Solitary retinal capillary hemangioma in a patient without von Hippel-Lindau disease. Note the prominent feeder vessels.
Singh AD, Shields JA, Shields CL. Solitary Retinal Capillary HemangiomaHereditary (von Hippel-Lindau Disease) or Nonhereditary?. Arch Ophthalmol. 2001;119(2):232-234. doi:10-1001/pubs.Ophthalmol.-ISSN-0003-9950-119-2-ecs90255
To estimate the probability of von Hippel-Lindau (VHL) disease in patients with solitary retinal capillary hemangioma.
Risk estimation was performed mathematically on the basis of the Bayes theorem using the published data on prevalence of VHL disease, prevalence of solitary retinal capillary hemangioma, and age-dependent penetrance of VHL disease.
The probability of VHL disease in patients with solitary retinal capillary hemangioma was estimated to be 46%. When adjusted for age at the time of diagnosis, the risk varied from 45% in the young age group (≤10 years) to 0.5% in the older age group (61-70 years).
Approximately half of the patients with solitary retinal capillary hemangioma are expected to have underlying VHL disease and the risk progressively diminishes with increasing age at diagnosis. Detailed clinical evaluation in such patients is recommended using standard screening protocols. Knowledge of the age-dependent risk for VHL disease can help the clinician modify recommendations regarding systemic and genetic testing.
SINCE THE original description by Fuchs in 18821 and subsequent observations by Collins,2 von Hippel,3 and Coats,4 the clinical features and familial nature of retinal capillary hemangiomas have been recognized. The association of cerebellar and retinal capillary hemangiomas as a heritable entity was first reported by Lindau.5 In 1964 Melmon and Rosen6 reported cases of von Hippel disease and Lindau disease with overlapping ophthalmic, central nervous system, and visceral manifestations, establishing the clinical spectrum and diagnostic criteria of "von Hippel-Lindau" disease (VHL).6 Since then, detailed studies on the clinical features and the natural history of cohorts with VHL disease have been published, improving our understanding of the disease.7- 9
Based on extensive clinical and genetic studies of VHL disease, it is now believed that tumor formation in VHL disease follows the "2-hit model," initially hypothesized by Knudson for retinoblastoma.10- 13 According to this model, 1 allele is constitutionally inactivated (first hit) and the second allele is inactivated in specific tissues (second hit).12 At present, clinical screening protocols and genetic testing for detection of VHL disease mutations are available.7,14,15 The detection of constitutional or germline mutation is not only helpful in establishing the diagnosis but also may have prognostic implications since there is a correlation of clinical features with the type of mutation (genotype-phenotype correlation).16
Retinal capillary hemangioma (RCH), the most frequent and the earliest manifestation of VHL disease, has a mean age at diagnosis of 25 years (Figure 1).7,17- 21 Solitary RCH may be the initial manifestation of VHL disease (first hit in the germline and second hit in the retinal cell) or may occur in a patient with no risk for VHL disease, owing to the occurrence of both first and second hits in the retinal cell.17,22 Are patients with solitary RCH without other systemic stigmata affected by the hereditary (VHL disease) form or the nonhereditary form? In this study, using published VHL disease prevalence data, age-dependent penetrance, and the Bayes theorem, we conclude that either can be the case, and we estimate the probability of the hereditary form.11,23- 25
The prevalence estimates for VHL disease have been collated and published by Maher and colleagues11,24,25 and are based on methodical study of the East Anglian population in the United Kingdom. The prevalence of VHL disease (PVHL) in the study population was estimated to be 19 per million (19 × 10−6).11 The prevalence of solitary RCH (PRCH)in the same population is reported to be 9 per million (9 × 10−6).24 Additionally, the prevalence of solitary RCH in patients with VHL disease P(RCH / VHL) is derived from a recently published study in which 35 (22%) of 156 patients with VHL disease in whom the RCH could be counted accurately had solitary RCH.25
In this report, we chose to mathematically estimate the risk for VHL disease in patients manifesting solitary RCH based on the above known parameters. The systemic features of VHL disease have a tendency to manifest later in life (age-dependent penetrance) with almost complete penetrance by age 60 years.11
To calculate the prevalence of VHL disease in patients with solitary RCH, P(VHL | RCH) from Bayes theorem,23 we know that
Equation 2 can be arranged into
Substituting this into equation 1 we get
Substituting previously published values in equation 4, we get
These calculations show that the probability of VHL disease in patients with solitary RCH is 46%. The estimated risk of 0.46 was adjusted for each age group (at time of diagnosis) by multiplying the estimated risk with the likelihood of not manifesting VHL disease by that age group (1 − penetrance). For young patients (≤10 years) the estimated risk was 0.45 and for the older age group (61-70 years) the estimated risk was only 0.005. The estimated risks for all age groups are presented in Table 1.
