Article Text


Incidence and treatment of infantile haemangioma in preterm infants
  1. Rangmar Goelz,
  2. Christian F Poets
  1. Department of Neonatology, University Children's Hospital Tuebingen, Tuebingen, Germany
  1. Correspondence to Dr Rangmar Goelz, University Children's Hospital Tuebingen, Department of Neonatology, Calwerstrasse 7, Tuebingen D-72076, Germany; Rangmar.Goelz{at}


Infantile haemangioma (IH) are vascular tumours with a unique growth dynamic, mostly absent at birth, growth in the first months followed by involution over several years, often resulting in residual skin changes. Immune-histologically, IH cells are exclusively glucose transporter protein-1 positive.The incidence of IH is increasing with decreasing gestational age, from 1–4% in term infants to 23% in those of <1000 g birth weight, with a female and Caucasian predominance. Discovery of systemic and topical beta blockers as an effective treatment option resulted in a rapid shift away from systemic steroids towards these drugs. For preterm infants, however, data on efficacy, pharmacokinetics and long-term safety are sparse or absent. Topical treatment without systemic side effects like cryotherapy may thus be an attractive alternative at an early growth stage (<10 mm). Indications for treatment with beta blockers, mostly propranolol systemically and timolol maleat 0.5% topically, are currently extrapolated from studies in older infants. Both seem effective, but adverse effects on sleep, circulation and metabolism are well described for propranolol. Long-term outcome data for either drug are missing. In conclusion, evidence on optimal IH treatment in preterms is lacking despite their high incidence; pharmacokinetic and clinical studies are warranted.

  • Infantile Haemangioma
  • preterm
  • incidence
  • treatment
  • Neonatology

Statistics from


Infantile haemangioma (IH) or haemangioma of infancy consists of proliferating vascular tumours. IH is typically absent at birth and has a characteristic growth pattern with rapid enlargement during the first 5–9 postnatal months, a period of most rapid growth at 5–8 weeks,1 and a slower regression or involution phase taking up to 5 years to complete, with residual deformities like skin atrophy, scarring, destroyed anatomical structures, a fibrofatty mass, redundant skin and telangiectasia in some cases.2–6

A landmark in defining IH was the differentiation between haemangiomas and vascular malformations introduced by Mulliken and Glowatzki.7–9 Based on a large case series, IHs were subsequently further classified as localised, representing the majority with 67% (1022/1530), segmental (13%), indeterminate (17%) and multifocal (4%),10 and additionally subdivided into superficial, mixed and deep11–13 resulting in a distinction by the pattern of distribution on the body surface and anatomic depth of involvement.14 Recently found histochemical markers like glucose transporter protein-1 are characteristic of IH and differentiate them from other vascular tumours or birthmarks, especially congenital haemangioma (CH).15–18

Differential diagnosis of IH includes CH, comprising rapidly involuting CH (RICH) and non-involuting CH (NICH), kaposiform haemangioendothelioma, tufted angioma, pyogenic angioma and multifocal lymphangioendotheliomatosis.17 Older nomenclature like capillary or strawberry haemangioma has been discouraged in recent years.2


The incidence of IH in preterm infants is probably higher than in term infants, but exact incidence data are missing, not least due to the inconsistent nomenclature used prior to the widespread acceptance of the above classification system.7 ,9 Data from existing studies are summarised in table 1 and indicate an increasing IH incidence with decreasing gestational age (GA) and birth weight, with up to 23% for those born at <1000 g compared with 1–4% for term infants, a female predominance and a higher rate in Caucasian infants compared with other ethnicities.19–21 Facial involvement seems to be less common in preterm infants.22

Table 1

Incidences of IH in preterm infants reported in the literature

Indication for treatment

Despite the high frequency of IH in preterm infants, there are no studies tailored especially to this high-risk group except for one small study using nitrogen-cooled cryocontact therapy (NCCT) exclusively in preterm infants.30 Thus, decisions can only be made in analogy to term infants, and therefore indications for treatment are mainly based on individual considerations. A Cochrane review on interventions for IH identified only four randomised controlled trials (RCTs) until March 2011 and concluded that evidence from RCTs is too limited to support any of the interventions existing at that time (pulsed dye laser and corticosteroids).31 For infants, there is agreement that haemangioma with life-threatening or function-threatening properties, ulcerations or a risk thereof, rapid growth or a risk for acute or chronic disfigurement must be treated. This applies to those in the face, scalp, neck, hands, feet and intertriginous and anogenital regions.5 ,32–36

