Editor—The diagnostic criteria for neurofibromatosis type 1 (NF1) was established in 1987 by the National Institutes of Health (NIH) in the consensus development conference and defined further in 1997 (table 1).1 2 The frequency of clinical findings in NF1 patients varies greatly between series, probably on account of ascertainment differences and population versus hospital based patient groups. In the present report, the clinical findings are derived from a Finnish population based study of NF. The clinical and in some cases risk haplotype analysis was extended to the first degree relatives of the patients.

The clinical assessment formed part of a population based study of neurofibromatosis in northern Finland in which 172 patients with NF1 or the segmental type of NF were examined by the author during the years 1989-1996 at the Department of Clinical Genetics, Oulu University Hospital, or at one of its four satellite clinics. Clinical evaluation of the patients included a review of their medical history, including hospital records, necropsy reports, and family histories. A detailed clinical examination was performed, followed by genetic counselling. A clinical assessment was made of cutaneous, neurological, ocular, orthopaedic, and cardiovascular manifestations. Endocrine, auditory, and dysmorphic findings were recorded systematically according to a modified examination schedule used by the NNFF (National Neurofibromatosis Foundation) International Database,3 and developmental and behavioural features and growth parameters were recorded. An ophthalmological or neuro-ophthalmological (n=123) examination and a CT or MRI scan of the brain (n=121) and MRI scan of the spine (n=77) were performed. In selected cases, the patients were seen by other specialists. The first degree relatives of the index cases were examined. The protocol was approved by the Ethics Committee of the Medical Faculty, University of Oulu.

All data were recorded on a standardised form. Values were expressed as means (SD). The chi-square test and the independent samplest test were used to examine the significance of differences in frequencies, and in the mean ages at clinical manifestation between the groups. Two sided p values were calculated at a significance level of 0.05.

A total of 197 NF1 patients were identified in the area.4Full clinical data were available in 164 cases (83%) who fulfilled the NIH diagnostic criteria for NF1 (table 1). Adequate clinical details were available on all eight segmental NF cases ascertained.

The 164 NF1 patients included 111 probands (68%) and 53 affected relatives (32%). Seventy cases were sporadic, 91 were familial, and in three cases, the mothers had segmental NF. A total of 85 patients were males and 79 females.

Café au lait spots (CLS) were the most common finding (96%), followed by freckles (87%), Lisch nodules (70%), and neurofibromas (69%) (table 2). The most common combination of diagnostic features was CLS and freckles, which were seen simultaneously in 87% of the cases. In seven patients (six children under 6 years of age and one adult), the diagnosis of NF1 was made by reference to the presence of CLS and a family history. The frequencies of the main diagnostic criteria in the different age groups are shown in table 3. The mean number of diagnostic criteria present in the NF1 patients was four (range from two to six).

Café au lait spots were detected in all the patients in the age groups from 3 months up to 30 years, but they were less frequent in the older age groups (table 3). A total of 129 of those who had CLS (82%) had between six and 50 macules, and 22 (14%) had more than 50. Of the seven patients who had between one and five full sized macules (⩾1.5 cm), six were under 11 years of age and also had between six and 50 macules of a diameter from 0.5 to 1.5 cm, thus fulfilling the diagnostic criteria for NF1. One of the patients was an adult male who did not fulfil this diagnostic criterion with his three CLS. Six adult NF1 patients did not show any CLS.

Freckles in the axillary or inguinal region were seen in all the patients aged 11 to 20 years, 64% of those under the age of 11 years, and 95% of those aged 11 years or older. Even in the youngest age group (under 6 years), 52% of the patients had freckles (table3).

