Findings of the current study should be interpreted in the light of the following limitations. First, a randomized experimental design could not be used because the introduction of cup feeding would have been a threat to internal validity through the possible diffusion of cup feeding to the control group. Consequently, randomization was not possible. Second, the principal investigator collected all of the data, including the assessments of breastfeeding behavior after discharge. Thus, the PI was not blinded to group assignment or to the purpose of the study. Further study using independent data collectors who are blind to group assignment and purpose is required to overcome these limitations. Third, it was intended that determination of breastfeeding practices would have continued longer after discharge. However, 56% of the mothers had either only some or no education so that long term written documentation was not feasible. Additionally, concerns regarding accuracy of maternal verbal recall were present. These concerns were minimized by recording breastfeeding practices for the first week after discharge only. Use of a simple feeding diary is recommended for future studies. Finally because of the loss to follow up, the sample was too small to adequately answer the primary research question. Future research needs to be planned using a larger sample size to account for attrition. Despite the limitations, the current study was the first to implement cup feeding for preterm infants in Egypt and was one of the few studies to use cup feeding as the only oral feeding method for preterm infants during hospitalization [5, 13, 14].
All 60 mothers intended to breastfeed after discharge. Even though, only 17 infants in the bottle feeding group and 20 infants in the cup feeding group had breastfeeding experiences during hospitalization. The low incidence of breastfeeding during hospitalization is most likely a result of many factors, including infrequent visiting and lack of encouragement of the mothers to be actively involved in their infants' care during visitation. Consequently, 38% of the infants were discharged with no breastfeeding experiences during hospitalization. Despite this, the overall mean proportion of feedings that were breast feeding for the entire group one week after discharge was 72%, with a significantly higher proportion in the cup feeding group when compared to the bottle feeding group. Infants in the cup feeding group had significantly more breastfeedings per day one week after discharge from the hospital than infants in the bottle feeding group, suggesting that the transition to breastfeeding progressed more quickly for cup feeding infants than for bottle feeding infants. The lack of exposure of cup fed infants to oral mechanisms used during bottle feeding, which are different than the oral mechanisms used during breastfeeding,  might facilitate adaptation to breastfeeding. However, the explanation for this finding is unclear .
Sipping and lapping used during cup feeding has been theorized to enhance development of tongue movements needed for breastfeeding . However, the mechanisms of sipping and lapping differ from those required during breastfeeding [7, 28]. Sipping and lapping require the lips to be closed, rather than open, as required during breastfeeding and to a lesser extent during bottle feeding. Differences in mouth contour, activity of the masseter, temporalis and buccinators muscles, and the position of the lips between cup, bottle and breastfeeding may contribute to subsequent breastfeeding difficulties . A recent electromyographic study carried out during cup feeding, bottle feeding and breastfeeding found that the range of contraction and mean contraction of the masseter muscle were greater during cup feeding than during bottle feeding . This finding supported the recommendation that if breastfeeding is not possible at certain times, cup feeding may be indicated, as it allows the participation of the masseter and temporalis muscles in a way that is similar to the participation of these muscles during breastfeeding . These differences in oral mechanisms underlie the differences found in breastfeeding patterns when alternative methods such as bottle feeding have been used for supplementation of the breastfeeding infant.
An increased prevalence of breastfeeding has been reported when bottle feeding was replaced by cup feeding for preterm infants [17, 20] as well as full term infants . The findings of the current study are consistent with a recent randomized controlled trial  that found that cup feeding significantly increased the odds of breastfeeding at discharge. Additionally, Collins et al. reported a significant increase in the prevalence of breastfeeding at three and six months after discharge for infants fed by cup during hospitalization when compared with bottle supplementation . In contrast, another randomized controlled trial provided unclear evidence of the effect of using cup or bottle for feeding preterm infants on breastfeeding . No differences in breastfeeding prevalence at the first return visit between infants fed by cup and infants fed by bottle during hospitalization were found. At the first visit, 56% of bottle fed infants and 57% of cup fed infants had already been weaned and both groups presented similar breastfeeding prevalence. However, the percentage of infants still breastfeeding was two times greater in the cup fed group.
Although there were no statistically significant differences in the current study between cup feeding and bottle feeding groups regarding their breastfeeding type (full or partial) one week after discharge, more infants were exclusively breastfed in cup feeding group than in the bottle feeding group (Figure 2). These results are consistent with previous cup feeding studies [13, 19] that reported that cup feeding significantly increased the likelihood that the preterm infants would be fully breast fed at hospital discharge. Additionally, it has been demonstrated that exclusivity of breastfeeding at one month after birth predicted the likelihood of continuing breastfeeding at six months . In the current study it is possible that the shorter duration of cup feeding (9.1 days ± 5.61) than bottle feeding (12.5 days ± 8.20) resulted in these more optimal findings, as infants in the cup feeding group had less exposure to a feeding method other than breastfeeding. However, these findings were not significantly different, suggesting that it is the process of cup feeding rather than the duration of the feeding method that contributed to the better breastfeeding for the cup feeding group. In contrast to the suggestion of a recent Cochrane review , the length of hospital stay was shorter for the cup fed infants.
The second research question examined the maturation of breastfeeding behaviors from one through six weeks after discharge for both cup and bottle groups. The study findings demonstrated statistically significant differences between the cup and bottle feeding groups in their total breastfeeding behavior scores from the first-to-the sixth week after hospital discharge, reflecting higher mean PIBBS scores at each time point for the cup fed infants than the bottle fed infants. Additionally, infants in both groups showed an increase in their PIBBS scores over the six weeks, indicating a maturation of breastfeeding behavior over time. These findings are consistent with those of Nyqvist .
There was a significant interaction effect between group and time, demonstrating that, although both groups demonstrated expected maturation of breastfeeding behaviors, cup fed infants were significantly more mature in their breastfeeding behaviors at all time points than bottle fed infants, despite the cup feeding infants having statistically significant younger ages at discharge. The finding of improved breastfeeding behavior maturation among cup fed infants may be related to the higher breastfeeding proportion for this group, in that more breastfeeding experience may promote the maturation of breastfeeding behaviors. Conversely, more mature breastfeeding behavior may promote the frequency of breastfeeding.
Most previous studies have been concerned only with descriptions of infants' sucking and swallowing behavior , physiologic responses [32, 33], and milk transfer . Nyqvist, (1996) developed the PIBBS and used the instrument to describe the behaviors of breastfeeding preterm infants. However, the current study is the first to use the PIBBS to compare groups of breastfeeding preterm infants in relation to the method of supplementation . A recent report used the PIBBS to evaluate breastfeeding behaviors for two groups of term infants to determine if epidural anesthesia had an effect on breastfeeding behaviors. There were no statistically significant differences found between the groups . However, the PIBBS was not an appropriate instrument for use in that study as it had been developed for use with preterm infants. Future research should focus on the use of the PIBBS to compare preterm infants in relation to a variety of different experiences this population may have during the transition to full oral feeding.