I read with interest the excellent report of Arnoldi and Tychsen (1995;32:296-30l). As the authors and Dr Appen in the accompanying guest editorial point out, however, this report employs a rather broad definition of spasmus nutans. It is possible, therefore, that dysconjugate nystagmus in some of these patients might actually represent other benign forms of nystagmus that are not associated with intracranial gliomas. Because the number of patients in this report is still relatively small, this could lead to an underestimation of the true prevalence of intracranial lesions in patients with spasmus nutans. Dutton reviewed the literature on optic pathway gliomas in 1994,1 and reported that nystagmus was seen in 23% of 264 patients on the initial examination, and that nystagmus in an optic nerve glioma may precede visual loss. Dutton reported 5 cases of intracranial glioma that presented with classic findings of spasmus nutans, and the author reviewed 24 cases of gliomas masquerading as spasmus nutans.
The authors provide a useful and important list of signs suggestive of tumor in patients with spasmus nutans that should mandate prompt neuroimaging. The determination of an accurate visual acuity in very young children, however, can be extremely difficult. Likewise, afferent pupillary defects and even mild optic disc pallor can be missed, especially in a non-cooperative child. Although I agree with the authors that the likelihood of an intracranial lesion in a child with spasmus nutans without other neurological or neuro-ophthalmological signs is low, the morbidity of a noninvasive study, such as magnetic resonance imaging or computed tomography, is even lower. In addition, there is a potential risk if the diagnosis of glioma is delayed. The prognosis for tumor confined to the optic nerve is good with a mortality rate of about 5%, but morbidity rates rise precipitously with extension to the chiasm and hypothalamus.1 Under the circumstances, it would seem prudent to err on the side of caution by performing neuroimaging. Although initial observation is a reasonable alternative, the patient should first undergo an extensive and complete neuro-ophthalmologic and/or pediatrie neurology assessment. The parents of the patient should be informed of the risk of an intracranial lesion, and should be allowed to participate in the decision for neuroimaging. Nevertheless, I commend the authors for an excellent review and discussion of this diagnostic dilemma.
ANDREW LEE, MD
1. Dutton JJ. Gliomas of the anterior visual pathway. Surv Ophihalmal. 1994;38:427-452.
1. It is tiresome to argue about whose cases of spasmus nutans are more narrow, real, pure, or authentic. Children with high frequency, small amplitude disconjugate nystagmus and titubation or torticollis have spasmus nutans, independent of whether they have or do not have other signs as well (eg, latent nystagmus, strabismus, or developmental delay). Be wary of a study in which the authors decide to throw out cases because they do not meet thenexpectations. Logicians call this "begging the question." Statisticians call it "cooking the data" or "cherry picking." Applied to our study this Une of argumentation would lead to the reductio ad absurdum: because many of the children had additional signs, but did not develop gliomas, many of the children in our study did not have "real" spasmus nutans.
2. Dutton's excellent review article does not present new cases of optic glioma and spasmus nutans. He cites five published reports by authors also cited in our paper. Dutton notes, as we do, that nystagmus may be present in children with glioma, but the nystagmus in these cases is accompanied by one or more other signs (listed in our Table 3). Clinicians should not be misled by the fact that 23% of children with gliomas have some type of nystagmus (in point of fact few of these had spasmus nutans). The much more important question is: taking all infants who present with spasmus nutans, how many have gliomas? That was the question our study addressed.
3. Dr Lee is specifically concerned about delayed diagnosis of a tumor confined to the optic nerve. A glioma of the optic nerve alone is three times less common than chiasmal glioma; is least likely to be associated with spasmus nutans; and is most likely to cause early disc edema/pallor or an obvious afferent pupillar defect. The tumor (ie, the optic nerve) would not be excised unless the eye was blind. Moreover, the majority of gliomas in young children appear to involve multiple loci along the optic pathway; it is a matter of some debate whether excising an orbital glioma prevents future growth of intracranial glioma.
4. We do not send patients with spasmus nutans to the pediatrie neurologist unless we suspect a neurological problem outside the visual pathway.
5. Dr Lee makes a good point about the role of the family in a decision to scan. We say to the family that there appears to be about a 1 in 100 risk that their child could be growing a glioma; that gliomas tend to grow slowly if they grow; that gliomas tend to be within the nerve tissue and can not be cut out. We tell them that their child must be followed carefully; their child has a much greater chance (about 1 in 2) of developing eye crossing that could cause lazy vision.
6. There are two schools of thought on neuroimaging of infants diagnosed with spasmus nutans.
One school says, "scan 'em all ."The assumptions underlying this strategy are: a) gliomas are "common" in spasmus nutans so I am bound to have a lot of positive scans; b) scanning the infants who have negative scans entails no risk, is not burdensome to the families, and is not a waste of resources; c) a negative scan means that I will not have to see them back or do as good an exam; d) gliomas grow rapidly and diagnosis of a glioma "early" (months before onset of optic disc pallor/edema or an afferent pupillary defect) will lead to a substantially better outcome; and e) early scanning is the only way to keep families happy and lawyers away.
We contend that each of these assumptions is false. We place ourselves in the other school, which says: "think before you scan, and scan fewer." The logic motivating this approach is: a) the prevalence of glioma in spasmus nutans is low, and overscanning is an imprudent use of the families time and medical resources; b) no one would treat a glioma until it was proven to have produced objective evidence of significant visual loss, so even if you detected an enlarged chiasm by early scanning, you would observe the child over a period of months, and in some cases years before recommending treatment; c) the value of earlier detection for timely treatment is not clear at all; most pediatrie visual pathway gliomas are multicentric and intrinsic, not exophytic; they can not be excised; they may or may not grow; and current treatment is lousy (chemotherapy before age 3 years); d) you have to follow them carefully anyway, because infants with spasmus nutans stand a good chance of developing garden variety strabismus or amblyopia; and e) scanning is no substitute for a simple, clear, and brief explanation of the importance of follow up, one reason for which is tumor development, but more likely for strabismus and amblyopia.
Disciples of the "scan 'em all" school are over-represented in the literature. Disciples of the "think before" school are under-represented in the literature. This is understandable, because authors tend to write and journals tend to publish reports of the exciting (few glioma cases) over the mundane (many no glioma cases).
LAWRENCE TYCHSEN, MD
KYLE ARNOLDI, CO, COMT
St. Louis, Missouri