A series of recent publications, however, emphasised the high proportion of residual deformity found in 69–88% of IHs.6 ,37 The characteristic growth pattern with a peak at 5–8 weeks suggests that therapy should be started before reaching this age.1 Moreover, the introduction of beta blockers since 2008 as a very effective therapy with less severe adverse events lowered the threshold for intervention.38 Thus, a paradigm shift can be observed in the timing of referral and initiation of IH treatment.1 A rapid increase in beta blocker and decrease in steroid use have already been reported.39


Pharmacological therapy of IH has been revolutionised by the observation of Léauté-Labrèze that propranolol resulted in a dramatic shrinking of large haemangiomas.38 It is now considered by most experts to be the first line drug if systemic therapy is indicated.5 ,16 ,40 The same is true for topical application of another beta blocker, timolol maleate 0.5%, following its first description by Guo.41

Therefore, only therapeutic studies published after 2008 are considered in the following. Traditionally used drugs like corticosteroids, interferon-alpha and vincristine13 are now regarded second or third line agents, mainly due to their severe side effects, including spastic diplegia and other neurological abnormalities.42–44 Additional therapeutic options include cryocontact, laser and surgical therapy, as well as topical and intralesional drug application.

A PubMed search to assess the effects of interventions for IH in preterm infants using the keywords ‘hemangioma, infantile, (therapy or treatment), (study or trial), controlled, preterm’ identified only a single study which used NCCT (see below). Thus, treatment for the patient group with the highest incidence (and the highest risk of side effects) is the one least well studied, as also reflected in national guidelines that exclude preterm infants from their recommendations34 or do not mention them at all.36

Repeating the above PubMed search after omitting ‘preterm’ brought 29 hits for studies published after 2008, seven of them being RCTs comparing propranolol versus steroids (two studies), propranolol versus atenolol,1 propranolol versus placebo3 or different application modes1 (table 2), but again none of these RCTs referred to preterm infants.

Table 2

Randomised controlled studies using beta blockers for treatment of IH

Treatment options for preterm infants

For topical therapy, data from preterm infants exist for cryocontact, laser and timolol maleat,16 while for systemic therapy, studies on propranolol and, as a second line drug, steroids (prednisolone), the former standard therapy, are available.

Cryocontact therapy

One prospective controlled study (quasi-randomised) could be identified,30 using NCCT (−196°C, 2–6 s application time) of IH (<10 mm in diameter) in infants ≤34 weeks GA compared with untreated intraindividual control-IH. NCCT led to fast regression with good cosmetic results. Limitations include the premature termination of the study, a follow-up period of only 2 years and mild scarring as a side effect in four of 17 infants (compared with residual IH in 14 of 17 controls). Further data on NCCT, with good results and rare scarring, are reported from other centres52 ,53; this treatment is also recommended in current German guidelines.34 The method is fast, easy to apply, well tolerated, cheap, available at the bedside and, most important, without systemic side effects. It is yet, however, limited to superficial IH of ≤10 mm in diameter.


There are no data on laser therapy from preterm infants. A single RCT compared pulsed dye laser in early childhood with no treatment. When evaluated at 1 year of age, there was no group difference in the number of children with complete regression or minimal residual signs.54 At present, there are insufficient data to recommend this treatment for preterm infants.2 ,55


Six RCTs have been published in the last 6 years (table 2), two comparing propranolol with prednisolone, of which one was stopped early due to severe adverse events in the steroid group.46 A combination of both propranolol and prednisolone was not superior to propranolol only.47 Two trials were placebo controlled and favoured treatment.50 ,51 Propranolol and atenolol were compared in one trial with similar results for complete and partial response, but atenolol had fewer side effects, possibly because of its higher beta-1 receptor selectivity.45 One study compared oral propranolol with topical and intralesional application and found best results for oral administration.48 None of these studies reported the inclusion of preterm infants. One large RCT, double blind and placebo controlled, finished recruitment in May 2014; first results are available and show better cosmetic results with propranolol at 3 mg/kg/d for 24 weeks. Preterm infants, however, could only be included after reaching term equivalent age (TEA) (Study to Demonstrate the Efficacy and Safety of Propranolol Oral Solution in Infants With Proliferating Infantile Hemangiomas Requiring Systemic Therapy. Trial Identifier: NCT01056341). Details can be seen in the release of the Committee for Medicinal Products for Human Use (CHMP) Assessment Report of the European Medicines Agency (EMA) for this drug (

