Abstract
To investigate the contributions of overall breastfeeding duration and exclusive breastfeeding in reducing the risk of hospitalisation for infectious causes, we analysed data from a three‐stage survey on infant feeding practices and health outcomes in over 10 000 UK women in 2010–2011. The main outcome measures were risk of overnight hospital admission in the first 8–10 months of infancy. A graded beneficial effect was found between longer duration of any breastfeeding and hospital admission for infectious causes and for respiratory tract infections, with a significantly lower risk in infants breastfed for at least 3 months compared with those never breastfed. The effects were stronger in the subgroup who was also exclusively breastfed. For example, among infants breastfed for 3–6 months, the reduction in risk for infectious causes for those who were also exclusively breastfed for at least 6 weeks was 0.42 (95% CI: 0.22–0.81) and for those not exclusively breastfed for 6 weeks 0.79 (95% CI: 0.49–1.26). Likewise, among infants breastfed for 6 months or more, the odds ratio for those who were also exclusively breastfed for at least 6 weeks was 0.48 (95% CI: 0.32–0.72) and for those not exclusively breastfed for 6 weeks 0.72 (95% CI: 0.48–1.08). The apparent protective effect of any breastfeeding for a long duration may in part be driven by a prolonged period of exclusive breastfeeding. Exclusive breastfeeding in the initial weeks after childbirth and continuing to breastfeed (either exclusively or partially) for at least 3 months, preferably 6 months, may reduce morbidity due to infectious illness in infants.
Keywords: breastfeeding, infant feeding, hospitalisation, respiratory tract infections, infections
Introduction
The benefits of breastfeeding, to infant and mother, are widely accepted and are supported by a comprehensive body of evidence. Numerous studies, both in developed and developing country settings, have shown positive effects of breastfeeding on rates of infectious illness in infants, in particular gastrointestinal infections (Quigley et al. 2007; Quigley et al. 2006; Fisk et al. 2011; Howie et al. 1990; Kramer et al. 2001; Ip et al. 2007; Baker et al. 1998; Monterrosa et al. 2008; Kramer et al. 2003; Ladomenou et al. 2010; Ajetunmobi et al. 2015) and respiratory infections (Quigley et al. 2007; Fisk et al. 2011; Ip et al. 2007; Ladomenou et al. 2010; Bachrach et al. 2003; Duijts et al. 2009; Tarrant et al. 2010; Chantry et al. 2006; Barsam et al. 2013; McNiel et al. 2010; Oddy et al. 2003), including significant cost benefits (Renfrew et al. 2012). Several studies have demonstrated statistically significant benefits of exclusive breastfeeding (EBF) (Kramer & Kakuma 2012; McNiel et al. 2010; Quigley et al. 2007; Paricio Talayero et al. 2006; Chantry et al. 2006; Kramer et al. 2003; Ladomenou et al. 2010; Rebhan et al. 2009); however, the evidence base is less conclusive regarding the effect of non‐EBF – with some recent studies showing a significant protective effect (Baker et al. 1998; Tarrant et al. 2010; Tarrant et al. 2012) and others showing no significant effect (Quigley et al. 2007; Ladomenou et al. 2010). However, few studies have attempted, or been large enough, to look at the separate effects of duration of any breastfeeding (defined as partial or EBF and referred to henceforth as ‘breastfeeding’) and duration of EBF (defined as infant receiving breast milk only) to understand the contributions of each of these elements to the potential benefit.
The World Health Organization (WHO) recommends that all infants should be exclusively breastfed for the first 6 months of life (WHO 2001). In the UK, only a very small proportion of women (1%) currently achieve this, with only 23% still exclusively breastfeeding at 6 weeks (McAndrew et al. 2012). In comparison, national surveys in Canada and Australia show that 52% and 56% of mothers respectively were still exclusively breastfeeding at 3 months (Chalmers et al. 2009; Baxter 2014).
While the goal of exclusively breastfeeding for 6 months remains desirable, with so few women in the UK exclusively breastfeeding for even 3 months, it is important to differentiate between the benefit of longer duration of EBF and longer duration of any breastfeeding. In this study, we aim to investigate the effect of breastfeeding on hospitalisation rates for infection among infants in the UK and to understand how the distinct components of ‘duration of breastfeeding’ and ‘duration of EBF’ contribute to any health benefits conferred.
Key messages.
In some developed countries, including the UK, few women achieve the World Health Organization recommendation of exclusive breastfeeding for 6 months. Exclusive breastfeeding is associated with reduced risk of gastroenteritis and respiratory infections in infants but the impact of partial breastfeeding in less clear.
Any breastfeeding for at least 3 months may reduce the risk of hospitalisation in infants and the population burden of hospital admission, but this risk reduction is greater with a prolonged period of initial exclusive breastfeeding
In settings where few women exclusively breastfeed for the recommended duration, shorter durations of breastfeeding may still confer significant health benefits.
Methods
Infant Feeding Survey
The Infant Feeding Survey (IFS) is a UK‐wide cross‐sectional survey conducted every five years. The survey was despatched (by post and online) in three stages, between September 2010 and August 2011, staggered so that the questionnaire reached mothers when their infants were around 6 weeks (stage 1), 4–6 months (stage 2) and 8–10 months old (stage 3). A sample of 30 760 births was selected from all births registered in the period August to October 2010. All births registered in Wales and Northern Ireland and a random sample of those registered in England and Scotland were selected, with oversampling of mothers from the lowest quintile for area deprivation to ensure adequate representation of these groups. Mothers were only sent the later stage questionnaire if they had responded to the previous stage.
The response rates at various stages of the survey were 51% at stage 1, 80% at stage 2 and 86% at stage 3. Overall, 68% of stage 1 responders completed all three stages of the survey. Data were weighted to correct for differential sampling, differential response rates and non‐response bias (McAndrew et al. 2012).
