Article Text

Download PDFPDF

Periorbital haemangiomas
  1. W V Good
  1. Correspondence to: W V Good Smith Kettlewell Eye Research Institute, 2318 Fillmore Street, San Francisco, CA 94115, USA;

Statistics from

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.

Paediatric ophthalmology’s stepchild?

Every paediatric ophthalmology practice cares for infants and young children with periorbital haemangioma, and yet there is very little evidence based research to guide treatment. Options for management generally include topical, intralesional or systemic steroids, surgical resection, or no treatment at all, as haemangioma usually involute after a period of growth. When tumours block the visual axis, aggressive intervention is mandatory. In almost every other case, management is based on anecdote and clinical judgment and, in many cases, is optional or elective.

The aetiology of periocular haemangioma is also debated. Current opinion holds that haemangiomas occur more frequently in premature infants, yet older research on this subject suggests that this finding is caused by ascertainment bias—that is, premature infants are observed more often than full term infants, leading observers to conclude that premature infants develop haemangioma more commonly.1,2 Why do haemangioma have a predilection for areas around the eyes, or do they? The astigmatism seen with periocular haemangioma is usually attributed to pressure effect of the tumour, but is there evidence to support this conclusion, or to validate that this is always the cause of astigmatism? Why does astigmatism often persist after the tumour regresses? Which locations for tumours are most likely to cause or contribute to amblyopia? Again, the answer to these questions is not fully known.

Such is the state of the practice of medicine as it pertains to these tumours: many questions with far too few answers. And so it is ironic and fitting that Ranschod and colleagues in their paper reported in this issue of the BJO (p 1134) shed no light on management of periorbital haemangioma, despite a thorough review of the literature and a thoughtful meta-analysis of research on the subject. Rather than criticise the authors for taking readers of the BJO down a blind alley, we should be grateful to them for identifying a very important and generally disregarded problem, the periorbital haemangioma. While this tumour continues to occur at a significant frequency, and to threaten vision and appearance, its epidemiology, cause, biology, and management are in need of further study.

Should the reader then conclude that the periocular haemangioma is, indeed, a sort of stepchild in paediatric ophthalmology research, largely ignored and second fiddle to other conditions such as amblyopia and strabismus? If the answer to this is affirmative, then it may be useful to explore the means to beat this condition. One reason might be that the pathology and effects of these tumours generally fall outside the purview of many of us. The periorbital haemangioma, after all, is a tumour, often responsible for cosmetic changes. Ophthalmologists generally do not have experience with laser treatments for haemangioma, and may be happy enough to defer systemic treatment implementation to paediatricians and dermatologists. Consider, too, that very serious side effects from ophthalmologist led treatment have been reported, including blindness associated with intralesional injection.

While this tumour continues to occur at a significant frequency, and to threaten vision and appearance, its epidemiology, cause, biology, and management are in need of further study.

I believe that a different reason exists for the relative scientific neglect of haemangioma and for the low number of controlled clinical trials on this subject. The tumour is protean, its biology and growth variable, and its location around the eyes highly variable in terms of extent, location, and size. These facts make it very difficult to compare treatments at all. How can treatment of a globular haemangioma of the upper outer lid be compared to the same treatment for the same type and size tumour of the upper inner lid? Never mind that the haemangioma could be diffuse, affect both lids, start at an earlier age in some infants, occur in a premature infant, or occur in conjunction with tumours in other locations. All of these factors could influence the effect of the haemangioma on refractive error, amblyopia, tumour growth, and so on. Scientists have understandably avoided controlled studies on this problem, because experimental design is difficult.

Consequently, case reports have provided the foundation for knowledge about periocular haemangioma and their management. Case reports and series are often maligned as second class science, anecdotal, uncontrolled, and failing to test a hypothesis. But without case reports noting that astigmatism sometimes improves with intralesional injections, or of side effects of treatment, including adrenal suppression and blindness, clinicians would have very little data at all to guide management of haemangioma. Case reports will continue to be important to describe effects of treatment for this condition.

Lastly, we should try to imagine the design of a controlled clinical trial to test treatment for periorbital haemangioma. This is the charge set forth by Ranshod, and it is not an easy one. Controlling time of onset of haemangioma, size, location, degree of astigmatism at outset, premature status, and presence or absence of intermittent occlusion are just a few of the issues. All of these variables could be controlled in a large enough trial with, perhaps, a photographic reading centre where masked investigators would interpret external findings.

The recent explosion of developmental and molecular biology research offers an intriguing and different venue for research on this subject, too. The haemangioma’s biological behaviour dictates its effects on vision. This is obvious in the case of occlusion. The large and rapid growing tumour is more likely to cause occlusion, amblyopia, and astigmatism, whether or not one ascribes to the pressure theory of haemangioma induced astigmatism. Tumour biology could influence astigmatism and cosmesis in more subtle ways, as well. Knowledge that a given haemangioma has biological characteristics linked to rapid growth, or astigmatism, or significant cosmetic changes, would help guide clinical intervention. An amalgamation of clinical and basic science research for the study of management of periorbital haemangioma could lead to earlier detection of problematic tumours.

Let’s hope that Ranshod and colleagues, and others, carry forth their concerns about this neglected problem and work to rectify the current situation, where a shortage of evidence is used to support clinical decision making. The periorbital haemangioma deserves full citizenship in the world of paediatric ophthalmology research.

Note in Proof

Paediatric ophthalmology’s stepchild?


Linked Articles