Tongue-Tie and Breastfeeding in Newborns—Mothers' Perspective

Introduction

Tongue-tie (TT) (ankyloglossia) is a condition in which the bottom of the tongue is tethered to the floor of the mouth by a membrane (frenulum), limiting tongue motion. TT may interfere with oral cavity development, breastfeeding, sucking, swallowing, or speech.^1–11 Usually, when the infant latches onto the mother's breast, the tongue extends over the edge of the lower gum and cups around the areola. Restricted tongue movement due to TT may disturb latching and breastfeeding leading to sore nipples, poor intake and weight gain, and eventually the abandonment of breastfeeding.^{1,5,6,12–23} TT is frequently familial. Although considered a minor malformation, sometimes a part of certain syndromes, it is an isolated finding in most cases.^9,24–26 The actual incidence of TT is unknown, but is estimated to be 1.7–4.8%.^{1,5,6,8,27,28} None of the various grading systems suggested for the definition of TT and its severity has become common practice.^1,6,27,29,30

Extensive research on the benefits of breastfeeding has led to breastfeeding encouragement in recent years, ³¹ reviving the controversy on frenotomies. Although some professionals advocate relatively early frenotomies to support breastfeeding,^32–37 others seriously question such routine procedures, even if they could be considered minor surgical procedures.^{1–3,5,6,8,10,21,22,28,32–38}

Our aim was to approach this debate from the maternal point of view, by trying to explore mothers' breastfeeding experience, problems they encountered, and possible association with TT.

Materials and Methods

This was a prospective observational study of infants with TT and without TT. Data were collected from telephone follow-up interviews with mothers of infants diagnosed with TT in the hospital after delivery and a control group 1–6 years later (January 1, 2010–December 31, 2012). All medical records of newborns delivered at Bnai Zion Medical Center at a gestational age of >37 weeks who were diagnosed with TT between January 1, 2005 and December 31, 2010 were retrieved using a computerized medical records manager system (version 7.6; Max 2000 Ltd., Haifa, Israel). The diagnosis of TT was clinically made by physicians, nurses, or lactation consultants as part of their routine primary assessment of the infant and was confirmed by the discharging senior neonatologist. However, its severity was not defined in the medical records. Although some additional data were found in some of the cases regarding the type and nature of the sublingual frenulum, sucking function, or breastfeeding, these data were inconsistent and could not have been used as a formal assessment of severity. Two gender-matched newborns without TT born on the same day were randomly assigned as controls. Exclusion criteria included neonatal morbidity, hospitalization for more than a week, congenital malformations or syndromes, or significant neurological signs or symptoms (i.e., generalized hypotonia). The study was approved by the Bnai Zion Medical Center hospital's ethics committee (protocol number BZ 45/06).

Basic demographic and medical data were retrieved from the medical records. The mother of each infant was then contacted and, after giving consent, was interviewed over the telephone by one of four female family physicians with experience in lactation consultation. A uniform structured questionnaire was used for the interview and documentation. Data collected during the interview included breastfeeding duration, breastfeeding problems in the first month, causes for discontinuation of breastfeeding, and lactation consultation and support. When the mothers were asked about problems with breastfeeding, first they were asked a general open question in order to record their spontaneous response. Only after this were they asked specific questions regarding common problems encountered during breastfeeding. All items of the questions are as translated from the original questionnaire in Hebrew. Mothers of infants with TT were also asked how and when was TT diagnosed, presence of breastfeeding problems, had the baby undergone frenotomy and, if so, the indication, age, improvement after the procedure, and family history of TT. No frenotomies were performed in the hospital in the immediate period after delivery. If the diagnosis was suggested and breastfeeding problems were recorded, the infant would have been invited to a clinic, either in the hospital or in other healthcare premises, for follow-up and decision regarding frenotomy there. Frenotomies were performed only by physicians.

Statistical analysis (SigmaStat version 2.03; SPSS, Inc., Chicago, IL) included descriptive statistics, chi-squared test, Student's t test, and analysis of variance for univariate analysis. Stepwise regression model (SAS version 9.2; SAS Institute, Cary, NC) was used for the multivariate analysis. Statistical significance was set at 5%.

The interobserver variation was evaluated by calculating the agreement among the four interviewers on a random sample of interviews, performed twice by different interviewers on different occasions (inter-rater reliability).

