Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
JAMA. Author manuscript; available in PMC 2014 March 13.
Published in final edited form as:
PMCID: PMC3752893

Effects of promoting increased duration and exclusivity of breastfeeding on adiposity and insulin-like growth factor-I at age 11.5 years: a randomized trial



Evidence that increased duration and exclusivity of breastfeeding reduces child obesity risk is based on observational studies that are prone to confounding.


To investigate effects of an intervention to promote increased duration and exclusivity of breastfeeding on child adiposity and circulating insulin-like growth factor (IGF)-I (which regulates growth).


Cluster-randomized controlled trial.


31 Belarusian maternity hospitals and their affiliated polyclinics, randomized to usual practices (n=15) or a breastfeeding promotion intervention (n=16).


17,046 breastfeeding mother-infant pairs enrolled in 1996/7, of whom 13,879 (81.4%) were followed-up between January 2008 and December 2010 at a median age of 11.5 years.


Breastfeeding promotion intervention modeled on the WHO/UNICEF Baby Friendly Hospital Initiative.

Main outcome measures

Body mass index (BMI), fat and fat-free mass indices (FMI and FFMI), percent body fat, waist circumference, triceps and subscapular skinfold thicknesses, overweight and obesity, and whole-blood IGF-I. Primary analysis was based on modified intention-to-treat (without imputation), accounting for clustering within hospitals/clinics.


The experimental intervention substantially increased breastfeeding duration and exclusivity (43% vs. 6% and 7.9% vs. 0.6% exclusively breastfed at 3 and 6 months, respectively) versus the control intervention. Cluster-adjusted mean differences in outcomes at 11.5 years between experimental vs. control groups were: 0.19 kg/m2 (95% 4 CI: −0.09, 0.46) for BMI; 0.12 kg/m2 (−0.03, 0.28) for FMI; 0.04 kg/m2 (−0.11, 0.18) for FFMI; 0.47% (−0.11, 1.05) for % body fat; 0.30 cm (−1.41, 2.01) for waist circumference; −0.07 mm (−1.71, 1.57) for triceps and −0.02 mm (−0.79, 0.75) for subscapular skinfold thicknesses; and −0.02 standard deviations (−0.12, 0.08) for IGF-I. The cluster-adjusted odds ratio for overweight / obesity (BMI ≥85th percentile vs <85th percentile) was 1.18 (1.01, 1.39) and for obesity (BMI ≥95th vs <85th percentile) was 1.17 (0.97, 1.41).

Conclusions and relevance

Among healthy term infants in Belarus, an intervention that succeeded in improving the duration and exclusivity of breastfeeding did not prevent overweight or obesity, nor did it affect IGF-I levels, at age 11.5 years. Breastfeeding has many advantages, but population strategies to increase the duration and exclusivity of breastfeeding are unlikely to curb the obesity epidemic.

Keywords: Breast feeding, lactation, adiposity, body mass index, randomized controlled trial, insulin-like growth factor-1, childhood


Observational studies suggest that greater duration and exclusivity of having been breastfed is inversely associated with adiposity,1;2 but positively associated with stature3;4 and later life serum insulin-like growth factor (IGF)-I,5;6 a regulator of childhood statural growth and body composition.7;8 Breastfeeding and growth are socially patterned in many settings,9 however, and observed associations between these variables are at least partly explained by confounding structures.1;2;10;11 Also, the relationship between breastfeeding and growth is dynamic and bi-directional, because infant behaviour and weight gain influence maternal feeding decisions, making reverse causation an alternative explanation for observed associations of breastfeeding duration and exclusivity with growth.12;13

To overcome limitations inherent in observational studies of breastfeeding’s long-term effects, we designed a follow-up of 17,046 children participating in the Promotion of Breastfeeding Intervention Trial (PROBIT).14 The intervention resulted in 2 groups with substantially different durations and exclusivity of breastfeeding, providing a unique opportunity to test, in an intention-to-treat (ITT) analysis, the extent to which breastfeeding causally influences growth and its regulation.

The breastfeeding promotion intervention had no measurable effect at age 6.5 years (PROBIT II) on stature or physical measures of adiposity,15 but adiposity was not directly measured and the findings could have been distorted by variations in timing of the adiposity rebound by duration of breastfeeding.16 The current study (PROBIT III) 6 provides experimental evidence on whether beneficial effects of increased duration and exclusivity of breastfeeding on growth develop later in childhood, based on direct measurements by bioimpedence of body fat and lean mass and on circulating IGF-I.


