Neurological complications of neurofibromatosis type 1 in adulthood

Brain (1999), 122, 473 481 Neurological complications of neurofibromatosis type 1 in adulthood A. Créange, 1,2,6 J. Zeller, 2,3 S. Rostaing-Rigattieri, 2,4 P. Brugières, 2,5 J.-D. Degos, 1 J. Revuz 3 and P. Wolkenstein 2,3 1 Service de Neurologie, 2 Réseau NF, 3 Service de Correspondence to: Dr A. Créange, Service de Neurologie, Dermatologie, 4 Service de Neuroradiologie, 5 Service Hôpital Henri Mondor, 94010 Créteil Cedex, Paris, France d Anesthésie-Réanimation and 6 Laboratoire GERMEN, E-mail: creange@univ-paris12.fr CHU Henri Mondor, Université Paris XII, Créteil, France Summary Neurofibromatosis type 1 (NF1) is a genetic disease with a wide range of neurological manifestations. To examine these, and to evaluate neurological morbidity in adulthood of patients with NF1, we studied a hospital-based series of 158 patients that included 138 adult patients aged >18 years and 20 children. NF1 evaluation included a multidisciplinary clinical and a clinically oriented radiological investigation. Neurological events occurring during childhood (in both children and adults of the series) and adulthood were recorded. One or several neurological manifestations have been observed in 55% of patients (adults and children) (n 87). These included: headache (28 patients); hydrocephalus (7); epilepsy (5); lacunar stroke (1); white matter disease (1); intraspinal neurofibroma (3); facial palsy (1); radiculopathy (5); and polyneuropathy (2). Tumours included: optic pathway tumours (20); meningioma (2); cerebral glioma (3); and malignant peripheral nerve sheath tumours (6). Lifethreatening complications were observed in five adults and included four malignant peripheral nerve sheath tumours and one meningioma. Pain was the leading symptom in 11 adults and was related to malignant peripheral nerve sheath tumours, complications of intraspinal neurofibromas, subcutaneous neurofibromas and peripheral nerve neurofibromas. NF1 in adults was not associated with other disabling or life-threatening neurological complications. Symptomatic optic pathway tumours, cerebral gliomas, symptomatic aqueductal stenosis and spinal compression due to intraspinal NF were observed exclusively during childhood. In this series, the predominant neurological features of adults with NF1 were chronic pain and malignant peripheral nerve sheath tumours. Keywords: neurofibromatosis type 1; adulthood; neurological complications Abbreviations: NF neurofibroma; NF1 neurofibromatosis type 1 Introduction Neurofibromatosis type 1 (NF1) is a common human genetic disease with an incidence of about 1 in 2500 3300, an autosomal dominant mode of inheritance and a high rate of new mutations (Riccardi, 1991). The NF1 gene is located on chromosome 17q11.2, and codes for a large tumour suppressor protein, called neurofibromin (Cawthon et al., 1990; Viskochil et al., 1990; Wallace et al., 1990). NF1 may exhibit a wide range of complications that can involve any of the body systems. Many of these features have an important age factor; for example, cutaneous neurofibromas do not usually develop before adolescence, and their frequency increases with age (Riccardi, 1992). In contrast, café-au-lait spots are usually present by the age of 5 years but their number declines with age after 50 years (Riccardi, 1992). Neurological Oxford University Press 1999 manifestations include involvement of the cerebrum, spine, cranial nerves and peripheral nerves. Although the overall severity of NF1 is considered to increase with age and life expectancy is reduced in NF1 patients (Zoller et al., 1995), the natural history of some potentially disabling or life-threatening NF1 complications remains unclear in adults. NF1 penetrance is virtually 100% by the age of 5 years (Huson et al., 1989a, b), and patients are usually investigated during childhood. Only one series, half of which were adult patients, has focused on neurological complications in adults (Huson et al., 1988). Thus, NF1- related neurological morbidity and prognosis in adult patients needs to be defined more precisely. In order to determine the specific morbidity related to neurological diagnoses in adults with NF1, we conducted a cross-