Cronbach's alpha was calculated for different sets of questions to estimate the reliability of our questionnaire (internal consistency).

Results

Two hundred thirty-nine newborns (1.1%) were diagnosed with TT out of the 21,424 newborns ≥37 weeks of gestation. There was a gradual although not steady increase in the rate of diagnosis of TT over the 6 years studied (0.7%, 0.8%, 1.1%, 1.1%, 1.7%, and 1.3%, respectively).

Of the 239 infants diagnosed with TT and 478 controls, 172 (72.0%) and 325 (68.0%) of the mothers were interviewed. Most of those who were not interviewed were lost to follow-up (52/67 [77.6%] and 119/153 [77.8%], respectively; difference not significant). Of the 325 controls, 11 (3.4%) were found to have TT after discharge from the hospital, as was discovered during the phone interview with their mothers. Thus, we had 183 infants with TT in the study group (172 [94.0%] diagnosed by us and 11 [6.0%] found in the controls) and 314 controls.

There were no significant differences in most demographic characteristics between the study and control groups (Table 1). However, TT was significantly more frequent in males and in infants of Jewish compared with Arabic decent (Table 1). Children with TT were significantly younger at the time their mothers were interviewed by us (median, 3.2 vs. 3.9 years, respectively) (Table 1).

There were no significant differences between the groups in rate or duration of exclusive/partial breastfeeding in the first 6 months (Table 2). Infants with TT did not lose more weight or required longer hospitalization after delivery (Table 1) and were not re-admitted more in the neonatal period (Table 2). Although the overall rate of reporting on breastfeeding problems in the first month was not significantly different in the TT and control groups (59% vs. 52.5%, respectively), TT mothers described more latching problems, prolonged breastfeeding beyond the first week, infant's exhaustion during feedings, and other breastfeeding problems (Table 2). TT mothers used more pumped breastmilk to supplement partial breastfeeding compared with control mothers who used formula more (Table 2). The higher maximal bilirubin levels found in the controls were clinically meaningless (Table 2).

No significant differences were found regarding reasons for discontinuation of breastfeeding (Table 2). Multivariate analysis revealed the following significant negative associations: between successful breastfeeding during the first 6 months of life and breastfeeding problems in the first month (p<0.0001), specifically problems with insufficient milk supply (p<0.0001), pain and sore nipples (p<0.0001), latching (p<0.05), and other unspecified problems (p<0.01); exclusive breastfeeding duration and breastfeeding problems in the first month (p<0.01), specifically insufficient milk supply (p<0.0001); total duration of breastfeeding and insufficient milk supply (p<0.0001); and partial breastfeeding duration and insufficient milk supply (p<0.05), as well as latching problems (p<0.005). None was significantly associated with the presence of TT.

Significantly more TT mothers sought lactation consultation after discharge. In accordance with problems they described with breastfeeding in the first month (Table 2), latching and prolonged breastfeedings also were more common causes for seeking help (Table 3). There were no significant differences between TT and controls in the authorities consulted regarding breastfeeding. Besides friends and relatives, mothers, especially of those of infants with TT, consulted physicians. More than half of the mothers in both groups eventually turned to certified lactation consultants (Table 3). Although the measures for breastfeeding support were not significantly different, significantly more TT mothers felt that lactation consultation helped (Table 3).

In the TT group, 86.9% of the mothers knew about the diagnosis of TT, yet only 49.7% were aware that TT might be associated with breastfeeding problems, and only 56.5% from those who reported on problems in breastfeeding related them to TT (Table 4).

Of the mothers in the TT group, 10.3% (19/183) reported speech problems in their children (Table 4). We also calculated what we have considered a truer rate of 11.9% (19/159) speech problems in TT infants, assuming that speech problems could be diagnosed with certainty only in children ≥2 years when the mothers were interviewed. Thirty-five percent of TT infants had at least one first-degree family member with TT (Table 4).