A detailed description of the cluster-based randomisation, experimental intervention and participant eligibility in PROBIT has been published.14 Briefly, the units of randomization (clusters) were maternity hospitals and their associated polyclinics (outpatient health clinics following up both well and ill children). These units were randomized to a control intervention (continuation of the breastfeeding practices and policies in effect at the time of randomisation) or an experimental intervention, based on the Baby Friendly Hospital Initiative developed by the World Health Organization (WHO) and United Nations Children’s Fund (UNICEF) to promote and support breastfeeding, particularly among mothers who have chosen to initiate breastfeeding.17 The trial results are based on a total of 17,046 healthy breastfed infants from 31 maternity hospitals/polyclinics, born at term (≥ 37 weeks gestation) in 1996-7 and enrolled during their postpartum stay. Trial inclusion criteria required the infants to be healthy, singleton, with birth weight ≥2500g and Apgar score ≥5 at 5 minutes, and their mothers to have initiated breastfeeding and to have no condition that would interfere with breastfeeding.14

Between January 2008 and December 2010 the children were followed-up at dedicated research visits by 39 specially-trained pediatricians, 1 in each of 23 polyclinics and 2 sharing visits at each of the remaining 8 high-volume clinics. Training and quality assurance procedures have been described in detail previously.18 We asked children to fast for at least eight hours prior to the visit, which included the following measurements in duplicate: standing and sitting height, measured with a wall-mounted stadiometer (Medtechnika, Pinsk, Belarus); triceps and subscapular skinfold thicknesses, measured using Lange spring-loaded skinfold calipers (Beta Technology, Santa Cruz, CA); head, mid-upper arm, waist and hip circumferences and upper arm length measured using non-stretchable measuring tapes; weight, percent body fat, fat mass and fat-free (lean) mass, measured by foot-to-foot bioelectrical impedance (Tanita TBF 300GS body-fat analyser, Tanita Corporation , Inc., Illinois, USA).

Body mass, fat mass and fat-free (lean) mass indices (BMI, FMI and FFMI) were defined as weight (kg)/height (m2), fat mass (kg)/height (m2) and lean mass (kg)/height (m2), respectively. We defined overweight as being between the 85th to 94th percentiles and obesity as ≥ 95th percentile of BMI, based on the Centers for Disease Control and Prevention (CDC) 2000 reference data.19 In our analysis we compared ≥ 95th percentile (obese) versus < 85th percentile and ≥ 85th percentile (overweight/obese) versus < 85th percentile. We computed leg length as standing height minus trunk length (sitting height minus stool height).

At the visit, finger-prick whole-blood spot samples were collected by the 39 pediatricians, who had received special training,2022 onto Whatman 903 filter paper cards21 as described previously.23 The dried blood spot cards were stored in a –20°C freezer at each of the 31 polyclinic sites until transport to the laboratory at the National Mother and Child Centre in Minsk, where they were stored at −80°C. The samples for IGF-I were stored at −20 °C for a median of 1.7 months (interquartile range, IQR 1.0–5.1) and at −80 °C for a median of 18.4 months (IQR 13.3–21.6).

We quantified circulating IGF-I from a single 3-mm diameter disc (≈ 3 µL of blood) per child, after a single thaw, using the validated method of Diamandi et al.24 Mean intra-assay coefficients of variation were 6%, 7% and 9% for ‘low’, ‘medium’ and 'high’ IGF values, respectively; the respective inter-assay values were 8%, 12% and 16%. The Spearman correlation coefficient between 50 paired whole-blood spot versus serum IGF-I samples, collected simultaneously was 0.93 (95% confidence interval: 0.87 to 0.96). Between 98–105% of known quantities (250, 300, 350, 450 and 550 ng/ml) of IGF-I prepared from recombinant human IGF-I was recovered from the dried blood spots. We also demonstrated that blood spot IGF-I was stable for at least 24 months at −80 °C (data on request). IGF-I was assayed from 2 lots of reagents between January 2010 and November 2011 and, as other authors have noted, assay kits of different lot number have been observed to cause some variation in measured IGF-I.25 To remove the potential effect of between-lot or between-run variation, we standardised values of IGF within each assay run (n=43) by computing z-scores ([IGF value – mean for each run]/ SD of the mean).

Audit visits were conducted to assess inter-observer reproducibility of the outcome data, an important step given that blinding of pediatricians to the experimental vs. control randomized group assignment was not feasible. For each of the 39 pediatricians, 1–5 children were randomly selected to return for re-measurement of all variables, for a total of 143 audited children (108 with baseline and repeat IGF-I values). So that all children seen in follow-up were eligible for selection, the repeated measurements were carried out after completion of primary data collection, an average of 1.3 years (range 0.2 to 2.4) after the initial clinic visit. The audit was carried out by 1 of 5 Minsk-based pediatricians not involved in primary data collection and blinded to the measures obtained at the initial visit but not to experimental or control status. Because of the time elapsed between the audit and initial visits, results were compared using Pearson correlation coefficients.

