Capillary hemangiomas in childhood commonly involve the eyelids and orbit, typically presenting during the first 2 months after birth.1'3 Their growth is most rapid during the first 6 months of life, after which they tend to stabilize and to involute gradually. Resolution with minimal residual skin changes has been noted in 60% of cases by age T.2"6 For unknown reasons, hemangiomas affect females more than males.1,3,4,7
Amblyopia occurs in association with adnexal hemangiomas when they occlude the pupillary axis, cause strabismus, or induce astigmatism.3'4·7'9 The last mechanism seems by far the most common, and cylinders of high degree are5-7'9 characteristically oriented with their plus axes directed at the lesion. From 20% to 43% of eyes with adnexal hemangiomas have anisometropia, and most of these are considered at risk of amblyopia during the critical period.4'7·9 In some cases, especially those with occlusion of the pupil and severe astigmatism, the amblyopia can be profound.
Chiefly because of the risk of visual loss, but also to relieve the disfigurement caused by more severe lesions, other authors have described various treatments to hasten their resolution. Early efforts included cryotherapy, irradiation, and intralesional sclerosing agents.2'8·10"12 These largely have been abandoned because of inadequate effectiveness or unacceptable side effects. Subsequently, intralesional steroid injection has been advocated as the preferred method of treatment. Among the advantages noted were effectiveness, simplicity, repeatability, and apparent safety.11
During the past decade, however, several reports of side effects have caused some to question the safety of intralesional steroids. Among these were retention of subcutaneous steroid deposits, foreign body granuloma formation, cutaneous telangiectasia, subcutaneous fat atrophy, skin depigmentation, adrenal suppression, full-thickness eyelid necrosis, and both unilateral and bilateral retinal artery embolization.3·11'16 Systemic steroids and surgical excision therefore have been considered recently, along with experimental treatments.2·14
We have found that the majority of children with adnexal hemangiomas can be managed conservatively, with glasses for anisometropia and patching for amblyopia. We present our experience with such patients and compare the visual, refractive, and cosmetic results of those managed with and without steroid injections.
PATIENTS AND METHODS
Between 1984 and 1993, 23 children with adnexal hemangiomas were examined on the pediatrie ophthalmology service at the Albany Medical College. Of these, 8 had no significant astigmatism, strabismus, or pupillary occlusion.
Clinical Data for Both Patient Groups
These patients were considered not at risk of amblyopia and received no treatment other than observation. One child had exotropia in the opposite eye and received patching of the eye with a small upper lid hemangioma. Because of threatened occlusion of the pupillary axis, 1 child received oral steroids, 3 received intralesional steroids, and 1 received steroids by both routes.
The remaining nine patients had astigmatism greater in the affected eye with the plus axes oriented toward the lesion. None had strabismus or pupillary occlusion. The pediatrie ophthalmologists caring for these children had, by chance, divergent indications for steroid injection of such patients. One chose conservative treatment with glasses and/or patching in all four cases he managed. Five other children, in addition to treatment with glasses and/or patching, received steroid injection. These nine cases were investigated to determine the effectiveness of steroid injection. As indicated in the Table, the two groups were similar in age at presentation, size and location of lesions, amount of induced astigmatism, and length of follow up.
In all cases, steroid injections were performed according to the method of Kushner.11 An equal mixture of betamethasone 6 mg/cc and triamcinolone 40 mg/cc was prepared in a syringe that was then fitted with a 27 gauge, 0.5-inch needle. Pupils were dilated before induction of anesthesia, and indirect ophthalmoscopy was performed immediately following injection. Multiple injection sites were used as needed, and approximately 2 cc of the mixture was injected. Attempts were made to concentrate the steroid in the thickest part of the lesion, but covering most of the mass anterior to the orbital septum. Before each injection, the needle was drawn back to prevent infusion into large vascular spaces. No recognized complications ensued.
All patients underwent standard clinical examination at intervals determined by age, patching intensity, and change in appearance of the mass or in refraction. Cycloplegic refractions were performed following instillation of one drop of proparacaine HCl 0.5% and a combination of tropicamide 1% and cyclopentolate 2%. Glasses were prescribed for greater than 1 diopter of induced cylinder, taken as the difference in cylinder between the two eyes. Patching was performed based on preferential looking or Teller acuity during the prescribed period and on recognition acuity in verbal patients.
As indicated in the Table, the initial induced cylinder ranged 0.50 to 5.00 D (mean 2.20 D) in the injection group and from 0.75 to 4.00 D (mean 2.00 D) in the conservative group. Glasses were prescribed for a total of four patients (3, 4, 6, and 7), beginning at 3 to 13 months of age depending on the age at presentation. Prescriptions were changed approximately annually. At the final examination, induced cylinders had decreased in four of the five patients receiving injections. The change in this group ranged from an increase of 0.50 D to a 3.50 D decrease (mean decrease of 1.10 D). In the conservatively treated group, induced cylinders decreased in three of the four patients. The change in this group ranged from an increase of 0.75 D to a decrease of 3.00 D (mean decrease of 1.40 D).
Patching was prescribed in all patients for 10% to 50% of waking hours, with similar patching regimens in injectiontreated and conservatively treated groups. In the injectiontreated group, final visual acuities indicated 3 patients with no amblyopia, one with two lines of amblyopia, and 1 with nine lines (mean 2.2 lines). In the conservative group, final visual acuities were equal in 2 patients, 1 had one line of amblyopia, and 1 had 1.5 lines (mean 0.6 lines). Patient 5, from the injection- treated group, was exceptionally noncompliant with both glasses and patching. Her final visual acuity was 20/400 in the affected eye.