Although the possibility of frenotomy was mentioned to 68.8% of the TT group, it was performed in only 35.5% of the infants with TT (Table 4). All frenotomies were performed by physicians: 38.5% by pediatric surgeons, 10.8% by pediatricians, 7.7% by ear-nose-throat specialist, 7.7% by dentists, 6.1% by dermatologists, and 3.1% by family practitioners. The 24.6% performed during Jewish circumcision were actually done by physicians licensed to do these ritual circumcisions, and the 1.5% (probably more) done by lactation consultants were also performed by physicians who were also trained and licensed to work as lactation consultants. Although no official training or license is required in order to perform frenotomies, as far as we know all physicians who perform them were trained and did it under the authority granted to them as physicians. Data on the procedure are presented in Table 4. Frenotomies for speech problems only were performed significantly later than those done for other causes (5.4±10.5 [median, 1.8] months, compared with 22.2±24.2 [median, 8.0] days). However, 12 of these 20 frenotomies were performed at an age that was young for such a diagnosis. The differences in the rates of reporting breastfeeding problems between those who had frenotomy and those who did not are presented in Table 5.

Only 40.0% of the mothers whose children had frenotomies reported improvement following the procedure. However, 88.5% of the mothers who had their infants undergo frenotomies to solve breastfeeding problems were satisfied with the procedure and reported improvement.

Regarding inter-rater reliability, the mean agreement score was 0.92±0.08 (range, 0.84–0.98). One-way analysis of variance showed that the means of the different interviewer pairs did not significantly differ from each other. Using 92% agreement and assuming 50% agreement rate expected by chance alone, we calculated a kappa of 0.84, which according to Landis and Koch criteria represents excellent agreement beyond chance. Using the stricter estimate of 66% agreement by chance would result in a kappa of 0.76, which is still beyond the 0.75 cutoff defining excellent agreement.

Regarding internal consistency, the standardized Cronbach's alpha for the set of questions regarding breastfeeding problems in the first month (Table 2) was 0.71, reflecting that they indeed explore problems associated with breastfeeding difficulties.

Discussion

The overall rate of reporting on early breastfeeding problems in the first month was not significantly different in the TT and control groups. Although some significant differences were noted in the early breastfeeding problems, they did not result in significant differences in breastfeeding rates or duration later on. We found no differences in the duration of exclusive or partial breastfeeding between mothers of infants with and without TT, possibly because of the different interventions used (Tables 2–5). However, we found that TT mothers who could not exclusively breastfeed more frequently supplemented feeds with pumped breastmilk as opposed to controls, who used more formula. We speculate that this might be attributed to greater exposure of these mothers to lactation consultants who stress the importance of breastmilk to the infant's health.

TT mothers seem to have had more specific breastfeeding problems in the first month, including difficulties with the infant's latching, prolonged breastfeedings, and exhaustion of the infant from the breastfeeding effort (Table 2). None of these breastfeeding problems in the first month was associated with long-term breastfeeding problems or earlier discontinuation of breastfeeding compared with controls (Table 2). It is surprising that pain and sore nipples, which are considered classical symptoms of TT and possible indications for frenotomy, were not more common among the TT group. This should be taken in account as one of the limitations of the study because on the phone interview mothers were not asked to try and scale the severity of their nipple pain and damage or its duration. It is possible that the simplistic measure of “sore nipples” might have underestimated the difference between the groups in this respect. TT mothers did not describe more breast engorgement from their infant's insufficient nursing, nor did they complain of insufficient milk supply. They more frequently reported poor weight gain, but the difference was not statistically significant. This is in agreement with the medical records' findings that there was no evidence of neonatal problems that might have been associated with poor breastfeeding in TT newborns (Tables 1 and 2). However, it is possible that the hospital stay was too short to detect this.

It is possible that the lack of differences in the duration and success of breastfeeding could be related to higher rates of turning to lactation consultation by TT mothers (Table 2). Although they were not different in the reasons that led them to seek help, TT mothers were significantly more satisfied with lactation consultation, possibly reflecting success of this modality in helping significant breastfeeding problems (Table 3).

No significant differences were found regarding reasons for discontinuation of breastfeeding (Table 2). Among the reasons for stopping breastfeeding is return of the mother to work (15–17% in both groups). It is important to mention in this regard that the official maternity leave in Israel is 3 months; however, working mothers can prolong it up to 6 months, which many mothers chose to do. This might explain the relatively high rates of full or partial breastfeeding at 6 months among the mothers in our study (Table 2).

