Mode of delivery directs the phagocyte functions of infants for the first 6 months of life


Minna-MaijaGrönlund MD Department of Paediatrics, Turku University Central Hospital, Kiinamyllynkatu 4–8, FI-20520 Turku, Finland.   E-mail:


Factors that direct the immune responsiveness of the newborn beyond the immediate post-natal period are not known. We investigated the influence of mode of delivery and type of feeding on the phagocyte activity during the first 6 months of life. Sixty-four healthy infants (34 delivered vaginally and 30 by elective Caesarean section) were studied at birth and at the ages of 2 and 6 months. Phagocyte functions were studied by measuring the chemiluminescence (CL) activity of whole blood and isolated leucocytes and by investigating the expression of phagocyte receptors (FcγRI (CD64), FcγRII (CD32), FcγRIII (CD16), CR1 (CD35), CR3 (CD11b) and FcαR (CD89)) on neutrophils, monocytes and eosinophils by using receptor-specific MoAbs and immunofluorescence flow cytometry. Infants born by elective Caesarean section had significantly higher CL activity than those delivered vaginally during the entire 6-month follow up. In addition, infants who received formula feeds had significantly higher CL activity at 6 months of age and higher expression of FcγRI-, Fcα- and CR3-receptors on neutrophils than infants exclusively breast-fed. We suggest that stress reaction associated with labour influences the phagocytic activity measured in the cord blood but later during infancy the intraluminal antigens, gut microflora and diet, become important determinants in immune programming of human individuals.


The increasing rate of Caesarean sections has evoked world-wild concern [ 1[2][3]–4]; between 1970 and 1990 the Caesarean section rate in the USA rose from 5% to 25% [ 4, 5]. Since it seems that the indications for Caesarean sections are changing to greater permissiveness [ 1, 2], this trend can be expected to continue. However, little is known about the impact of this phenomenon on the overall immune responsiveness of the newborn infant.

There are data to suggest that the method of delivery modifies the immunological balance of the newborn infant. Labour leads to leucocytosis, which is selective for neutrophils [ 6[7][8]–9], monocytes and natural killer (NK) cells [ 9, 10]. Consequently, infants delivered by Caesarean section without labour have significantly lower leucocyte and neutrophil counts than vaginally delivered neonates [ 6, 8, 9, 11]. Furthermore, contradictory results exist concerning the effect of method of delivery on the functional activity of phagocytes. Phagocytes in vaginally delivered neonates have been reported to be more active by some investigators [ 12[13]–14] and less active by others [ 11], while some [ 8, 15] have observed no differences. No studies have been conducted, to our knowledge, on the duration of the influence of method of delivery on the phagocyte functions during the critical period of immune development of infancy.

The purpose of this study was to evaluate the impact of mode of delivery and type of feeding on phagocyte functions in 64 young infants during the first 6 months of life.



The study consisted of 64 full-term deliveries of healthy mothers at the Department of Obstetrics and Gynecology, University Central Hospital, Turku, Finland. Thirty-four infants were delivered vaginally and 30 by elective Caesarean section undertaken for reasons of feto-maternal disproportion or malpresentation. After delivery the newborns were randomly admitted to one of the two maternity wards for healthy newborn infants. Parents were informed about the study at the maternity wards and all infants whose parents chose to take part in the study were included after written informed consent was obtained. The mothers delivered by Caesarean section received intravenously 2 g of ampicillin as selective intrapartum chemoprophylaxis 2 h before delivery. None of the mothers had received anti-microbial agents within 1 month prior to delivery. The study was approved by the Joint Commission on Ethics of the Turku University and the Turku University Central Hospital.

The mothers kept a diary on the infants' nutrient intake (breast feeding/quality and volume of formula feeding), use of any anti-microbial treatment, frequency of normal/colicky cry, and quality and number of stools and vomiting for the first 2 months. Further information on feeding, health, growth and anti-microbial treatment between the ages of 2 and 6 months was collected at a scheduled visit at 6 months of age. The children were clinically examined at the ages of 2 and 6 months, always by the same investigator (M.-M.G.). All 64 infants remained in the study during the whole study period. Blood samples were taken from 62 infants at 2 months of age (in two infants the blood drawing did not succeed) and from 64 infants at 6 months of age. However, because of the small amount of blood, not all assays could be done from every sample ( Table 1). In addition, 11 babies received anti-microbial treatment between 2 and 6 months of age for upper respiratory tract infections, and subsequent results pertaining to these children at 6 months of age were excluded from the analysis.