PROBIT III follow-up was approved by the Belarussian Ministry of Health and received ethical approval from the McGill University Health Centre Research Ethics Board; the Human Subjects Committees at Harvard Pilgrim Health Care; and the Avon Longitudinal Study of Parents and Children (ALSPAC) Law and Ethics Committee. A parent or legal guardian provided written informed consent in Russian at enrollment and at the follow-up visits, and all children provided written assent at the 11.5 year visit.

Statistical analysis

We conducted all analysis using SAS version 9.3 (SAS Institute, Cary, NC), unless otherwise stated. Our main outcomes were measures of adiposity (BMI, FMI, FFMI, percent body fat, waist circumference, triceps and subscapular skinfold thicknesses, overweight and obesity as defined above) and IGF-I. We also explored the effect of the intervention on other anthropometric measurements: standing height, leg length, hip circumference, waist:hip ratio, head circumference and mid upper-arm circumference. Comparisons between the experimental and control groups were based on a modified intention-to-treat analysis without imputation for missing outcome data (i.e. based on the 13,879 children with observed outcomes).

We accounted for possible non-independence of measurements within individual hospital/polyclinic sites (clustering) using random effects models, which permit inference at the level of the individual child rather than at the level of the cluster (maternity hospital and polyclinic). The MIXED procedure was used for continuous outcomes (to estimate mean differences and 95% CIs) and the GLIMMIX procedure for binary outcomes (to estimate odds ratios and 95% CIs) in SAS. The results are presented for the simple cluster-adjusted model, as well as after additional adjustment for stratum-level variables (urban versus rural and East versus West Belarus residence), and for child age at follow-up, sex, birth weight, and maternal and paternal education. Controlling for maternal and paternal height (for standing height, leg-length and IGF-I) and for measured maternal and mother-reported paternal BMI (for adiposity measures) made little material difference to the effect-estimates and so are not presented in the main results table. The results for IGF-I were additionally controlled for the time between blood sampling to assay, as this has previously been reported to influence levels25 (although we found that IGF-I from dried blood spots was stable to freezing). To determine whether results differed in boys versus girls, we also analyzed mixed models that included terms for the sex of each child and a multiplicative sex*trial arm interaction term.

In a sensitivity analysis, we investigated whether loss to follow-up influenced the results by undertaking multiple imputation to generate plausible values of missing 11.5 year outcomes and thereby including all 17,046 randomized participants in the intention-to-treat analysis.26;27 We used SAS imputations (Proc MI) to impute 5 values for each missing observation and combined multivariable modeling estimates using Proc MI ANALYZE in SAS.

The intention-to-treat analysis may underestimate the effect of the true exposure of interest (breastfeeding exclusivity and duration), owing to overlap in breastfeeding between the randomized groups (many intervention mothers did not exclusively breastfeed for 3 or 6 months, whereas some control mothers did). In a secondary analysis, we applied instrumental variable methods to estimate unbiased associations of the difference in breastfeeding exclusivity and duration achieved between the 2 randomized groups on our outcomes.28 In this approach, we used randomization status as the ‘instrument’ that is independent of any confounders of the exposure-outcome relationship, and is related to the outcomes only via the exposure (breastfeeding duration and exclusivity). We performed instrumental variable estimation of continuous outcomes using the generalized 2-stage least squares estimator implemented in the xtivreg command in Stata version 12.1 (Stata Corp., College Station, TX), while accounting for clustering by study site. We performed instrumental variable estimation of binary outcomes using a random effects version of the ratio estimator of the causal odds ratio.29

To assess whether we could reproduce the inverse associations of increased duration and exclusivity of breastfeeding reported in previous observational studies, we conducted observational analyses (i.e., disregarding randomization status), in which we estimated the effects of the duration of any breastfeeding and of the duration of exclusive breastfeeding on the same outcomes, also accounting for clustering and the same baseline characteristics as in the expanded mixed models described above, using multiple linear regression for continuous outcomes and multiple logistic regression for the binary outcomes. Duration of breastfeeding was classified as < 3 months (reference), ≥ 3 to < 6 months and ≥ 6 months.

Power calculations

A priori we calculated detectable differences in outcomes based on following up 14,000 children in 31 clusters, an intention-to-treat analysis and the design effect based on a realistic value (0.01) of the intra-cluster correlation coefficient (ICC).30 The mean detectable differences (5% significance, 80% power) in intention-to-treat analyses were 0.24 kg/m2 for BMI (assuming a plausible effect of longer versus shorter duration of exclusive breastfeeding of 0.6 kg/m2) and 4.25 ng/ml for IGF-I (assuming an effect of longer duration of exclusive breastfeeding of 10.63 ng/ml).


