Incidence and Clinical Characteristics of Periocular Infantile Hemangiomas

Objective To report the incidence, demographics, and clinical findings among a population-based cohort of children with periocular infantile hemangiomas.

Methods The medical records of all patients (<19 years of age) diagnosed as having periocular infantile hemangiomas while residing in Olmsted County, Minnesota, from January 1, 1965, through December 31, 2004, were retrospectively reviewed.

Results Forty-three children were diagnosed as having periocular infantile hemangiomas during the 40-year period, yielding an incidence of 5.4 per 100 000 individuals younger than 19 years (95% CI, 3.8-7.1) or a birth prevalence of 1 in 1586 live births. Thirty children (70%) were female (P < .001). There was a history of maternal infertility in approximately 1 in 5 children and premature birth in 1 in 8 children. Twenty-six children (61%) had other abnormalities, including secondary hemangiomas in 9 (21%). Forty-one patients (95%) had unilateral disease, and 37 hemangiomas (86%) were located on the upper eyelid.

Conclusions In this population-based study, periocular infantile hemangiomas occurred in 1 in 1586 live births and were most prevalent on the unilateral upper eyelid of white female patients. Prevalent associations included maternal infertility and premature birth. Other abnormalities, including secondary hemangiomas in 1 in 5 children, were common in this cohort.

Infantile hemangiomas are the most common tumor of childhood, consisting of benign vascular proliferations of endothelial cells.¹ These abnormal proliferations affect approximately 1 in 10 children younger than 1 year² and have a predilection for the head and neck region.³ Risk factors for their development include female sex, white race, low birth weight or premature birth, and mothers who underwent chorionic-villus sampling during the prenatal period.^4-8

The reported prevalence of infantile hemangioma has been observed to be steadily increasing with the increasing rate of low-birth-weight infants, presumably due to escalated use of assisted reproductive technologies.^8,9 There is, however, no known population-based study on periocular infantile hemangiomas from the United States. The purpose of this study is to describe the incidence, demographics, and clinical characteristics among a population-based cohort of children diagnosed as having periocular infantile hemangioma during a 40-year period.

The medical records of all patients younger than 19 years who were diagnosed as having periocular infantile hemangiomas from January 1, 1965, through December 31, 2004, while residing in Olmsted County, Minnesota, were retrospectively reviewed. Institutional review board approval was obtained from Mayo Clinic and Olmsted Medical Group. Periocular infantile hemangioma was defined as any hemangioma that was located within the orbit or eyelids, inferior to the glabella, lateral to the nasal bone, and medial to the zygomatic arch. Potential cases were identified using the resources of the Rochester Epidemiology Project, a medical record linkage system designed to capture data on any patient-physician encounter in Olmsted County.¹⁰The population of Olmsted County is relatively isolated from other urban areas, and almost all medical care is provided to its residents by the Mayo Clinic, Olmsted Medical Group, and their affiliated hospitals.

A list of potential cases from the 2 institutions, generated by a comprehensive diagnostic code search, identified 649 patients. The codes searched included the terms hemangioma or benign neoplasm associated with the following descriptors: eyelid, lid, canthus, orbit, head, face, scalp, forehead, glabella, temporal region, zygomatic region, malar region, parotid region, cheek, skin, subcutaneous tissue, and unspecified site. Each record was meticulously reviewed by an ophthalmologist (B.G.M.) for confirmation of a periocular infantile hemangioma based on the listed criteria. Cases were excluded if the hemangioma was not located within the specified region, the patient was found to have a diagnosis other than the infantile hemangioma, or was diagnosed outside the period of this study. Forty-three patients met the inclusion criteria and were included in the study.

The 43 medical records were reviewed for demographics, including sex, race, date of diagnosis, and perinatal, developmental, medical, and familial histories. Clinical characteristics of the periocular hemangioma, including size, location, and final outcome, were reviewed.

The birth prevalence of periocular infantile hemangiomas was calculated using the number of live births occurring in Olmsted County from January 1, 1965, to December 31, 2004. The incidence of periocular infantile hemangiomas in this county was determined by using annual age- and sex-specific white population figures obtained from the US Census. The Poisson distribution was used in calculating the 95% CIs.

