*3.1. Univariate Analysis*

Due to the small number of breastfed infants (11.3%) who had been exclusively breastfed whilst in hospital we investigated the association of covariates with the duration of full and any breastfeeding for the first six months of life. In the univariate analysis (Table 2), the duration of any and/or full breastfeeding was associated with socio-demographic factors (maternal age, country of birth and level of education), parity, prelacteal feeding, introduction of a pacifier, whether a mother had roomed-in for 24 h with her infant and psychosocial factors (partner's and maternal grandmother's support for breastfeeding).

### *3.2. Multivariate Analysis*

Table 3 shows the results of the multivariate analysis. After adjustment, maternal level of education and country of birth were independently associated with both duration of full and any breastfeeding. Mothers with 12 or more years of education were less likely to stop any (Adjusted Hazard Ratio (AdjHR) = 0.68) and full (AdjHR = 0.74) breastfeeding during the six month follow-up period compared with mothers with less than 12 years of education. Similarly, mothers born in other Arab countries were less likely to stop any (AdjHR = 0.53) and full (AdjHR = 0.65) breastfeeding compared with women born in Kuwait. Mothers who did not intend to return to work within six months post-partum (AdjHR = 0.76) and those who did not experience breastfeeding problems in hospital (AdjHR = 0.80) were less likely to have stopped full breastfeeding. Conversely, women who did not feed on demand while in hospital (AdjHR = 1.28) or whose partner preferred formula feeding or was ambivalent as to how his child was fed (AdjHR = 1.33) were more likely to stop full breastfeeding. Multiparous women (Adj HR = 0.63) were less likely to cease any breastfeeding while those women who introduced a pacifier to their infant before four weeks (AdjHR = 1.66) or whose own mother preferred formula feeding or was ambivalent as to how her grandchild was fed (AdjHR = 2.11) were more likely to stop breastfeeding during the six month follow-up period.

#### *3.3. Reasons for Discontinuing Breastfeeding*

The reasons given by mothers for stopping breastfeeding are given in Table 4. The majority of women (86.8%) indicated that they were concerned about the adequacy of their breast milk in terms of either quantity or quality. Almost half (49.1%) indicated that their baby had either weaned them self, preferred a bottle or were ready for solids. A notable proportion stopped breastfeeding because they had returned to work or study. Only a small number of women cited mother-centered reasons like inconvenience and dislike of breastfeeding. The reasons for cessation did not vary markedly according to the infant age at which women stopped breastfeeding.

#### **4. Discussion**

While breastfeeding initiation is virtually universal amongst women living in Kuwait, targets for breastfeeding duration are not being met, with no woman in this study exclusively breastfeeding to six month of age and only 2% of women fully breastfeeding their infants to this age. As not all women had ceased breastfeeding by 26 weeks, it was not possible to estimate mean duration of breastfeeding however, the median duration of any breastfeeding was slightly more than three months and the mean duration of those women who had stopped breastfeeding before 26 weeks was six weeks. This suggests that the mean duration of any breastfeeding in Kuwait is likely to be less than the 4.9 months reported in 1997 [28] and that breastfeeding duration is declining.

In this study, women born in other Arab countries were less likely to have discontinued any and full breastfeeding than women born in Kuwait or other countries. These women are likely to be the wives of Middle Eastern guest workers employed in the oil and construction industries and the infant feeding practices of these women likely reflect those of their home country where children are breastfed for longer than infants in Kuwait. For instance, contemporaneous studies have reported a mean duration of breastfeeding of 8.6 months in the UAE [16] and 7.6 months in Bahrain [35], and a median duration of 12.4 months in Jordan [36]. A recent study in Kuwait reported that Kuwaiti mothers use bottle feeding more than the non-Kuwaiti mothers [37] which is consistent with our findings. Al Fadli *et al.* [37] proposed that such practice could be explained by the lifestyle changes that occurred in Kuwait due to oil revenue and through using modern technology similar to what happened in Western countries in the 1960s and 1970s.



**Table 2.***Cont.*


**Table 2.***Cont.*


**Table 3.** Factors independently associated with the risk for discontinuing any or full breastfeeding in a cohort of mother-infant dyads followed


a Women may have given more than one reason for stopping; b

*N* = number of subjects.