Exclusions
Analysis was limited to those who completed all three stages of the survey (n = 10 768). Exclusions were made for infants who were preterm (<37 weeks gestational age) (n = 577); 130 non‐singleton births; and those for whom data were missing, including where duration of breastfeeding could not be ascertained (n = 42), without 6 months of follow‐up (n = 18) or for whom a complete dataset for potential confounders was unavailable (n = 253). Imputation for missing data was not considered necessary as data were only missing for 3% of the total sample, which is well below the accepted threshold (Schafer 1999).
Breastfeeding
Duration of breastfeeding was derived from questions at all three stages. An infant was considered breastfed if it was given any breast milk, either exclusively or in addition to other milk, drinks or food. EBF was defined as breastfeeding with no other milk, drinks (including water) or solid food given, in line with WHO and UK guidance. We defined exclusivity at 6 months in accordance with the WHO definition, for example, if an infant was introduced to solids and/or formula at 26 weeks (or 180 days) or later, it was considered as breastfed exclusively for 6 months (WHO 2015). This is in contrast to the IFS report which only defines an infant as exclusively breastfed for 6 months if solids and/or formula/drinks were introduced after 6 months (McAndrew et al. 2012), meaning that an exclusively breastfed infant introduced to solids at 26 weeks was not considered exclusively breastfed for 6 months. This results in slightly more women being categorised as ‘EBF at 6 months’ in our analysis than the IFS report (5% vs. 1%). Where data were collected at multiple stages of the survey, the earliest completed data were used. Responses to the same question at subsequent stages were used to fill in for non‐response in earlier stages to specific questions. Duration of breastfeeding and EBF were measured in days or weeks – at later stages of the survey, data were only collected in multiples of weeks. Data was therefore analysed by duration category.
Hospital admission
Mothers were asked at both stages 2 (infants aged 4–6 months) and 3 (infants aged 8–10 months) if their infant had ever been hospitalised overnight and if so, what the cause was. For analysis, causes were categorised to derive the outcome measures ‘hospitalised for infectious causes’ (IC) and the subset ‘hospitalised for respiratory tract infections (RTI)’ (including both upper and lower respiratory tract infection) (Table 1). Data on the number of hospital admissions and infant's age at time of hospitalisation were not captured.
Table 1.
Definition of outcome measures*
| Hospitalisation due to infectious causes (IC) | Hospitalisation due to respiratory tract infection (upper or lower) (RTI) |
|---|---|
| Chest problems/infection | Chest problems/infection |
| Bronchitis/bronchiolitis | Bronchitis/bronchiolitis |
| Croup | Croup |
| Sore throat/throat infection | Sore throat/throat infection |
| A cold | A cold |
| A cough | A cough |
| Pneumonia | Pneumonia |
| Respiratory syncytial virus | Respiratory syncytial virus |
| Sickness or vomiting | |
| Diarrhoea | |
| Tummy bug/gastroenteritis | |
| Ear problems/infection | |
| Fever | |
| Viral infection | |
| Bacterial infection |
Other causes of hospitalisation listed but not included in either outcome measure: constipation, urinary tract infection, colic, thrush, lack of weight gain, too much weight gain, eczema, rash, haemangioma, milk allergy, jaundice, reflux, pyloric stenosis, hernia, congestion, whooping cough, asthma, breathing difficulties, conjunctivitis, meningitis, heart problems, kidney problems, infected umbilicus, chickenpox, measles, weight loss, blood in excrement, reaction to immunisation, teething, joint/bone problems, ‘something else’.
Statistical analysis
Statistical analyses were conducted using STATA version 13.1 (Statacorp 2013). Multivariable logistic regression was used to determine associations between duration of breastfeeding (and EBF) and each outcome measure (in separate analyses). Data were collected for a number of variables, which are potentially associated with breastfeeding and/or the outcomes (Table 2 and associated footnote), and these variables were tested individually for association with the outcomes; with only those demonstrating a significant association (Wald P‐value <0.10) included in the multivariable model. All means, proportions and odds ratio (OR) presented are weighted. Population attributable fractions (PAFs) were estimated for hospitalisation associated with breastfeeding for less than 6 weeks (compared with 3–6 months) and for breastfeeding less than 3 months (compared with breastfeeding for at least 6 months with EBF at least 6 weeks).
Table 2.
Characteristics of the sample according to breastfeeding duration†
| BF duration | |||||||
|---|---|---|---|---|---|---|---|
| Never BF (n = 1874) | <1 week (n = 1083) | 1–6 weeks (n = 1303) | 6 weeks to 3 months (n = 873) | 3–6 months (n = 1301) | ≥6 months (n = 3379) | ||
| n, (%) | n, (%) | n, (%) | n, (%) | n, (%) | n, (%) | ||
| Maternal characteristics* | |||||||