Von Hippel-Lindau disease is a multisystem disorder with varied features such as RCH, spinal/cerebellar hemangioma, pheochromocytoma, renal cell carcinoma, multiple renal and pancreatic cysts, and endolymphatic sac tumors manifesting over many years.26 Because RCH is the most common and earliest manifestation of VHL disease, often as a result of new germline mutations,26 patients with RCH must be evaluated for VHL disease.18- 21 When evaluating a patient with RCH, the presence of multiple retinal RCH lesions(2 or more), other manifestations of VHL disease, or positive family history indicate the presence of underlying VHL disease. However, even in patients with only solitary RCH and lacking other systemic or family stigmata of VHL disease, the possibility of underlying VHL disease with its associated morbidity must be considered.
The prevalence of RCH in the VHL disease population in various studies has been observed to be approximately 60%.7,18- 20 The risk of underlying VHL disease in patients with solitary RCH has not been previously quantified until now. In this report, we have mathematically estimated the risk of underlying VHL disease in patients with solitary RCH. Based on these parameters, we found that the probability of VHL disease in patients with solitary RCH was 46%, but varied from 45% in preteen patients to less than 1% in those older than 60 years (Table 1).
All prevalence estimates used in this study were based on a comprehensive study of patients derived from the East Anglian population in the United Kingdom.11,24 Because of the ethnic and racial differences in various populations, the risk estimates derived in our study may not be applicable to populations elsewhere. The estimated prevalence of solitary RCH in the general population is perhaps an underestimate since some of the asymptomatic RCH cases (in the absence of VHL disease) may not be detected. Additionally, our calculations do not account for reduced survival in patients with VHL disease.7 In solitary RCH cases without VHL disease (which account for about half of all solitary RCH cases), it is assumed that there is no effect on patient survival.
Our results indicate that approximately half of the patients with solitary RCH are expected to have underlying VHL disease and that detailed clinical evaluation in patients with solitary RCH is recommended using standard screening protocols.14 Genetic testing with high detection rate is currently available for both sporadic and familial cases of VHL disease and should be offered to patients after detailed counseling.15,27 Knowledge of the age-dependent risk for VHL disease in a patient manifesting solitary RCH can help the clinician modify recommendations regarding systemic and genetic testing.
Accepted for publication June 23, 2000.
This investigation was supported by the Sarah B. Kant Fund (Dr Singh) Philadelphia, Pa, Paul Kayser Award of Merit in Retinal Research, Houston, Tex (Dr J. A. Shields), the Macula Foundation, New York, NY (Dr C. L. Shields), and the Eye Tumor Reseach Foundation Inc, Philadelphia (Dr C. L. Shields).
We are grateful to Larry A. Donoso, MD, and Alfred G. Knudson Jr, MD, PhD, for reviewing the manuscript.
Since submission of our manuscript, a study with similar results has been published.28
Corresponding author and reprints: Arun D. Singh, MD, Oncology Service, Wills Eye Hospital, 900 Walnut St, Philadelphia, PA 19107 (e-mail: firstname.lastname@example.org).