Introduction

Neuroblastoma is the most common extracranial solid tumor among children under the age of 5 years, with a published population-based incidence of one in 6000 children.1 Neuroblastic tumors, derived from primordial neural crest cells, are characterized by defective catecholamine synthesis, frequently resulting in the accumulation and excretion of the catecholamine metabolites homovanillic acid (HVA) and vanillylmandelic acid (VMA).2, 3 Landmark papers concerning the use of urine catecholamine samples to detect neuroblastoma and related neural crest tumors have estimated that urinary levels of HVA and VMA were elevated in approximately 90–95% of cases.3, 4 However, current reports regarding the detection of neuroblastoma in cases of pediatric Horner syndrome have detected normal catecholamine levels at presentation in a number of cases despite the presence of neural crest tumors.5, 6

The important role ophthalmologists have in the detection and management of neuroblastoma is underscored by the numerous ocular manifestations of this disease, including proptosis,1, 7, 8, 9 periorbital ecchymosis,1, 7, 8, 9 Horner syndrome,5, 6, 10, 11, 12, 13, 14 opsoclonus/myoclonus,15 ocular motility defects,7, 16 ptosis,16 and blindness8, 17 Of special ophthalmic interest has been the detection of underlying neuroblastoma in cases of pediatric Horner syndrome, including the role urine catecholamine levels have in the recommended ophthalmic work-up.1, 5, 6, 10, 11, 12, 14, 18, 19 The purpose of this paper is to report the percentage of neuroblastoma cases presenting with elevated catecholamine levels among a cohort of patients <19 years diagnosed as residents of Olmsted County, Minnesota, over a defined 40-year period, and to suggest an ophthalmic work-up in cases of idiopathic pediatric Horner syndrome.

Materials and methods

The medical records of all pediatric (<19 years) patients residing in Olmsted County when diagnosed with neuroblastoma from 1 January 1969 through 31 December 2008 were retrospectively reviewed. Potential cases of neuroblastoma 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, Minnesota.20 The racial distribution of Olmsted County residents in 1990 was 95.7% Caucasian, 3.0% Asian-American, 0.7% African-American, and 0.3% each Native American and other. The population of this county (106 470 in 1990) is relatively isolated from other urban areas, and essentially all medical care is provided to residents by Mayo Clinic or Olmsted Medical Group and their affiliated hospitals.

All diagnoses were entered into the Rochester Epidemiology Project database and residency status was verified by specially trained personnel. Children not living in Olmsted County at the time of their diagnosis were excluded. Neuroblastoma was defined in this study by clinical diagnosis, based on tumor biopsy results, catecholamine levels, and other systemic findings. Data collected included age at diagnosis, tumor stage, HVA, and VMA levels at presentation, and final outcome.

Results

A total of 14 patients <19 years of age were diagnosed with neuroblastoma as residents of Olmsted County, Minnesota, during the 40-year study period. All 14 cases presented before the age of 5, with a mean age at diagnosis for all 14 of 22.5 months (range, 0.4–42.6 months). Urinary VMA or HVA levels were elevated above twice the upper limit of normal in 10 (71%) of 14 cases of neuroblastoma. Urinary VMA levels were greater than twice the upper limit of normal in 8 (57%) of 14 cases, whereas HVA levels were greater than two times the upper limit of normal in 10 (71%) of the 14 cases. Three (75%) of the four cases without significant urinary VMA or HVA elevation were diagnosed with stage IV disease and died within a year of diagnosis, whereas the fourth case (25%) had stage II neuroblastoma and recovered.

Clinical information, including age at diagnosis, tumor stage, VMA, and HVA levels at presentation, and final outcome for all 14 patients is shown in Table 1 . The adrenal gland was the primary tumor site in 10 (72%) of the 14, three (21%) were in the abdomen, and one patient (7%) had tumors located in both the mediastinum and thoracic spine. Eight (57%) of the 14 tumors were diagnosed as stage IV, with two (14%) each of stage II, III, and IVS. Seven (88%) of the eight cases diagnosed with stage IV disease died, whereas all four cases diagnosed before 1 year of age where in remission 5 years after diagnosis. Horner syndrome and opsoclonus were not observed in this cohort.

Table 1 Urine catecholamine levels and clinical characteristics of 14 patients <19 years diagnosed with Neuroblastoma in Olmsted County, Minnesota, 1969–2008
Full size table

Discussion

This population-based study found age-adjusted, clinically significant elevations of urinary VMA or HVA catecholamine levels at presentation in 10 (71%) of 14 neuroblastoma cases over the 40-year study. Urinary HVA levels were found to be elevated at greater than twice the upper limit of normal for age in 10 (71%) of 14 cases, whereas VMA levels were similarly elevated in only eight (57%) of the 14 neuroblastoma cases.