Forty-three new cases of childhood periocular infantile hemangiomas were diagnosed during the 40-year study period, yielding an annual incidence of 5.4 (95% CI, 3.8-7.1) per 100 000 patients younger than 19 years or a birth prevalence of 1 in 1586 live births. The demographic characteristics of the 43 patients are summarized in Table 1. There were 30 female patients (70%) (P = .01). Of the 30 patients reporting race, 28 (93%) were white and 2 (7%) were Hispanic. A history of premature birth occurred in 5 patients (12%), maternal infertility in 9 (21%), and amniocentesis in 3 (7%). The mean maternal and paternal ages recorded at childbirth for each patient were 27.6 and 29.0 years, respectively. A positive family history of vascular malformations was observed in 2 patients (5%), both of which had secondary hemangiomas in addition to the periocular lesion. The affected family member in both cases was a first-degree relative, with one having a history of an eyelid hemangioma and the other having a facial port wine stain.

Most patients had their conditions diagnosed by their pediatrician (35 [81%]) and treated by an ophthalmologist (21 [50%]) (Table 2). The locations of the hemangiomas in the 43-patient cohort are given in Table 3. Nineteen patients (44%) had hemangiomas of the right eyelid, 22 (51%) had hemangiomas of the left eyelid, and 2 (5%) had bilateral hemangiomas. Thirty-seven hemangiomas (86%) were located on the upper eyelid. In the 21 hemangiomas (49%) that had clinical measurements, the mean size of the hemangioma at diagnosis was 1.66 cm² (range, 0.04-7 cm²). Only 1 of the 43 patients (2%) reported bleeding of the hemangioma, whereas no patients had ulcerative changes or other adverse sequelae.

The occurrence of other, nonocular disorders among the 43 study patients is given in Table 4. Seventeen (40%) of the 43 patients had unremarkable medical histories, 3 (7%) had a secondary hemangioma in the head-neck-nonfacial region, 3 (7%) had a secondary hemangioma in the facial region, 2 (5%) had a secondary hemangioma in the buttock region, and 1 (2%) had a secondary hemangioma in the abdominal region.

This report, to the best of our knowledge, is the first population-based study on periocular infantile hemangiomas in the United States. In this population-based cohort of children diagnosed as having infantile hemangiomas during a 40-year period, periocular involvement occurred in 1 in 1586 live births. One in 5 had a history of maternal infertility, and 1 in 8 were born prematurely. Nine (21%) had an additional hemangioma outside the periocular region. Most patients were white females with unilateral upper eyelid involvement.

Infantile hemangiomas, located anywhere on the body, are reported to occur in 2.6% of newborns younger than 2 weeks and in 10% of children younger than 1 year.^2,11-14 Although most appear within the first year of life, 30% are evident at birth,¹² as was similarly observed in this cohort. Finn and coauthors³ and others¹⁵ further subclassified hemangiomas by location, reporting that 60% are found in the head and neck region, 25% in the truncal area, and 15% in the extremities. Haggstrom et al¹⁶ found similar distributions in a large prospective cohort study. Although most infantile hemangiomas are found within the head and neck region, the incidence from this study (5.4 children per 100 000 patients younger than 19 years) demonstrates the infrequency of an infantile hemangioma being located solely within the periocular region. Although previous studies^8,9 have reported a steadily increasing prevalence of infantile hemangiomas associated with an increasing rate of premature births, no such rise in prevalence was observed in this cohort.

The clinical features of this cohort confirm that females are more likely to develop hemangiomas than males by a ratio of 3:1. Both Drolet and coauthors¹⁷ and Stigmar and coauthors¹⁸ report a similar sex predominance, with the study by Stigmar et al reporting female involvement in 75% of their cohort. It is unclear why infantile hemangiomas are more common among females. Schwartz and coauthors¹⁹ hypothesize that this predominance may be an overestimate, stating that both their study and the study by Stigmar et al¹⁸ found females to be more likely than males to have larger periocular hemangiomas with more severe amblyopia, therefore requiring ophthalmic referral earlier and more frequently. However, because the current study is population based and includes all patients diagnosed as having periocular hemangioma in any department at our institution, our findings confirm that females are, unrelated to referral or other bias, more likely than males to develop periocular infantile hemangiomas.

Other risk factors for the development of infantile hemangiomas include race, prematurity and low birth weight, maternal infertility, clomiphene use, and chorionic-villus sampling with no correlation to gestational age at the time of invasive sampling.^4-8,20-23 Similar to other studies,^19,24 most patients from this cohort were white (28 patients of 30 reporting race). In this cohort, maternal infertility (9 [21%]) and premature birth (5 [12%]) were commonly associated risk factors, but low birth weight (n = 0), use of clomiphene (n = 1), and chorionic-villus sampling (n = 0) were not. Except for race, only 15 of our 43 patients (35%) had 1 of the previously reported risk factors.