| Mother's age | Under 20 years (n = 241) | 119 (49) | 44 (18) | 36 (15) | 11 (5) | 10 (4) | 21 (9) |
| 20–24 years (n = 1075) | 372 (35) | 164 (15) | 177 (16) | 101 (9) | 86 (8) | 175 (16) | |
| 25–29 years (n = 2611) | 521 (20) | 328 (13) | 400 (15) | 238 (9) | 328 (13) | 796 (30) | |
| 30–34 years (n = 3489) | 520 (15) | 337 (10) | 418 (12) | 293 (8) | 542 (16) | 1379 (40) | |
| 35 years or over (n = 2383) | 339 (14) | 208 (9) | 270 (11) | 227 (10) | 334 (14) | 1005 (42) | |
| Age left full‐time education | 16 years or under (n = 1342) | 517 (39) | 197 (15) | 175 (13) | 101 (8) | 131 (10) | 221 (16) |
| 17 or 18 years (n = 2641) | 715 (27) | 389 (15) | 419 (16) | 237 (9) | 285 (11) | 596 (23) | |
| Over 18 years (n = 5830) | 642 (11) | 497 (9) | 709 (12) | 535 (9) | 885 (15) | 2562 (44) | |
| SEC | Managerial and professional (n = 4366) | 453 (10) | 380 (9) | 550 (13) | 407 (9) | 718 (16) | 1858 (43) |
| Intermediate occupations (n = 2066) | 457 (22) | 279 (14) | 299 (14) | 184 (9) | 237 (11) | 610 (30) | |
| Routine and manual occupations (n = 2200) | 645 (29) | 304 (14) | 329 (15) | 179 (8) | 207 (9) | 536 (24) | |
| Never worked (n = 473) | 177 (37) | 43 (9) | 37 (8) | 29 (6) | 54 (11) | 133 (28) | |
| Not classified (n = 708) | 142 (20) | 77 (11) | 88 (12) | 74 (10) | 85 (12) | 242 (34) | |
| Smoking status (pre and during pregnancy) | Never smoked or gave up >1 year before pregnancy (n = 7553) | 1187 (16) | 767 (10) | 947 (13) | 660 (9) | 1067 (14) | 2925 (39) |
| Gave up during pregnancy or within before getting pregnant (n = 1234) | 294 (24) | 174 (14) | 188 (15) | 129 (10) | 156 (13) | 293 (24) | |
| Continued to smoke during pregnancy (n = 847) | 357 (42) | 120 (14) | 143 (17) | 66 (8) | 50 (6) | 111 (13) | |
| Missing (n = 179) | 36 (20) | 22 (12) | 25 (14) | 18 (10) | 28 (16) | 50 (28) | |
| Ethnicity | White (n = 7195) | 1289 (18) | 782 (11) | 1003 (14) | 638 (9) | 952 (13) | 2531 (35) |
| Mixed (n = 93) | 13 (14) | 3 (3) | 6 (6) | 10 (11) | 15 (16) | 46 (49) | |
| Asian or Asian British (n = 344) | 12 (3) | 16 (5) | 31 (9) | 37 (11) | 67 (19) | 181 (53) | |
| Black or Black British (n = 188) | 7 (4) | 0 (0) | 8 (4) | 11 (6) | 39 (21) | 123 (65) | |
| Chinese or other (n = 78) | 4 (5) | 3 (4) | 5 (6) | 7 (9) | 10 (13) | 49 (63) | |
| Missing (n = 1915) | 549 (29) | 279 (15) | 250 (13) | 170 (9) | 218 (11) | 449 (23) | |
| Deprivation (IMD) | 1 (most deprived) (n = 1839) | 536 (28) | 232 (13) | 214 (12) | 133 (7) | 176 (10) | 548 (30) |
| 2 (n = 1872) | 416 (22) | 228 (12) | 258 (14) | 172 (9) | 237 (13) | 561 (30) | |
| 3 (n = 2025) | 391 (19) | 236 (12) | 289 (14) | 182 (9) | 252 (13) | 675 (33) | |
| 4 (n = 2041) | 299 (14) | 204 (10) | 277 (14) | 178 (9) | 320 (16) | 763 (37) | |
| 5 (least deprived) (n = 2029) | 232 (11) | 183 (9) | 263 (13) | 207 (10) | 315 (16) | 829 (41) | |
| Country | England (n = 4471) | 548 (12) | 375 (8) | 622 (14) | 420 (9) | 667 (15) | 1839 (41) |
| Scotland (n = 1931) | 404 (21) | 204 (11) | 211 (11) | 151 (8) | 250 (13) | 711 (37) | |
| Wales (n = 1678) | 389 (23) | 234 (14) | 237 (14) | 150 (9) | 191 (11) | 477 (28) | |
| Northern Ireland (n = 1733) | 533 (31) | 270 (16) | 233 (13) | 152 (9) | 193 (11) | 352 (20) | |
| Infant/birth characteristics | |||||||
| Gestational age (categories) | 37–39 weeks (n = 3827) | 804 (21) | 414 (11) | 543 (14) | 364 (10) | 478 (12) | 1224 (32) |
| 40–42 weeks (n = 5986) | 1070 (18) | 669 (11) | 760 (13) | 509 (9) | 823 (14) | 2155 (36) | |
| Birth weight | <2.5 kg (n = 177) | 38 (21) | 15 (8) | 27 (15) | 16 (9) | 23 (13) | 58 (33) |
| 2.5–2.99 kg (n = 1188) | 256 (22) | 122 (10) | 170 (14) | 107 (9) | 144 (12) | 389 (33) | |
| 3–3.49 kg (n = 3534) | 695 (20) | 378 (11) | 472 (13) | 316 (9) | 474 (13) | 1199 (34) | |
| 3.5 kg+ (n = 4853) | 882 (18) | 561 (12) | 629 (13) | 428 (9) | 650 (13) | 1703 (35) | |
| Birth order | First born (n = 4800) | 714 (15) | 586 (12) | 778 (16) | 452 (9) | 645 (13) | 1625 (34) |
| Not first born (n = 5013) | 1160 (23) | 497 (10) | 525 (10) | 421 (8) | 656 (13) | 1754 (35) | |
| Special care | Special care or lamp required (n = 669) | 103 (15) | 79 (12) | 115 (17) | 67 (10) | 77 (12) | 228 (34) |
| No special care or lamp (n = 9144) | 1771 (19) | 1004 (11) | 1188 (13) | 806 (9) | 1224 (13) | 3151 (34) | |
| Mode of delivery | Normal vaginal (n = 5863) | 1131 (19) | 602 (10) | 722 (12) | 500 (9) | 784 (13) | 2124 (36) |
| Delivery with ventouse (n = 618) | 94 (15) | 64 (10) | 86 (14) | 60 (10) | 91 (15) | 223 (36) | |
| Delivery with forceps (n = 863) | 142 (16) | 116 (13) | 125 (14) | 82 (10) | 122 (14) | 276 (32) | |
| Caesarean (n = 2456) | 506 (21) | 299 (12) | 369 (15) | 231 (9) | 301 (12) | 750 (31) | |
| Childcare setting (stage 2) | None (n = 9423) | 1788 (19) | 1039 (11) | 1244 (13) | 826 (9) | 1239 (13) | 3287 (35) |
| Crèche or nursery (n = 53) | 10 (19) | 4 (8) | 9 (17) | 6 (11) | 8 (15) | 16 (30) | |
| Nanny or childminder (n = 46) | 8 (17) | 2 (4) | 5 (11) | 9 (20) | 11 (24) | 11 (24) | |
| Friend or relative (n = 282) | 68 (24) | 38 (13) | 45 (16) | 28 (10) | 41 (15) | 62 (22) | |
SEC, Socio‐economic classification. IMD, Index of Multiple Deprivation. BF, breastfeeding.
