Prolonged Duration of Breastfeeding Does Not Affect Lipid Profile in Adulthood

Abstract

Background:

The Hertfordshire study suggested that age of weaning and methods of infant feeding may influence adult serum low-density lipoprotein cholesterol (LDL-C) and mortality from ischemic heart disease in men. Although breastfeeding <1 year appeared to be protective compared with formula, prolonged breastfeeding (>1 year) was associated with increased LDL-C and mortality. This study tested the effect of breastfeeding duration on lipid profile in young adults.

Materials and Methods:

Adult volunteers whose mothers could recall the age at which their child was weaned were recruited. We excluded patients with known dyslipidemia, dyslipidemia in a first-degree relative, obesity (body mass index [BMI] >30 kg/m²), pregnant or within 3 months of a previous pregnancy, subjects taking medications that may affect lipid profile, or subjects with a chronic medical condition. Nutritional questionnaires and number of weekly hours of exercise were recorded. Lipid profiles were obtained after a night fast.

Results:

The three groups (no breastfeeding, breastfeeding 0–6 months, or breastfeeding >9 months) did not differ in terms of age, gender, BMI, level of education, amount of exercise, and dietary style. By stepwise background multiple regression analysis taking into account exercise, nutritional habitus, age, BMI, gender, and socioeconomic status, blood lipids were not affected by duration of breastfeeding.

Conclusions:

Duration of breastfeeding does not affect lipid profiles in young adults. These findings do not support negative messages on the long-term effect of prolonged lactation generated by the Hertfordshire study.

Introduction

The Hertfordshire retrospective study suggested that age of weaning and methods of infant feeding may influence adult serum low-density lipoprotein (LDL) cholesterol (LDL-C) and mortality from ischemic heart disease in men.¹ In that study, breastfeeding of less than 1 year appeared to be protective compared with formula. Paradoxically, prolonged breastfeeding above 1 year was associated with increased LDL-C levels.¹ Whether or not the protective effects of human milk feeding on adult ischemic vascular disease are related to variations in lipid metabolism or through other mechanisms is unknown. On the one hand, human milk–fed infants have a much lower endogenous cholesterol synthesis rate than formula-fed infants, presumably because human milk is much richer in cholesterol than cow's milk–based formulas.² Moreover, human milk–fed infants have much higher serum cholesterol and LDL-C concentrations than formula-fed infants.² We have recently shown that during prolonged breastfeeding beyond 1 year of lactation, human milk had extremely high fat content compared with mature milk produced during the first 6 months of lactation.³ Finally, in a large primate model of breastfeeding, Mott et al.⁴ showed that adult baboons breastfed as infants had an approximately 8% lower cholesterol production rate than formula-fed animals and a 20% higher bile cholesterol saturation index. Moreover, in this study, the level of cholesterol in the infant formulas influenced cholesterol metabolism in adulthood but not serum lipoprotein concentrations. Paradoxically, as young adults, breastfed baboons had more extensive arterial fatty streaks than formula-fed baboons.⁴

We conducted the following study to test the null hypothesis that lipid profiles in young adults are not affected by breastfeeding duration, while taking into account lifestyle (exercise, nutritional habitus), age, gender, and socioeconomic status.

Materials and Methods

This study was designed as a pilot one that would enable us to calculate sample size necessary to detect the effect if any of breastfeeding duration on lipid profiles of young adults. We aimed to recruit 50 healthy volunteers during a 2-month period. All volunteers were adults (>18 years of age and up to 40 years of age) and were all physicians, medical students, or nurses or healthy parents of children admitted at the Dana-Dwek Children's Hospital in the Tel Aviv Medical Center. Per protocol, all participants had to have a living mother who could recall exactly the age at which her child had been weaned from breastfeeding. This age was defined as breastfeeding duration and expressed in full months. We excluded all participants with known dyslipidemia (or with a known dyslipidemia in a first-degree relative), obese participants (body mass index [BMI] >30 kg/m²), pregnant women or those within 3 months of a previous pregnancy, subjects taking medications that may affect lipid profile, or subjects with chronic medical conditions.

