Table 1.
Responsivity of breast milk fatty acid content to maternal diet
| Ref | Participants Risk of bias |
Study | Ethnicity | Breast milk collection timing | Effects on breast milk PUFA content |
|---|---|---|---|---|---|
| Experimental studies | |||||
| Argaw, 2020 | Healthy lactating women 6–12 months PP Ethiopia RCT Some concerns |
215 mg DHA and 285 mg EPA per d (n 72) or placebo (control maize oil without DHA or EPA, n 82) for 6 months | African | Not provided | Supplementation increased BM DHA, by 39 % (95 % CI 20·6, 57·5 %, P < 0·001) and EPA by 36 % (95 % CI 16·0, 56·4 % P < 0·001) compared with controls |
| Boris, 2004 | Healthy lactating women enrolled at 30 weeks GA Denmark RCT Some concerns |
Fish oil supplementation (900 mg DHA and 300 mg EPA per d) until delivery (n 12), or further 30 d (n 11), or placebo (olive oil, n 13) | Danish women supposedly White Caucasian | Morning milk from one breast only | BM DHA levels 2·1, 3·6 and 2·8 times higher at 4, 16 and 30 d, respectively, in extended supplementation group v. control group (all P < 0·001) |
| Craig-Schmidt, 1984 | Healthy lactating women 2 months PP USA RCT Some concerns |
BM samples (n 8). Participants provided with a two 5 d diets with an intervening 2 d period. Diets for the two periods were identical except that source of hydrogenated fats were used in the first period and non-hydrogenated fats is the second period | Not provided | Milk samples collected after first nursing of the day by manual expression for both breasts | BM myristic acid (14:0), palmitic acid (16:0) and palmitoleic acid (16:1n-7), all significantly lower following diets with hydrogenated v. non-hydrogenated fats, whereas elaidic acid (18:1 trans-9) and oleic acid (18:1n-9) were significantly higher |
| Fougere, 2021 | Healthy lactating women, < 72 h PP France RCT Low risk of bias |
1·2 g DHA per d (n 196) or placebo (maize oil and soyabean oil, (n 193) for 14 d | Not provided | Not provided | After 14 d, significantly higher levels of BM mean (sd) DHA and EPA in supplemented (DHA: 0·95 % (0·44 %); EPA: 0·08 %, (0·05 %)) v. control (DHA: 0·34 %, (0·20 %); EPA: 0·07 %, (0·07 %)), both P < 0·0001 |
| Hawkes, 2002 | Healthy lactating women 3 d PP Australia RCT Some concerns |
300 mg DHA and 70 mg EPA per d (n 26), or 600 mg DHA and 140 mg EPA per d (n 28), or placebo (sunflower oil, n 27) for 4 weeks | Not provided | Hand-expressed morning milk | Mean (sd) BM DHA content increased in a linear manner in response to dietary DHA (placebo DHA: 0·26 %, (0·08 %); LoDHA: 0·39 %, (0·09 %); HiDHA: 0·66 %, (0·18 %); all P < 0·05). BM EPA only significantly increased in HiDHA group (placebo EPA: 0·11 %, (0·02 %); LoDHA: 0·11 %, (0·02 %); HiDHA: 0·14 %, (0·03 %); P < 0·05). No significant effects on ARA |
| Lauritzen2002 | Healthy lactating women 4 months PP Denmark RCT Some concerns |
BM samples (n 12). Lactating women were given fish oil (2–8 g) for breakfast and delivered 6–12 BM samples during the following 24 h | Not provided | Morning milk, after first feeding | Mean (sd) BM DHA of the fish-eating mothers was 0·57 %, (0·28 %) and non-fish-eating mothers was 0·42 %, (0·15 %); P = 0·05. Fish oil supplementation resulted in a 2-fold increase in BM DHA levels, peaked after 10 h and lasted for 24 h |
| Mazurier, 2017 | Lactating mothers 1–4 months PP France RCT Low risk of bias |