Additional variables considered as potential confounders: smoking status post birth; skin‐to‐skin contact post birth.
Missing values shown only where total missing values is greater than 1%.
This study used anonymised, publically available data from the UK Data Archive, and therefore, ethical approval was not required.
Results
Breastfeeding
The average age of the infants at stage 3 was 260 days (SD: 25; range: 180–392). Of the 9813 infants included in the analysis, 81% (7939) were breastfed at least once, 11% (1083) were breastfed for less than one week and 34% for at least 6 months (Fig. 1a). Just over a quarter of infants (28%) were still breastfed exclusively at 6 weeks, a fifth at 3 months and only 5% at 6 months (Fig. 1b). Of those infants who were breastfed for between 6 weeks and 3 months, 84% were EBF for less than 6 weeks. Even among infants who were breastfed for longer durations, many were EBF for less than 6 weeks – of those who were breastfed for 3–6 months, almost two‐thirds were EBF for less than 6 weeks, and among those who were breastfed for at least 6 months, over one‐third were EBF for less than 6 weeks.
Figure 1.
(a) Breastfeeding duration and hospitalisation. Percentage of infants breastfed and percentage hospitalised by breastfeeding duration. (b) Exclusive breastfeeding duration and hospitalisation. Percentage of infants exclusively breastfed and percentage hospitalised by exclusive breastfeeding duration.
Duration of any breastfeeding varied greatly according to maternal characteristics. Infants whose mothers were less than 20 years old, who left full‐time education before 16 and who had never worked were more likely to have never breastfed (Table 2).
Hospital admission
Eleven percent of infants (1104) had an overnight hospital stay over the reporting period. The most common individual cause of hospitalisation was ‘chest problems/infection’ for which 360 infants were hospitalised. Other common causes were ‘sickness or vomiting’ (n = 206), ‘insufficient weight gain’ (n = 155), ‘diarrhoea’ (n = 103) and ‘bronchitis/bronchiolitis’ (n = 79).
Six hundred and sixty‐four infants (7%) were hospitalised for infectious causes, with 410 of these being for upper or lower RTI (or symptoms of). Factors that were significantly associated with hospital admission in univariable analysis comprised birth order, gestational age, mother's level of education and whether special care or a lamp for jaundice were required post birth.
Association between breastfeeding and hospital admission
A graded beneficial effect was found between longer duration of breastfeeding and both the outcome measures of hospital admission for IC and for admission for RTI, before and after adjustment for confounders (Table 3). For IC, breastfeeding for at least 6 months was significantly protective (OR: 0.58, 95% CI: 0.41–0.82) compared with never breastfeeding. Further analysis showed that the significant protective effect was maintained and of greater magnitude for those who breastfed for at least 6 months and also EBF for at least 6 weeks (OR: 0.48 95% CI: 0.32–0.72), but not for those who EBF for less than 6 weeks (OR: 0.72, 95% CI: 0.48–1.08) (Table 3, Fig. 2). Breastfeeding for 3–6 months was also significantly associated with a lower risk of hospitalisation for IC compared with never breastfeeding (OR: 0.64, 95% CI: 0.42–0.99); again, this effect was stronger in those who were EBF for at least 6 weeks (OR: 0.42, 95% CI: 0.22–0.81) than in those who were not EBF for at least 6 weeks (OR: 0.79, 95% CI: 0.49–1.26). The same patterns were observed for RTI (Table 3, Fig. 2).
Table 3.
Associations between breastfeeding duration with hospitalisation for ‘all infectious causes’ and for ‘respiratory tract infections’, with unadjusted and adjusted* odds ratios
| Hospitalised (%) | IC (UNADJ. OR, 95% CI) | IC (ADJ. OR, 95% CI) | Hospitalised (%) | RTI (UNADJ. OR, 95% CI) | RTI (ADJ. OR, 95% CI) | |
|---|---|---|---|---|---|---|
| Breastfeeding duration | ||||||
| Never breastfed (n = 1874) | 8 | 1 | 1 | 5 | 1 | 1 |
| 1.1 (0.71–1.7) | 1.16 (0.75–1.8) | 0.81 (0.48–1.36) | 0.89 (0.53–1.5) | |||
| <1 week (n = 1083) | 9 | P = 0.664 | P = 0.504 | 6 | P = 0.425 | P = 0.649 |
| 0.89 (0.6–1.31) | 0.92 (0.62–1.38) | 0.83 (0.51–1.36) | 0.93 (0.57–1.54) | |||
| 1–6 weeks (n = 1303) | 8 | P = 0.551 | P = 0.693 | 5 | P = 0.46 | P = 0.779 |
| 0.79 (0.51–1.24) | 0.79 (0.5–1.24) | 0.69 (0.39–1.2) | 0.74 (0.42–1.31) | |||
| 6 weeks to 3 months (n = 873) | 7 | P = 0.304 | P = 0.301 | 4 | P = 0.188 | P = 0.306 |
| 0.65 (0.43–1.0) | 0.64 (0.42–0.99) | 0.57 (0.34–0.95) | 0.59 (0.36–1.0) | |||
| 3–6 months (n = 1301) | 5 | P = 0.048 | P = 0.044 | 3 | P = 0.033 | P = 0.048 |
| 0.81 (0.51–1.29) | 0.79 (0.49–1.26) | 0.61 (0.35–1.09) | 0.63 (0.36–1.13) | |||
| <6 weeks EBF (n = 777) | 6 | P = 0.374 | P = 0.313 | 4 | P = 0.094 | P = 0.121 |
| 0.41 (0.21–0.80) | 0.42 (0.22–0.81) | 0.50 (0.24–1.06) | 0.53 (0.25–1.13) | |||
| ≥6 weeks EBF (n = 524) | 3 | P = 0.009 | P = 0.01 | 2 | P = 0.07 | P = 0.099 |
| 0.59 (0.42–0.84) | 0.58 (0.41–0.82) | 0.45 (0.29–0.71) | 0.47 (0.3–0.73) | |||
| 6 months or more (n = 3379) | 5 | P = 0.003 | P = 0.002 | 3 | P = 0.001 | P = 0.001 |
| 0.75 (0.5–1.13) | 0.72 (0.48–1.08) | 0.58 (0.34–0.98) | 0.59 (0.35–1.01) | |||
| <6 weeks EBF (n = 1295) | 7 | P = 0.174 | P = 0.11 | 4 | P = 0.042 | P = 0.053 |
| 0.48 (0.33–0.72) | 0.48 (0.32–0.72) | 0.37 (0.22–0.61) | 0.38 (0.23–0.63) | |||
| ≥6 weeks EBF (n = 2084) | 5 | P < 0.001 | P < 0.001 | 3 | P < 0.001 | P < 0.001 |
CI, confidence interval; UNADJ. OR, unadjusted odds ratio; ADJ. OR, adjusted odds ratio; RTI, respiratory tract infections; EBF, exclusive breastfeeding.
