Prevalence

There is wide variation in prevalence rates reported in different series, from 0.02% to 10.7%.15–18 Most professionals would accept that the likely rate is 2–5%. The variation in prevalence is due to the lack of a uniform definition. Some assessments concentrate on the visual appearance of the frenulum and tongue while others place greater emphasis on the tongue function. The latter is more important because it is what will determine whether the baby is able to breast feed effectively.

It is more common in male infants in a ratio 1.5–2.6:1, with high family correlations. Most cases are sporadic but mutations in the T-box transcription factor TBX22 may cause a hereditable ankyloglossia with an association with cleft lip and palate.19 In relation to environmental factors, there is an increased association with maternal cocaine abuse.20

Clinical and functional assessment of tongue-tie

For the busy clinician the key point to appreciate is that, in a case of tongue-tie, the frenulum is attached close to the tip of the tongue.21 The tongue appearance should be examined when the tongue is lifted.22 In rare cases the frenulum attachment is to the proximal tongue base which is shortened and this can cause similar restrictions in tongue movements. When examined visually, the ‘free tongue’ length in newborns should be >16 mm; measurements of <11 mm indicate moderate ankyloglossia and <7 mm indicates severe ankyloglossia.23 In practical terms, however, this measurement is difficult to perform in a wriggling newborn.

Another measurement-based assessment has been proposed by Ruffoli et al.24 They state that, in a normal infant, the frenulum length is >2 cm and the intrinsic distance >2.3 cm.

One should be familiar with the tools used to assess function when one is assessing studies on tongue-tie and effects of frenotomy. The most comprehensive clinical assessment is the Hazelbaker Assessment Tool for Lingual Frenulum Function (HATLFF).25 It consists of five appearance criteria and seven function criteria each scored on a 2/1/0 system. Seven movements are evaluated: lateralisation, lift, extension, cupping, peristalsis, spread of anterior tongue and snap back. Tongue-tie is diagnosed if the appearance score is ≤8 or the function score is ≤11. A study of 58 infants found the tool to be highly reliable.26 Subsequently a study of over 140 babies found that 55.2% were not categorised using the HATLFF score.27 Ricke et al reported that the inter-rater agreement using HATLFF was moderate and many infants did not conform to the categories designed.28 It has been reported that the last four function items do not have good inter-rater reliability.29 Hazelbaker has pointed out that HATLFF is a screening tool and is insufficient to be used as a predictor of breastfeeding outcomes.30

In determining whether or not there are difficulties with breast feeding, a number of objective tools are used. The LATCH score,31 which is similar in design to the APGAR score, consists of five items marked on a 0–2 basis. The items are latch, audible swallowing, type of nipple, comfort and hold positioning. The LATCH score was found by Riordan et al31a to be a useful identifier for mothers at risk of early weaning. It provides a useful assessment of breastfeeding success and is used as an outcome in many studies.

The short-form McGill Pain Questionnaire (SF-MPQ)32 is used in many studies to rate the mother's nipple pain. It takes 2–5 min to administer and has three sections. It is a short version of one of the most globally recognised pain assessments.

The Infant Breastfeeding Assessment Tool (IBFAT)33 has been noted to correlate with breastfeeding competence and maternal satisfaction with breast feeding. While the LATCH score includes maternal nipple pain, the IBFAT tool does not. One small descriptive study found that the IBFAT and LATCH tools were not sufficiently reliable at that stage in their development to be valid for clinical use. They recommended revision and retesting before use in clinical practice to identify breastfeeding mother–infant pairs who need intervention.34

Efficacy of frenotomy

Four systematic reviews and five RCTs have assessed the efficacy of frenotomy. Webb et al35 in their methodological review confined their analysis to the 20 studies that met level 4 evidence or above. The objective improvements following frenotomy were LATCH scores (3 studies), SF MPQ Index (2 studies), IBFAT (1 study), feeding characteristics (3 studies). The subjective parameters were improvement in the maternal perception of breastfeeding (14 studies) and maternal pain scores (4 studies). Ito36 has recently published a systematic review that included 4 RCTs and 12 observational studies. The two most important outcomes emphasised in the review were latching and nipple pain. The conclusion was a moderate quality of evidence for the effectiveness of frenotomy.