Table 1.  . Number of different assays done at follow-up visits * Chemiluminescence measured from whole blood. † Chemiluminescence measured from isolated leucocytes.Thumbnail image of

Preparation of blood samples and counting of leucocytes

Umbilical cord blood samples (mixed blood) and peripheral venous blood samples at the ages of 2 and 6 months were collected for analyses. A 1–5-ml sample was taken into a 5-ml EDTA tube. Total and differential counts of leucocytes were determined from the EDTA blood samples with a Technicon H-2 automatic blood cell analyser (Technicon Instruments Co., Tarrytown, NY). Subsequently, blood erythrocytes were lysed with ammonium chloride (1.5 ml blood, 8.5 ml of 0.83% ammonium chloride) at 20°C for 15 min. After lysing, leucocytes were centrifuged (400 g for 10 min at 4°C) and the leucocyte pellet was resuspended in 500 μl of ice-cold Ca2+- and Mg2+-free Hank's balanced salt solution supplemented with 0.1% gelatin (gHBSS; Sigma Chemical Co., St Louis, MO).

Measurement of respiratory burst activity by chemiluminescence


Luminol (5-amino-2,3-dihydro-1,4-phthalazinedione) and zymosan (Zymosan A from Saccharomyces cerevisiae) were purchased from Sigma Chemical Co. A stock solution of 10 mmol/l luminol was prepared in 0.2 mol/l sodium borate buffer pH 9.0 and stored at −20°C. Zymosan was prepared by boiling for 20 min and washing twice with gHBSS. For opsonization, the zymosan suspension in gHBSS was incubated in 65% pooled serum from healthy volunteers for 40 min at 37°C and washed twice with gHBSS. Unopsonized and opsonized zymosan were resuspended (20 mg/ml) in gHBSS and stored at −70°C.

Measurement of chemiluminescence

For zymosan-induced chemiluminescence (CL), reaction mixtures contained a total volume of 0.5 ml of gHBSS including 0.4 mmol/l luminol, 1 mg of opsonized/unopsonized zymosan and 100 nl of blood for CL measurement of whole blood (WBCL), or 0.4 mmol/l luminol, 50 μg of opsonized/unopsonized zymosan and 100 000 leucocytes for CL measurement of isolated leucocytes (ILCL). In control tubes zymosan was replaced with gHBSS. The reaction was monitored at 37°C for 60 min. Chemiluminescence was measured with a BioOrbit 1251 luminometer with MultiUse 2.01 software (Bio-Orbit Ltd, Turku, Finland). The capacity of phagocytes to emit CL was expressed as the peak value (mV) during the 60-min period and the time to reach the peak CL value as peak time (min). All CL assays were performed in duplicate.

For milk-induced CL, milk protein-coated polystyrene vials were prepared by incubating 100 μl of gHBSS-diluted low-fat milk (homogenized and pasteurized) with a total protein content of 66 μg/vial or 100 μl gHBSS as control for 30 min at 37°C. The liquids were removed by vacuum and the vials were stored at −20°C. Similarly, for gliadin-induced CL, the vials were incubated in the presence of 4 μg of gliadin (Sigma) in 100 μl of gHBSS. Milk- and gliadin-induced CL were detected by adding 0.4 mmol/l luminol and 50 μl of whole blood or 100 000 isolated leucocytes in 0.5 ml of gHBSS. CL was measured at 37°C for 60 min. The reaction was monitored and measured, and the results expressed as in the zymosan-induced CL above.

Measurement of phagocyte receptor expression


FITC-labelled anti-FcγRI (CD64), anti-FcγRII (CD32), and anti-FcγRIII (CD16) were purchased from Medarex Inc. (Annandale, NJ). FITC-conjugated anti-CR1 (CD35) was obtained from CLB Immune Reagents (Amsterdam, The Netherlands), and PE-conjugated anti-CR3 (CD11b) and anti-FcαR (CD89) from PharMingen (San Diego, CA). Isotype-matched MoAbs directed to an irrelevant antigen were used as controls (anti-mouse IgG1–FITC conjugate, anti-mouse IgG1–PE conjugate, anti-mouse IgG2b–PE conjugate; Caltag Labs, San Francisco, CA).

