RESEARCH LETTERS

Breastfeeding: it is worth trying with the second baby

Jennifer Ingram, Michael Woolridge, Rosemary Greenwood Mothers who experience breastfeeding difficulties with their first babies and give up breastfeeding are less likely to breastfeed subsequent babies than mothers who do not experience such difficulties. We carried out a longitudinal study of 22 mothers in which milk output was measured at 1 week and 4 weeks after giving birth to their first and second babies. Significantly more breast milk was produced at 1 week for the second lactation (an increase of 31% [95% CI 1151%]) and the net increase was greatest for those with the lowest milk output on the first occasion (90% [30149%]). They spent less time feeding their second baby (a decrease of 20% [ 34 to 5%]). This increased efficiency of milk transfer was also evident at 4 weeks. Health professionals should encourage women to breastfeed all their children, whatever their experience with their first child.

Quartile for milk volume for the first baby (n) Lowest quartile (5) Second quartile (6) Third quartile (6) Highest quartile (5)

Baby 1 Baby 2 Difference in % change milk volume milk volume milk volume (95% CI)* mL/day mL/day mL/day 357 515 605 716 654 688 675 678 297 173 70 38 +90 (30 to 149) +33 (12 to 54) +11 ( 14 to 37) 5.6 ( 31 to 20)

*Average of the percentage increases for each mother.

Table 2: Mean milk output at 1 week by quartiles for first and second babies

Lancet 2001; 358: 98687

Many mothers give up breastfeeding in the first 6 weeks because of feeding difficulties, the most common of which (48%) is a perception of insufficient milk or the baby being hungry.1 Those who give up feeding early are less likely to breastfeed their subsequent children. In a previous study of 91 women of mixed parity (54 primiparous, 37 multiparous), in which we investigated the maternal predictors of milk output,2 we found that multiparous mothers produced significantly more breast milk (around 140 mL) at one week than the primiparous mothers. We have been able to validate this finding (based on crosssectional data) by following up the primiparous women into their second pregnancy, and asking them to repeat the measurements they had undertaken previously. The 54 primiparous women in our first study were contacted and asked if they would take part in a similar study with their second baby. 14 (26%) had either moved away or did not reply, and a further ten (19%) did not take part (six had already had their second baby and four declined to take part). The remaining 30 were willing to take part. We included 22 of these women in the study because they had had a second baby within the 2 years May, 1997, to April, 1999. They each completed two 24 h test-weighing periods (at 1 week and 4 weeks postnatally), as described previously.2 At the time of the second delivery, the women were aged 2243 (mean 335) years, four delivered by caesarean section, and the mean gestation was 40 weeks. The babies mean birthweight was 376 kg and there were 10 girls and 12 boys. On average, the second baby was heavier than the first baby at birth (an increase of 220 g) and significantly heavier at
Variable Weight (kg) Birth 1 week Volume of milk (mL) 1 week 4 weeks Total feeding time (min) 1 week 4 weeks Intake (mL/kg) 1 week* 4 weeks* Baby 1 354 354 549 771 249 229 156 182 Baby 2 376 385 674 823 191 168 176 175

1 week (an increase of 310 g). The correlation between milk volume for the first and second babies was not significant at 1 week (r=019) or at 4 weeks (r= 029). Significantly more breast milk was produced at 1 week for the second lactation (an increase of 125 mL, 31%) although similar volumes were taken at 4 weeks (an increase of 52 mL, not significant; table 1). Women were grouped into quartiles for milk production with their first baby at 1 week and the change in production with the second baby compared by these quartiles. The women who were lowest producers first time around increased their milk output (by 297 mL), significantly more than higher producers (ANOVA p=0005; as shown in table 2). To control for the babys weight, milk intake was corrected for infant weight (volume/kg) and this was also found to be significantly greater for second babies at 1 week but not at 4 weeks. Total time spent feeding in 24 h was significantly less at 1 week and 4 weeks for the second babies (table 1); a 20% reduction in feeding time in the early weeks, totalling a cut of about 1 h per day. These data indicate an overall, graded upward shift in milk output from first baby to second baby, which is demonstrably greatest (a 90% increase) for women whose milk output was lowest with their first baby (table 2). Regression to the mean may contribute in part to these results, but this is not solely a statistical entity; it is a real phenomenon, which clearly benefits women whose milk output was lowest with their first baby. The absence of a strong correlation in milk volume between babies suggests that feeding the second baby is independent of feeding the first. Although this a small study, as far as we know it is the only one of its type, and the results are clear and unambiguous. The message is that even if a mother felt that she had insufficient milk to feed her first baby, it is worth breastfeeding a second baby, and that health professionals can reassure women that it is well worth them trying a second time.
We thank the 22 mothers who gave their time so willingly. Difference (95% CI) 022 ( 002 to 046) 031 (007 to 055) 125 (49 to 201) 52 ( 62 to 165) 58 ( 103 to 61 ( 101 to 12) 21) t test 194 269 342 095 262 317 215 081 p value 0066 0014 0003 035 0016 0005 0043 043

20 (07 to 38) 7 ( 26 to 12)

*Controlled for 1 week weight and 4 week weight, respectively.

Table 1: Mean breastfeeding-associated differences between first and second babies

986

THE LANCET Vol 358 September 22, 2001

For personal use. Only reproduce with permission from The Lancet Publishing Group.