474 A. Créange et al. sectional study of neurological manifestations and a 2- year follow-up evaluation in a hospital-based series of adult patients. Patients Patients were referred to our multidisciplinary neurofibromatosis clinic by paediatricians, dermatologists, neurologists, pain specialists and the French Neurofibromatosis lay group. Patients were prospectively ascertained from June 1995 to September 1996. A clinical and/or paraclinical follow-up evaluation was performed from the time of inclusion until December 1997. Diagnosis of NF1 followed the criteria of the National Institutes of Health Consensus Conference (1988), which includes patients with two or more of the following: (i) six or more café-au-lait spots 5 mm in greatest diameter in prepubertal individuals and 15 mm in greatest diameter in postpubertal individuals; (ii) two or more neurofibromas of any type or one plexiform neurofibroma; (iii) freckling in the axillary or inguinal regions; (iv) optic pathway tumour; (v) two or more Lisch nodules (iris hamartomas); (vi) a distinctive osseous lesion such as a sphenoid dysplasia or thinning of the longbone cortex with or without pseudoarthrosis; (vii) a first-degree relative (parent, sibling or offspring) with NF1 by the above criteria (National Institutes of Health Consensus Development Conference Statement, 1988). Method A systematic multidisciplinary clinical investigation and a familial enquiry were performed for each patient, and included at least dermatological, neurological and ophthalmological examinations. Other specialists were consulted as required, based on clinical findings. Before June 1995, the screening policy to investigate patients with possible NF1 included chest X-ray, abdominal ultrasonography and cerebral imaging (CT scan or MRI). Since June 1995, investigations have been directed by abnormal clinical findings rather than by systematic inquiry about the most frequently encountered complications of NF1 (Wolkenstein et al., 1996). For some patients, data were collected from other centres that had undertaken systematic screening investigations. Non-neurological symptomatic complications of NF1 in the present series were compared with those observed in the Neurofibromatosis Institute Clinical Research Program (Riccardi, 1992), the Southeast Wales study (Huson et al., 1989a, b) and the study of Friedman and Birch study (1997) (Table 1). Neurological complications were classified according to the presence or absence of tumours. Those not associated with tumours were further classified as having cerebral involvement, spinal involvement, cranial nerve involvement or peripheral nerve involvement. Age of onset and age of treatment of neurological complications were recorded. Childhood was defined as the period under 18 years and adulthood as the period above 18 years. In our series, events occurring during childhood (in both children and adults of the series) and adulthood were recorded. Neurological features were compared with those of the Southeast Wales series (Huson et al., 1988), the Neurofibromatosis Institute Clinical Research Program series (Riccardi, 1992), the Ferner and Hughes series (Hughes, 1994) and the series of Friedman and Birch (1997) (Table 1). Results Our series included 158 patients [92 females (58%) and 66 males (42%)]; 138 patients were adults and 20 were children. Mean age ( SD) of patients at the time of the study was 33 14 years (range 11 77 years); that of the adults was 36 13 years and that of the children was 14.7 0.7 years. Patients were determined by history to be familial cases (43%) or cases resulting from new mutations (57%). Age of the patients at the time of NF1 diagnosis was 14 12.7 years (range 2 75 years). One or several neurological manifestations were observed in 87 (55%) of all patients (adults and controls) (Table 2 and Fig. 1). Tumours Cerebral tumours were observed in 24 patients (11 adults, 13 children). These included five brain tumours and 20 optic pathway tumours. Cerebral imaging (CT scanning or MRI) was available in 87.4% of patients (n 138). Brain tumours Clinical presentation. Brain tumours included three CNS gliomas (two brainstem gliomas and one temporal lobe glioma without histological proof) and two meningiomas. Childhood cases: two patients presented with acute hydrocephalus, in whom brainstem gliomas were diagnosed at 13 and 15 years of age. Adulthood cases: a temporal lobe glioma (on radiological criteria) was diagnosed incidentally by systematic cerebral imaging. It remains asymptomatic at the time of writing (the patient was 25 years old at the time of diagnosis, and 28 at the time of the study). Age at diagnosis of the two meningiomas was 40 and 55 years. The younger patient was asymptomatic and the meningioma was revealed by a systematic CT scan examination after a generalized seizure. Headache was the presenting symptom of the other meningioma. Treatment and clinical course. Childhood cases: the two patients with brainstem gliomas were treated by chemotherapy, radiotherapy, and ventricular shunt. At the time of examination for this study, neither patient had clinical deterioration or radiological progression (2 and 5 years of follow-up). Adulthood cases: one meningioma required surgery at the time of presentation (55 years of age). No treatment was considered for the other meningioma or for the probable temporal lobe glioma.