As previously reported,14 the randomization produced 2 groups with similar distributions of baseline sociodemographic and potential confounding factors. The intervention substantially increased breastfeeding duration and exclusivity (based on WHO definitions17) versus the control arm: e.g., at 3 months, intervention infants were 7 times more likely to be exclusively (43.3% vs. 6.4%) and twice as likely to be predominantly (51.9 vs. 28.3%) breastfed, and were breastfed to any degree at higher rates throughout infancy, although at 6 months both exclusive (7.9% vs. 0.6%) and predominant breastfeeding (10.6% vs. 1.6) were low.14 Comparing the experimental vs. the control group, 72.7% vs. 60.0%, respectively, were still breastfeeding to any degree at 3 months, 49.8% vs. 36.1%, respectively, were still breastfeeding at 6 months, and 19.7% vs. 11.4%, respectively, were still breastfeeding at 12 months.

A total of 13,879 children were examined at a median age of 11.5 years (standard deviation, interquartile range: 0.50, 11.3–11.8 years), representing 81.4% of the 17,046 originally randomized (Figure 1). Of the 3,167 children randomized but not followed up at 11.5 years, 97 had died since randomisation, 2,645 were lost to follow-up, and 425 were unable or unwilling to come for their visit (Figure 1). Follow-up rates were similar overall in the experimental (83.5%) and control (79.1%) polyclinics, although they varied from 48 to 98%. The children followed up at 11.5 years in the experimental and control groups were similar in baseline characteristics, with small differences paralleling those seen (and previously reported14) at randomisation (Table 1). The groups were also virtually identical in mean parental height and BMI (measured for mothers at 11.5 years follow-up and reported by the mothers for the fathers at the 6.5 year follow up).

Figure 1
Flow diagram of progress of clusters and individuals through PROBIT recruitment and follow-up phases I, II and III
Table 1
Baseline and follow-up characteristics

eTable 1 summarizes the audit results, showing high correlations (Pearson r >0.80) between initial clinic results and blinded repeat (audit) measures of weight, fat mass, fat free mass, percent fat, subscapular skinfold thickness, hip circumference, standing height, 14 and mid-upper arm circumference, and substantial correlations (r = 0.73–<0.80) for waist circumference, triceps skinfold thickness, leg length and upper arm length, but only modest correlations for head circumference (r = 0.50) and IGF-I (r = 0.37).

The main results are shown in Table 2. There was a moderate degree of within-polyclinic clustering (the tendency for measurements on children attending the same polyclinic to be more similar to each other than to children attending other polyclinics18) for triceps skinfold thickness and head circumference (intraclass correlation coefficients, ICCs ≥0.10), but a low degree of clustering for the other measures. Mean BMI, FMI, percent body fat, waist circumference, and the prevalence of overweight and obesity, were slightly higher in the experimental versus control groups, but the cluster adjusted confidence intervals were consistent with chance and rule out any important protective effect (lower values) on adiposity. The cluster-adjusted mean differences in our main 11.5 year outcomes between experimental versus control groups were: 0.19 kg/m2 (−0.09, 0.46) for BMI; 0.12 kg/m2 (−0.03, 0.28) for FMI; 0.04 kg/m2 (−0.11, 0.18) for FFMI; 0.47% (−0.11, 1.05) for body fat; 0.30 cm (−1.41, 2.01) for waist circumference; −0.07 mm (−1.71, 1.57) for triceps and −0.02 mm (−0.79, 0.75) for subscapular skinfold thicknesses; and −0.02 standard deviations (−0.12, 0.08) for IGF-I. The cluster-adjusted odds ratio for overweight / obesity (BMI ≥85th percentile vs <85th percentile) was 1.18 (1.01, 1.39) and for obesity (BMI ≥95th vs <85th percentile) was 1.17 (0.97, 1.41). In exploratory analyses, there was little evidence that the intervention affected standing height, leg length, waist:hip ratio, head or mid upper-arm circumference. A weak positive effect on hip circumference was observed in the fully-adjusted (mean difference: 0.81 cm; 95% CI: 0.09, 1.53), but not cluster-adjusted, model.

Table 2
Modified intention-to-treat analysis (without imputation) showing differences in adiposity measures, IGF-I, circumferences and height comparing intervention vs. control groups (n=13,879)

These conclusions were unaltered after further adjusting for baseline potential confounders (Table 2) or using multiply imputed outcomes (eTable 2). There was little evidence of interaction by sex (all interaction p-values > 0.10, except for subscapular (p = 0.03) and triceps (p = 0.05) skinfold thickness).

In observational analyses (Table 3 and eTable 3), increased duration of exclusive breastfeeding was positively associated with body mass, fat mass and fat-free mass indices, hip circumference, head and mid upper-arm circumference, and overweight/obesity. Results were similar for duration of any breastfeeding (eTable 4). The results of the instrumental variable analyses (Table 4), which provide estimates of the unbiased associations of exclusive breastfeeding for ≥ 3 to < 6 months and ≥ 6 months versus < 3 months (and therefore directly comparable to estimates from observational studies), are in line with the inference that increased duration and exclusivity of breastfeeding provides no important beneficial effects on the study outcomes.