The number of neurofibromas increased from 4% in the age group under 6 years to 100% in that over 41 years (table 3). Neurofibromas were recorded more often in females (75%) than in males (64%), but the difference was not significant (p=0.7). Cutaneous and subcutaneous neurofibromas were both seen in about 70% of cases, and pendulous neurofibromas in about 40% of the affected patients. The number of cutaneous neurofibromas ranged from 10 to 100 in 45% of cases and exceeded 100 in 18%. A total of 77 patients underwent an MRI scan of the spine and spinal neurofibromas were observed in 26% of them. These patients included one family with familial spinal neurofibromatosis.16

A plexiform neurofibroma was diagnosed in 20% (33) of the patients, the youngest being a boy aged 15 months who had a tumour at the back of the neck. In nine cases, the skin was found to be hyperpigmented on top of the plexiform neurofibroma, and in two the skin was also hairy.

A total of 123 NF1 patients were studied by a neuro-ophthalmologist, yielding findings of Lisch nodules (⩾2) in 75% (92) of cases. In six patients, Lisch nodules were found only in one eye. Lisch nodules were seen in 81% of cases in the age group older than 5 years and in 89% of those over the age of 20 years. The number of Lisch nodules ranged from two to 10 in 47% of cases and exceeded 10 in 50%. In addition, there were four patients (3%) in whom only one Lisch nodule was seen, three of these being under 5 years old, while the fourth was a 12 year old boy.17

The presence of optic glioma was evaluated from a CT or MRI scan of the brain (n=121) paying special attention to the optic nerves. Optic glioma was found in 20% (n=24) of the patients. The mean age at diagnosis was 11 (SD 11) years (range 2-40 years), and in children under 16 years (n=18) it was 6 (SD 2.8) years (range 2-12 years). A total of 79% (n=19) of the cases were symptomatic on neuro-ophthalmological examination and 21% (n=5) asymptomatic.

Pseudarthrosis was seen in five patients, all with the tibial bone affected. Pseudarthrosis was diagnosed before or at the age of 1 year in all cases.18

Additional dermatological findings included haemangiomas in nine patients, hypopigmented macules in 16, and hyperpigmented regions with a dark “dirty skin“ impression in 50 cases. Troublesome itching was reported by 30 subjects (18%), 21 of whom suffered from atopy. A total of 23 (77%) out of these 30 patients experienced itching of the skin above the neurofibroma or in areas where neurofibromas appeared later. Abnormalities of the teeth were seen in 42% (n=68) of the cases, including 39 with increased caries, nine with positional abnormalities, and 20 cases reported to have a prosthesis by the age of about 30 years. Two patients had cysts in the jaw bone.

Short stature (⩽−2 SD for age and sex in Finnish children) was recorded in 18% of the patients, 22% of the females and 14% of the males (p=0.3). A markedly asymmetrical face was seen in 13 patients and other dysmorphic features in seven. Five patients had features characteristic of Noonan syndrome, although other members of these families with NF1 did not have any signs of Noonan syndrome. Macrocephaly (head circumference ⩾+2 SD for age and sex) was diagnosed in 29% (46), this being more common in the males (36%) than in the females (22%) (p=0.08). Five hydrocephalus patients were excluded. The MRI scan of the brain showed high signal intensity lesions in T₂ weighted images in 68% of the 31 cases of macrocephaly. Twelve patients had macrocephaly and learning difficulties and nine showed high intensity lesions. A total of 35/72 NF1 patients with a normal head circumference (47%) showed high signal intensity lesions in T₂ weighted MR images. These findings were significantly more frequent in the younger patients, being seen in 84% of cases under 16 years of age and in 24% aged 16 years or older (p=<0.001). Other major MRI findings included one previously undiagnosed patient with hydrocephalus and another with an adenoma of the hypophysis.

Malignant tumours were found in 11 cases (7%), the most common being malignant peripheral nerve sheath tumour, which occurred in five patients at the site of a previous plexiform neurofibroma. Of the total of 164 NF1 patients in the series, seven died during the seven year monitoring, and in three of them the cause of death was malignant peripheral nerve sheath tumour and its metastases, at the ages of 23, 26, and 56 years.19

Learning difficulties at school and in adulthood and delayed developmental milestones at preschool age were observed in 28% (46) of the patients, 72% (33) of whom were males and 28% (13) females (p=0.001). Fine motor coordination problems were seen in 39% (64) of cases, and the IQ was under 85 in 13% (22), eight of these being mentally retarded (IQ<70) and the remaining 14 having a borderline intelligence with an IQ in the range 70-85. Only 5% of the 91 adults (older than 20 years) had passed the school matriculation examination, including one who was studying at university and one who had a degree in sociology. Forty six percent of the adults were employed, 12% were unemployed, 4% were housewives, 29% were receiving a disability pension, and 9% were receiving a retirement pension. One third of the adults had never worked outside their home. A psychiatric diagnosis had been made in 19% (17) of the adult patients, including chronic psychosis in five and chronic depression in 12.