Adjusted for birth order, special care after delivery, age mother left full‐time education, gestational age and infant age at survey stage 3.
Figure 2.
Odds ratio for hospitalisation compared with ‘never breastfed’ by breastfeeding duration.
Population attributable fractions
We estimated that 31% of overnight hospital admissions for IC were attributable to breastfeeding for less than 3 months rather than breastfeeding for at least 6 months including at least 6 weeks of EBF. Furthermore, 19% of admissions for IC were attributable to breastfeeding for less than 6 weeks or not at all, instead of breastfeeding for 3–6 months (Table 4).
Table 4.
Adjusted OR and PAF for breastfeeding duration and hospitalisation for infectious causes
| Duration | % of infants breastfed | % of cases breastfed | Hospitalisation adjusted OR* (95% CI) | P‐value | PAF† | Hospitalisation adjusted OR* (95% CI) | P‐value | PAF† |
|---|---|---|---|---|---|---|---|---|
| Never breastfed | 19 | 22 | 2.07 (1.38–3.11) | <0.001 | 11%a | 1.55 (1.01–2.37) | 0.046 | 8%d |
| BF <6 weeks | 24 | 30 | 2.12 (1.48–3.04) | <0.001 | 16%b | 1.58 (1.09–2.31) | 0.017 | 11%e |
| BF 6 weeks to 3 months | 9 | 10 | 1.64 (1.05–2.57) | 0.029 | 4%c | 1.23 (0.77–1.95) | 0.385 | — |
| BF 3–6 months | 13 | 10 | 1.34 (0.88–2.03) | 0.167 | — | 1 | — | reference |
| BF ≥6 months (with EBF <6 weeks) | 13 | 13 | 1.49 (0.99–2.23) | 0.053 | — | 1.11 (0.73–1.70) | 0.62 | — |
| BF ≥6 months (with EBF ≥6 weeks) | 21 | 15 | 1 | — | reference | 0.75 (0.49–1.13) | 0.167 | — |
OR, odds ratio; PFA, population attributable fraction; CI, confidence interval; BF, breastfeeding; EBF, exclusive breastfeeding.
Models from Table 2 have been refitted with different reference groups.
Overall PAF for breastfeeding less than 3 months (a + b + c) = 31%; overall PAF for breastfeeding less than 6 weeks (d + e) = 19%. PAFs only calculated where the OR is significant (P‐value <0.05).
PAF = [0.22 × (1.07/2.07)] =11%.
PAF = [0.3 × (1.12/2.12)] =16%.
PAF = [0.1 × (0.64/1.64)] =4%.
PAF = [0.22 × (0.55/1.55)] =8%.
PAF = [0.3 × (0.58/1.58)] =11%.
Discussion
The main finding of this study is that any breastfeeding for at least 3 months is protective against hospital admission for IC and particularly RTI, but the magnitude of this effect is greater in infants who were also exclusively breastfed for a prolonged period (≥6 weeks). Seven percent of infants were hospitalised for IC during the study, and about one‐third of these hospital admissions were attributable to breastfeeding for less than 3 months rather than breastfeeding for at least 6 months including at least 6 weeks of EBF.
We estimated that 31% of overnight hospital admissions for IC could be avoided if those mothers who never breastfed or breastfed for less than 3 months continued to breastfeed for at least 6 months including at least 6 weeks exclusively and that 19% of admissions could have been avoided if mothers who breastfed less than 6 weeks continued to breastfeed for just 3–6 months. Infections, in particular respiratory tract infections, are among the most common causes of hospitalisation in infants in the UK (DH 2013; Hull et al. 2000). Here, we have demonstrated that relatively small incremental increases in breastfeeding duration can have a substantial impact on the population level burden of hospitalisation because of infection.
The mechanism by which breastfeeding reduces risk of infection in infants is thought to be both through the active immunological protection from breast milk itself and from avoidance of contamination introduced through alternative feeding (Quigley et al. 2007; Fisk et al. 2011), presenting a biological argument that an extended period of breastfeeding is likely to provide a longer period of protection against infection. Further strengthening this argument, a previous study in the UK showed the protective effect of breastfeeding wears off soon after cessation of breastfeeding (Quigley et al. 2007). However, it is important to understand that a long duration of breastfeeding on its own may not necessarily provide protection without a prolonged period of EBF.