Table 3
Observational associations of duration of exclusive breastfeeding with adiposity measures, IGF-I, circumferences and height at 11.5 years (n = 13,879)
Table 4
Instrumental variable estimates of the unbiased associations of duration of exclusive breastfeeding on BMI (continuous), waist circumference, overweight and obesity (binary) and standing height at 11.5 years using randomized treatment as the instrumental ...


The results from this large cluster-randomized trial indicate that the experimental intervention to promote increased duration and exclusivity of breastfeeding did not reduce continuous measures of adiposity, nor reduce the prevalence of overweight or obesity, at age 11.5 years despite causing large increases in the duration and exclusivity of breastfeeding. The results are similar to those we obtained at age 6.5 years.15 The new data thus extend our observations to older children and include more direct measures of fat and lean mass.

Our findings also concur with the evidence provided by other authors attempting to systematically assess unbiased and unconfounded associations of breastfeeding on adiposity. An individual-participant meta-analysis provided empirical evidence that previously reported associations of having been breastfed with both continuous measures of adiposity and overweight/obesity may have arisen as a result of residual confounding, selective reporting and/or publication bias.1;2 In low- and middle-income countries9;11;31 or older ‘Western’ cohorts,3;4;32;33 with confounding structures that are neutral or opposite to those currently seen in high-income countries, inverse associations of breastfeeding with adiposity are not consistently observed. An inverse association of breastfeeding with obesity was not seen in a matched sibling-pair analysis that was restricted to comparisons of infants from the same families (and hence socio-economic background),34 while a larger study reported effect-estimates consistent with the null from both the sibling-pair and full cohort analysis.35

We emphasize that the breastfeeding promotion intervention was designed to increase the duration and exclusivity of breastfeeding, not its initiation (initial breastfeeding was an inclusion criterion). Our findings may not, therefore, apply to comparisons of initiating 17 breast- versus formula feeding, the comparisons most frequently described in the literature. Nonetheless, many previous studies have reported dose-response inverse associations of the duration and/or exclusivity of breastfeeding with adiposity.36 Our findings do not support those associations.

We observed a positive association of the intervention with overweight/obesity, although the magnitude was small. As previously reported, the PROBIT intervention was associated with faster weight gain during the first 3 months of infancy, although these differences disappeared by 12 months of age.13 We also reported a non-significant positive association with overweight/obesity at age 6.5 years.15 It is possible that PROBIT mothers randomized to receive the breastfeeding promotion intervention, knowing their infants were entirely dependent nutritionally on their breast milk, deliberately increased the frequency and duration of feeds, leading to the faster weight and length gains we observed in the first 3 months of life. Explaining the disappearance of these differences by 12 months and their reappearance (for weight) in later childhood, however, would require a combination of metabolic “programming” in infancy, followed by a prolonged latent period. In our view, a more likely explanation is that the positive association observed in later childhood arose by chance.

Our 2 trial cohorts were created by randomization at the time of birth (not by the mother’s choice), which resulted in substantial differences between those cohorts in the duration and degree of breastfeeding. Coupled with our high rates of follow-up over 11.5 years, the intention-to-treat analysis minimizes the confounding and reverse causality biases that plague observational studies.37 We also minimized measurement bias: by assessing infant feeding contemporaneously, strictly adhering to WHO definitions17 of breastfeeding duration and exclusivity, and utilizing specific measures of body fat distribution (waist and hip circumferences) and body composition (skinfold thicknesses, fat mass and fat free mass measured by bioimpedance).38;39

To estimate the unbiased effects of the experimental breastfeeding promotion intervention, we used an intention-to-treat rather than a ‘per-protocol’ analysis, as it is well known that ‘per-protocol’ analyses of randomized trials, in which participants are grouped according to the intervention they received, rather than that to which they were randomized, may be seriously biased. The estimates provided by the intention-to-treat analysis are the most robustly estimated expected average effects on the anthropometric and adiposity outcomes of the experimental breastfeeding promotion intervention. However, because of the substantial overlap in breastfeeding duration and exclusivity in the 2 randomized groups, these average effects may considerably underestimate the differences in outcome caused by increased duration and exclusivity of breastfeeding. We therefore estimated the magnitude of the unbiased and unconfounded associations with anthropometry and adiposity using instrumental variables analysis, which supported our inference that increased duration and exclusivity of breastfeeding was not associated with the outcomes of interest. The confidence intervals of the instrumental variables analyses were wide, however, so cannot exclude the estimates reported in observational studies.1;2;40

Despite our large sample size, the precision of the observed differences (the width of the confidence intervals) was only modest, because of clustering of both the intervention (children were clustered within polyclinic, and so would tend to have more similar outcomes) and measurement (a tendency for measurements made within a clinic to be more similar than measurements between clinics).41 Nevertheless, for most measurements, we could rule out differences as small as 0.15 to 0.20 standard deviations. For triceps skinfold thickness and head circumference, with ICCs ≥0.1 (indicating an important degree of within-polyclinic clustering18), we could not rule out differences as large as 0.35 standard deviations. The point estimates for all adiposity measures were in the opposite direction to those hypothesised, however, and for overweight and obesity, the lower limit of the confidence interval excludes an important protective effect of breastfeeding.