Measurement of receptor expression

Leucocytes (3  ×  105) were incubated with MoAbs in polystyrene vials for 30 min at 4°C. The incubation volume was 50 μl. After incubation, the cells were washed with cold Ca2+- and Mg2+-free gHBSS. Flow cytometric analysis was performed with an EPICS XL flow cytometer with an argon ion laser (Coulter, Hialeah, FL) with an excitation wavelength of 488 nm. Emitted light was collected through 550-nm and 600-nm dichromic and 525-nm and 575-nm bandpass filters, respectively. The fluorescence of 5000 cells was measured by means of logarithmic amplification. A relative measure of receptor expression was obtained by determining the mean log fluorescence intensity after two-colour spectral compensation network adjustment.

Statistical analysis

Because of the skewed distribution of the CL data, the peak value and the peak times of CL were transformed to their natural logarithmics. The CL data are reported as geometric means with 95% confidence intervals (95% CI). Other parameters are reported as means with s.d. or range. When the effect of mode of delivery and the effect of feeding practices on CL activities were studied separately, analysis of variance ( anova) for repeated measurements was used. When anova indicated a significant difference, pair-wise comparisons with Fisher's LSD test were carried out to identify which group differences accounted for the overall P value. When the effect of mode of delivery and the effect of type of feeding on the CL activities were assessed at the same time, two-way anova was used. The clinical characteristics between the groups were compared with the two-sample t-test and Fisher's exact test. P   <   0.05 was considered statistically significant.


Clinical characteristics

The mean duration of pregnancy was 39 weeks (range 37–42 weeks). The mean (range) birth weight was 3575 g (2640–4480 g) and the mean (range) Apgar scores at 1, 5 and 15 min: 9 (3–10), 9 (7–10) and 9 (8–10), respectively. Infants born by vaginal delivery had a slightly higher gestational age than those born by Caesarean delivery (40 and 39 weeks, respectively; P = 0.04), but the mean birth weight and the mean Apgar scores did not differ between the delivery groups.

At 2 months of age 39 infants were exclusively breast-fed, 14 partially breast-fed and six exclusively formula-fed. At 6 months of age 26 infants were exclusively breast-fed, three partially breast-fed and 30 exclusively formula-fed. In addition, all infants received solid foods at 6 months of age. In five cases the information about breast feeding was missing. To compare the phagocyte functions in infants with different feeding practices the infants were divided into two groups at 2 and 6 months of age: 1, exclusively breast-fed, and 2, partly/totally formula-fed. There were no statistical differences in the mean gestational age, the mean birth weight or the mean Apgar scores between the feeding groups at 2 and 6 months of age.

Total and differential counts of leucocytes

As shown in Table 2, the total leucocyte and neutrophil counts as well as the differential count of neutrophils were higher and the differential count of lymphocytes lower in cord blood after vaginal delivery compared with Caesarean section. These differences were transient and had levelled off by 2 months of age. There were no statistically significant differences between the delivery groups in the differential counts of monocytes, basophils and eosinophils. Type of feeding did not affect the total or differential counts of leucocytes at 2 and 6 months of age (data not shown).

Table 2.  . Total and differential count of leucocytes at 0, 2, and 6 months of age in infants delivered by Caesarean section or by vaginal delivery Values are given as mean (s.d.). * P   <   0.0001 between the delivery groups. † Large unstained cells. Thumbnail image of

Zymosan-induced CL activities

The WBCL activity was significantly higher in the infants born by Caesarean section than in those born by vaginal delivery; this difference persisting during the entire 6-month follow-up period. This was the case in both opsonized and unopsonized zymosan-induced WBCL ( Fig. 1a, b). There were no statistical differences between the delivery groups in the peak times of WBCL (data not shown). Likewise, there were no differences in peak values or in peak times (data not shown) of the ILCL measurements between the delivery groups for either opsonized or unopsonized zymosan as stimulant.

Figure 1.

. Chemiluminescence activities of whole blood (WBCL) with opsonized (a) and unopsonized (b) zymosan as stimulant in infants delivered by vaginal delivery (—–) or by Caesarean section (- - - - -) at 0, 2 and 6 months of age. The values are given with geometric means with 95% CI. Statistical differences between the delivery groups at different time points: (a) * P = 0.0007; **P = 0.07; ***P = 0.03; (b) *P = 0.003; ** P = 0.02; *** P = 0.002.

When both the effect of mode of delivery and the effect of type of feeding on CL activity were analysed, both factors were seen to exert an independent effect on the CL activity at 6 months of age. Infants who were born by Caesarean delivery had higher WBCL activities compared with those born by vaginal delivery ( Table 3). Further, infants who received formula feeds had higher opsonized zymosan-stimulated WBCL activity at 6 months of age than breast-fed infants ( Table 3). No differences were detected in ILCL activities (data not shown).

