[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
December 15, 2008

Infantile Hemangiomas With Unusually Prolonged Growth Phase: A Case Series

Author Affiliations

Copyright 2008 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2008

Arch Dermatol. 2008;144(12):1632-1637. doi:10.1001/archderm.144.12.1632

Background  Most infantile hemangiomas (IHs) complete their proliferative growth phase before 9 months of age, but those with unusually prolonged growth create unique clinical challenges. We performed a retrospective case series of IHs with prolonged growth to further characterize these lesions and their treatment.

Observations  We identified 23 patients as having IHs with prolonged growth after 9 months of age, with growth to a mean age of 17 months. All of the IHs had a deep dermal to subcutaneous component, all had either segmental or indeterminate morphologic characteristics, and 39% involved the parotid gland. A total of 20 of 23 received prolonged treatment with systemic corticosteroids (mean duration of treatment, 11 months), and 9 of 20 received additional systemic therapies (vincristine sulfate and/or interferon alfa-2a or alfa-2b).

Conclusions  Prolonged growth was observed primarily in IHs with a deep component and segmental morphologic characteristics. Recognition of this subset of hemangiomas is important for clinicians, and further study of IHs may provide clues to their pathogenesis.

Infantile hemangiomas are the most common tumor of infancy and have a well-described natural history of rapid growth during early infancy followed by gradual involution. Although there are other types of vascular tumors sometimes referred to under the umbrella term hemangiomas (eg, lobular capillary hemangioma [pyogenic granuloma]), hereafter in this article the term hemangioma will refer to infantile hemangiomas specifically. The duration of the growth phase can be variable, but it has long been reported that most hemangiomas reach their maximal size within the first 6 to 10 months of life.1-3 The growth characteristics of hemangiomas were more rigorously evaluated in a recent large prospective study,4 which found that the early proliferative growth phase was essentially complete by 5 months of age, and overall growth was nearly always complete before 9 months of age. Although hemangiomas with growth after 9 months of age represent only a small subset, their continued growth can cause therapeutic challenges, functional sequelae, and considerable parental anxiety.

The classification of infantile hemangiomas has been based on their depth of soft-tissue involvement described as superficial, deep, or combined. Deep hemangiomas have been observed to have a longer growth phase compared with superficial hemangiomas.2,5 More recently, hemangiomas have also been characterized by their spatial configurations as segmental, localized, or indeterminate. Segmental hemangiomas have a geographic pattern over a cutaneous territory that frequently corresponds to a developmental segment.6 Previous studies7-9 have found that segmental hemangiomas are more likely to result in complications, such as ulceration, and are more likely to require treatment than localized hemangiomas. Segmental hemangiomas may also have a greater propensity to grow after 6 months of age.4 In addition, anatomic location may be associated with a greater propensity for aggressive growth and recalcitrance to therapy. Hemangiomas involving the parotid gland have been reported to involute more slowly and be less responsive to treatment, although the latter is controversial.9-12 This retrospective case series focusing on hemangiomas with unusually prolonged growth was undertaken to further characterize this small but important subset of hemangiomas.


Prolonged growth was defined as a proliferative growth phase past 9 months of age without treatment, or rebound growth after 1 year of age with treatment. Patients meeting inclusion criteria were identified by clinicians at the following participating institutions: the University of California, San Francisco; Texas Children's Hospital, Baylor College of Medicine, Houston; Hospital de la Santa Creu i Sant Pau, Barcelona, Spain; Children's Hospital Medical Center of Cincinnati, Cincinnati, Ohio; and Hospital de Pediatria, J. P. Garrahan, Buenos Aires, Argentina. Approval for the study was obtained from the University of California, San Francisco, committee on human research and the equivalent review boards at participating institutions as required.