The trial was carried out in Belarus, rather than North America or Western Europe, because at the time of randomization, maternity hospital practices in Belarus and other former Soviet republics were similar to those in North America and Western Europe 30–40 years ago and thus provided a greater potential contrast between intervention and control study sites. However, although different in many socioeconomic, cultural and economic respects from North America and Western Europe, Belarus is a relatively developed country, with strict hygienic standards, high immunization rates, low incidence of infection, low rates of infant and child mortality, similar types of formula feeds and accessible health care services. The prevalence of childhood obesity in the U.S. (and some other Western countries) is much higher (>15% compared to approximately 5% in Belarus based on the CDC 2000 reference data42). It is possible that our results may not generalize to settings with a much higher prevalence of obesity than Belarus, although such a lack of generalizability implies an interaction between method of infant feeding and some (unknown) environmental factor in more obesogenic environments.

IGF-I was measured from dried blood spots but, in addition to the validation data presented in this article, circulating IGF-I has previously been reported to be stable when dried on filter paper (for 40 days at room temperature24;43 and 5 months at −20°C44) and validly and reliably measured from dried blood spots using either a commercially available radioimmunoassay44 or ELISA kits from Diagnostic Systems Laboratories,24;43 including the kit used in this study.24


Among healthy term infants in Belarus, an intervention to improve the duration and exclusivity of infant breastfeeding did not prevent overweight or obesity, nor did it affect IGF-I levels, among these children when they were aged 11.5 years. Nevertheless, breastfeeding has many health advantages for the offspring, including beneficial effects demonstrated by our PROBIT trial on gastrointestinal infections and atopic eczema in infancy14 and improved cognitive development at age 6.5 years.45 Although breastfeeding is unlikely to stem the current obesity epidemic, its other advantages are amply sufficient to justify continued public health efforts to promote, protect, and support it.


We are grateful to the cohort members, their parents and the study pediatricians and auditors who participated so willingly in the study. We acknowledge Jennifer Thompson MPP, Harvard Medical School and Harvard Pilgrim Health Care Institute, for assistance with data analysis (uncompensated).

RMM had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. RMM, MSK, GDS and MG developed the hypotheses and secured funding; RP, NB, NS and LG ran the fieldwork under the direction of RMM, MSK, EO and KV. RP, NB and LG were responsible for the database and data cleaning. NG and YF developed the methods for the dried blood spot assays for IGF-I and NG directed the laboratory analyses. RP, TP and SLR undertook statistical analyses. RMM wrote the first draft of the paper. All authors critically commented on and approved the final submitted version of the paper.

EO has given an invited talk for Nestle Nutrition Institute on secular trends in birth weight. RMM gave an invited talk for the Nestle Nutrition Institute in 2010 on the role of the insulin-like growth factor system in growth and chronic disease risk.

Funding: Supported by: European Union, Early Nutrition Programming Long-term Efficacy and Safety Trials grant no. FOOD-DT-2005-007036; Canadian Institutes of Health Research (MOP - 53155); and the USA National Institutes of Health (R01 HD050758). Dr. Oken was supported by US National Institute of Child Health and Development (K24 HD069408). Professors Martin and Davey Smith, and Dr Palmer work within the CAITE centre, which is supported by the MRC (G0600705) and the University of Bristol. The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.