These results are consistent with findings from earlier studies, which have shown that partial BF, that is, non‐EBF may be beneficial in reducing infectious morbidity in infants. Earlier studies in Scotland (Howie et al. 1990) and England (Baker et al. 1998) concluded that benefit was conferred from any breastfeeding for at least 3 months, and a recent Irish study found reduced risk of incidence of gastrointestinal illness from as little as 6 weeks of any breastfeeding (Tarrant et al. 2012), albeit with small numbers and low rates of breastfeeding. Some authors have concluded that partial breastfeeding provides no significant protection against gastrointestinal or respiratory infections (Quigley et al. 2007; Quigley et al. 2006; Ladomenou et al. 2010); however in these studies, the total duration of breastfeeding among those exclusively or partially breastfeeding was not investigated. It has been demonstrated elsewhere that EBF is associated with a longer duration of breastfeeding overall (Feinstein et al. 1986; Hill et al. 1997); therefore, it is possible that a lack of protection demonstrated from partial breastfeeding may be due to shorter overall duration of breastfeeding among these mothers. There is potential that the lack of exclusion of EBF in the group defined as ‘any breastfeeding’ in our analysis – in contrast to the studies by Quigley et al. (2007) and Ladomenou et al. (2010) where infants were categorised as either EBF or partially breastfed – may have biased our results away from the null, indicating a significant association between long duration of any breastfeeding and hospitalisation; however, we know that only a small proportion of mothers EBF for longer than 3 months, so any impact would be minimal. Only Duijts et al. (2010) investigated the separate contribution of EBF, concluding that EBF for 4 months and partial thereafter was significantly protective against incidence of RTIs but that breastfeeding up to 6 months with no EBF or EBF for 4 months then no continued breastfeeding was not beneficial – a similar pattern to our findings, although with a smaller sample size.
Our definition of ‘any breastfeeding’ whereby an infant is defined as ‘breastfed’ if it receives any breast milk at all categorises infants who are exclusively breastfed, predominantly breastfed but not exclusively and those who receive predominantly formula and only a small amount of breast milk, together in one group. Although we know that only a small proportion is still exclusively breastfed beyond 3 months, we are otherwise unable to determine the quantity or frequency of breastfeeding among this group. However, research indicates that frequent breastfeeding (even if supplemented with formula) and predominant breastfeeding are both correlated with increased duration of breastfeeding (Feinstein et al. 1986), and physiologically, frequent suckling is a key driver in milk production (Jonas & Woodside 2015). Future research in this area should attempt to robustly quantify mixed feeding patterns to better understand the potential benefits from non‐EBF.
The strengths of this study are the large sample size, allowing for identification of large and significant effects even with a relatively uncommon outcome and that the primary outcome measure of overnight hospitalisation is both well defined and an indicator of severity of illness, therefore being more robust to recall bias. There are inherent weaknesses in the observational study design, notably misclassification bias, non‐response bias and potential residual confounding. Although this survey questionnaire was not validated, studies have shown maternal reporting of breastfeeding duration to be reliable (Li et al. 2005; Herrmann et al. 2011) and parental recall of hospitalisation in infants to be highly accurate (Spencer & Coe 2000; D'Souza‐Vazirani et al. 2005). The frequency of the surveys, requiring only short‐term recall at each stage, reduces the risk of recall error. Furthermore, hospitalisation rates in our study are comparable with those found elsewhere, suggesting limited impact of misclassification (Quigley et al. 2007; Ladomenou et al. 2010). There was an indication of slight under‐representation in the sample of younger mothers and mothers living in the most deprived areas; therefore, survey data were weighted for differential response rates among different groups of the population to mitigate non‐response bias and for potential selection bias from the sampling methodology (McAndrew et al. 2012).
It is well established that multiple physiological, psychological and environmental factors impact on both infant feeding patterns and infant illness. Our results were adjusted for several variables known to have an association with breastfeeding and infant morbidity; however, residual confounding, for example, maternal behavioural characteristics intrinsic in the motivation to breastfeed, cannot be ruled out. Moreover, because of the lack of data on the timing of hospitalisation and the impact of hospitalisation on infant feeding, reverse causality cannot be entirely ruled out. There is little published evidence of the impact of infant hospitalisation on the continuation of breastfeeding; therefore, it is difficult to estimate the impact this may have had on our results. UK guidelines published in 2009 by the Royal College of Nursing (Royal College of Nursing 2013) encourage continuation of breastfeeding for hospitalised infants and support for mothers to do so; therefore, there is no indication that hospitalisation itself should necessarily lead to interruption or termination of breastfeeding. However, it is reasonable to assume that in some instances, hospitalisation, in particular for a prolonged period, may lead to interruption of breastfeeding or loss of EBF, and this may have biased our odds ratios away from null. Future studies of this nature should attempt to collect data on timing of hospitalisation to mitigate this issue.
Our results are widely generalisable to other industrialised countries. The low number of infants in our sample who were exclusively breastfed for at least 6 months, reflective of UK rates described elsewhere,(Quigley et al. 2007; McAndrew et al. 2012) are not unusual and have been reported in other developed countries (Duijts et al. 2010), therefore ought not to limit generalisability. The findings of this study further add to the evidence base by examining the contributions of overall breastfeeding duration and EBF to the reduction in morbidity for infectious diseases in infancy, in a large population‐based sample.
Very few women in the UK reach the recommended level of EBF, for a variety of reasons, both structural and personal. Although 6 months of EBF may provide optimal benefit to mother and baby, we have demonstrated that significant benefits may be attainable from an extended period – at least 3 months – of breastfeeding, not necessarily exclusively, when combined with a short initial duration of EBF (at least 6 weeks). From this analysis, however, we cannot quantify how much breast milk is the minimum requirement. While we continue to strive to increase rates of EBF, parents, front‐line staff and health service providers need to be aware of the opportunity for benefits from non‐EBF but that EBF in the early weeks is likely a key factor in realising these benefits and ought to be strongly encouraged and supported. Significant morbidity, and its associated costs, may potentially be avoided with a population‐level increase in exclusive and prolonged breastfeeding among mothers in the UK, and further research is warranted in understanding effective interventions to help mothers breastfeed for longer and in particular to exclusively breastfeed for a prolonged initial period.