All groups received 350 mg DHA and 210 mg EPA per d but varied in ALA and LA content. n 19–22 per group, with 15 d washout period followed by 15 d supplementation | White Caucasian | Human milk collected at the first infant feeding of the morning | Significant dose response following in ALA (P < 0·003), but no significant effects on LA. No significant differences in DHA or EPA between groups. ARA content significantly increased only in intermediate ALA and LA dose, and n-3 PUFA-enriched rapeseed oil group |
| Mellies, 1979 | Overweight and normal weight lactating women 1 month PP USA RCT Some concerns |
BM samples (n 14), 2 weeks baseline nutrition history collected by a nutritionist, mothers randomly assigned to one diet followed by the other; diet 1: cholesterol-poor phytosterol-rich, PUFA-rich (PUFA:SFA ratio 1·8); diet 2: cholesterol-rich, phytosterol-poor, PUFA-poor (PUFA:SFA ratio 0·12) | Not provided | Samples collected at the beginning or end of the second nursing period of the day, through manual expression or breast pump | Mean (s em) for baseline diet v. diet 1 v. diet 2; milk cholesterol (mg/g milk fat) 2·4 (0·4) v. 2·4 (0·1) v. 2·5(0·2): milk phytosterols (mg/g milk fat) 0·17 (0·03) v. 2·2 (0·3) v. 0·7 (0·1); total milk fat, 3·58 (0·56) v. 2·69 (0·16) v. 2·66 (0·16) (P < 0·001) |
| Nasser, 2010 | Healthy lactating women between 2 and 6 months PP, vegetarian excluded Canada RCT Some concerns |
BM samples (n 14), low-fat diet or high-fat diet for 4 d in randomised order | Not provided | Milk collected on the last 2 d of each 4-d period between 1 and 14.00 using a manual breast pump | Significant differences in mean (sem) BM composition in low fat v. high-fat diet for lauric acid (12:0) 5·38 (1·16) v. 3·98 (0·37) (P = 0·01), palmitoleic acid (16:1n-7) 1·95 (0·29) v. 1·31 (0·23) (P = 0·046), ALA 1·22 (0·04) v. 0·69 (0·06) (P = 0·01), ARA 0·34 (0·01) v. 0·30 (0·02) (P = 0·02) |
| Park, 1999 | Healthy lactating women between 1 and 26 months PP USA RCT Some concerns |
Three-week crossover study. Week 1, minimal rumenic acid (18:2 cis-9, trans-11) foods (depletion), then either high-fat dairy food or low-fat dairy food intake for 1 week, then crossover (n 8 per group). BM samples (n 16), and dietary records during last 3 d of each period and FFQ | Not provided | Not provided | Significantly higher BM (mean (sem), μmol/g lipid) in high-fat dairy v. low-fat dairy groups for: rumenic acid (13·5 (1·1) v. 8·2 (0·4)), myristic acid (264·7 (34·2) v. 195·2 (11·0)), palmitic acid (707·0 (51·5) v. (511·3 (16·4)), stearic acid (1055·0 (103·4) v. 874·3 (33·3)), oleic acid (1055·0 (103·4) v. (874·3 (33·3)) and significantly lower ALA (10·7 (3·6) v. 17·6 (1·1)), all P < 0·05) |
| Smithers, 2010 | Mothers of preterm infants born 33 weeks GA Australia RCT Low risk of bias |
900 mg DHA, 195 mg EPA, and 54 mg ARA per d, (n 69) or placebo (soyabean oil, n 74) for 2 weeks | 98 % Caucasian | Not provided | DHA significantly higher in supplemented (M = 1·0 % sd = 0·4 %) v. placebo (M = 0·3 % sd = 0·1 %) groups, (P < 0·05). No significant differences in EPA or ARA between groups |
| Storck lindholm, 2012 | Obese and normal-weight lactating women after delivery Sweden Obs/RCT Some concerns |
Control group (BMI < 25 kg/m2, n 26), Group O (BMI > 30 kg/m2, n 25) and Group I (BMI > 30 kg/m2, n 25) were given dietary advice (e.g. eat fish 2–3 times a week) and increase physical activity. BM measured 3 and 10 d, 1 and 2 months | White Caucasian | Not provided | Group O had low fish intake and at baseline had the lowest BM ALA, EPA and DHA (all P < 0·01), which was continued across subsequent samples. The ARA: EPA + DHA ratio was significantly higher in Group O across repeated samples (P < 0·01), compared with the other groups. Group I levels approached those seen in control group |