A total of 65% (106) of the NF1 patients suffered from one or more symptoms related to NF1 and needed either medical intervention, rehabilitation, or follow up. In 38% (63), the patient's medical problems had been treated before NF1 was actually diagnosed. The mean number of symptoms was 2 (SD 1), with a range of one to five per patient. The findings are shown in table 4, where they are compared with those of other large series. About half of the patients needed admission to hospital because of NF1, and 68% (111) had undergone related surgery. A total of 65% (107) had used the primary health care service three (SD 6) times (range 1-60) in the two years preceding the first genetic visit, and 89% (146) had had four (SD 7) (range 1-80) visits to a specialist hospital outpatient department. At present, 82% (50) of the patients who are alive and under 16 years and 38% (36) of the older ones attend for regular follow up examinations, at least once a year for children and once every second year for older patients (p=<0.001).

Constant, symmetrical spinal neurofibromas were discovered in one family, in two generations with a phenotype significantly different from classical NF1. This family has been described in detail elsewhere.16 In two additional families, all nine affected subjects had only cutaneous findings.

Altogether, 64% of the adults (n=91) were or had been married and 88% (n=51) of these had a total of 132 children (mean 2.6, range 1-11). Out of 132 offspring, 42% (56) were affected. The ratio does not differ significantly from the expected 0.5 (p=0.08). Seventy three percent of the 34 affected women with a history of pregnancy had reported new neurofibromas and growing of the existing cutaneous neurofibromas during their pregnancies, and four had had toxaemia of pregnancy. Three families had asked for prenatal diagnosis, but none had eventually considered the disease to be severe enough for any further prenatal analysis, although all the cases were indicative in terms of linkage analysis.

Segmental NF was found in eight adults, three of whom were mothers of daughters suffering from classical NF1. In six cases, the segmental area was on the left side of the body and in four of them it was in the upper quadrant. In four patients, the lesions were CLS or freckles or both, and in one these were accompanied by neurofibromas. Three additional patients had only a few neurofibromas and one patient had multiple Lisch nodules as the only lesion (table 5).

This clinical series comprises 164 NF1 and eight segmental NF cases. All the NF1 patients fulfilled the NIH diagnostic criteria. The 33 remaining cases out of the original population of 1974 who were not included in the clinical study were mostly adults who were excluded because of either inadequate information for clinical analysis or refusal.

The mean age of diagnosis of NF1 has been continuously dropping in the area concerned, from 20 years among patients born in the 1960s to less than six years for patients born in the 1980s. This is best explained by a growing awareness of the disease and by better knowledge of its diagnostic features. Also, the systematic study of first degree relatives has improved the detection of new affected cases. The youngest patient was only 3 months of age, but there were no cases in the discharge records in which NF1 was detectable or reported in a newborn infant. NF1 seems to be easily diagnosable before the age of 6 years according to the present data, as has been reported previously by others.2 10 21-23 All 27 children belonging to the youngest age group (under 6 years) presented with a diagnostic number of CLS, about half had freckles, and one out of five had Lisch nodules. It is interesting that 70% of the children in this age group had an affected parent.

An additional diagnostic feature which could be used in controversial cases proved to be hyperintense T₂ lesions in an MRI scan of the brain. These were found in 94% of the children under the age of 6 years who had had an MRI and in 84% of those under 16 years who had had an MRI head scan. These T₂ weighted hyperintense spots have been reported in 63-93% of published cases of NF1 children.24 25

The frequencies of the various diagnostic features in this population were commensurable with those reported in other large studies (table2). In particular, current data as well as other recent clinical NF1 reports13 14 have confirmed the value of axillary/inguinal freckles as an independent clinical diagnostic feature.