Conclusions
Breastfeeding for at least 3 months, with a substantial component of initial EBF may reduce hospital admissions for infectious illness in infants in a developed country. Large and significant public health benefit can be achieved by encouraging more mothers to breastfeed, and exclusively breastfeed, beyond the first weeks after childbirth.
Source of funding
None.
Conflicts of interest
The authors declare that they have no conflicts of interest.
Contributions
SP contributed to the conception and design of the study, undertook the analysis and drafted the paper. MQ contributed to the conception and design of the study, contributed to the interpretation of the data and critically reviewed the paper. Both authors have approved the final version to be published.
Acknowledgements
We acknowledge the work of the Health and Social Care Information Centre and thank all the parents who completed the Infant Feeding Survey. We would also like to thank Claire Carson for helpful comments on an earlier draft.
Payne S., and Quigley M. A. (2017) Breastfeeding and infant hospitalisation: analysis of the UK 2010 Infant Feeding Survey, Maternal & Child Nutrition, 13, e12263. doi: 10.1111/mcn.12263.
References
- Ajetunmobi O.M., Whyte B., Chalmers J., Tappin D.M., Wolfson L., Fleming M. et al. (2015) Breastfeeding is associated with reduced childhood hospitalization: evidence from a Scottish birth cohort (1997–2009). The Journal of Pediatrics 166 (3), 620–625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bachrach V.R., Schwarz E. & Bachrach L.R. (2003) Breastfeeding and the risk of hospitalization for respiratory disease in infancy: a meta‐analysis. Archives of Pediatrics & Adolescent Medicine 157 (3), 237–243. [DOI] [PubMed] [Google Scholar]
- Baker D., Taylor H. & Henderson J. (1998) Inequality in infant morbidity: causes and consequences in England in the 1990s. ALSPAC study team. Avon Longitudinal Study of Pregnancy and Childhood. Journal of Epidemiology and Community Health 52 (7), 451–458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Barsam F.J., Borges G.S., Severino A.B., de Mello L.M., da Silva A.S. & Nunes A.A. (2013) Factors associated with community‐acquired pneumonia in hospitalised children and adolescents aged 6 months to 13 years old. European Journal of Pediatrics 172 (4), 493–499. [DOI] [PubMed] [Google Scholar]
- Baxter J. (2014) Growing up in Australia: the longitudinal study of Australian children annual report 2006–2007 . Growing Up in Australia. Available from: http://www.growingupinaustralia.gov.au/pubs/ar/ar200607/breastfeeding (Accessed 18 August 2014).
- Chalmers B., Levitt C., Heaman M., O'Brien B., Sauve R., Kaczorowski J. et al. (2009) Breastfeeding rates and hospital breastfeeding practices in Canada: a national survey of women. Birth (Berkeley, Calif.) 36 (2), 122–132. [DOI] [PubMed] [Google Scholar]
- Chantry C.J., Howard C.R. & Auinger P. (2006) Full breastfeeding duration and associated decrease in respiratory tract infection in US children. Pediatrics 117 (2), 425–432. [DOI] [PubMed] [Google Scholar]
- DH (2013) Infant feeding profiles 2010–2011. Department of Health. Available from: https://www.gov.uk/government/publications/infant-feeding-profiles-2010-to-2011 (Accessed 15 September 2014).
- D'Souza‐Vazirani D., Minkovitz C.S. & Strobino D.M. (2005) Validity of maternal report of acute health care use for children younger than 3 years. Archives of Pediatrics & Adolescent Medicine 159 (2), 167–172. [DOI] [PubMed] [Google Scholar]
- Duijts L., Jaddoe V.W., Hofman A. & Moll H.A. (2010) Prolonged and exclusive breastfeeding reduces the risk of infectious diseases in infancy. Pediatrics 126 (1), e18–e25. [DOI] [PubMed] [Google Scholar]
- Duijts L., Ramadhani M.K. & Moll H.A. (2009) Breastfeeding protects against infectious diseases during infancy in industrialized countries. A systematic review. Maternal & Child Nutrition 5 (3), 199–210. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Feinstein J.M., Berkelhamer J.E., Gruszka M.E., Wong C.A. & Carey A.E. (1986) Factors related to early termination of breast‐feeding in an urban population. Pediatrics 78 (2), 210–215. [PubMed] [Google Scholar]
- Fisk C.M., Crozier S.R., Inskip H.M., Godfrey K.M., Cooper C., Roberts G.C. et al. (2011) Breastfeeding and reported morbidity during infancy: findings from the Southampton Women's Survey. Maternal & Child Nutrition 7 (1), 61–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Herrmann D., Suling M., Reisch L., Siani A., De Bourdeaudhuij I., Maes L. et al. (2011) Repeatability of maternal report on prenatal, perinatal and early postnatal factors: findings from the IDEFICS parental questionnaire. International Journal of Obesity (2005) 35 (Suppl 1), S52–S60. [DOI] [PubMed] [Google Scholar]
- Hill P.D., Humenick S.S., Brennan M.L. & Woolley D. (1997) Does early supplementation affect long‐term breastfeeding? Clinical Pediatrics 36 (6), 345–350. [DOI] [PubMed] [Google Scholar]
- Howie P.W., Forsyth J.S., Ogston S.A., Clark A. & Florey C.D. (1990) Protective effect of breast feeding against infection. BMJ (Clinical research ed.) 300 (6716), 11–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Hull J., Thomson A. & Kwiatkowski D. (2000) Association of respiratory syncytial virus bronchiolitis with the interleukin 8 gene region in UK families. Thorax 55 (12), 1023–1027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ip S., Chung M., Raman G., Chew P., Magula N., DeVine D. et al (2007) Breastfeeding and Maternal and Infant Health Outcomes in Developed Countries. AHRQ: United States. [PMC free article] [PubMed] [Google Scholar]