Using medical records, including photographs of all patients, a retrospective medical chart review was performed to collect data on the clinical characteristics, treatment, and outcome of patients with infantile hemangiomas identified as having prolonged growth. Hemangiomas were diagnosed based on medical history and clinical appearance, aided by imaging only when clinically indicated. Size was reported as the surface area of involvement in centimeters squared using the longest diameter and the measurement perpendicular to it. Lesions were classified as superficial, deep, or combined and as segmental, localized, or indeterminate by unblinded author agreement using previously defined criteria.6,8,9 Segment numbers for facial lesions were assigned according to the patterns identified by Haggstrom et al6 with segment 1 (S1) corresponding to the frontotemporal region; segment 2 (S2), the maxillary region; segment 3 (S3), the mandibular region; and segment 4 (S4), the frontonasal region. Response to treatment was based on the clinical assessment of the physician(s) caring for these infants.


We identified 23 patients with infantile hemangiomas meeting our criteria for prolonged growth. The characteristics of these patients and their hemangiomas are summarized in the Table. Eighteen (78%) of the patients were female. The hemangiomas were all located on the head and neck region, although 3 (cases 3, 19, and 23) had lesions extending onto the trunk. Absolute size measurements were not available for all lesions because 6 were noted as being too large to accurately measure. Of the 17 cases with measurements available, the mean (SD) size was 126 (103) cm2 (range, 5-315 cm2). Eighteen (78%) were segmental lesions, and the other 5 (22%) were indeterminate. Segment 3 was the most commonly represented, involved in 11 cases (48%). Six (26%) were defined as deep lesions and 17 (74%) as combined lesions.

Characteristics of Infantile Hemangiomas With Prolonged Growth (PG)
Characteristics of Infantile Hemangiomas With Prolonged Growth (PG)

The parotid gland was involved in 9 patients (39%). Three (13%) had clinically significant airway involvement, but only 1 (4%) (case 23) required a tracheostomy. Other complications included ulceration in 12 cases (52%) and growth failure in 5 cases (22%). Three patients (13%) (cases 15, 21, and 23) required gastrostomy tube placement. One patient (4%) (case 3) had congestive heart failure, and 1 patient (4%) (case 21) was found to have a cardiomyopathy thought to be related to high cardiac output demand of the hemangioma or possibly secondary to treatment. None of the patients had hypothyroidism.

The age at last documented growth ranged from 10 to 44 months, with a mean (SD) age of 17.3 (8.0) months. Figure 1 demonstrates growth in case 4 at 6, 12, and 16 months of age; Figure 2, growth in case 14 at 9 and 15 months of age; and Figure 3, growth in case 18 at 4, 16, and 24 months of age.

Figure 1.
Case 4. Combined (superficial and deep), segmental infantile hemangioma involving left facial segment 2 at 6 months of age during taper of systemic corticosteroids (A), at 12 months of age off treatment (B), and at 16 months of age showing prolonged growth in the deep component (C).

Case 4. Combined (superficial and deep), segmental infantile hemangioma involving left facial segment 2 at 6 months of age during taper of systemic corticosteroids (A), at 12 months of age off treatment (B), and at 16 months of age showing prolonged growth in the deep component (C).

Figure 2.
Case 14. Deep, segmental infantile hemangioma involving left facial segment 3 at 9 months of age during the taper of systemic corticosteroids (A), and at 15 months of age after the steroid taper showing rebound growth, most notably in the parotid gland (B).

Case 14. Deep, segmental infantile hemangioma involving left facial segment 3 at 9 months of age during the taper of systemic corticosteroids (A), and at 15 months of age after the steroid taper showing rebound growth, most notably in the parotid gland (B).

Figure 3.
A patient with combined (superficial and deep) segmental infantile hemangioma involving bilateral facial segment 3. The patient at 4 months of age shortly after initiation of systemic corticosteroids (A), at 16 months of age after stopping treatment with systemic steroids and showing growth in the lower lip and left parotid gland (inset) (B), and at 24 months of age prior to initiation of interferon therapy with continued growth in the lower lip (C).