The option of frenotomy to solve breastfeeding problems was mentioned to approximately two-thirds of TT mothers (Table 4). Frenotomy for solving speech problems was mentioned at least as frequently, which is quite surprising in view of the current approach not to recommend early frenotomies to prevent possible speech articulation problems, before such problems could be diagnosed. Nevertheless, in our study 11.9% of TT mothers reported speech problems in infants ≥2 years, which is a relatively high rate.

Frenotomy was performed on slightly more than a third of the infants with TT; of these, two-thirds were performed for solving breastfeeding problems alone or in combination with other indications (Table 4). Indeed, breastfeeding problems, especially latch and infant's exhaustion, were reported significantly more commonly among TT mothers whose infants had frenotomy (Table 5). These mothers also turned more to lactation consultation. The median age for performing frenotomies in our population was 8 days, which is unique to our society and probably reflects the widespread practice of performing it with the Jewish circumcision. The satisfaction rate with frenotomies was relatively low (40%); however, among the satisfied mothers most did it for solving breastfeeding problems. This probably suggests that this procedure should be performed for the right indication, usually significant breastfeeding problems, and at the right time (i.e., after a qualified lactation consultation failed to help).

This study is the first to actively explore the issue of TT and breastfeeding from the mothers' point of view. The strength of this study lies in the large number of interviews and the strict methodology, as depicted by the agreement between the interviewers and the quality of the questionnaire itself. The relatively comfortable nature of the interviews, taken by female family physicians with previous experience in lactation consultation, along with the temporal distance from the immediate breastfeeding period enabled frank, truthful responses. The structure of the interview, first inquiring about general breastfeeding experience, problems, and support and only later advancing to specific questions related to TT, could also enhance unbiased responses regarding possible breastfeeding problems without immediately creating the association to TT.

Among the limitations of this study are its single-center and retrospective nature. One of the main limitations is the subjective bias associated with any interview-based study, especially when the interviewers were aware of the diagnosis of TT. The structured interview and the high agreement among interviewers decrease, although do not abolish, these limitations. It is important to mention that none of our interviewers was related to these mothers or infants. They did not know them before, nor did they take care of the infants. They did not make the initial diagnosis of TT and were not involved in the follow-up, lactation support, or treatment.

Another limitation is related to the temporal distance that could have been associated with recall bias. Children in both study and control groups were 1–6 years at the time of the interviews. However, it turned out that those in the TT group were significantly younger (by 7 months, on average) at the time their mothers were interviewed (Table 1). We believe this was due to faulty design on our behalf, as we had first started interviewing mothers of infants diagnosed with TT and only after 6 months started interviewing mothers of control group infants as well, which eventually may have led to this untoward difference. When mothers are asked retrospectively about breastfeeding duration, the reports become increasingly inaccurate with increasing time since cessation.^39,40 Thus, recall bias is an inherent limitation of this study because interviews were conducted 1–6 years later. Although the difference in the age of the children at the time of the interviews was significant between the TT and control groups, we do not believe that these 7 months, on average, could have caused a major difference between the groups regarding the extent of the recall bias.

The diagnosis of TT, which is clinical and, unfortunately, subjective, although verified by a senior neonatologist, is also a limitation. Another limitation is that this study has not considered the contribution of maternal anatomy to breastfeeding problems when an infant has TT. In practice, infant TT might not be a problem for a mother with easily graspable nipples, whereas another mother with flatter nipples may struggle to attach her infant without a frenotomy or use of a nipple shield.

The retrospective design and the resultant bias related to a procedure that needs to be justified are also limitations. It could be argued that the lack of effect of TT or frenotomy could be related to lack of definite diagnostic criteria and indications for the procedure, which resulted in lumping together mild and severe cases. It is difficult to appreciate in retrospect whether the lack of differences in breastfeeding duration rates resulted from no effect of TT or from positive effects of lactation consultation or frenotomy.

In summary, this study provides further insight on breastfeeding problems in general and the importance of lactation consultation. As long as we have no better ways of defining and grading TT, the best way of assessing its significance will continue to be clinical evaluation of symptoms, of which breastfeeding problems are the most common and important. Carefully identifying infants with TT in whom lactation support fails to assist and who could benefit from frenotomy should result in successful breastfeeding and increased satisfaction rate.