Neurological complications of NF1 in adults 475 Table 1 Complications of NF1 in the present series (adults and children) compared with four other series Series Créange et al. Huson et al. Riccardi Friedman and Birch Ferner and Hughes (1999) (1988) (1992) (1997) (1994) (n 158) (n 135) (n 953) (n 1728) (n 103) NF1 complication (%) Plexiform NF 43.0 30 39 27.6 NA Learning disabilities 67.1 33 67 39.6 NA Pseudarthrosis 2.5 3.7 3 2 NA Pheochromocytoma 1.3 0.7 0.1 0.1 NA Xanthogranuloma 1.3 0.7 2 2.5 NA Neurological complication (%) Headache 18 2 10 23.4 NA Hydrocephalus 4 1 2 3.6 1.5 Epilepsy 3.8 1 5 6 8.1 Cerebral tumour 3.1 3 1 1.9 0.7 MPNT 3.8 1.5 3.7 NA 3.5 Intraspinal NF 1.9 1 5 NA 1.5 OPT 12.6 4 15 3.8* 0.7 Neurological complications differing from two or more other series (P 0.05) are in bold characters. MPNT malignant peripheral nerve sheath tumour; NA not available; OPT optic pathway tumour. *Symptomatic cases only. Learning disability refers to an educational level below 14 years. Table 2 Neurological complications in patients (adults and children) with NF1 Complications Patients (n) Cerebral Headache 28 Migraine 2 Hydrocephalus 7 Epilepsy 6 Lacunar stroke 1 Possible multiple sclerosis 1 Spinal Intraspinal neurofibroma 3 Meningocoel 2 Peripheral nerve Facial palsy 1 Radiculopathy 5 Polyneuropathy 2 Tumoral Optic pathway tumour 20 Malignant peripheral nerve sheath tumour 6 Cerebral meningioma 2 Glioma 3 Miscellaneous Pain 18 Optic pathway tumours Clinical presentation. Childhood cases: optic pathway tumours were diagnosed in 12 children at 11.4 4.8 years (mean SD). Diagnosis coincided with the time of inclusion in the study in one case in a patient who had not complained of visual symptoms but was referred to our centre after a cerebral imaging study. Adulthood cases: optic pathway tumours were diagnosed in eight adults at 29.8 10.6 years. Diagnosis coincided with time of inclusion in the study in five cases. Fig. 1 Principal neurological manifestations according to age of onset or discovery and severity. MPNST malignant peripheral nerve sheath tumour; OPT optic pathway tumour. Treatment and clinical course. Childhood cases: three of the cases of optic pathway tumour were symptomatic. These patients were treated by chemotherapy (n 1), surgery (n 1) and radiotherapy (n 1) at 4 years (n 2) and 13 years (n 1) for progressive visual dysfunction and/or radiological extension. Adulthood cases: all patients had a

476 A. Créange et al. systematic survey (an annual ophthalmological examination including visual acuity and campimetric examination) before inclusion in the study and since then; it did not show any significant decrease in visual function. One 28-year-old patient was referred for evaluation in our centre with only 40% visual acuity in her left eye, related to a previously undiagnosed optic pathway tumour. No decrease in visual acuity has been detected since this time. Other adult patients with optic pathway tumours had not experienced progression of visual symptoms 10 years after diagnosis of their optic tumour (mean 7.6 years; range 1 24 years). One 26-year-old patient had a retrochiasmatic progression of the lesion without symptomatic visual deterioration (Fig. 2). Malignant peripheral nerve sheath tumours Six patients had a malignant peripheral nerve sheath tumour. Clinical presentation. Childhood cases: presenting symptoms occurred at 15 and 18 years of age. Adulthood cases: presenting symptoms occurred at ages 19, 20, 31 and 44 years. First symptoms were characterized in five adult cases by an increase in volume of a previously known subcutaneous neurofibroma (NF), exacerbation of pain or the appearance of a motor deficit. One adult case was discovered by histological examination of tissue after surgical removal of a pseudarthrosis. Treatment and clinical course. Childhood cases: the evolution of one malignant peripheral nerve sheath tumour was characterized by local spreading of the tumour from the sciatic nerve and death after 6 years despite surgery, radiotherapy and chemotherapy. One patient is alive after extensive surgery (6 months of follow-up). Adulthood cases: the evolution of the four malignant peripheral nerve sheath tumours was characterized by local spreading of the tumour from the cauda equina and cervical roots. Evolution was associated with pelvic organ compressions, cervical spinal cord compression and intractable pain. The four patients died 1, 2 and 3 years after treatment that included surgery, radiotherapy and chemotherapy. One patient is alive after limb amputation (2 years of follow-up). Other neurological diagnoses Spinal compression Five patients have had spinal cord or cauda equina compressions. Clinical presentation. Childhood cases: spinal cord compression (n 2) due to intraspinal NF occurred at 6 and 18 years of age. Adulthood cases: spinal cord compression related to intraspinal extension of malignant peripheral nerve sheath tumours (n 2) occurred at 31 and 46 years of age. One patient had mild sensory symptoms in all four limbs, Fig. 2 Asymptomatic optic pathway tumour in a 26-year-old patient. Thickening of the optic chiasma and of both distal portions of the optic nerve on T 1 -weighted images (A) and backspreading of the tumour along the posterior optic pathways on T 2 - weighted images (B).