Weight and height were carefully recorded using a digital scale, and the subjects' BMIs were calculated and expressed as kg/m². A 72-hour dietary diary (including 1 weekend day) was prospectively recorded in order to determine nutritional lifestyle of each participant. The weekly number of exercise hours was collected from each subject, and the number of years of education was used as a proxy for socioeconomic status.⁵

Lipid profiles included the measurement of total cholesterol, LDL-C, high-density lipoprotein (HDL) cholesterol (HDL-C), and triglycerides.⁶ All blood samples were obtained by venipuncture in all subjects after an 8-hour night fast.^6–8

The study was approved by our local institutional review board, and a written informed consent was obtained from each one of the participants.

Statistical analysis

The statistical analyses were conducted using Minitab version 16.4.1 software (Minitab Inc., State College, PA). Each of the lipids measured from the profile was first correlated with lactation duration in a univariate regression analysis. Analysis of variance was used to detect differences in lipids among three groups of lactation duration (no breastfeeding, breastfeeding 0–6 months, breastfeeding >9 months). We used backward stepwise logistic (using lactation duration by group) and multiple (using lactation duration as a continuous variable) regression analyses in order to study the effect of breastfeeding duration (independent variable) on a given lipid concentration (dependent variable) while taking into account potentials effects of BMI, physical activity, socioeconomic status, and cholesterol and saturated fat daily intake. Continuous variables were expressed as median and range when they were not normally distributed and as mean and standard deviation when they were normally distributed. Dichotomous variables were expressed as a percentage/ratio. A value of p≤0.05 was considered significant.

Results

Table 1 depicts selected demographic and clinical characteristics of the subjects, classified as three groups by breastfeeding duration: no breastfeeding, breastfeeding 0–6 months, and breastfeeding >9 months. In brief, by analysis of variance of means (for continuous data) and chi-squared test (for categorical data), there were no differences among groups in age, gender, BMI, level of education, amount of exercise per week, and dietary style. Table 2 depicts the lipid profiles of the three groups. There were no significant differences in any of the variables studied among the groups.

Table 1.

Demographic and Clinical Characteristics of the Subjects

	No breastfeeding (n=16)	Breastfeeding <6 months (n=21)	Breastfeeding >9 months (n=18)
Age (years)	29.9±3.3 (30, 23–36)	28.8±3.0 (28, 24–33)	27.0±4.7 (29, 18–34)
Gender (F:M)	14:2	10:11	13:5
BMI (kg/m²)	21.1±1.9 (21.0, 17.8–24.8)	22.5±2.6 (22.2, 16.5–26.8)	22.2±2.1 (22.2, 18.8–26.0)
Education (years)	17.4±2.2 (18, 13–21)	18.0±2.9 (18, 12–24)	16.4±2.9 (16.5, 12–21)
Exercise (hours per week)	2.7±2.4 (3, 0–7)	1.6±2.2 (1, 0–7)	3.17±3.6 (1.5, 0–14)
Breastfeeding duration (months)	0	3.8±1.6 (4, 0.5–6)	16.4±9.8 (13, 9–48)
Diet style (1:2:3)^a	13:1:2	19:2:0	11:6:1

Data are mean±standard deviation (median, range) values or number of subjects, as indicated.

1=omnivorous diet, 2=vegetarian diet, and 3=vegan diet.

BMI, body mass index; F, female; M, male.

Table 2.

Lipid Profile of the Participants

	No breastfeeding (n=16)	Breastfeeding <6 months (n=21)	Breastfeeding >9 months (n=18)	p
Total cholesterol (mg/dL)	174±38 (173, 108–229)	160±22 (155, 125–213)	168±34 (163, 116–252)	NS
Triglycerides (mg/dL)	78±30 (72, 34–156)	81±29 (75, 41–178)	81±42 (71, 30–186)	NS
HDL-C (mg/dL)	73±20 (68, 52–130)	64±19 (59, 38–106)	65±18 (62, 43–113)	NS
LDL-C (mg/dL)	85±25 (82, 42–133)	79±21 (82, 36–111)	88±30 (84, 38–157)	NS

Data are expressed as mean±standard deviation (median, range) values.

HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; NS, not significant.

In stepwise backward multiple regression analysis taking into account exercise, nutritional lifestyle, age, BMI, gender, and years of education, none of the measured blood lipids was affected by the duration of breastfeeding (whether expressed as a continuous variable or a categorical variable of the three groups).

Discussion

The results of our current study show that in young adults 18–36 years of age, lipid profiles (total cholesterol, LDL-C, HDL-C, and triglycerides) are unaffected by the mode of feeding during the first 2 years of life (whether or not they had been breastfed for any period of time) or by breastfeeding duration.

This study is in sharp contrast to the study of Fall et al.,¹ published 20 years ago, whereby a population of men born from 1911 to 1930 had its mortality and lipids profile characteristics determined in relation to the mode of feeding in early childhood and in particular whether or not they had been breastfed and whether or not they had been weaned by 1 year of age. This previous study supported strongly the hypothesis that infant feeding has a lasting effect on cardiovascular mortality and lipid metabolism. Indeed, in this previous study men who had been exclusively breastfed but weaned by 1 year of age had lower standardized mortality ratios, lower standardized mortality ratio for ischemic heart disease, lower serum concentrations of total cholesterol and LDL-C, and a lower LDL/HDL ratio, in both fasting and nonfasting samples, than those who had received mixed feeding (bottle feeding and breastfeeding) during the first year of life and those who had been exclusively breastfed and weaned after 1 year of age. It is important that the highest mortality ratios, serum concentrations of total cholesterol and LDL-C, and LDL/HDL ratio were found in the exclusively bottle-fed group.² Thus, in the study of Fall et al.,¹ although human milk feeding appeared to have a long-term protective effect on mortality from ischemic heart disease and a favorable effect on lipid profiles (compared with formula feeding), paradoxically, these effects appeared to be “lost” when human milk feeding lasted more than 1 year.

The possible reasons for the discrepancy between our results and those of Fall et al.¹ are multiple:

(1) In the previous study the sample size of 5,718 men (nearly all the population of men living in Hertfordshire, United Kingdom, during those years) may have helped detecting differences that could not be found in our small study (type 2 error). However, in the current study there is not even a trend for higher serum cholesterol and LDL-C levels and LDL/HDL ratio in former formula-fed adults compared with breastfed ones.

(2) The Hertfordshire findings cannot be applied to our times for multiple reasons: the bottle foods available 100 years ago included patent preparations of dried cow's milk, unmodified cow's milk, diluted condensed milk, and even patent foods made from wheat flour or arrowroot.¹ Modern formulas are strikingly different in that they are fortified with iron and vitamins,^9,10 they have been defatted and their fat has been replaced by fat of vegetable origin,^9,11 and they are devoid of cholesterol (leaving only traces of cholesterol in the formulas²) and composed mostly of unsaturated fatty acids.^9,11 Thus, if lipid composition of the formula is to have long-term, lasting effects on lipid metabolism in later life, we should expect striking differences between the Hertfordshire men and contemporary men born 18–26 years ago.

(3) The definition of prolonged breastfeeding in the study of Fall et al.¹ strikingly differs from ours. In our study weaning meant that the child would stop having any amount of breastmilk. In contrast, in the previous study the health visitors recorded that each baby was or was not weaned at 1 year. The term “weaned” implied that either breastfeeding has stopped or that solid food had been introduced. Thus, the group categorized as weaned prior to 1 year of age may have contained an unknown percentage of children who continued to breastfeed (partially) beyond 1 year of age. In the study of Fall et al.¹ 6% of exclusively formula-fed infants were categorized as weaned by 1 year of age (presumably because they had started solids). Thus it is not possible in this previous study to clearly identify who belonged to each group, which adds much background “noise” to the results.

