Tongue-tie and frenotomy: what evidence do we have and what do we need?

We remain uncertain about which infants will benefit from surgical treatment

Ankyloglossia in newborns with breastfeeding difficulties has been diagnosed and treated in North America with increasing frequency over the past ten years.1,2 Kapoor and colleagues3 report in this issue of the MJA that frenotomy rates in Australia increased by 420% between 2006 and 2016, with differences between states and territories in the magnitude of the change.

We are pleased that our observations and concerns in the United States are shared by our Australian colleagues. We may be observing the results of global increases in breastfeeding rates and the availability of breastfeeding support services, changes in cultural expectations about breastfeeding, and increased recognition of the ankyloglossia diagnosis. However, overdiagnosis and unnecessary surgery by well meaning clinicians may also play roles. Confusing, low quality, and sometimes misleading information propagated by social media and in online resources about ankyloglossia and its effect on breastfeeding possibly also increase the demand for treatment. Personal, societal and cultural expectations of the breastfeeding process are also likely to play roles; ankyloglossia is diagnosed more frequently in first-born infants and in infants from families of higher socio-economic status.1,2

High quality evidence that frenotomy improves breastfeeding is sparse. Systematic reviews of randomised controlled trials have concluded that surgery for ankyloglossia reduced maternal pain caused by feeding, but the design flaws, selection and assessment biases, inadequate controls, and limited follow-up periods in the included trials dampen confidence in the findings.4-6 Several case series reports have concluded that frenotomy improves breastfeeding, but problems with patient selection, the lack of controls, and measuring treatment effects mean that assessing the benefit of surgical treatment is difficult.5

Despite the growing clinical enthusiasm for diagnosing and treating ankyloglossia, we should recognise that only some breastfeeding difficulties are caused by tongue-tie, and that some infants with anatomically restrictive tongue-tie can breastfeed quite well. Pain with feeding is a common complaint by mothers of tongue-tied infants, but 34–96% of all mothers experience pain when they start breastfeeding.7 Interestingly, pain and discomfort during feeding is reported for many different cultures, despite wide variations in breastfeeding success rates and ankyloglossia treatment;8-10 while about 80% of children are still breastfed at 12 months of age in much of Africa, 30% or fewer are breastfed this late in the United States.7 Cynicism regarding this disparity should be tempered by acknowledgement that the consequences of failed breastfeeding differ significantly between cultures. Mothers in countries with high breastfeeding rates often have limited access to health care, few alternatives to breast milk, and high infant mortality rates.

The present state of research into ankyloglossia, and perhaps into newborn feeding in general, is limited by subjective diagnostic assessments without rigorous validation, lack of data about anatomic variability, and the uncertain natural history of feeding difficulties. For example, many common diagnostic assessments, including the Hazelbaker Assessment Tool of Lingual Frenulum Function (HATLFF), can be unwieldy, and clinicians treating ankyloglossia are often not trained in the use of HATLFF.11 As our knowledge of normal infant feeding and anatomy — including lip, tongue, jaw, and neuromuscular control — improves, the quality of research will improve. The ideal study of newborn ankyloglossia would include objective assessment of feeding behaviour, oral anatomy and function, recognition of medical co-factors that affect feeding, randomisation to a standardised surgical treatment or observation, and precise assessment of changes in feeding, with an adequate follow-up period.

Evidence-based practice is based on the synthesis of the best medical evidence, clinical experience, and the preferences and values of patients.12 For some interventions, risks may be so low, the perceived benefits so great, and patient preferences so strong, that a randomised controlled trial may become hard to complete, if not impossible. Surgery for ankyloglossia in newborns may be such an intervention, but we need to ensure that patient assessment and treatment practices indeed minimise the risks and associated costs of frenotomy.13

Attributing breastfeeding difficulties to ankyloglossia is complex. We remain uncertain about which newborn infants should undergo frenotomy, and we question whether some of our observed treatment success is just the result of natural history. We, like our Australian colleagues, cautiously ride this wave of increasing diagnosis and treatment of ankyloglossia, educating parents and clinicians about the limitations of our knowledge, and reminding ourselves to offer the best evidence-based care for our patients, until we refine our understanding and assessment of newborn feeding. Perhaps we will then not only know whether frenotomy helps many infants, but will also be able to precisely select which infants can be helped.