Neurological complications of NF1 in adults 477 related to a cervical intraspinal NF that appeared at 26 years and did not progress. Clinical course and treatment. Childhood cases: intraspinal NF led to cervical spinal cord compression. The two patients were quadriplegic despite surgery at 6 and 18 years of age. Adulthood cases: the two patients with malignant peripheral nerve sheath tumours had severe pain and complete spinal or cauda equina compressions characterized by quadriplegia and neurological sphincter dysfunction. One died from complications of cervical cord compression. The other died from pelvic organ compression. The third patient with intraspinal NF had not experienced progression of spinal sensory symptoms until the time of the study (7 years of follow-up). Epilepsy Six patients had epilepsy. Clinical presentation. Childhood cases: one patient had symptomatic epilepsy associated with sequelae of aqueductal stenosis at 14 years, and one patient had an idiopathic epilepsy that had begun at 18 years. Adulthood cases: four patients had their first manifestation of epilepsy during adulthood. Three of them had a symptomatic epilepsy associated with post-traumatic subdural haematoma (n 1), cerebral hamartoma (n 1) and meningioma (n 1). The first crisis in patients with symptomatic epilepsy occurred at 20, 32 and 40 years of age. One patient had a cryptogenic epilepsy for which the first crisis occurred at 25 years. Treatment and clinical course. Idiopathic epilepsy was controlled by monotherapy in all cases. The patient with epilepsy related to a hamartoma had a complex partial epilepsy; the others had generalized epilepsy. In all cases epilepsy was controlled by medication. Hydrocephalus Seven patients had hydrocephalus. Clinical presentation. Childhood cases: diagnosis was made in six cases after acute intracranial hypertension syndrome (n 3), motor dysfunction (n 1) and systematic cerebral imaging (n 2). Hydrocephalus secondary to idiopathic aqueductal stenosis (n 4) was associated with intracranial hypertension syndrome (n 2), motor dysfunction (n 1) or was asymptomatic (n 1). In one of the cases, hydrocephalus was confined to the lateral ventricles. One case of hydrocephalus was secondary to a brainstem glioma. Diagnosis was made between 11 and 14 years in the patients with aqueductal stenosis and brainstem glioma, and at 16 and 18 years in the patients with absence of symptoms or mild symptomatic manifestations (headache). Adult cases: headache was the presenting symptom of hydrocephalus in one 21-year-old. Hydrocephalus was secondary to a brainstem benign cyst and aqueductal stenosis. Clinical course and treatment. Childhood cases: four patients had acute intracranial hypertension (n 3) or motor dysfunction (n 1) at the time of diagnosis (11 14 years of age) and underwent a ventricular shunt. The two other patients were considered asymptomatic and were not treated. They are still asymptomatic after follow-up of 2 and 10 years, respectively (ages at the time of the study were 18 and 28 years). Adult cases: the headache was considered to be unrelated to the hydrocephalus; the patient is still virtually asymptomatic after a 9-year follow-up period and has not been treated by ventricular shunt. Peripheral nerve involvement Peripheral nerve complications were documented in nine patients. These included cranial nerve involvement, radicular involvement and polyneuropathies. Clinical presentation. Childhood cases: one patient experienced a unilateral upper limb motor predominant pluriradicular deficit at 15 years of age, related to a surgical excision of a painful plexiform NF. A secondary increase in the motor deficit at 16 years was due to a malignant peripheral nerve sheath tumour degeneration. Another patient (aged 16 years) had a painful radicular involvement related to a paravertebral NF. One patient had had a very slowly progressive and disabling distal, symmetrical, motorpredominant neuropathy of the lower limb since the age of 16 years. Adulthood cases: one patient had a regressive idiopathic facial paralysis. Three patients had upper or lower limb painful radicular manifestations related to paravertebral NF (n 2) or degenerative spinal abnormalities (n 1) (ages at the time of the study were 28, 42 and 52 years). Two patients (29 and 30 years old) had asymptomatic decreases of nerve conduction velocities on electrophysiological examination. Clinical course and treatment. All these manifestations (except the malignant peripheral nerve sheath tumours) were stable or slowly progressive. Treatment was symptomatic in all cases except for the patient with malignant peripheral nerve sheath tumours, who underwent extensive surgery and radiotherapy. Pain Painful manifestations were classified as headache- and nonheadache-associated pain. Headache: clinical presentation. Headaches were recorded in 28 patients. These included tension-type headaches, chronic idiopathic headaches, analgesic-abuse headaches and migraines. Migraines were reported by patients