Table 3.  . Effect of mode of delivery and type of feeding on the activity of whole blood chemiluminescence (mV/1000 neutrophils) at 6 months of age * Geometric mean. † Two-way anova: P = 0.02 for mode of delivery and P = 0.02 for type of feeding. ‡ Two-way anova: P = 0.05 for mode of delivery and P = 0.34 for type of feeding. Thumbnail image of

Milk- and gliadin-induced CL activities

Breast-fed infants had lower milk-induced WBCL activities both at 2 and 6 months of age (P = 0.07 and P = 0.002, respectively) than formula-fed infants, while mode of delivery had no further effect (P = 0.97 and P = 0.71, respectively). The geometric mean (95% CI) of milk-induced WBCL activities (mV/500 000 neutrophils) in the different feeding groups were: 2.8 (1.0–7.8) in breast-fed (n = 27) and 14.8 (2.9–74.8) in formula-fed infants (n = 16) at 2 months of age and 20.9 (8.3–52.3) (n = 20) and 92.5 (72.7–117.5) (n = 21) at 6 months of age, respectively.

Mode of delivery and type of feeding had no influence on gliadin-induced WBCL activities or on milk- or gliadin-induced ILCL activities (data not shown).

Phagocyte receptors

Since binding and phagocytosis of particles coated with immune complexes is mediated by phagocyte receptors, we further investigated if mode of delivery or feeding practices influenced the expression of these receptors on neutrophils, monocytes and eosinophils.

There were no statistical differences in the expression of CR1, CR3, FcγRI, FcγRII, FcγRIII or Fcα receptors on neutrophils, monocytes or eosinophils between the delivery groups (data not shown). In contrast, formula feeding increased the expression of phagocyte receptors on neutrophils. Formula-fed infants had higher cell surface expression of FcγRI and Fcα on neutrophils at 2 months of age and CR3 receptors both at 2 and 6 months of age ( Fig. 2a, b, c).

Figure 2.



Our results show that the zymosan-induced CL response of cord blood leucocytes is significantly higher in infants born by Caesarean section than in infants born vaginally. Physiological changes associated with stress during labour could explain these differences observed in phagocyte functions. Overall, stress related to different neonatal disorders (i.e. sepsis, respiratory distress syndrome, pneumonia, meconium aspiration) has been shown to depress neutrophil functions and CL activity [ 16[17]–18]. Likewise, the stress caused by labour has been held to explain the lower lymphocyte and the higher neutrophil count in the cord blood [ 7, 9], which we also detected.

Both the mode of delivery and the prophylactic antibiotic administered to the mother before Caesarean delivery may have influenced the phagocyte activity of the neonates born by Caesarean section. Since ampicillin has been shown to inhibit the phagocytic activity of leucocytes in vitro and in vivo [ 19, 20], the prophylactic ampicillin used here seems an unlikely explanation for the higher CL activity observed in the cord blood of these infants. In contrast, without ampicillin the CL activity in the cord blood could have been even higher in infants born by Caesarean section. In addition, it is unlikely that the effect of the ampicillin therapy was apparent any more at 2 and 6 months of age, when significant differences were still seen between the delivery and feeding groups.

Some previous papers have reported that leucocytes from vaginally delivered infants are functionally more active, while others have found them less active than those from neonates delivered by elective Caesarean section. Closer consideration, however, reveals that the phagocyte functions observed to be more active in vaginally delivered babies are involved in the secretory functions of phagocytes: increased respiratory burst upon activation with a chemotactic peptide n-formyl methionyl leucyl phenylalanine [ 12, 21], higher extracellular superoxide anion production [ 13] and higher leucocytic pyrogen (obviously IL-1 in modern terms) secretion [ 14]. On the other hand, functions found to be less active in vaginally delivered infants seem to be involved in phagocytosis. Frazier et al. [ 11] measured phagocytosis-related phenomena (zymosan-stimulated oxygen consumption, hexose monophosphate shunt activity and quantitative nitroblue tetrazolium dye reduction) and found them to be significantly lower in the cord blood polymorphonuclear leucocytes of vaginally delivered infants when compared with infants delivered by elective Caesarean section.