1. Owen CG, Martin RM, Whincup PH, Davey Smith G, Cook DG. The effect of infant feeding on the risk of obesity across the lifecourse; a quantitative review of published evidence. Pediatrics. 2005;115:1367–1377. [PubMed]
2. Owen CG, Martin RM, Whincup PH, Davey Smith G, Gillman MW, Cook DG. The effect of infant feeding on mean body mass index throughout the lifecourse; a quantitative review of observational evidence. Am J Clin Nutr. 2005;82:1298–1307. [PubMed]
3. Martin RM, Davey Smith G, Mangtani P, Frankel S, Gunnell D. Association between breast feeding and growth: The Boyd-Orr cohort study. Arch Dis Child Fetal Neonatal Ed. 2002;87:F193–F201. [PMC free article] [PubMed]
4. Wadsworth ME, Hardy RJ, Paul AA, Marshall SF, Cole TJ. Leg and trunk length at 43 years in relation to childhood health, diet and family circumstances; evidence from the 1946 national birth cohort. Int J Epidemiol. 2002;31:383–390. [PubMed]
5. Martin RM, Holly JMP, Davey Smith G, et al. Could associations between breastfeeding and insulin-like growth factors underlie associations of breastfeeding with adult chronic disease? The Avon Longitudinal Study of Parents and Children. Clin Endocrinol. 2005;62:728–737. [PubMed]
6. Larnkjaer A, Ingstrup HK, Schack-Nielsen L, et al. Early programming of the IGF-I axis: Negative association between IGF-I in infancy and late adolescence in a 17-year longitudinal follow-up study of healthy subjects. Growth Hormone & IGF Research. 2009;19:82–86. [PubMed]
7. Ong KK, Langkamp M, Ranke MB. Insulin-like growth factor I concentrations in infancy predict differential gains in body length and adiposity: the Cambridge Baby Growth Study. The American Journal of Clinical Nutrition. 2009;90:156–161. [PubMed]
8. Hoppe C, Molgaard C, Thomsen BL, Juul A, Michaelsen KF. Protein intake at 9 mo of age is associated with body size but not with body fat in 10-y-old Danish children. The American Journal of Clinical Nutrition. 2004;79:494–501. [PubMed]
9. Kwok MK, Schooling CM, Lam TH, Leung GM. Does breastfeeding protect against childhood overweight? Hong Kong's 'Children of 1997' birth cohort. International Journal of Epidemiology. 2010;39:297–305. [PubMed]
10. Nelson MC, Gordon-Larsen P, Adair LS. Are Adolescents Who Were Breast-fed Less Likely to Be Overweight?: Analyses of Sibling Pairs to Reduce Confounding. Epidemiology. 2005;16:247–253. [PubMed]
11. Brion MJ, Lawlor DA, Matijasevich A, et al. What are the causal effects of breastfeeding on IQ, obesity and blood pressure? Evidence from comparing high-income with middle-income cohorts. International Journal of Epidemiology. 2011 [PMC free article] [PubMed]
12. Kramer MS, Moodie EEM, Dahhou M, Platt RW. Breastfeeding and Infant Size: Evidence of Reverse Causality. Am J Epidemiol. 2011;173:978–983. [PMC free article] [PubMed]
13. Kramer MS, Guo T, Platt RW. Breastfeeding and infant growth: biology or bias? Pediatrics. 2002;110:343–347. [PubMed]
14. Kramer MS, Chalmers B, Hodnett ED. Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus. JAMA. 2001;285:413–420. [PubMed]
15. Kramer MS, Matush L, Vanilovich I, et al. Effects of prolonged and exclusive breastfeeding on child height, weight, adiposity, and blood pressure at age 6.5 y: evidence from a large randomized trial. American Journal of Clinical Nutrition. 2007;86:1717–1721. [PubMed]
16. Chivers P, Hands B, Parker H, et al. Body mass index, adiposity rebound and early feeding in a longitudinal cohort (Raine Study) Int J Obes. 2010;34:1169–1176. [PubMed]
17. WHO/UNICEF. Protecting, Promoting and Supporting Breastfeeding: The Special Role of Maternity Services. Geneva, Switzerland: World Health Organisation; 1989.
18. Guthrie LB, Oken E, Sterne JAC. Ongoing monitoring of data clustering in multicenter studies. BMC Medical Research Methodology. 2012;12 [PMC free article] [PubMed]
19. Ogden CL, Kuczmarski RJ, Flegal KM. Centers for Disease Control and Prevention 2000 Growth Charts for the United States: Improvements to the 1977 National Center for Health Statistics Version. Pediatrics. 2002;109:45–60. [PubMed]
20. Warnick GR, Leary ET, Ammirati EB, Allen MP. Cholesterol in fingerstick capillary specimens can be equivalent to conventional venous measurements. Archives of Pathology & Laboratory Medicine. 1994;118:1110–1114. [PubMed]
21. Mei JV, Alexander JR, Adam BW, Hannon WH. Use of Filter Paper for the Collection and Analysis of Human Whole Blood Specimens. Journal of Nutrition. 2001;131:1631S–1636S. [PubMed]