478 A. Créange et al. aged 11 66 years (3 children and 25 adults) and occurred over periods of many years. Headache: clinical course and treatment. Severe migraines occurred in only two adult patients. Migraine and other headache types were otherwise controlled (defined as a decrease of 50% in migraine severity). Non-headache pain: clinical presentation. Childhood cases (C) and adulthood cases (A): Pain was a predominant neurological feature in 18 patients. Pain began in adulthood in 11 cases and in childhood in seven cases. The origin of pain was considered as certain (12/18; C, 4; A, 8), uncertain (3/18; C, 1; A, 2) or inorganic (3/18; C, 2; A, 1). Definite causes included subcutaneous NF (n 2; C, 2), nodular plexiform NF (n 2; C, 1; A, 1), peripheral nerve and root NF (n 7; C, 2; A, 5), malignant peripheral nerve sheath tumours (n 3; C, 1; A, 2), arthrosis (n 1; A, 1), decubitus complications (n 1; C, 1) and pain secondary to surgery (n 4; C, 2; A, 2). Pain was recorded in a single site (n 8; C, 3; A, 5), in multiple sites (n 7; C, 4; A, 3) or diffuse (n 3; A, 3). The majority of patients (n 9; C, 3; A, 6) had truncular or radicular pain. Six other patients (A, 6) had mild pain related to subcutaneous NF that was not considered as disabling by the patients. Clinical course and treatment. Childhood cases (C) and adulthood cases (A): Pain duration varied from 15 days to 20 years. Pain was permanent in half of the patients and intermittent in the other half (both adults and children). It was complicated by spontaneous paroxysmal increases and exacerbated by movement or by contact with the painful NF. When multifocal or diffuse, some painful sites were not associated with a definite cause. Optimal pain management (analgesics, antidepressants, anticonvulsants or associations) was associated with complete remission (n 1; C, 1), partial remission (n 14; C, 5; A, 9) or absence of modification of pain (n 3; C, 1; A, 2). Remission of pain and one case with absence of modification of pain were associated with inorganic pain. In two cases, surgical removal of an NF (C, 1) and a plexiform NF (C, 1) was associated with an increase in pain. The majority of patients were treated by a combination of common analgesic drugs, tricyclic antidepressants or anticonvulsants. Three patients required chronic systemic delivery of opioids, including intrathecal morphine injections (C, 1; A, 2). Three patients were treated by neurostimulation (one cutaneous, one spinal cord, one cortical) (C, 1; A, 2), and two patients with malignant peripheral nerve sheath tumours had surgical treatment of pain (one spinal root section, one anterolateral cordectomy) (A, 2). Other complications Clinical presentation. Childhood cases: no other neurological complication was recorded. Adulthood cases: two patients had a meningocoel. One meningocoel (lumbar) was revealed by lumbar pain at 24 years of age; the other (dorsal) remained asymptomatic. One patient (aged 19 years) had a left hemiparesis related to a lacunar stroke of unknown origin. One patient has begun, at 35 years of age, to have a slowly progressive spastic paraparesis with multiple high signal hyperintensities on a T 2 -weighted MRI. Clinical course and treatment. The lacunar stroke and the lumbar pain associated with the meningocoel spontaneously resolved. The spastic paraparesis was resistant to steroids but was virtually non-progressive over a 20- year period. Discussion This study describes the neurological problems in a series of 158 patients with NF1 in which major defining features and complications of NF1 (Table 1) were similar to those found in other large NF1 series (Huson et al., 1989a, b; Riccardi, 1992), suggesting that major bias towards more severely affected patients was avoided in the recruitment of patients of this hospital-based series. This sample was characterized by a majority of adult patients (138 patients), in contrast to previous large series, which included younger individuals (Riccardi, 1992), and highlights several specific features of the adult NF1 neurological complications spectrum. Tumours included those of the CNS and peripheral nerves. Brainstem gliomas and half of the optic pathway tumours were discovered during childhood. All five cases of brainstem and optic pathway tumours were treated during childhood on the basis of clinical and radiological progression. Adult patients did not require any treatment since the optic pathway tumours and brainstem gliomas remained asymptomatic or non-progressive (up to 10 years of follow-up for optic pathway tumours). The absence of cerebral gliomas in adults in this series is probably a bias related to a cross-sectional study. However, it may emphasize a higher risk of cerebral glioma in children than in adults. This result suggests that the concept of low progression potential of brainstem glioma and optic pathway tumour observed in children (Pollack et al., 1996) could also be applicable to adults, and in this context should be taken into account in the follow-up and routine surveillance of adult patients. In many institutions, routine surveillance scanning of patients with NF1 is performed. Up to 20% of patients harbour CNS gliomas which are mainly asymptomatic (DiMario et al., 1993; Pollack et al., 1996). Only 2% of patients with intracranial masses require treatment in childhood (Pollack et al., 1996). A follow-up that includes a full ophthalmological examination (including visual acuity, colour vision, visual fields, slit-lamp examination) rather than repetitive neuroimaging examinations should be evaluated in adults; it has been recommended for children (Listernick et al., 1997). A similar protocol could also be evaluated for long-term follow-up in adults with brainstem glioma, where treatment is