A patient with combined (superficial and deep) segmental infantile hemangioma involving bilateral facial segment 3. The patient at 4 months of age shortly after initiation of systemic corticosteroids (A), at 16 months of age after stopping treatment with systemic steroids and showing growth in the lower lip and left parotid gland (inset) (B), and at 24 months of age prior to initiation of interferon therapy with continued growth in the lower lip (C).

Twenty-one patients (87%) received systemic corticosteroids (CS), and 12 (52%) had additional therapies, including vincristine sulfate, interferon alfa-2a or alfa-2b, pulsed-dye laser, embolization, and debulking surgery. Two patients (9%) (cases 10 and 11) did not receive any treatments, and 1 patient (4%) received only a very brief course of treatment with discontinuation after 1 month because of problems with health insurance. The mean (SD) age at initiation of CS treatment was 3.2 (2.6) months (range, 3 weeks to 1 year). Starting dosages of CS ranged from 2.5 to 3.1 mg/kg/d, and 1 patient's dosage (case 17) was increased to 4 mg/kg/d shortly after initiation of treatment owing to poor response. The mean total duration of CS treatment was 11.2 (2.9) months, but CS treatment was intermittent rather than continuous in 10 patients. Thirteen of 19 patients (68%) had an initial decrease in size or growth cessation of their lesion, but the others had slowed or continued growth. (Data documenting initial response was lacking in 2 patients.) All patients had rebound growth during or after the CS taper, except 1 patient (case 3) who was given vincristine during the CS taper after no initial response. In several cases in which alternate-day steroids were used as part of a CS taper, parents reported growth on the “off” days when steroids were not administered. Common complications of CS treatment included Cushingoid features, irritability, and difficulty sleeping. Two patients (cases 17 and 23) had elevated blood pressure. Five patients had somatic growth delay that could have been the result of CS therapy, but conceivably may have been the result of the increased metabolic demands of a very large vascular tumor or difficulties in feeding because of a mass effect.

Vincristine was the most common second-line treatment, used in 6 patients. Two patients had central catheter infections, 1 caused by Candida, and 1 resulting in bacterial sepsis. Two patients received interferon therapy, both initiated after 18 months of age, without neurologic sequelae. The mean (SD) age at which all systemic therapy was complete was 19.9 (8.1) months.


This case series highlights the clinical characteristics and treatments of 23 infants with infantile hemangiomas with unusually prolonged growth. To our knowledge, there have been no previous case series focusing on the “outliers” with late, ongoing growth. A recent prospective study4 on the growth characteristics of 526 hemangiomas found that patients reached the late proliferative stage at a mean (SD) age of 6.2 (2.5) months, whereas the lesions reported herein grew until a mean (SD) age of 17.3 (8.0) months (range, up to 44 months).

All of the hemangiomas in this series were classified as segmental or indeterminate, a contrast to their incidence in a large prospective study9 in which only 453 of 1530 (29.6%) were classified as segmental or indeterminate. Segmental hemangiomas have been found to have higher rates of complications and need for treatment, and segmental morphologic characteristic is the single greatest predictor of risk of complications and need for treatment.7-9 Segmental hemangiomas are on average 10 times larger than localized hemangiomas,4 and our cases were no exception, with a mean (SD) area of involvement of 125 (103) cm2. The size and patterns of segmental hemangiomas on the face respect anatomic boundaries corresponding to embryologic neuroectodermal derivatives, suggesting an error in development as early as 6 to 8 weeks of gestation.6 Their more aggressive clinical behavior may hypothetically be attributed to a more pervasive developmental defect not present in localized hemangiomas.