22. Blood Collection on Filter Paper for Neonatal Screening Programs, approved standard, National Committee for Clinical Laboratory Standards Document A4A3. 3rd edition. Wayne, PA: National Committee for Clinical Laboratory Standards; 1997. National Committee for Clinical Laboratory Standards (NCCLS)
23. Martin RM, Patel R, Zinovik A, et al. Filter Paper Blood Spot Enzyme Linked Immunoassay for Insulin and Application in the Evaluation of Determinants of Child Insulin Resistance. PLoS ONE. 2012;7:e46752. [PMC free article] [PubMed]
24. Diamandi A, Khosravi MJ, Mistry J, Martinez V, Guevara-Aguirre J. Filter paper blood spot assay of human insulin-like growth factor I (IGF-I) and IGF-binding protein-3 and preliminary application in the evaluation of growth hormone status. Journal of Clinical Endocrinology & Metabolism. 1998;83:2296–2301. [PubMed]
25. Rowlands MA, Holly JMP, Gunnell D, et al. Circulating Insulin-Like Growth Factors and IGF-Binding Proteins in PSA-Detected Prostate Cancer: The Large Case-Control Study ProtecT. Cancer Res. 2012;72:503–515. [PMC free article] [PubMed]
26. Horton NJ, Kleinman KP. Much Ado About Nothing: A comparison of missing data methods and software to fit incomplete data regression models. The American Statistician. 2007;61:79–90. [PMC free article] [PubMed]
27. Rubin DB. Multiple Imputation for Non Response In Surveys. New York: J. Wiley & Sons; 1987.
28. Angrist JD, Imbens GW, Rubin DB. Identification of causal effects using instrumental variables. Journal of the American Statistical Association. 1996;91:444–472.
29. Palmer TM, Sterne JAC, Harbord RM, et al. Instrumental Variable Estimation of Causal Risk Ratios and Causal Odds Ratios in Mendelian Randomization Analyses. Am J Epidemiol. 2011;173:1392–1403. [PubMed]
30. Ukoumunne OC, Gulliford MCCS, Sterne JAC, Burney PGJ, Donner A. Evaluation of health interventions at area and organisational level. BMJ. 1999;319:376–379. [PMC free article] [PubMed]
31. Victora CG, Barros F, Lima RC, Horta BL, Wells J. Anthropometry and body composition of 18 year old men according to duration of breast feeding: birth cohort study from Brazil. BMJ. 2003;327:901–904. [PMC free article] [PubMed]
32. Parsons TJ, Power C, Manor O. Infant feeding and obesity through the lifecourse. Arch Dis Child. 2003;88:793–794. [PMC free article] [PubMed]
33. Michels KB, Willett WC, Graubard BI, et al. A longitudinal study of infant feeding and obesity throughout life course. Int J Obes. 2007;31:1078–1085. [PubMed]
34. Nelson MC, Gordon-Larsen P, Adair LS. Are Adolescents Who Were Breast-fed Less Likely to Be Overweight?: Analyses of Sibling Pairs to Reduce Confounding. Epidemiology. 2005;16:247–253. [PubMed]
35. Gillman MW, Rifas-Shiman SL, Berkey CS, et al. Breast-feeding and Overweight in Adolescence. Epidemiology. 2006;17:112–114. [PMC free article] [PubMed]
36. Harder T, Bergmann R, Kallischnigg G, Plagemann A. Duration of Breastfeeding and Risk of Overweight: A Meta-Analysis. American Journal of Epidemiology. 2005;162:397–403. [PubMed]
37. Davey Smith G, Ebrahim S. Data dredging, bias, or confounding. BMJ. 2002;325:1437–1438. [PMC free article] [PubMed]
38. Kipping RR, Jago R, Lawlor DA. Obesity in children. Part 1: Epidemiology, measurement, risk factors, and screening. BMJ. 2008;337:a1824. [PubMed]
39. Goran MI. Measurement Issues Related to Studies of Childhood Obesity: Assessment of Body Composition, Body Fat Distribution, Physical Activity, and Food Intake. Pediatrics. 1998;101:505–518. [PubMed]
40. Monasta L, Batty GD, Cattaneo A, et al. Early-life determinants of overweight and obesity: a review of systematic reviews. Obesity Reviews. 2010;11:695–708. [PubMed]
41. Kramer MS, Martin RM, Sterne JAC, Shapiro S, Dahhou M, Platt RW. The double jeopardy of clustered measurement and cluster randomisation. BMJ. 2009;339:b2900. [PubMed]
42. Ogden CL, Caroll MD, Kit BK, Flegal KM. Prevalence of obesity in the United States, 2009–2010. Hyattsville, MD: 2012.
43. Jones JS. Insulin-like growth factor I measurement on filter paper blood spots. Hormone Research. 2001;55(Suppl 2):80–83. [PubMed]
44. Schutt BS, Weber K, Elmlinger MW, Ranke MB. Measuring IGF-I, IGFBP-2 and IGFBP-3 from dried blood spots on filter paper is not only practical but also reliable. Growth Hormone & IGF Research. 2003;13:75–80. [PubMed]
45. Kramer MS, Aboud F, Mironova E, et al. Breastfeeding and Child Cognitive Development: New Evidence From a Large Randomized Trial. Archives of General Psychiatry. 2008;65:578–584. [PubMed]