Neurological complications of NF1 in adults 479 Table 3 Life-threatening complications and their complications associated with functional disability tabulated according to time of onset and clinical severity (only disabling pain was recorded) Symptomatic/treated Untreated/asymptomatic, mildly symptomatic Childhood Adulthood Childhood Adulthood OPT 3 0 9 8 Cerebral gliomas 2 0 1 0 Meningioma 0 1 0 1 MPNST 2 4 0 0 Intraspinal NF with cord 2 0 0 1 compression Hydrocephalus 4 0 2 1 Epilepsy 0 0 4 2 Pain 7 11 Cause definite 4 8 Cause uncertain 1 2 Inorganic 2 1 MPNST malignant peripheral nerve sheath tumour; NF neurofibroma; OPT optic pathway tumour. recommended only in cases of significant clinical deterioration (Pollack et al., 1996). Finally, considering the low and unpredictable risk of cerebral glioma in adults with NF1, systematic CNS imaging does not appear to be useful. Malignant peripheral nerve sheath tumours were observed in six cases, four of them in adults. Three of the four adult patients died. Malignant peripheral nerve sheath tumour was the only cause of death in adults related to neurological complications in this series. In three studies that included a total of 1158 patients [from a 12-year follow-up study in adults (Zoller et al., 1995), a population-based series (Huson et al., 1988) and a large adult NF1 series (Riccardi, 1992)], deaths related to neurological complications were mostly secondary to malignant peripheral nerve sheath tumours [except for four cases which were secondary to intraspinal NF (Riccardi, 1992)]. Malignant peripheral nerve sheath tumour is a highly aggressive neoplastic disease which often arises from plexiform NFs, as in this series. Rapid change in tumour size and/or symptoms, particularly pain, should be immediately investigated to rule out the possibility of malignant transformation of a benign lesion (Gutmann et al., 1997). The frequency in adults of plexiform NF, for which imaging characteristics cannot reliably distinguish benign from malignant lesions (Tonsgard et al., 1998), support the importance of clinical monitoring of plexiform NF. In these three studies, the only causes of death in adulthood were four cases which were secondary to intraspinal NF (Riccardi, 1992) and two related to complications of partial resection of astrocytomas (Huson et al., 1988). In the present series, life-threatening complications were observed in 18 cases (Table 3). These included 12 tumours and their complications (three optic pathway tumours, two gliomas, six malignant peripheral nerve sheath tumours and one meningioma) and five other manifestations (symptomatic aqueductal stenosis, and spinal compression due to intraspinal NF). Except for one meningioma (where the relationship to NF1 is not known) and four malignant peripheral nerve sheath tumours, these complications occurred during childhood. Hydrocephalus was symptomatic in four children but was largely asymptomatic in three adults. In a series of six patients with symptomatic hydrocephalus in NF1 (Senveli et al., 1989), all patients were under 19 years of age. Hydrocephalus may result from intrinsic tectal mass, retrochiasmatic progression of optic pathway tumours and focal proliferation of periaqueductal subependymal glial cells (Radhakrishnan et al., 1981; Pollack et al., 1996; Alshail et al., 1997; Bilaniuk et al., 1997). The slow spreading behaviour of optic pathway tumours, the favourable outlook of brainstem intrinsic tumours in patients with NF1, the extremely low rate of growth of the periaqueduct subependymal glial tumours (Boydston et al., 1992) and the non-progressive behaviour of periaqueductal glial cell with hydrocephalus suggest that a close systematic radiological follow-up is probably unnecessary in adults with hydrocephalus. Spinal and paraspinal NF may be complicated by major neurological disability but appear to be rare in adults. No other adult patients in our series or in the South Wales series experienced progressive or disabling spinal complications. Moreover, one might speculate on the presence of asymptomatic spinal NF in adults, since no systematic extensive spinal imaging is usually performed in adults with NF1. However, Riccardi s series included four deaths of adult patients (Riccardi, 1992), and a potential evolution of mildly symptomatic lesions in young adults cannot be excluded by our cross-sectional study. In adults, apart from malignant peripheral nerve sheath tumours, pain related to neurological symptoms was the most frequent disabling concern (Table 3). Pain is often overlooked in studies on NF1. While frequency of headache is often stressed (10%) (Riccardi, 1992), its frequency in the present series (18%) was similar to the frequency of migraine (6% in men and 20% in women) (Stewart et al., 1992) and lower than the estimated prevalence of headache of all types in the