Usmani et al. [ 15] found no difference between infants delivered vaginally and by elective Caesarean section in the non-opsonized latex-stimulated CL of isolated polymorphonuclear leucocytes. They found, however, a negative correlation between the peak CL value and the duration of labour in infants delivered by Caesarean section after labour. Because the measurements were made at room temperature and the peak times of the CL were short, this work obviously did not measure phagocytosis at all, rather adherence-related events.

Previous findings about the effect of method of delivery are, thus, in accordance with our results; zymosan-induced, luminol-amplified WBCL was significantly lower in the cord blood of vaginally delivered infants. This particle-induced luminol-amplified CL measures mainly intracellular events which depend on the release of myeloperoxidase from primary granules into phagosomes [ 22].

The impact of the mode of delivery was of much longer duration than previously thought, lasting for the entire 6-month follow-up period. After the abatement of the stress reaction caused by labour in vaginally delivered infants, what could have been the reason for the elevated activity of phagocytes in Caesarean section babies lasting for at least 6 months? During the first months of life an intense process of bacterial colonization takes place in the intestine [ 23[24]–25]. Caesarean section infants have delayed intestinal colonization with anaerobic bacteria, especially with Bacteroides sp., lactobacilli and bifidobacteria [ 23, 26[27][28]–29]. Bacterial colonization is shown to promote the intestine's barrier function, in both the immunological and the non-immunological sphere [ 30, 31]. Delayed intestinal colonization might result in increased uptake of antigens, from intestinal microflora and diet, across the intestinal wall in infants born by Caesarean section with ensuing priming of the neutrophils.

In support of the hypothesis of neutrophil priming by intestinal antigens was the recorded higher zymosan-induced CL activity in infants who had received cow milk-based formula feeds compared with exclusively breast-fed infants. The observed higher zymosan-induced WBCL implies that differences at the cellular level might be involved between the two feeding groups. This could be related to the fact that formula-fed babies had higher cell surface expression of FcγRI, Fcα and CR3 receptors than breast-fed babies, while there were no differences in receptor expressions between the delivery groups. Thus, the factors behind the priming effect of neutrophils in the feeding and delivery groups of infants might act via different mechanisms.

Since we used a low dilution of blood in the milk- and gliadin-induced WBCL (i.e. 1:10; serum factors included) compared with zymosan-induced WBCL (i.e. 1:5000; serum factors excluded), the milk-induced WBCL represented probably the activation of phagocytes by immune complexes consisting of milk protein and specific antibodies to milk protein, and/or complement factors [ 32]. The higher milk-induced WBCL results in formula-fed infants as opposed to breast-fed ones suggest that the cow's milk constituents of formula feeds caused secretion of milk protein-specific antibodies in plasma, which in turn, together with antigens, may have primed the phagocytes via Fc-receptors as described by Hallett & Lloyds [ 33].

We observed the primed activity of phagocytes only in the whole blood measurements but not in the measurements with isolated leucocytes. This observation is in harmony with recent findings by Brown et al. [ 34], who demonstrated that the CL activity of primed, isolated neutrophils was very unstable compared with that of primed neutrophils in the whole blood. Blood-borne components were suspected to stabilize or modify neutrophil priming.

We conclude that at birth the phagocytic activity in vaginally delivered infants is depressed by stress related to labour, and later in life, during the development of the adaptive immune system, phagocytes in infants delivered by Caesarean section and in those formula-fed may be primed by intraluminal antigens (i.e. gut microflora and diet).

At present we do not know whether the higher phagocyte activity observed in infants born by Caesarean delivery and in those formula-fed is beneficial for the immune protection of the neonates, or whether it includes a danger of tissue injury due to over-zealous production of possibly cytotoxic reactive oxygen intermediates [ 35[36]–37]. Early infancy is a critical period of life when a great variety of new antigens is encountered via mucosal surfaces and the immunological memory directed to these antigens is consolidated [ 38]. Over-zealous production of reactive oxygen intermediates by primed neutrophils could result in significant tissue injury at mucosal surfaces and consequently in a more permeable gut mucosal barrier. Enhanced presentation of dietary antigens to the immature immune system entails a risk of sensitization. Since Caesarean section rates have shown a marked tendency to rise during the past few decades [ 3, 39], the immunomodulative role of the mode of delivery deserves to be further investigated.


We thank Tuija Poussa, MSc, for able statistical consultations. This study was supported by The South-West Finnish Fund of Neonatal Research, The Technology Development Centre of Finland and The Finnish Cultural Foundation.