- Jonas W. & Woodside B. (2015) Physiological mechanisms, behavioral and psychological factors influencing the transfer of milk from mothers to their young. Hormones and Behavior. DOI: 10.1016/j.yhbeh.2015.07.018. [DOI] [PubMed] [Google Scholar]
- Kramer M.S., Chalmers B., Hodnett E.D., Sevkovskaya Z., Dzikovich I., Shapiro S. et al. (2001) Promotion of Breastfeeding Intervention Trial (PROBIT): a randomized trial in the Republic of Belarus. JAMA, the Journal of the American Medical Association 285 (4), 413–420. [DOI] [PubMed] [Google Scholar]
- Kramer M.S., Guo T., Platt R.W., Sevkovskaya Z., Dzikovich I., Collet J.P. et al. (2003) Infant growth and health outcomes associated with 3 compared with 6 mo of exclusive breastfeeding. The American Journal of Clinical Nutrition 78 (2), 291–295. [DOI] [PubMed] [Google Scholar]
- Kramer M.S. & Kakuma R. (2012) Optimal duration of exclusive breastfeeding. The Cochrane Database of Systematic Reviews 8 CD003517. DOI: 10.1002/14651858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Ladomenou F., Moschandreas J., Kafatos A., Tselentis Y. & Galanakis E. (2010) Protective effect of exclusive breastfeeding against infections during infancy: a prospective study. Archives of Disease in Childhood 95 (12), 1004–1008. [DOI] [PubMed] [Google Scholar]
- Li R., Scanlon K.S. & Serdula M.K. (2005) The validity and reliability of maternal recall of breastfeeding practice. Nutrition Reviews 63 (4), 103–110. [DOI] [PubMed] [Google Scholar]
- McAndrew F., Thompson J., Fellows L., Large A., Speed M. & Renfrew M.J. (2012) Infant Feeding Survey 2010. Health and Social care Information Centre: UK. [Google Scholar]
- McNiel M.E., Labbok M.H. & Abrahams S.W. (2010) What are the risks associated with formula feeding? A re‐analysis and review. Birth (Berkeley, Calif.) 37 (1), 50–58. [DOI] [PubMed] [Google Scholar]
- Monterrosa E.C., Frongillo E.A., Vasquez‐Garibay E.M., Romero‐Velarde E., Casey L.M. & Willows N.D. (2008) Predominant breast‐feeding from birth to six months is associated with fewer gastrointestinal infections and increased risk for iron deficiency among infants. The Journal of Nutrition 138 (8), 1499–1504. [DOI] [PubMed] [Google Scholar]
- Oddy W.H., Sly P.D., de Klerk N.H., Landau L.I., Kendall G.E., Holt P.G. et al. (2003) Breast feeding and respiratory morbidity in infancy: a birth cohort study. Archives of Disease in Childhood 88 (3), 224–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Paricio Talayero J.M., Lizan‐Garcia M., Otero Puime A., Benlloch Muncharaz M.J., Beseler Soto B., Sanchez‐Palomares M. et al. (2006) Full breastfeeding and hospitalization as a result of infections in the first year of life. Pediatrics 118, e93–e99. [DOI] [PubMed] [Google Scholar]
- Quigley M.A., Kelly Y.J. & Sacker A. (2007) Breastfeeding and hospitalization for diarrheal and respiratory infection in the United Kingdom Millennium Cohort Study. Pediatrics 119 (4), e837–e842. [DOI] [PubMed] [Google Scholar]
- Quigley M.A., Cumberland P., Cowden J.M. & Rodrigues L.C. (2006) How protective is breast feeding against diarrhoeal disease in infants in 1990s England? A case–control study. Archives of Disease in Childhood 91 (3), 245–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Rebhan B., Kohlhuber M., Schwegler U., Fromme H., Abou‐Dakn M. & Koletzko B.V. (2009) Breastfeeding duration and exclusivity associated with infants' health and growth: data from a prospective cohort study in Bavaria, Germany. Acta Paediatrica (Oslo, Norway: 1992) 98 (6), 974–980. [DOI] [PubMed] [Google Scholar]
- Renfrew M., Pokhrel S., Quigley M.A., McCormick F., Fox‐Rushby J., Dodds R. et al. (2012) Preventing Disease and Saving Resources: The Potential Contribution of Increasing Breastfeeding Rates in the UK. UNICEF: UK. [Google Scholar]
- Royal College of Nursing (2013) Breastfeeding in Children's Wards and Departments. Guidance for Good Practice. Royal College of Nursing: London. [Google Scholar]
- Schafer J.L. (1999) Multiple imputation: a primer. Statistical Methods in Medical Research 8 (1), 3–15. [DOI] [PubMed] [Google Scholar]
- Spencer N.J. & Coe C. (2000) Validation of the Warwick Child Health and Morbidity profile in routine child health surveillance. Child: Care, Health and Development 26 (4), 323–336. [DOI] [PubMed] [Google Scholar]
- Statacorp (2013) Stata Statistical Software: Release 13. StataCorp LP: College Station, TX. [Google Scholar]
- Tarrant M., Kwok M.K., Lam T.H., Leung G.M. & Schooling C.M. (2010) Breast‐feeding and childhood hospitalizations for infections. Epidemiology (Cambridge, Mass.) 21 (6), 847–854. [DOI] [PubMed] [Google Scholar]
- Tarrant R.C., Sheridan‐Pereira M., Younger K.M. & Kearney J.M. (2012) The positive role of breastfeeding on infant health during the first 6 weeks: findings from a prospective observational study based on maternal reports. Irish Medical Journal 105 (3), 75–78. [PubMed] [Google Scholar]
- WHO (2015) Up to What Age Can a Baby Stay Well Nourished by Just Being Breastfed? WHO; Available from: http://www.who.int/features/qa/21/en/ (Accessed 19 August 2014). [Google Scholar]
- WHO (2001) Global Strategy for Infant and Young Child Feeding. The Optimal Duration of Exclusive Breastfeeding. WHO: Geneva. [Google Scholar]