| Valentine, 2013 | Milk donors mean lactational age 19 weeks USA RCT Some concerns |
1 g DHA/d (n 69), or placebo (soyabean oil, n 74) for 14 d | Not provided | Not provided | Supplementation significantly increased DHA content when expressed as mol wt%, but not in absolute amounts. No significant effect of supplementation on EPA or ARA content |
| Valenzuela, 2015 | Lactating women at delivery and 6 months PP Chile RCT Some concerns |
10·1 g ALA per d (chia oil, n 19) or untreated control group (n 21) for 9 months | Hispanic | Not provided | Significant increase in ALA and significant decrease in LA following supplementation v. control at all time points (all P < 0·05). DHA significantly increased in first 3 months (P < 0·05), then no effect, no change in EPA or ARA v. control at any time point. |
| Yang, 2022 | Healthy lactating women between 30 and 50 d PP China RCT Some concerns |
200 mg DHA per d (n 77) or placebo (n 60) for 8 weeks | Asian | Breast milk samples collected between 07.30 and 09.00 | Absolute GLA (18:3n-6), ARA and DHA significantly decreased over the study in control group (P < 0·001, P = 0·001 and P < 0·037, respectively), whereas GLA and DHA were maintained in supplemented group, although there was a significant decrease in ARA (P = 0·03). DHA content was significantly higher in supplemented v. control group at the end (P = 0·012. Similar trends were found when expressed as relative content |
| Observational studies | |||||
| Aitchison, 1977 | Healthy lactating women between 4 and 6 months PP USA Obs Fair Quality 6 |
BM samples (n 11), recorded food intake for 1 week and saved duplicated food portions consumed on 3 d | Not provided | Experiment 1: five subjects took morning and evening milk samples. Experiment 2: six additional subjects collected morning milk only |
Correlation coefficient (P) between PUFA to SFA ratio in maternal diet and milk (% of total FA): 0·46 considering diet and milk in the same evening; 0·43 (P < 0·05) considering milk in the next morning In nine of eleven subjects, fluctuation of percent total trans acids in the milk appeared to follow dietary trans changes after a 12–36-h lag period |
| Antonakou, 2013 | Healthy lactating women 1 month PP Greece Obs Fair Quality 6 |
BM samples 1-month PP (n 64), 3-month PP (n 39), 6-month PP (n 24). Three-day dietary record at 1st, 3rd and 6th month PP | Caucasian | Morning milk collected after at least 2 h after previous breast-feeding | BM fat ranged from 26·3 and 30·2 g/l (P < 0·05). Strong positive effect found during first month lactation between mother’s PUFA intakes and BM PUFA concentration, r = 0·25, P < 0·05; n-3 fatty acids, r = 0·26, P < 0·05; DHA r = 0·27, P < 0·05 and LA, r = 0·26, P < 0·05, while MUFA intake was strongly correlated with concentration of PUFA, r = 0·29, P < 0·05; n-6, r = 0·27, P < 0·05 and LA,, r = 0·25, P < 0·05 |
| Bzikowska, 2019 | Healthy lactating women 1 month PP Poland Obs Fair quality 6 |
