Familial neurofibromatosis type I: Clinical experience with DNA testing K a r e n J. H o f m a n , MBBCh, a n d C o r i n n e D. B o e h m , MS From the Center for Medical Genetics, Department of Pediatrics, Johns Hopkins University School of Medicine, Baltimore, Maryland

To determine how DNA testing for familial neurofibromatosis type 1 (NF-I) would be used in a clinical setting by patients and physicians, we performed confirmatory DNA testing on 24 individuals with a family history of NF-I and on nine couples who requested DNA testing for current or future prenatal diagnosis. A further eight families were unsuitable for DNA linkage testing because of their pedigree structure. For the majority of persons the certainty of the test result was 95% to 99%. In five individuals, only one of whom was less than 6 years of age, the DNA-based diagnosis was discrepant with the clinical diagnosis at the time of referral. In all five cases, results of subsequent clinical reexaminations were consistent with the DNA diagnosis. We conclude that DNA testing by linkage analysis may be most useful as an adjunct to the clinical diagnosis of familial NF-I (I) in children less than 6 years of a g e in whom the full manifestations may not yet be apparent, (2) in NF-I families interested in prenatal testing, and (3) when the resources available for a complete clinical examination are limited. (J PEDIATR1992;120:394-8)

Neurofibromatosis type 1 is a chronic, progressive disease characterized b y abnormalities of growth of the nervous system and other organs, disfigurement resulting from neurofibromas that gradually increase in size and number with age, and diverse complications including learning disabilities, malignancies, and skeletal deformities. The prevalence is approximately 1:4000 in the general population. It is an autosomal dominant condition thought to be 100% penetrant 1-3;each individual who carries the gene will eventually manifest some clinical features. Mildly affected patients may not come to medical attention until the third or fourth decade, or may be identified only by extended family screening. The gene for NF-1 has been mapped to chromosome 17ql 1.24, 5; portions of it have now been cloned.6, 7 Since 1989, closely linked DNA markers have made accurate diSupported by a grant from the National NeurofibromatosisFoundation, Inc., New York, N.Y. Submitted for publication April 5, 1991; accepted Oct. 3, 1091. Reprint requests: Karen J. Hofman, MBBCh, Blalock 1008, Johns Hopkins University School of Medicine, 600 N. Wolfe St., Baltimore, MD 2120.

9/20/34170

394

agnosis by linkage analysis possible in most cases of familial NF- 1. Even before the availability of intragenie probes, DNA linkage studies in families with an already identified member with NF-1 were shown to be highly predictive of N F status. 8 The purpose of our study was to evaluate how DNA testing for NF-1 would be used by patients and physicians for prenatal diagnosis and for confirmatory testing. METHODS Notification of test availability was made to potentially interested patients through the 70 N F centers in the United States and through the National Neurofibromatosis FounCAL cM NF NF-1 RFLP

Caf6 au lait Centimorgan Neurofibromatosis Neurofibromatosis type 1 Restriction fragment length polymorphism

dation (NNFF) mailing list, which includes approximately 8000 NF-I families or persons with an NF-l-affected family member. The families and couples in this study were identified through an NF center staffed by a physician with expertise in NF-1. A clinical evaluation based on the 1987

Volume 120 Number 3

National Institutes of Health (NIH) Consensus Conference criteria for NF-19 (Table I) was requested for each affected and at-risk individual whose DNA was analyzed. Blood samples were sent alternately to either one of two laboratory centers chosen to perform DNA testing through the NNFF. The results that follow are based on samples sent to the DNA diagnostic laboratory at Johns Hopkins Hospital. Nine restriction fragment length polymorphism markers were chosen on the basis of proximity to the gene and ability to provide informativeness on both sides of the NF-1 gene. The RFLPs used in haplotype construction included Bgl II and Tag I sites (probes EW301 l0 and F9.811), a Pst I site (probe 21212), a Bgl II site (probe 17.1911), a Barn HI site (probe 2C11.7iI), an Msp I site (probe EW2061~ and Bgl II and Hin dlII sites (probe 20710). For calculation of the diagnostic error rate resulting from recombination, data from Goldgar et al. 13 and Fain et al. 1~ were used for the EW301, EW206, EW207, and 212 loci. Localization data and estimates of recombination rates from a smaller number of informative meioses were used for loci detected by probes 2Cll.7, 17.19, and F9.8.11 RESULTS The DNA testing was performed on 25 families referred from 20 different NF-1 centers. An additional nine families who were referred either elected not to proceed with testing (one family) or were unsuitable for analysis because of pedigree structure (eight families). Confirmatory testing. Twenty-four persons from 15 families requested testing to confirm or exclude NF-1. The age range of the persons undergoing confirmatory testing was 9 months to 68 years. The majority of persons (16/24; 70%) who requested confirmatory testing were 6 years of age and older. Clinical features of NF- 1 were absent in 11 (46%) of 24 persons, of whom six were younger than 6 years of age. The most common reason for requesting DNA testing was insufficient number of caf6 au lait spots to establish a clinical diagnosis; this was the case in 8 (33%) of the 24 persons tested. In one patient, aged 68 years, the only clinical feature of NF-1 was the presence of three Lisch nodules. Finally, in a 9-year-old girl, clinical features included hemihypertrophy, macrocephaly, and learning disabilities. In these two cases the results of DNA testing was positive for NF-1 in the former and negative in the latter; t h e s e findings were therefore consistent with the clinical diagnoses made on the basis of the NIH consensus criteria. 9 In 20 (83%) of the 24 persons tested, the DNA diagnosis correlated with the initial clinical impression, and in four (patients 1 to 4, Table II) the clinical diagnosis was changed after DNA testing. One other patient in whom there was lack of concordance between the DNA diagnosis and the