The indications for steroid injection of adnexal hemangiomas in infancy are unclear. Some studies follow Kushner's original criteria, including only patients with pupillary occlusion, strabismus, and anisometropia threatening to cause amblyopia.1'3'5*11·16·17 Other authors have included cosmetic indications.5·16 Still others have been discouraged by the incidence of serious complications and have recommended withholding injections in favor of alternative treatments.2·14 Reasoning that all intralesional steroid injections are intravascular and have access to the orbital circulation, Deans and associates2 prefer surgical excision in selected cases. Ruttum et al14 recommend surgery, oral steroids, or alpha interferon.
Our experience suggests that conservative management, with glasses and patching as indicated, may be successful in most cases. Steroid injection, systemic steroid treatments, and surgical excision are all reasonable alternatives for lesions that threaten the visual axis. The deprivation amblyopia, that can be expected in such cases likely will be severe and untreatable unless the eye is opened. Astigmatism, on the other hand, can be eliminated effectively with glasses in most infants. Anisometropic amblyopia may necessitate steroid injections. We have found preferential looking and Teller acuity measurements helpful in titrating occlusion in these infants. It is noteworthy that 8 of our 23 patients had no evidence of amblyopia and needed no intervention at all.
We recognize that our four conservatively treated and five injected anisometropes were not randomized before treatment. This is a small series, potentially biased in ways we cannot identify. However, the age, induced astigmatism, and length of follow up of the two groups are all similar and their treatment was roughly concurrent. The two groups differed only in that, coincidentally, their ophthalmologists had different indications for steroid injection. The visual results and decrease in induced cylinder were similar in the two groups. In all patients, the hemangiomas gradually decreased in size, and in older children, were scarcely apparent.
We endorse steroid injection as a reasonable treatment for children with threatened occlusion of the visual axis and for children with refractory amblyopia. Conservative management of induced astigmatism and anisometropic amblyopia is likely to be successful.
1. Boyd MJ, Collin JR. Capillary hemangiomas: an approach to their management. Br J Ophthalmol. 1991;75:298-300.
2. Deans RM, Harris GJ, Kivlin JD. Surgical dissection of capillary hemangiomas: an alternative to intralesional corticosteroide. Arch Ophthalmol. 1992; 110: 1743-1747.
3. Morrell AJ, Willshaw HE. Normalisation of refractive error after steroid injection for adnexal hemangiomas. Br J Ophthalmol. 1991;75:301-305.
4. Stigoear G, Crawfbrd JS, Ward CM, Thomson HG. Ophthalmic sequelae of infantile hemangiomas of the eyelids and orbit. Am J Ophthalmol. 1978;85:806-813.
5. Nelson LB, Melick JE, Harley RD. Intralesional corticosteroid injections for infantile hemangiomas of the eyelid. Pediatrics. 1984;74:241245.
6. Weise AH. Adrenal suppression after corticosteroid injection of periocular hemangiomas. Am J Ophthalmol. 1989;107:518-22.
7. Haik BG, Jakobiec FA, Ellsworth RM, Jones IS. Capillary hemangioma of the eyelids and orbit: an analysis of the clinical features and therapeutic results in 101 cases. Ophthalmology. 1979;86:760-789.
8. Boga S, Simon JW, Krohel GB, Nelson LB. Astigmatism associated with adnexal masses in infancy. Arch Ophthalmol. 1987;105:13681370.
9. Robb RM. Refractive errors associated with hemangioma of the eyelids and orbit in infancy. Am J Ophthalmol. 1977;83:52-58.
10. niingsworth RS. Thoughts on treatment of strawberry naevi. Arch Dis Child. 1976;51: 138-140.
11. Kushner BJ. The treatment of periorbital infantile hemangioma with intralesional corticosteroid. Ophthalmic Plast Recensir Surg. 1985;76:517-524.
12. Cogen MS, Elees FJ. Eyelid depigmentatíon following corticosteroid injection for infantile ocular adnexal hemangioma. J Pediatr Ophthalmol Strabismus. 1989;26: 35-38.
13. Droste PJ, Ellis FD, Sondhi N, Helveston EM. Linear subcutaneous fat atrophy after corticosteroid injection of pe riocular hemangiomas. Am J Ophthalmol. 1988; 105: 65-69.
14. Ruttum MS, Abrams GW, Harris GJ, Ellis MK Bilateral retinal embolization associated with intralesional corticosteroid injection for capillary hemangioma of infancy. J Pediatr Ophthalmol Strabismus. 1993;30:4-7.
15. Sutula FC, Glover AT. Eyelid necrosis following corticosteroid injection for capillary hemangioma. Ophthalmic Surg. 1987; 18: 103- 105.
16. Shorr N, Seiff SR. Central retinal artery occlusion associated with periocular corticosteroid injection for juvenile hemangioma. Ophthalmic Surg. 1986:17:229-231.
17. Brown BZ, Huflaker G. Local injection of steroids for juvenile hemangiomas which disturb the visual axis. Ophthalmic Surg. 1982; 13:630-636.
Clinical Data for Both Patient Groups