BM samples (n 32) and dietary information 3-d dietary record and FFQ | Not provided | Foremilk and hindmilk collected from four time periods: 06.00–12.00, 12.00–18.00, 18.00–00.00 and 00.00–06.00 | Significant positive correlation between fatty fish consumption and DHA (τb = 0·25, P = 0·049), EPA (τb = 0·27, P = 0·03) and ALA (τb = 0·28, P = 0·02). ALA positively correlated with intakes of linseed oil (τb = 0·3, P = 0·01), coconut oil (τb = 0·29, P = 0·02) and milk (τb = 0·26, P = 0·04). EPA positively correlated with pork consumption (τb 0·29, P = 0·02) |
| Daud, 2013 | Healthy lactating women between 15 d and 6 months PP Malaysia Obs Fair Quality 6 |
BM samples (n 101). Participants provided a 1-year period FFQ. Sub-experiment, BM samples (n 18). Participants provided a 3-d FFQ | Asian | Not provided | The most abundant BM trans-fatty acid was linoelaidic acid (mean = 1·44%, sem = 0·60 % fatty acids), which was also the most consumed (mean = 0·07 sem = 0·01, g/100 g of food). Ten food items had an effect on the total BM trans-fatty acids (buns, chicken burgers, cheeseburgers, shortening, powdered milk, sweetened milk blended oil mayonnaise, maize oil and ice cream). No association between consumption and BM trans-fatty acid levels |
| De la Presa-Owens, 1996 | Healthy lactating women < 1 PP Spain Obs Poor Quality 2 |
BM samples (n 40), dietary questionnaire | Not provided | Not provided | Lower BM LA observed between mothers consuming olive oil (n 15) or sunflower (n 6) as the preferred source of fat (P < 0·001). Significant differences in BM DHA and EPA between mothers reporting high, low or no fish consumption (P < 0·001) |
| Freitas, 2019 | Healthy lactating women < 3 PP Brazil Obs Fair quality 2 |
Diet quality assessed through a semi-structured questionnaire (n 106) | Most participants black/multiracial (82 %) | Sample collected after the first breast-feeding of the morning | Long-chain n-3 and n-6 PUFA not analysed separately. Total fruits and whole fruits, r = −0·302, r = 0·283, respectively, both P < 0·05 |
| Juber, 2017 | Healthy new mothers > 1 week PP USA Obs Fair quality 7 |
BM at baseline (n 84), subject received analysis of BM DHA and dietary recommendations (n 60) had second sample at 1 month | 99 % White Caucasian | Not provided | Those reporting taking DHA supplements (n 43) had higher levels than those who did not (0·23 % v. 0·15 %, P < 0·0001). In second sample, median breast milk DHA content increased from 0·19 % to 0·22 % (P < 0·01) |
| Liu, 2015 | Healthy lactating women 22–25 d PP China Obs Good quality 8 |
Dietary intake of lactating women assessed with 24-h dietary recall questionnaire (n 514) | Asian | Morning milk, manual expression between 09.00 and 11.00 | Significant negative correlation between dietary ALA and BM GLA (18:3n-6, r2 = −0·201, P = 0·03) and adrenic acid (22:4n-6, r2 = −0·197, P = 0·03), and dietary LA and BM DGLA (20:3n-6, r2 = −0·182, P = 0·03) |
| Makela, 2013 | Overweight and normal-weight lactating women 3 months PP Finland Obs Good quality 8 |
BM samples (n 100), self-administrated dietary record of the day before milk samples collection every day for 1 week | Not provided | Morning milk, manual expression | Mean (sd) BM from overweight v. normal weight women: SFA (46·3 % (4·4) v. 43·6 % (6·0), P = 0·012), n-3 PUFA (2·2 % (0·79) v. 2·7 % (1·1), P = 0·010), ratio of unsaturated to saturated FA (1·1 (0·2) v. 1·3 (0·4), P = 0·008). Pearson’s correlation coefficient between the high-fat dairy products and breast milk SFA: 0·21 (0·04). |
| Olafsdottir, 2006 | Healthy lactating women between 2 and 4 months PP Iceland Obs Fair quality 6 |