NF1 has been considered to be a slowly progressive disorder with hamartomatous lesions increasing in number and size with age, as shown by the appearance of diagnostic features in different age groups in the patients in this material and in the Welsh and French series (table 3). The only exception was CLS, which were present in all patients under the age of 16 years in these three series, while the older age groups included patients without them.8 9 12 In the present series there were no recorded cases of the disappearance of CLS.

Learning difficulties (28%) were less common than in the reports of Riccardi (70%)10 or Huson et al (33%)8 9 or in the NNFF international database (40%)14 (table 4). The difference is partly explained by the fact that most of Riccardi's cases were derived from a tertiary care hospital in Houston, Texas and most of the NNFF international database cases were sent from special NF clinics. Another reason may be that the older NF1 patients in this series had not been psychologically tested at school age and mild learning difficulties may have gone undiagnosed. This finding is supported by the fact that 50% of the patients younger than 16 years were found to have learning difficulties, while the corresponding figure in the older age groups was 15%. The overall frequency of mild intellectual handicap (IQ<85) was twice as high as in previous reports, while the frequency of mental retardation (IQ<70) was 5%, that is, comparable to that in other studies.9 10 14

From the 33 plexiform neurofibromas identified, five, that is, one out of every seven, showed malignant transformation to malignant peripheral nerve sheath tumour (MPNST).19

Eight new cases of segmental NF were identified. At least 150 segmental NF cases have been published.26 The first definition of a localised form of the disease was published by Gammel27 in 1931 and Crowe et al 6 published further cases in 1956 and started to use the term sectorial NF. Later on, Riccardi10 described segmental neurofibromatosis in 1982 as CLS and/or neurofibromas in a single unilateral segment of the body, without crossing the median line, with no familial history and with no systemic involvement. Five years later, Rothet al 28 divided segmental NF into four subtypes, described as a true segmental type (Riccardi's NF5), a localised type with deep involvement, a hereditary type, and a bilateral type. The present eight patients with segmental NF included three with true segmental NF and two with bilateral segmental NF. In the three familial cases, the daughter of an affected mother with segmental NF had classical NF1. This finding is exceptional and has been described previously in 15 published cases.10 29-33The findings in these families suggest that the disorder in the parent-child pairs are of the same origin and that, in addition to segmental NF, the affected parent has mosaicism of the NF1 gene. If so the patients with segmental NF represent mosaicism for the NF1 gene.2 9 10 32 33 It is sometimes difficult to discern true segmental changes from a variant form of classical NF1 and more research is also needed to evaluate the role of mosaicism.

Owing to the complexity and wide variability of NF, all affected subjects and their families need a proper clinical examination and genetic counselling, preferably by an experienced clinical genetic team. Early diagnosis is needed to alleviate unnecessary worries, and collaboration between specialists such as a paediatrician, neuropaediatrician, dermatologist, neurologist, ophthalmologist, oncologist, pathologist, neurosurgeon, paediatric surgeon, internist, and clinical geneticist is needed to recognise those NF cases that need immediate medical attention and to organise monitoring. In agreement with earlier reports,2 9 10 23 34 35 it was found that patients and their families appreciated genetic counselling and follow up visits, during which the whole spectrum of NF with its variability and unpredictability could be clearly explained.

The results reported here show that individually scheduled monitoring is needed from the view point of both the patient and the family. In the diagnostic phase, a follow up examination may be needed half yearly or yearly to confirm the diagnosis and to help the family adapt to the situation. Later on, monitoring is needed in cases of NF1 in particular, for early detection of possible developmental delay, macrocephaly, or optic glioma.2 11 13 15 23 34-36 In young adulthood or adolescence, patients may need individual counselling on the inheritance and genetic aspects of NF. Adult patients should be instructed to note any changes in their condition, such as a sudden increase in the size of the existing tumour, and for this purpose individually scheduled monitoring is needed.