Additionally, numerous reports evaluated the efficacy of propranolol. In a systematic review of all studies reporting on >10 patients published between June 2008 and June 2012, the average response rate for treatment with propranolol was 98%. Most common adverse effects were changes in sleep (like insomnia, nightmares, restlessness and sleep disturbances), acrocyanosis, hypotension, bradycardia, hypoglycaemia and respiratory as well as gastrointestinal symptoms. Rebound growth was reported in 17% after stopping initial treatment.56 A second large systematic review and meta-analysis compared the data available from 1965 through March 2012 to compare propranolol and corticosteroids in the treatment of IH. Again, only studies with >10 patients and systemic application were included. There was only one RCT in each group. The difference concerning any improvement was 97.3% for propranolol versus 71.0% for corticosteroids (p<0.0001).57 Both publications did not mention preterm infants (table 3).

Table 3

Systematic reviews on the treatment of IH

In a case series, nine very low birthweight (VLBW) infants were treated with propranolol (2 or 3 mg/kg/d) at a corrected age of −5 to+15 weeks. Growth was not impaired and no relevant side effects were reported.58 A larger case series of 99 infants treated with propranolol included 7% preterm infants. IH improved in 99%, but no additional information on the preterm group was given.59 A unique severe side effect, hyperkalaemia, has been reported in a girl of 28 weeks gestation with a large and ulcerating IH 72 h after starting propranolol at TEA. No electrocardiographic changes were noted; maximum K+ concentration was 6.5 mmol/L with therapy.60

In summary, currently available data for IH treatment of preterm infants are scarce, both for treatment in the first postnatal weeks or beyond reaching TEA. There is also no information about long-term neurocognitive outcomes. Therefore, systemic propranolol, a vasoactive drug, should be indicated cautiously in immature infants. That former preterms having passed their TEA can be considered similar to term infants may seem plausible, but needs to be confirmed in clinical studies.

Topical timolol maleat solution or gel

There is one double-blind placebo-controlled RCT comparing topical timolol 0.5% solution with placebo. After 24 weeks of therapy twice daily, the endpoints colour and volume of IH favoured treatment (p<0.003 and p<0.002, respectively) (table 2).49 A PubMed search to assess the efficacy of topical application of timolol solution or gel-forming solution revealed 11 hits reporting study results for >10 enrolled infants/study (table 4). Most studies used either drug, applied 2–4 times/day by gently rubbing in with a fingertip. One study used a standardised dose of 0.05 mL timolol 0.5% gel with occlusive dressing through Finn Chambers.61 The duration of therapy was variable, in the only RCT it was 24 weeks. In all studies, good or excellent improvement in the majority of superficial IH was reported. While results were worse for deep IH (table 4), results improved using a combination of a special laser technique and topical timolol.62 A double-blind RCT currently recruiting patients compares topical timolol 0.5% and placebo, inclusion criteria are VLBW and diagnosis of IH; estimated completion date is September 2015 (

Table 4

Clinical studies with >10 study participants using topical timolol maleat

No pharmacokinetic (PK) data are available on the transdermal absorption of timolol. Systemic adverse effects rarely appear (see table 4). However, the skin of preterm infants has much less barrier function than that of older children; therefore, timolol absorption might increase considerably. Important to know, timolol is about six times more potent than propranolol.73 Up to now, no safety data are available for either preterm or term infants. PK of topical timolol should be evaluated urgently in these age groups.

Summarising currently available RCT data for topical timolol application to preterm infants is limited to one small study,61 both for treatment in the first postnatal weeks or beyond reaching TEA. There are neither PK data available nor data on long-term outcomes. However, it seems reasonable to consider treatment with topical timolol as a method with a considerably lower risk profile than systemic propranolol.


For preterm infants, evidence for IH treatment is lacking despite their high incidence. PK and clinical studies are warranted. Given the unknown long-term cognitive outcome of systemically used vasoactive beta blockers, local treatment like NCCT should be revisited as an alternative early intervention before the rapid growth phase starts. At present, most treatment decisions must be extrapolated from studies in older infants.


View Abstract


  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.