480 A. Créange et al. general population (78% of men and women of all ages) (Waters and O Connor, 1975). Evidence is lacking to support migraine as a specific NF1-related complication. However, 18 patients (11.3%) had pain as a leading neurological manifestation. A similar percentage was observed by Riccardi (1992). Large nodular plexiform NFs, or NFs developed in deep peripheral nerves or roots, were usually associated with permanent disabling pain, and subcutaneous NFs were sometimes associated with episodes of paroxysmal pain. When diffuse or multifocal, the cause of pain was mainly uncertain, despite the presence of organic lesions as a cause of one painful site. When present, the shooting and electric character of pain and its truncular topography suggest the presence of an underlying undetected NF. Although it has not been previously emphasized in NF1, pain related to peripheral nerve sheath tumours is the presenting manifestation in most patients with schwannomatosis, a peripheral nerve tumour syndrome (MacCollin et al., 1996). A large proportion of patients ( 70%) required combinations of analgesic drugs including tricyclic antidepressants and anticonvulsants. A small proportion of patients required the same drugs despite the absence of an objective cause of pain, suggesting that a psychological and/or depressive component should always be considered in patients with NF1 and pain. Pain should be considered as significant a cause of psychological difficulties in NF1 as are its dysmorphic consequences. Surgical removal of the tumours should be carefully discussed, since an increase in local pain may occur after peripheral painful schwannoma removal (Buenger et al., 1993), as it did in two of our patients. Cancer pain related to metastatic malignant peripheral nerve sheath tumours was the most severe neurological complication requiring systemic opioid delivery in several cases, surgery for pain in two cases and major expenditure of medical staff time (one patient was seen 104 times in 20 months). It is generally considered that in NF1 the period of risk is lifelong for epilepsy, CNS tumours, peripheral nerve malignancy, rhabdomyosarcoma and spinal NF (Huson et al., 1989a, b). The period of risk comprises the first three decades for aqueductal stenosis, and is limited to the first 6 years of life for the development of symptomatic optic pathway tumours (Listernick et al., 1997). Moreover, the difference observed between birth incidence and the decreased prevalence observed in the older age group (Huson et al., 1988; Clementi et al., 1990; Huson, 1994) is consistent with an increase in mortality in the younger patients. The absence of difference in the frequencies of the main neurological complications between the present sample and other large series (Table 1), including cerebral tumours, optic pathway tumours, malignant peripheral nerve sheath tumours, intraspinal NF, hydrocephalus and epilepsy, is consistent with the more aggressively progressive behaviour of NF1 in childhood than in adulthood for all neurological complications. Interestingly, both malignant and benign NF are believed to follow a common pathogenesis related to the two-hit hypothesis, in which one allele is constitutionally inactivated while the other is subsequently inactivated at the somatic level (Colman et al., 1995). In other diseases resulting from a tumour suppressor gene and dysfunction through a two-hit phenomenon (such as in retinoblastoma and Wilms tumour), 90% of complications occur very early in life. A similar pattern of complications with onset during life in these diseases is consistent with our observations, suggesting that adult patients with NF1 are usually devoid of common NF1 disabling and life-threatening neurological complications, except malignant peripheral nerve sheath tumours, and are dominated by chronic painful complications. Acknowledgement The authors express their gratitude to Patricia Birch for helpful criticisms and for improving the English. References Alshail E, Rutka JT, Becker LE, Hoffman HJ. Optic chiasmatichypothalamic glioma. [Review]. Brain Pathol 1997; 7: 799 806. Bilaniuk LT, Molloy PT, Zimmerman RA, Phillips PC, Vaughan SN, Liu GT, et al. Neurofibromatosis type 1: brain stem tumours. Neuroradiology 1997; 39: 642 53. Boydston WR, Sanford RA, Muhlbauer MS, Kun LE, Kirk E, Dohau FC Jr, et al. Gliomas of the tectum and periaqueductal region of the mesencephalon. Pediatr Neurosurg 1992; 17: 234 8. Buenger K, Porter NC, Dozier SE, Wagner RF Jr. Localized multiple neurilemmomas of the lower extremity. [Review]. Cutis 1993; 51: 36 8. Cawthon RM, Weiss R, Xu GF, Viskochil D, Culver M, Stevens J, et al. A major segment of the neurofibromatosis type 1 gene: cdna sequence, genomic structure, and point mutations [published erratum appears in Cell 1990; 62: following 608]. Cell 1990; 62: 193 201. Clementi M, Barbujani C, Turolla L, Tenconi R. Neurofibromatosis 1: a maximum likelihood estimation of mutation rate. Hum Genet 1990; 84: 116 8. Colman SD, Williams CA, Wallace MR. Benign neurofibromas in type 1 neurofibromatosis (NF1) show somatic deletions of the NF1 gene. Nat Genet 1995; 11: 90 2. DiMario FJ Jr, Ramsby G, Greenstein R, Langshur S, Dunham B. Neurofibromatosis type 1: magnetic resonance imaging findings. J Child Neurol 1993; 8: 32 9. Friedman JM, Birch PH. Type 1 neurofibromatosis: a descriptive analysis of the disorder in 1,728 patients. Am J Med Genet 1997; 70: 138 43. Gutmann DH, Aylsworth A, Carey JC, Korf B, Marks J, Pyeritz RE, et al. The diagnostic evaluation and multidisciplinary management of neurofibromatosis 1 and neurofibromatosis 2 [see comments]. [Review]. JAMA 1997; 278: 51 7. Comment in: JAMA 1997; 278: 1493 4.