RESEARCH LETTERS

1 2

Foster K, Lader D, Cheesbrough S. Infant Feeding 1995. Office for National Statistics 1997; HM Stationery Office, London. Ingram JC, Woolridge MW, Greenwood RJ, McGrath L. Maternal predictors of early breast milk output. Acta Paediatrica 1999; 88: 49399.

Institute of Child Health, Education Centre, Royal Hospital for Children, Bristol, BS2 8BJ, UK (J Ingram PhD); Mother and Infant Research Unit, University of Leeds, Leeds (M Woolridge DPhil); and Research and Development Support Unit, United Bristol Hospitals NHS Trust, Bristol (R Greenwood MSc) Correspondence to: Dr Jenny Ingram (e-mail: jenny.ingram@bristol.ac.uk)

Non-invasive distension of the small bowel for magnetic-resonance imaging

Michael A Patak, Johannes M Froehlich, Constantin von Weymarn, Marc A Ritz, Christoph L Zollikofer, Klaus-Ulrich Wentz Magnetic resonance enteroclysis is a promising technique that allows assessment of the small bowel but needs invasive nasoduodenal intubation. We propose a noninvasive distension method for magnetic-resonance imaging (MRI) in which ispaghula, dissolved in an aqueous solution with meglumine gadoterate taken orally over 4 h forms a viscous hydrogel within the intestinal lumen. MRI results from ten volunteers showed good luminal distension, constant signal homogeneity, optimum demarcation of the bowel content from surrounding tissues, and a low rate of artefacts. Our method permits non-invasive high quality MRI of the small bowel.

Lancet 2001; 358: 98788

Sufficient distension of the gut is crucial for intestinal imaging. In conventional imaging of the small bowel nasoduodenal intubation is used. Magnetic-resonance imaging (MRI) with its high soft-tissue contrast and threedimensional imaging capability is ideal for diagnosing intestinal pathologies.1 However, most studies of MRI adhere to radiographically controlled nasoduodenal intubation for bowel distension.13 We propose that ispaghula (Metamucil, Procter and Gamble, Phoenix, Arizona, USA) taken orally can provide a non-invasive method of smallbowel distension. Ten healthy volunteers (five women and five men), median age 32 (range 1849) years, median body-mass index 22 (range 1929) without any history of gastrointestinal disease or surgery, except appendectomy, participated in the study. Four doses, each of 5 mL 05 mol/L meglumine gadoterate (Dotarem, Guerbet, Roissy, France) in combination with 02 g per kg body-weight ispaghula dissolved in 250 mL water were administered orally every hour starting 4 h before MRI. Except for additional water, no food or drinks were allowed. Imaging was done with a 15 T unit (Philips Gyroscan, NT Intera, Best, Netherlands). Fat-saturated three dimensional gradient echo imaging with an isotropic resolution of 15 mm (repetition time/echotime 40/11 ms, flip-angle 25, 80 slices, matrix 256 256, field of view 400 mm, zerofilling) was done in apnoea over 30 s, with the patient in a prone position. Bowel motion was reduced by intravenous administration of 40 mg scopolamine butyl bromide (Buscopan, Boehringer Ingelheim, Germany).

Consensus assessment was done by following these criteria: semiquantitative measurement of the distension ranging from full (>20 mm), to intermediate (1020 mm), to slight (<10 mm) to collapsed regions, expressed as percentage of analysed small bowel length; qualitative classification of homogeneity of the intraluminal signal intensity ranging from five for completely homogeneous to one for inhomogeneous; demarcation of enhanced lumen against surrounding tissue ranging from five for clear demarcation to one for nondistinguishable; presence of artefacts ranging from one for no artefacts to five for severe image degradation. The mixture was well tolerated apart from slight abdominal discomfort and a sensation of being full reported by four volunteers. In eight volunteers, who had followed the protocol for ingestion and imaging exactly, contrast spread was excellent, beginning at the duodenum (two individuals) or proximal jejunum (six individuals), reaching the terminal ileum or beyond. In the remaining two the contrast column was carried further distally because of a 1 h delay. 32% (SD 27) of small bowels were fully distended (>20mm), 49% (25) had an intermediate width (1020 mm), 6% (4) showed slight distension (<10 mm), and 10% (2)mainly at bends were collapsed. Signal homogeneity was rated as 44 (06). In two individuals small structures with low signal intensity were seen over 15% and 18% of the contrast enhanced bowel length. Another volunteer had a 25 cm segmental signal drop because of excessive intake of water between the first and second ispaghula dose. Demarcation of the bowel content from extraluminal tissue was rated 5 (0) over the entire enhanced bowel length. Faint signal inhomogeneities in the two-phase encoding directions because of residual bowel motion did not impede image quality although rated 24 (08). Delayed imaging after administration of the spasmolytic agent led to pronounced motion artefacts in one volunteer. Sufficient luminal distension is crucial for diagnostic smallbowel imaging.2 Enteroclysis provides optimum distension and is well approbated in conventional imaging. The combination of enteroclysis and MRI is a promising technique but lacks acceptance by patients4 and the laborious radiographically controlled tube placement impedes its wider use. Earlier approaches to non-invasive MRI of the bowel,

Figure 1: Coronal T1 weighted three-dimensional gradient echo image of the abdomen after ingestion of ispaghula and meglumine gadoterate
Distension of the entire small bowel with homogeneous intraluminal signal enhancement and clear delineation to surrounding tissue is shown.

THE LANCET Vol 358 September 22, 2001

987

For personal use. Only reproduce with permission from The Lancet Publishing Group.