This study found no independent association with maternal age but, consistent with the findings of studies of women in Western countries [11,12], breastfeeding duration was positively associated with level of maternal education and parity, and negatively associated with maternal employment. Women with 12 or more years of education were less likely to have discontinued any or full breastfeeding compared with women with less than 12 years of education. Multiparous women were less likely to discontinue any breastfeeding than primiparous women which is consistent among women from Western [11,12] and other Middle-Eastern [14,18,22,23] countries. Previous breastfeeding success is a strong predictor of breastfeeding duration [12] and in general women breastfeed for longer with each successive pregnancy.

While there was no association with the duration for any breastfeeding, women who did not intend to return to work within 26 weeks were less likely to have discontinued full breastfeeding than those who planned to return to work. This suggests that women supplement breastfeeding with formula feeding either on return to work or in preparation for a return to work. This negative association between early return to work and breastfeeding duration has been reported in studies of women from other Middle-Eastern countries [17,19,20,22] and is consistent with studies of women from Western countries [31,38–40], and suggests that women everywhere have difficulty combining working with exclusive breastfeeding.

Other factors negatively associated with the continuation of full breastfeeding were whether prior to discharge the mother had experienced breastfeeding problems or her infant had been fed on demand, both of which may be inter-related. Milk production is directly related to suckling frequency [41] and there is evidence that fixed feeding schedules lead to insufficient milk supply and breastfeeding problems [42]. These findings highlight the importance of unrestricted breastfeeding in the early postpartum period to the successful establishment of breastfeeding. Hospital staff should encourage demand feeding and support and encourage women to persevere when they are experiencing difficulty establishing breastfeeding rather than resorting quickly to supplementing breast milk with formula.

This was the first study to investigate the association between pacifier use and breastfeeding duration among Middle Eastern women. The incidence of pacifier use amongst breastfeeding women in Kuwait (40%) was approximately half that reported for women in Australia [43,44] and the USA [45], and introduction of a pacifier before four weeks of age was found to be negatively associated with any breastfeeding duration which is consistent with the international literature [46]. While the mechanism remains unclear, it has been suggested that the non-nutritive sucking on a pacifier reduces the frequency of nutritive sucking from the breast, thereby leading to less stimulation of the breast and consequently less milk production [43,45].

Finally, this study highlights the importance of social support for breastfeeding. Support can come from a woman's partner, family and friends, and the degree to which each of these groups influences a woman's decision to breastfeed varies according to the mother's age, social class and cultural or ethnic background [47]. In traditional societies, women rely more on the advice and support of their mother, whereas in Western cultures they are more likely to identify their husband as their main source of support [48]. This and other studies of Muslim women have highlighted the importance of grandmothers both in providing practical support and as major influences on infant feeding decisions [8,49]. Advice received from their mother and mother-in-law can have both a negative and positive affect on a woman's breastfeeding practices. For instance, on one hand, breastfeeding is promoted in the Quran (*Al Baqara*, *233*) and by elders as the desired way to feed an infant, and the mean duration of breastfeeding is longer in most Muslim countries than in Western countries. On the other hand, there is a common perception amongst older women that the heavier the baby the healthier he or she is. There is anecdotal evidence that Kuwaiti grandmothers often encourage topping up with formula to ensure the baby is satiated and to stop hunger cries, which explains in this study the positive association with any breastfeeding but not full breastfeeding.

This study also showed that a husband's preference for breastfeeding over formula feeding was positively associated with breastfeeding initiation [30] and longer duration of full-breastfeeding which is consistent with Western studies [31,50,51]. To the best of our knowledge, no Middle-Eastern study has investigated previously the association of paternal attitudes and breastfeeding duration. There is, however, some evidence from Middle Eastern studies that support from a woman's husband is important for breastfeeding success and a study of women in Saudi Arabia found that mothers were more likely to initiate breastfeeding if their partners supported breastfeeding [25]. A Turkish intervention study reported the positive effects of an antenatal education program for fathers on their reproductive health knowledge, attitudes and behaviours, and women whose husbands attended these classes reported that their husbands became more supportive and communicative [52].