Neurological complications of NF1 in adults 481 Hughes RAC. Neurological complications of neurofibromatosis 1. In: Huson SM, Hughes RAC, editors. The neurofibromatoses: a pathogenetic and clinical overview. London: Chapman and Hall; 1994. p. 204 32. Huson SM. Neurofibromatosis 1: a clinical and genetic overview. In: Huson SM, Hughes RAC, editors. The neurofibromatoses: a pathogenetic and clinical overview. London: Chapman and Hall; 1994. p. 160 203. Huson SM, Harper PS, Compston DA. Von Recklinghausen neurofibromatosis: a clinical and population study in south east Wales. Brain 1988; 111: 1355 81. Huson SM, Compston DA, Clark P, Harper PS. A genetic study of von Recklinghausen neurofibromatosis in south east Wales. I. Prevalence, fitness, mutation rate and effect of parental transmission on severity. J Med Genet 1989a; 26: 704 11. Huson SM, Compston DA, Harper PS. A genetic study of von Recklinghausen neurofibromatosis in South East Wales. II. Guidelines for genetic counselling. [Review]. J Med Genet 1989b; 26: 712 21. Listernick R, Louis DN, Packer RJ, Gutmann DH. Optic pathway gliomas in children with neurofibromatosis 1: consensus statement from the NF1 Optic Pathway Glioma Task Force. Ann Neurol 1997; 41: 143 9. MacCollin M, Woodfin W, Kronn D, Short MP. Schwannomatosis: a clinical and pathologic study [see comments]. Neurology 1996; 46: 1072 9. Comment in: Neurology 1997; 48: 787 8. National Institutes of Health Consensus Development Conference. Neurofibromatosis. Conference statement. Arch Neurol 1988; 45: 575 8. Pollack IF, Shultz B, Mulvihill JJ. The management of brainstem gliomas in patients with neurofibromatosis 1. [Review]. Neurology 1996; 46: 1652 60. Radhakrishnan K, Kak VK, Sridharan R, Chopra JS. Adult aqueductal stenosis with Recklinghausen s neurofibromatosis. Surg Neurol 1981; 16: 262 5. Riccardi VM. Neurofibromatosis: past, present, and future. N Engl J Med 1991; 324: 1283 5. Riccardi VM, editor. Neurofibromatosis: phenotype, natural history, and pathogenesis. 2nd ed. Baltimore: Johns Hopkins University Press; 1992. Senveli E, Altinörs N, Kars Z, Arda N, Turker A, Cinar N, et al. Association of von Recklinghausen s neurofibromatosis and aqueduct stenosis [see comments]. [Review]. Neurosurgery 1989; 24: 99 101. Comment in: Neurosurgery 1989; 25: 318 9. Stewart WF, Lipton RB, Celentano DD, Reed ML. Prevalence of migraine headache in the United States. Relation to age, income, race, and other sociodemographic factors. JAMA 1992; 267: 64 9. Tonsgard JH, Kwak SM, Short MP, Dachman AH. CT imaging in adults with neurofibromatosis-1: frequent asymptomatic plexiform lesions. Neurology 1998; 50: 1755 60. Viskochil D, Buchberg AM, Xu G, Cawthon RM, Stevens J, Wolff RK, et al. Deletions and translocation interrupt a cloned gene at the neurofibromatosis type 1 locus. Cell 1990; 62: 187 92. Wallace MR, Marchuk DA, Andersen LB, Letcher R, Odeh HM, Saulino AM, et al. Type 1 neurofibromatosis gene: identification of a large transcript disrupted in three NF1 patients [published erratum appears in Science 1990; 250: 1749]. Science 1990; 249: 181 6. Waters WE, O Connor PJ. Prevalence of migraine. J Neurol Neurosurg Psychiatry 1975; 38: 613 6. Wolkenstein P, Frèche B, Zeller J, Revuz J. Usefulness of screening investigations in neurofibromatosis type 1. A study of 152 patients [see comments]. Arch Dermatol 1996; 132: 1333 6. Comment in: Arch Dermatol 1997; 133: 655 6. Zoller M, Rembeck B, Akesson HO, Angervall L. Life expectancy, mortality and prognostic factors in neurofibromatosis type 1. A twelve-year follow-up of an epidemiological study in Göteborg, Sweden. Acta Derm Venereol 1995; 75: 136 40. Received September 17, 1998. Accepted October 7, 1998