Segal et al37 in a methodological review included seven articles describing the effectiveness of frenotomy. They stated that most of the studies were of poor methodological quality, the mean quality score being 24.4 out of a possible 47 points, but on balance frenotomy is probably an effective treatment. None of the studies prospectively compared its method of diagnosis against a proposed criterion standard. Furthermore, the studies used different outcome measures including nipple pain, successful breast feeding, tongue mobility and infant growth. Nipple pain was constantly highlighted.

Finigan et al,16 in a systematic literature review consisting of 5 RCTs and 8 case studies, concluded that frenotomy appears to offer long-term breastfeeding improvement for more than 50% of cases. The blinding in the RCTs raised a number of reservations. It can fail if there is a small blood stain on the swab in the intervention cases and also the blinding can only be maintained for a short period of time.

A total of 316 infants have been enrolled in 5 frenotomy randomised trials and the main findings are summarised in table 1.

Hogan et al38 performed a study on 57 babies (40 breastfed and 17 bottlefed), in which the index group had an outpatient frenotomy and the control group had 48 h intense lactation support followed by the offer of a frenotomy. The researchers were blinded. A 96% improvement was demonstrated in the frenotomy group (n=27) and a 3% improvement was shown in the control group (n=1). At 48 h the control group was offered frenotomy, which they all accepted. The authors themselves stated that one of the limitations was the lack of an objective measure of improvement.

Dollberg et al39 recruited 25 infants with ankyloglossia whose mothers had sore nipples. The sequences in the two groups were (1) frenotomy, breast feeding, sham, breast feeding; and (2) sham, breast feeding, frenotomy, breast feeding. Both mothers and caregivers were blinded. An immediate significant decrease was noted in the pain score but not in the LATCH score after frenotomy compared with the sham procedure.

Buryk et al40 performed a single-blind controlled trial of 58 infants randomised either to frenotomy (n=30) or sham procedure (n=28). The outcome variables were nipple pain and the IBFAT tool. The issue of breastfeeding duration could not be answered because all but one of the mother/baby dyads in the sham procedure limb had a frenotomy performed at 2 weeks.

A double-blind trial by Berry et al41 randomised 57 babies to division or non-division. They found that 78% (21/27) of the mothers in the division group reported an immediate improvement in breast feeding compared with 47% (14/30) in the non-intervention group. This is less striking than the results of non-blinded trials, perhaps reflecting the placebo effect. After the intervention, those infants allocated to non-division had their tongue-tie divided.

Emond et al17 in a study of 107 infants with mild to moderate tongue-tie compared frenotomy with controls. Those with severe tongue-tie were not included as they were offered immediate frenotomy. The primary outcome was the LATCH score at 5 days. The main finding was improved self-efficacy with feeding. The obvious reservation with the study was the omission of infants with severe tongue-tie. Longer follow-up was not possible because many of the mothers in the control limb opted for frenotomy after the end of the 5 day period.

There are a number of case–control studies. Ballard et al1 devised a study in which the breastfeeding mothers acted as their own controls. Frenotomy was performed on 123 infants resulting in improved latch in all cases and maternal pain was significantly reduced. Ricke et al29 enrolled 49 infants with tongue-tie and 98 control infants. All the infants were breastfed. Both groups were followed up and assessed at 1 week and 1 month. At 1 week the tongue-tie group was three times more likely to be bottle feeding (RR 3.11 CI 1.21 to 8.03), but they pointed out that 80% did successfully breast feed. At 1 month tongue-tie babies were as likely as controls to be bottlefed only. The authors did not find the ATLFF useful in determining which tongue-tie infants would develop breastfeeding problems. Messner et al10 undertook a case–control study of 41 infants; 83% of tongue-tie infants and 92% of control infants were still breast feeding at 2 months.

