What Professor Carlsen actually said about infant feeding

Ben Goldacre has reproduced the press release from Professor Carlsen that generated the 'Breast NOT Best' news news stories this week. See:

The press release is headlined "Breastfeeding is not as beneficial as once thought". As I explained earlier, Professor Carslen's hypothesis is that mothers use formula because of excess male hormones, that also have a negative impact on the child in the womb, and it is the male hormones that explain the poorer health of formula-fed babies, rather than the method of feeding. To quote Professor Carlsen from the opening of the press release:

---Quote begins
It is true that breastfed infants are slightly healthier than bottle-fed babies. But apparently it is not the milk that makes the difference. Instead, the baby’s overall health is all determined before he or she is born. So why do so many studies associate breastfeeding with better health for young babies. The answer is simple: If a mother is able to breastfeed, and does so, this ability is essentially proof that the baby has already had an optimal life inside the womb.
---quote ends

His research was about the impact of hormones on infant health and feeding method, but the articles reporting on this have implied that breastfeeding has no impact on infant health compared to formula feeding.

UNICEF Baby Friendly Initiative in the UK has commented: "the claims made in relation to these findings do not account for the large differences in breastfeeding rates between countries, with some having 99% of mothers successfully breastfeeding. They are also contradicted by the large body of evidence which shows that levels of successful breastfeeding can be increased by a range of improved support interventions."

UNICEF Baby Friendly has also taken issue with the suggestion that breastfed infants, as a population, are not healthier than formula-fed infants. While Professor Carlsen reviewed only 50 studies to state breastfed infants are 'slightly healthier', UNICEF cites a systematic review that examined evidence from 400 studies on infant feeding, after ensuring they were well designed, and "found that breastfeeding is associated with a significant reduction in the incidence of: acute otitis media, non-specific gastroenteritis, severe lower respiratory tract infections, atopic dermatitis, obesity, type 1 diabetes, type 2 diabetes, childhood leukaemia, sudden infant death syndrome, necrotizing enterocolitis, maternal breast cancer and ovarian cancer." See:

Professor Carlsen wasn't looking at the impact of feeding methods on health, but the impact of hormones during pregnancy on a mother's breastfeeding behaviour. He found increased use of formula (ie less breastfeeding) amongst mothers with higher levels of male hormones. However, it is interesting to note that the results state there was "correction for maternal age, education and smoking", apparently missing other factors such as birth weight, delivery methods, separation of mother and child after birth etc. which may have been relevant, particularly given the small sample size.

UPDATE: The UK National Health Service has analysed the research and pointed out today that Professor Carlsen did not even look at the impact of male hormones on infant health. It states:

---NHS analysis extract
This study did not compare levels of male hormones during pregnancy to the baby’s health, but only to whether the mother breastfed after giving birth. The researchers have supplied no evidence to back up their claim that there is no benefit from breast milk.
---extract and Update ends

Here is the abstract for the research, available at:

---Professor Carlsen research
Mid-pregnancy androgen levels are negatively associated with breastfeeding

Objective. Breastfeeding depends on endocrine changes during pregnancy. The association between gestational hormones and lactation has been sparsely investigated. Previously, androgens were used for lactation inhibition. We investigated a possible association between second trimester maternal androgen levels and breastfeeding. Design. Prospective observational study.

Setting. University hospital setting.

Population. Women from a random sample of pregnancies (n = 63) and from a group with an increased risk for giving birth to a small-for-gestational age newborn (n = 118) were included. All participants had singleton pregnancies and one or two previous births.

Methods. Maternal androgen levels were measured in gestational week 25. The association with reported breastfeeding was explored by univariate and multivariate linear regression analyses. Analyses were adjusted for factors known to be associated with breastfeeding.

Main outcome measures. Breastfeeding at six weeks, three months, and six months postpartum.

Results. In the random group, breastfeeding at three and six months was negatively associated with maternal testosterone, androstendione, and free testosterone index levels. After correction for maternal age, education and smoking, breastfeeding at both three and six months was negatively associated with the free testosterone index. In the group of women with an increased risk for giving birth to a small-for-gestational age newborn, breastfeeding at six weeks and three months was associated negatively with maternal dehydroepiandrosterone and this association persisted after correction for maternal age, education, and smoking.

Conclusions. Maternal androgen levels in mid-pregnancy are negatively associated with breastfeeding.
---quote ends

Some journalists do like rushing out 'Breast NOT Best' headlines, which do harm around the world and misrepresent or distort the real story. We faced a similar situation last year. You may recall a scientist, Professor Kramer, was quoted in literally thousands of articles around the world as questioning the benefits of breastfeeding. He soon criticised the journalists who wrote the original article for misrepresenting what he had said - but few journalists and media outlets that had run the earlier articles issued follow-up corrections. See:

No comments:

Post a Comment