Most of this patient cohort (37 [86%]) had lesions of the upper eyelid. Previous studies^24-27 show similar findings, although the reason for this predilection has yet to be determined. Nine of the 43 patients (21%) had secondary hemangiomas in addition to their periocular lesion. Twenty-nine percent of patients in a study by Haik et al¹² had secondary hemangiomas outside the periocular region, whereas Drolet et al¹⁷ reported that 20% of affected children, in general, will have multiple hemangiomas. Other comorbidities observed in this population include a family history of vascular malformations in 2 patients (5%). It has been previously reported that 32% of patients with infantile hemangioma have a first-degree relative with a vascular anomaly, with 12% specifically having a first-degree relative with an infantile hemangioma.²⁸ This strong genetic component was not observed in this cohort. Another recently described comorbidity is retinopathy of prematurity (ROP). Praveen et al²⁹ found that 16.8% of their cohort had infantile hemangiomas and ROP compared with 6.7% with hemangioma but no ROP. This finding was independent of birth weight and postnatal corticosteroid use. No patients in this cohort had ROP.

Ulceration is the most common complication of hemangiomas^16,20 and has been reported to occur in 5% to 16% of cases, more frequently in larger, segmental lesions that outgrow blood supply.^16,30,31 The location of the hemangioma is also a predictor of ulceration. Chamlin et al³¹ found that hemangiomas of the upper eyelid were the least likely to ulcerate. This finding, in combination with the smaller sizes of hemangiomas in this cohort, may explain why ulceration was not observed in any patients. In addition, the previous studies reporting ulceration rates of 5% to 16% were not population based and thus likely to be an overestimate.

Infantile hemangiomas are associated with PHACES (Posterior fossa malformations, Hemangiomas, Arterial anomalies, Cardiac defects and coarctation of the Aorta, Eye abnormalities, and Sternal abnormalities or ventral developmental defects) syndrome, a rare condition in which a patient has a segmental-type hemangioma along with structural and developmental anomalies.^32,33 Ophthalmic involvement occurs in anywhere from 16% to 21% of affected patients.^32-36 In contrast to the risk factors for infantile hemangiomas, PHACES syndrome is observed more frequently in singleton, term infants with normal birth weight and at an even higher female to male ratio of 8 to 9:1.^32,35 No patients within this cohort were diagnosed as having PHACES syndrome.

Besides the standard set of complications that can affect hemangiomas of any location, periocular hemangiomas are particularly of concern because of the risk for ocular complications, including amblyopia, strabismus, ptosis, proptosis, and optic nerve compression.^17,18,24,26 Because of the potential to develop these complications, it is recommended that patients with periocular hemangiomas be closely observed by an ophthalmologist. In this patient cohort, primary care physicians diagnosed most of the hemangiomas (35 [81%]), but only half (21 [49%]) were referred to ophthalmology for management. The ramifications of this low referral rate include unnecessary adverse visual outcome for patients that could have been prevented by earlier referral and diagnosis. Although some patients with periocular hemangiomas may appear to be at a low risk for adverse visual sequelae, an ocular evaluation is important to ensure the best outcome for the patient.

The findings of this study have several limitations. Its retrospective design is limited by nonstandardized and incomplete data collection. In addition, some infantile hemangiomas may be asymptomatic or subcutaneous, thereby going unnoticed by the patient's caretaker or physician. Although most patients in Olmsted County are treated by the 2 medical systems within the community, some residents may have sought care outside Olmsted County, thereby further underestimating the true incidence in this population. Our ability to generalize these findings to other populations is limited by the demographics of Olmsted County, a relatively homogeneous, semiurban, white population. Finally, the study was limited by incomplete medical records relating to the size, morphologic subtype, depth, and status of the hemangioma at initial and final examination.

In this population, periocular infantile hemangiomas occurred in 1 in 1586 live births and were most prevalent on the upper eyelid of white female patients. One in 5 children had a history of maternal infertility, and 1 in 8 were born prematurely. Most children had other, nonocular abnormalities, including a secondary hemangioma in 9 (21%).

Correspondence: Brian G. Mohney, MD, Department of Ophthalmology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 (mohney@mayo.edu).

Submitted for Publication: July 1, 2011; final revision received December 23, 2011; accepted January 2, 2012.

Funding/Support: This study was made possible in part by Rochester Epidemiology Project grant R01-AG034676 from the National Institute of Arthritis and Musculoskeletal and Skin Diseases and by an unrestricted grant from Research to Prevent Blindness, Inc (Drs Griepentrog and Mohney).