Schwartz et al have reported that, for every day of maternal pain during the initial 3 weeks of breast feeding, there is a 10–26% risk of cessation of breast feeding.42 Geddes et al43 recruited 24 mother–infant pairs experiencing breastfeeding problems. A combination of clinical and submental ultrasound parameters were measured before and after frenotomy. All clinical measures improved following the procedure. Before the surgery the ultrasound showed that the infants compressed either the tip or the base of the nipple. After the frenotomy both these ultrasound patterns resolved.

Sethi et al44 reported an improvement in breast feeding at 2 weeks after frenotomy in 77% of mothers but not in the other 23%. The authors point out that tongue-tie is not the only breastfeeding problem encountered in these infants. Steehler et al45 followed up 302 infants who underwent frenotomy for ankyloglossia. 80.4% of mothers strongly believed that the procedure benefited their ability to breast feed. Ochi46 surveyed 20 mothers of infants with ankyloglossia before and 2 weeks after frenotomy; 15 controls that also had their score measured at a similar time interval. The breastfeeding symptom score decreased in the frenotomy group but remained unchanged in the control group. The reservation of the study is that neither the patients nor the doctors were blinded.

Hall and Renfrew14 in a comprehensive review almost 10 years ago made many of the points that are still valid today. Maternal pain during feeding and difficulty in latching the baby to the breast are the main problems attributed to tongue-tie. They point out that these difficulties can be caused by factors unconnected with tongue-tie. It is difficult get an objective assessment measure after frenotomy because, in many studies, the primary outcome measure is pain during breast feeding.

Isaacson et al47 in their review of ankyloglossia point out that there is still a lack of consensus among clinicians about frenotomy. They suggest that, if there are continuing problems with breast feeding despite lactation support, frenotomy should be considered.

Timing of frenotomy and use of analgesia

It is evident from the studies that the breastfeeding difficulties in infants with ankyloglossia become apparent in the first few days after birth. Nipple pain and difficult latching lasting for >3 weeks after birth result in a 10–26% cessation in breast feeding.34 The timing of the frenotomy across a series of studies was as follows: 6 days,48 1–21 days,32 20 days,14 <2 weeks,39 18 days.14

In young infants frenotomy is usually performed at the outpatient clinic. When the infant is <3 months old it is undertaken without anaesthesia.34 Some operators administer either sucrose or paracetamol. Griffiths49 in a study of 200 infants undergoing frenotomy without analgesia found that 18% cried during the procedure and 60% after the procedure. The mean crying time was 15 s. A questionnaire study50 has reported that frenotomy was safe with no or local analgesia. Ovental et al48 in an RCT of topical benzocaine in 21 infants found that it was not beneficial. The average crying times in the benzocaine and control groups was 21.6 s and 13.1 s. Barrington51 states that some form of analgesia such as sucrose should always be administered to the infant.

Complications of frenotomy and their prevention

The effective way to minimise complications associated with the procedure is to ensure that the operator is skilled and properly trained. This is not always the case. A survey of 425 North American physicians6 reported that 22% had performed the procedure but only 10% had received formal training. An Australian study52 reported 11 midwives trained in the technique of frenotomy over 5 years, and further staff being recruited and trained.

Complications following frenotomy are uncommon.53 ,54 The commonest is bleeding which usually stops quickly with local pressure. A local haematoma is another potential risk.55 Care needs to be taken to avoid the lingual vein. Damage to the salivary ducts is a rare complication. Subacute massive oedema of the submandibular has been reported.56 Two per cent of infants49 developed an ulcer under the tongue and 2.6% of infants required a repeat procedure.15

The complications of frenotomy may be minimised with a check list before embarking on the procedure.

• Assessment by a lactation nurse to confirm that the tongue-tie is the cause of the breastfeeding difficulties.

• Oral and systemic examination by a paediatrician or GP to exclude other causes of poor feeding such as a urinary tract infection.

• Confirmation that the infant received vitamin K after birth.

• Note any bleeding disorders in the family.

• Parents should sign a consent form and be informed of the potential complications of the procedure.

• Ensure that the frenotomy is undertaken by an appropriately trained professional.

• Follow-up of the infant; if there is no improvement there may be another medical cause for the feeding problem.