acad Sci USA. 2010;107(26):11971-5. Available from: www.pubmedcentral.nih.gov/articlerender.fc gi?artid=2900693&tool=pmcentrez&render type=abstract.
21. Madan JC, Hoen AG, Lundgren SN, et al. Association of Cesarean Delivery and Formula Supplementation With the Intestinal Microbiome of 6-Week-Old Infants. JAMA Pediatr. 2016;170(3):212- 9. Available from: www.ncbi.nlm.nih.gov/pubmed/ 26752321.
22. Bokulich NA, Chung J, Battaglia T, et al. Antibiotics, birth mode, and diet shape microbiome maturation during early life. SciTransl Med. 2016;8(343):1-14.
23. Bäckhed F, et al. Dynamics and Stabilization of the Human Gut Microbiome duringthe First Year of Life. Cell Host Microbe. 2015;17:852.
24. Penders J, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics. 2006;118:511-521.
25. Vatanen T, et al. Variation in Microbiome LPS Immunogenicity Contributes to Autoimmunity in Humans. 2016;165:842-853.
26. Azad MB, et al. Gut microbiota of healthy Canadian infants: profiles by mode of delivery and infant diet at 4 months. Can Med Assoc J. 2013;185:385-394.
27. Yassour M, Vatanen T, Siljander H, et al. Natural history of the infant gut microbiome and impact of antibiotic treatment on bacterial strain diversity and stability.Sci Transl Med. 2016;8(343):343-81. Available from: www.ncbi.nlm.nih.gov/ pubmed/27306663.
28. Stewart C, Skeath T, Nelson A, et al. Preterm gut microbiota and metabolome following discharge from intensive care. Sci Rep. 2015;5:17141.
29. Ma J, Prince AL, Bader D, et al. High-fat maternal diet during pregnancy persistently alters the offspring microbiome in a primate model. Nat Commun. 2014;5:3889. Available from: www.ncbi.nlm.nih.gov/ pubmed/24846660.
30. Chu DM, Antony KM, Ma J, et al. The early infant gut microbiome varies in association with a maternal high-fat diet. Genome Med. 2016;8(1):77. Available from: http://www. ncbi.nlm.nih.gov/pubmed/27503374.
31. Dominguez-Bello MG, De Jesus-Laboy KM, Shen N, et al. Partial Restoration of the microbiota of cesarean-born infants via vaginal microbial transfer. Nat Med. 2016;(Feb):1-4. Available from: http://dx.doi.org/10.1038/nm.4039.
32. Dominguez-Bello MG, De Jesus-Laboy KM, Shen N, et al. Partial Restoration of the microbiota of cesarean-born infants via vaginal microbial transfer. Nat Med. 2016;(Feb):1-4. Available from: http:// dx.doi.org/10.1038/nm.4039.