Women everywhere doubt the adequacy of their milk supply [53] and in this study more than eight in 10 women gave this as one of their reasons for discontinuing breastfeeding. Perceived breast milk insufficiency or *insufficient milk syndrome* (IMS) is frequently associated with the premature introduction of complementary foods [9] and with the cessation of breastfeeding [14,16] in Middle Eastern countries. It has been proposed [53] that IMS is increasing with "aspects of 'modernization': urbanization, education, and female employment—factors that are repeatedly found to be inversely associated with both the prevalence and duration of breastfeeding" (p42). It has been suggested that 'insufficient milk' is given as a socially acceptable reason for discontinuing breastfeeding when a mother decides she no longer wishes to breastfeed [54] and that claims of IMS should not be taken literally when they occur in cultural contexts that present the use of infant formula as an acceptable, if not preferred, alternative [55].

As we have previously identified [30], there are a number of limitations to this study. Firstly, the sample size is relatively small and this is reflected in the wide confidence intervals around some of the adjusted hazard ratios reported. Secondly, we may have underestimated the rate of prelacteal feeding, which in this study was defined as within the first three days after birth. The average length of post-partum stay for Kuwaiti public hospitals is a maximum of two nights for uncomplicated deliveries and five nights for a caesarean section. Therefore, it is possible that some mothers discharged within 48 h may have gone on to supplement breastfeeding with formula following discharge from hospital and within this 72 h period. Given, however, that almost nine out of 10 infants received prelacteal feeding in hospital, any underestimation of prelacteal feeding is likely to have had only a negligible effect on the results. Finally, the number of women who delivered by caesarean section is three times that of the national average. While every attempt was made to recruit mothers within 72 h and in most cases 48 h, women who had undergone a caesarean section had a greater chance of being recruited because of their longer hospital stay.

The major strength of this study is that it was the first prospective study of infant feeding practices in Kuwait, all other previously reported studies being cross-sectional. Mother-infant dyads were followed from birth to 26 weeks with data being collected at five time points during this period, thus minimizing the potential for maternal recall bias [56] as women were recalling events close to the time at which they occurred. The findings of the study are consistent with those of other studies in the region and, in most instances, studies of Western women, and can be used to inform infant feeding policy, hospital practices and the design of breastfeeding promotion interventions.

#### **5. Conclusions**

The duration of breastfeeding amongst women in Kuwait, particularly Kuwaiti born women, appears to be declining. Exclusive breastfeeding is virtually non-existent with almost nine in 10 infants receiving prelacteal feeds within the first three days of birth. Full breastfeeding is also relatively uncommon, and by four weeks, less than one third of infants were fully breastfed, while at six months, only four in 10 infants were receiving some breast milk. This study identified a number of areas for intervention. Hospitals should follow the 10 Steps for Successful Breastfeeding [57] and, in particular, promote the early initiation of breastfeeding, encourage women to feed on demand and avoid the unnecessary practice of prelacteal feeding. Women and health professionals need to be alerted to the negative consequences of early pacifier use on breastfeeding duration. The role of family members should not be underestimated in planning breastfeeding interventions. Community-based interventions are needed to support women to breastfeed and to provide a supportive environment. Close family members, especially husbands and maternal grandmothers, should be targeted in these interventions to ensure higher rates of exclusive breastfeeding and prolonged duration.

### **Acknowledgments**

We sincerely appreciate the assistance given by mothers in our study and the enthusiastic support from the Kuwait Ministry of Health and hospital staff in all participating hospitals. MD was supported by a PhD Scholarship from the Civil Services in Kuwait. We would like to acknowledge the statistical support given by Rosie Meng of Curtin University.

#### **Author Contributions**

MD participated in the design of the study, collected the data, performed the statistical analysis and co-wrote the first draft of the manuscript. JAS conceived of the study, developed the original questionnaires on which the study instruments were based, assisted with statistical analysis and co-wrote the first draft of the manuscript. CAE advised on the statistical analysis and commented on drafts of the manuscript, and MAS provided assistance with the on-site coordination of the study and commented on drafts of the manuscript. All authors read and approved the final manuscript.

#### **Conflicts of Interest**

The authors declare no conflict of interest.