In a recent prospective study,9 53% of hemangiomas had only a superficial component (Anita N. Haggstrom, MD, e-mail communication, November 2007), yet in our study all were either deep or both superficial and deep. Previous studies2,4,5 have suggested that deep hemangiomas may have a longer proliferative phase. In our cases, the component subjectively assessed to have prolonged growth was typically the deep component (Table and Figures 1-3). Our series also had an overrepresentation of hemangiomas involving the parotid gland (39%). Some authors10-12 have asserted that parotid hemangiomas involute more slowly and are more resistant to systemic therapy than other hemangiomas. The findings in our cases support the concept of true differences in behavior in deep hemangiomas, including those with parotid involvement, rather than simply being an observational bias, and raise the question of whether the milieu in deeper sites is more permissive for growth than in more superficial sites. Parotid glands may have a higher-than-normal rate of blood flow,11 and Blei and Rutkowski13 noted high blood flow in hemangiomas at certain sites, particularly the parotid gland, but also the upper extremity, upper lip, and scalp, and suggested that they are “transiently arterialized,” causing slower involution than at other anatomic sites.

Nearly all patients were treated with systemic CS, and a substantial minority received other systemic treatments. Most responded initially to CS but had rebound growth during or after steroid taper. This occurred despite treatment with CS for prolonged periods of time; the mean duration of treatment was 11 months—substantially longer than in other reported case series in which the mean duration of treatment was 6 to 8 months.11,14,15 Although the mechanisms of action of CS are only partially understood,16,17 Ritter et al18 proposed that infantile hemangioma involution stems from immune recognition of the aberrant hemangioma vascular phenotype, inciting an immune response targeted at hemangioma endothelial cells. Based on this model and our findings, we hypothesize that segmental hemangiomas represent a more fundamental developmental defect than localized hemangiomas and that those involving deeper tissues, particularly the parotid gland, are more impervious to immune recognition and sustained involution. Identifying a therapy that would function by triggering immune recognition of hemangioma endothelial cells may be a possible therapeutic strategy.

This study has several limitations related to its retrospective design. There is a potential for case selectivity based on recall bias. Data were accrued via medical chart review and were not documented in a systematic manner, and several patients were initially treated at other institutions. Some data required subjective clinical assessment, such as response to treatment and the characterization of lesions as deep or superficial, and as segmental or localized. Previous studies,7,9 however, have shown good interobserver agreement for these classifications. Further large prospective studies following lesions at regular intervals are needed to more rigorously document growth and treatment of infantile hemangiomas, including those with unusually prolonged growth.


Most infantile hemangiomas follow a well-described natural history of proliferation within the first 5 to 6 months of life, followed by slow, gradual involution. Some hemangiomas, however, demonstrate a more prolonged growth pattern and pose distinct clinical challenges. In this study, prolonged growth was observed primarily in hemangiomas with a deep component and segmental morphologic characteristics, types of hemangiomas that, although not rare, are present in a minority of cases. Although most infantile hemangiomas with these characteristics have typical growth patterns, this case series emphasizes that a small but clinically important subset will continue to grow. Recognition of this subset can alert parents and health care providers to consider more aggressive treatment for these hemangiomas. In addition, infantile hemangiomas with prolonged growth may provide clues to the pathogenesis of hemangiomas and the mechanisms of their involution, and therefore warrant further investigation.

Correspondence: Ilona J. Frieden, MD, Department of Dermatology, University of California, San Francisco, Box 0316, San Francisco, CA 94143-0316 (FriedenI@derm.ucsf.edu).

Accepted for Publication: April 1, 2008.

Author Contributions: Drs Brandling-Bennett and Frieden had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Brandling-Bennett and Frieden. Acquisition of data: Brandling-Bennett, Metry, Baselga, Lucky, Adams, Cordisco, and Frieden. Analysis and interpretation of data: Brandling-Bennett, Metry, Lucky, and Frieden. Drafting of the manuscript: Brandling-Bennett and Frieden. Critical revision of the manuscript for important intellectual content: Brandling-Bennett, Metry, Baselga, Lucky, Adams, and Cordisco. Statistical analysis: Cordisco. Study supervision: Adams, Cordisco, and Frieden.

Financial Disclosure: None reported.