BM samples (n 77), 24-h recalls and food questionnaire on fish consumption and dietary habits. Two groups: women consuming (n 18) v. not consuming cod liver oil (n 59) | Not provided | Collection of four times per d | Proportion of PUFA in the diet is significantly higher among women consuming cod liver oil. It also gives higher % of DHA, EPA and DPA n-3 in BM. Correlation coefficient between maternal diet and milk FA composition (% of total FA): maternal PUFA: SFA and milk ALA, 0·336 (P = 0·003): maternal PUFA and milk ALA, 0·432 (P < 0·001): maternal PUFA and milk EPA, 0·302 (P = 0·008): maternal protein and milk EPA, 0·362 (P = 0·001): maternal protein and milk DPA, 0·373 (P = 0·001): maternal protein and milk DHA, 0·346 (P = 0·002) |
| Perrin, 2019 | Healthy lactating women ≥ 2 weeks PP USA Obs Good quality 7 |
Single BM sample from vegan lactating women (n 26), vegetarian lactating women (n 22) and omnivore lactating women (n 26) | Not provided | Sample collected in the morning during first and second breast-feeding of the day and at least 2 h since previous breast-feeding in a dimly lit room to protect light-sensitive nutrients | Vegan, vegetarian and omnivores median (IQR) unsaturated fatty acids were 66·0 % (6·5 %), 57·8 % (9·8 %) and 56·2 % (8·5 %), respectively (P < 0·001). Total n-3 PUFA were 2·29 % (0·77 %) for vegans 1·55 % (0·56 %) for vegetarians and 1·46 % (0·94) for omnivores (P < 0·001), with significant difference driven by higher ALA (P < 0·001). Ratio of LA to ALA was significantly lower (P < 0·001) in BM vegans 9·3 % (2·1 %) compared with vegetarians 12·2 % (4·9 %) and omnivores 12·7 % (6·2 %). No significant differences in DHA, but over 80 % had levels below 0·30 %. Reports of n-3 PUFA supplement and seafood consumption were limited |
| Sanders, 1978 | Vegan and omnivore healthy lactating women between 2 and 6 months PP UK Obs Fair Quality 6 |
BM samples (n 8) | Caucasian | Sample collected at the start of the morning | BM of vegans contained lower proportions of 16:0, 16:1, 18:0 and 20:4 n-3 and higher proportions of 18:2 n-6 (P < 0·05) |
| Sanders, 1992 | Healthy lactating mothers < 14 weeks PP UK Obs Fair quality 6 |
Milk samples from (n 45); (n 19 vegans, n 5 vegetarians and n 21 omnivores); 3-d food dietary | White vegetarians and vegan, Indian vegetarians and Whit e omnivores | Not provided | In comparisons to omnivores, vegan’s BM contains higher proportion of SCFA (C10–C14) and lower proportion of medium-chain FA (C16–C18); (P < 0·01). Same proportion of ARA is in all groups, and proportion of BM DHA is lower in vegans than in omnivores and vegetarians (P < 0·01). The n-6/n-3 FA ratio was higher in the vegan group than in the others |
| Scopesi, 2001 | Healthy lactating women < 1 month PP Italy Obs Fair quality 6 |
BM samples (n 34) 1 d PP, 4 d PP, 14 d PP, 21 d PP and 28 d PP. Dietary questionnaire referred to the day prior milk extraction for sampling |
Not provided | Not provided | Pearson’s correlation coefficient between maternal dietary intake and corresponding BM concentrations (% of total FA): SFA, 0·60 (P < 0·01) in transitional milk; MUFA, 0·63 (P < 0·01) in transitional milk; PUFA, 0·65 (P < 0·01) in mature milk |
RCT, randomised control trial; BM, breast milk; PP, postpartum; AM, ante meridiem ; PM, post meridiem; M, median; GA, gestational age; Obs, observational study; ALA, α-linolenic acid; ARA, arachidonic acid; ALA, α-linolenic; LA, linoleic acid; DGLA, dihomo-γ-linolenic acid; DTA, docosatetraenoic acid; GLA, γ-linolenic acid.