Neuroblastoma accounts for 8–10% of all childhood cancers, with a gender and race-adjusted published incidence of 9.8 annual cases per million children <15 years.21 Prognosis is age and stage-dependent, with those diagnosed before their first birthday having superior outcomes.2, 22, 23 Ocular manifestations have an important role in the diagnosis and staging of this neoplasm. Orbital involvement, occasionally signaled by proptosis and periorbital ecchymosis, signals stage IV disease and is characterized by decreased length of survival,1, 24 whereas Horner syndrome is more frequently associated with localized neuroblastoma and a better prognosis.9 Both advanced orbital disease and Horner syndrome can present as the first sign of neuroblastoma, emphasizing the integral role ophthalmologists have in the management of this disease.6, 7

Historically, assessing urine catecholamine levels has had a principle role in the ophthalmic work-up of idiopathic pediatric Horner syndrome.5, 6, 10, 11, 12, 14, 18, 19 Elevated urinary HVA/VMA levels have been reported in 90–95% of neural crest tumor cases,3, 4 supporting the use of this test to rule out neuroblastoma. However, De Bernardi et al25 describing a cohort of 144 patients with stage I and II neuroblastoma, found urinary VMA or HVA levels elevated to greater than two standard deviations above the upper limit of normal in only 68% of their cases. This finding is similar to that of LaBrosse et al4 who reported significantly lower levels of VMA excretion in patients diagnosed with stage I, II, and III disease. It is surprising that, in this study, only one patient with stage II disease demonstrated HVA and VMA levels in the normal range, whereas three (38%) of eight cases with stage IV disease had VMA and HVA levels that were less than twice the upper limit of normal.

In 1988, Woodruff et al6 recorded two cases of neuroblastoma in a cohort of 10 patients with pediatric Horner syndrome, one of which demonstrated normal urine catecholamine levels. More recently, Mahoney et al5 reported underlying masses in six cases initially presenting with idiopathic Horner syndrome. Physical examination and imaging were used to detect all the masses, as all six cases had normal HVA/VMA levels at presentation. After surgical resection, pathologic examination demonstrated neuroblastoma in four (67%) of the six. These reports combined with the findings of this study reinforce the current belief that urine studies alone are not sufficient to rule out neuroblastoma in cases of idiopathic pediatric Horner syndrome.

Due in part to the ineffective urine studies in their cohort, Mahoney et al suggested that Horner syndrome in any patient without a surgical history warrants brain, neck, and chest magnetic resonance imaging with and without contrast, as well as urinary catecholamine metabolite testing. However, in 2009, we conducted a population-based study of pediatric Horner syndrome, concluding that the lack of underlying neuroblastoma in our cohort, combined with other published estimates regarding the frequency of Horner syndrome caused by neuroblastoma, cast doubt on the value of extensive imaging in all cases of idiopathic pediatric Horner syndrome.19 We suggested that a thorough physical exam combined with initial urine studies was sufficient in most cases of idiopathic pediatric Horner syndrome, while imaging could be appropriately reserved for cases that demonstrated signs of worsening disease, including acquired or increasing iris heterochromia.19

It is unclear why this and other recent reports found a lower percentage of elevated urine catecholamines in patients with neuroblastoma. A possible explanation could be the lack of sensitivity of the more modern spot urine collections versus the traditional 24-h urine analysis; however, 24-h collections were predominantly used in this study, as all but one case were diagnosed before 1998. Of the three cases of stage IV disease without significantly elevated urine catecholamine levels, two were diagnosed in the late 1970's and one was diagnosed in 1981 and all three were evaluated with 24-h collections.

The findings of this and other studies5, 6, 14, 25, 26 confirm the current belief that negative urine screens in cases of idiopathic Horner syndrome in children are not sufficient to rule out underlying malignancy. The findings of recent reports of pediatric Horner syndrome19 and neuroblastoma26 suggest that greater emphasis be placed on the physical examination in cases of idiopathic pediatric Horner syndrome. Specifically, careful palpation of the neck, abdomen, and axilla, with close follow-up monitoring should be undertaken, especially in patients with normal urine studies. The decision to use imaging, as suggested by Mahoney et al,5 should be made if there are signs of worsening disease, including the presence of a new neck mass or lymph adenopathy.

There are several limitations to the findings of this study, including the small sample size. It is possible that a larger study would have found elevated VMA and HVA levels in a higher percentage of patients; however, the findings of this study agree with those of larger studies,5, 25 and serve to highlight the importance of care in evaluating patients with idiopathic pediatric Horner syndrome. In addition, the size of the cohort makes it impossible to reevaluate the sensitivity and specificity of urine catecholamine studies in cases of neuroblastoma with statistical accuracy. Another possible weakness of this report is the predominant use of 24-h urinary collections to analyze catecholamines as opposed to the more modern use of spot or random urine collections, possibly explaining the comparatively low percentage of significantly elevated catecholamines in neuroblastoma patients. However, LaBrosse et al4 also used 24-h collections to obtain their urine samples, reporting elevated catecholamines in 90% of their neural crest tumor cases.

Urine catecholamine levels were found to be significantly elevated in 10 (71%) of 14 cases of neuroblastoma over a 40-year period in this population-based study. This finding underscores the importance of a thorough physical examination in cases of idiopathic pediatric Horner syndrome with negative urine catecholamine levels. Imaging studies should be utilized on the basis of the physical exam findings, including cases of acquired or increasing iris heterochromia or the detection of lymph adenopathy.