A limitation of our study is that breastfeeding data (mode of feeding and duration) were obtained in a retrospective manner, leaving room to the imprecision of human recollection. We took care of this limitation in the best possible manner, by limiting participation to only adults with a living mother who could recall the age at which her child had been totally weaned from breastfeeding. In contrast, a strength of our study is that we looked at breastfeeding duration as a continuous variable (rather than a dichotomous variable of less or more than 9 months), which allowed us to test whether or not there was a “dose–response” relationship between breastfeeding duration and lipid profiles.

Another strength of the study was that we took into account important variables that are known to possibly affect lipid profiles, such as those linked to lifestyle (exercise, nutritional habitus), and others (age, gender, and socioeconomic status). We also, in order to reduce confounders as much as possible, excluded all participants with known dyslipidemia or with a known dyslipidemia in a first-degree relative, obese participants (BMI >30 kg/m²), pregnant women or those within 3 months of a previous pregnancy, subjects taking medications that may affect lipid profile (including birth control pills), subjects with a chronic medical condition, and smokers.

Our group previously showed that the amount of cholesterol present in the food offered to infants (human milk or formula) during the first few months of life has striking effects on the infant fractional cholesterol synthesis rates (FCSR).² The FCSR can easily be “manipulated” by dietary changes; in brief, the least amount of dietary cholesterol, the highest endogenous FCSR. Because during prolonged lactation beyond 1 year, fat content of human milk (and presumably cholesterol content) can be extraordinarily high,³ this could possibly and paradoxically be protective by depressing FCSR. In a rat model, Reiser and Sidelman¹² showed that such manipulations during suckling are long lasting and that their effects persist in the adult rat. Whether or not lipid profiles after a cholesterol challenge may or may not be affected in adults by the way they were fed as children remains to be studied.

Conclusions

In conclusion, the results of the current study suggest that the duration of breastfeeding does not affect in any major way lipid profiles in young adults. We speculate that these findings do not support any negative messages on the long-term effect of prolonged lactation, generated by the previous Hertfordshire study of Fall et al.¹

Footnotes

Disclosure Statement

No competing financial interests exist.

References

Fall

, Barker

, Osmond

, et al. Relation of infant feeding to adult serum cholesterol concentration and death from ischaemic heart disease. BMJ, 1992; 304:801–805.

Cruz

, Wong

, Mimouni

, et al. Effects of infant nutrition on cholesterol synthesis rates. Pediatr Res, 1994; 35:135–140.

Mandel

, Lubetzky

, Dollberg

, et al. Fat and energy contents of expressed human breast milk in prolonged lactation. Pediatrics, 2005; 116:e432–e435.

Mott

, Lewis

, McGill

Jr . Programming of cholesterol metabolism by breast or formula feeding. Ciba Found Symp, 1991; 156:56–66.

Lien

, Friestad

, Klepp

. Adolescents' proxy reports of parents' socioeconomic status: How valid are they?. J Epidemiol Community Health, 2001; 55:731–737.

Wikipeida.. Lipid profile. Available at http://en.wikipedia.org/wiki/Lipid_profile (accessed July 22, 2014).

Langsted

, Freiberg

, Nordestgaard

. Fasting and nonfasting lipid levels: Influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction. Circulation, 2008; 118:2047–2056.

Sidhu

, Naugler

. Fasting time and lipid levels in a community-based population: A cross-sectional study. Arch Intern Med, 2012; 172:1707–1710.

O'Connor

. Infant formula. Am Fam Physician, 2009; 79:565–70.

10.

Lubetzky

, Mandel

, Mimouni

. Vitamin and mineral supplementation of term infants: Are they necessary?. World Rev Nutr Diet, 2013; 108:79–85.

11.

Koletzko

, Bremer

. Fat content and fatty acid composition of infant formulas. Acta Paediatr Scand, 1989; 78:513–521.

12.

Reiser

, Sidelman

. Control of serum cholesterol homeostasis by cholesterol in the milk of the suckling rat. J Nutr, 1972; 102:1009–1016.