Drolet  BAEsterly  NBFrieden  IJ Hemangiomas in children.  N Engl J Med 1999;341 (3) 173- 181PubMedGoogle ScholarCrossref
Esterly  NB Cutaneous hemangiomas, vascular stains and malformations, and associated syndromes.  Curr Probl Pediatr 1996;26 (1) 3- 39PubMedGoogle Scholar
Bowers  REGraham  EATomlinson  KM The natural history of the strawberry nevus.  Arch Dermatol 1960;82 (4) 667- 679Google ScholarCrossref
Chang  LCHaggstrom  ANDrolet  BA  et al.  The growth characteristics of infantile hemangiomas: implications for management.  Pediatrics 2008;122 (2) 360- 367PubMedGoogle ScholarCrossref
Nakayama  H Clinical and histological studies of the classification and the natural course of the strawberry mark.  J Dermatol 1981;8 (4) 277- 291PubMedGoogle Scholar
Haggstrom  ANLammer  EJSchneider  RAMarcucio  RFrieden  IJ Patterns of infantile hemangiomas: new clues to hemangioma pathogenesis and embryonic facial development.  Pediatrics 2006;117 (3) 698- 703PubMedGoogle ScholarCrossref
Waner  MNorth  PEScherer  KAFrieden  IJWaner  AMihm  MC  Jr The nonrandom distribution of facial hemangiomas.  Arch Dermatol 2003;139 (7) 869- 875PubMedGoogle ScholarCrossref
Chiller  KGPassaro  DFrieden  IJ Hemangiomas of infancy: clinical characteristics, morphologic subtypes, and their relationship to race, ethnicity, and sex.  Arch Dermatol 2002;138 (12) 1567- 1576PubMedGoogle ScholarCrossref
Haggstrom  ANDrolet  BABaselga  E  et al.  Prospective study of infantile hemangiomas: clinical characteristics predicting complications and treatment.  Pediatrics 2006;118 (3) 882- 887PubMedGoogle ScholarCrossref
Metry  DWHebert  AA Benign cutaneous vascular tumors of infancy: when to worry, what to do.  Arch Dermatol 2000;136 (7) 905- 914PubMedGoogle ScholarCrossref
Blei  FIsakoff  MDeb  G The response of parotid hemangiomas to the use of systemic interferon alfa-2a or corticosteroids.  Arch Otolaryngol Head Neck Surg 1997;123 (8) 841- 844PubMedGoogle ScholarCrossref
Greene  AKRogers  GFMulliken  JB Management of parotid hemangioma in 100 children.  Plast Reconstr Surg 2004;113 (1) 53- 60PubMedGoogle ScholarCrossref
Blei  FRutkowski  M Transiently arterialized hemangiomas: relevant clinical and cardiac issues.  Lymphat Res Biol 2003;1 (4) 317- 320PubMedGoogle ScholarCrossref
Boon  LMMacDonald  DMMulliken  JB Complications of systemic corticosteroid therapy for problematic hemangioma.  Plast Reconstr Surg 1999;104 (6) 1616- 1623PubMedGoogle ScholarCrossref
George  MESharma  VJacobson  JSimon  SNopper  AJ Adverse effects of systemic glucocorticosteroid therapy in infants with hemangiomas.  Arch Dermatol 2004;140 (8) 963- 969PubMedGoogle ScholarCrossref
Frieden  IJHaggstrom  ANDrolet  BA  et al.  Infantile hemangiomas: current knowledge, future directions: proceedings of a research workshop on infantile hemangiomas, April 7-9, 2005, Bethesda, Maryland, USA.  Pediatr Dermatol 2005;22 (5) 383- 406PubMedGoogle ScholarCrossref
Sadan  NWolach  B Treatment of hemangiomas of infants with high doses of prednisone.  J Pediatr 1996;128 (1) 141- 146PubMedGoogle ScholarCrossref
Ritter  MRMoreno  SKDorrell  MI  et al.  Identifying potential regulators of infantile hemangioma progression through large-scale expression analysis: a possible role for the immune system and indoleamine 2,3 dioxygenase (IDO) during involution.  Lymphat Res Biol 2003;1 (4) 291- 299PubMedGoogle ScholarCrossref