Familial neurofibrornatosis and DNA testing

395

Table I. NIH Consensus Conference criteria for NF-1 Two or more of the following criteria establish a clinical diagnosis of NF-I: 1. Six or more CAL macules (>5 mm in prepubertal patients and >15 mm in postpubertal patients) 2. Two or more neurofibromas, or one plexiform neurofibroma 3. Freckling in the axilla or inguinal region 4. Optic glioma 5. Two or more Lisch nodules 6. A distinct osseous lesion 7. A first-degree relative who meets the preceding criteria for NF-1 Data from Stuinpf DA, Alksne JF, Annegers JF, et al. Arch Neurol 1988;45: 575-8.

clinical diagnosis (patient 5) was referred for informativeness testing and will be discussed in the following paragraph. Among those referred for confirmatory testing, the diagnosis of NF-1 was excluded in one sibling (patient 1), previously considered to be affected, and was confirmed in the other sibling (patient 2), previously considered not affected. A third patient (No. 3) was reported to have Lisch nodules, although a slit-lamp examination had not been done. When DNA testing excluded the diagnosis of NF-1, a search for Lisch nodules was undertaken by slit-lamp examination; none was found. In a fourth patient (No. 4), NF- 1 was clinically diagnosed on the basis of six CAL spots greater than 1.5 cm in diameter, but DNA testing resulted in a reversal of the diagnosis. On a subsequent examination the patient was found to have only one typical CAL spot and two other areas of atypical hyperpigmentation. Informativeness and prenatal testing. A test for informativeness establishes whether DNA markers in a particular family will allow prenatal diagnosis in the future. This type of test was requested by 9 (36%) of 25 families. Among these nine couples, one patient (No. 5, Table II) withdrew from testing once informativeness had been established because DNA analysis reversed the prior clinical diagnosis of NF-1. Another patient in this group had a spontaneous pregnancy loss, and five other couples have not yet achieved pregnancy. The remaining two couples requested prenatal testing. In both cases, amniocentesis was performed and the fetus was found to be affected. In one case the parents elected to continue the pregnancy; the phenotype of the affected pregnant mother was mild, whereas her N F - I affected sister had died of a brain tumor at age 16 years. At age 3 months the infant has no clinical features of NF- 1. In the other case, the couple decided to terminate the affected pregnancy; the mother and two living children have NF-1. Predicted accuracy of the DNA test result. With the exception of three families, the certainty of the DNA test result was 95% to 99%. In one family studied for informative-

396

Hofman and Boehm

The Journal of Pediatrics March 1992

T a b l e II. Patients whose clinical diagnosis was changed by D N A testing

Patient No.

A g e (yr)

Sex

1

7

F

2

9

F

3

32

4

5

Clinical features in addition to family history of NF-I

Clinical diagnosis

DNA diagnosis (predicted accuracy of test result)

Outcome

Unaffected

Affected (99%)

Reexamined; 5 CAL spots found

Affected

Unaffected (99%)

NF-I excluded

F

1 CAL spot, abdominal freckling Hemihypertrophy, macrocephaly, learning disability Lisch nodules

Affected

Unaffected (99%)

3

M

6 CAL spots

Affected

Unaffected (98%)

36

M

1 CAL spot, 2 axillary freckles, Lisch nodules possible

Affected

Unaffected (99%)

Second slit-lamp examination--no Lisch nodules Second examination-only 1 CAL, 2 areas of atypical hyperpigmentation Slit-lamp examination--no Liscb nodules

ness, there was a 25% chance that a prenatal test would be inconclusive. The reasons were that (1) the coupling phase could be established through only one meiotic event and (2) the affected mother's DNA, although informative at flanking RFLPs, was informative at sites separated by 13 cM in female subjects. In two other families, NF-1 was ruled out with an accuracy of 84%. In both families the mother was the affected parent and her D N A was informative at only one site, 8 cM away from the NF-1 locus. With the discovery of intragenic RFLPs, D N A diagnosis with a higher degree of certainty will be possible in these cases. DISCUSSION Our experience with D N A analysis supports the conclusion that a clinical diagnosis is possible in the majority of patients with familial NF-1 by age 6 years. ~4'15 Even though D N A testing changed the diagnosis in five persons, in four of them the correct diagnosis could have been established by clinical examination alone had the N I H consensus criteria 9 been strictly applied. In the 7-year-old girl judged affected by D N A analysis, although we do not have information abou t Lisch nodules, the clinical impression of N F - i is Strongly suggestive, with five CAL spots and a family history of N F - I . In the 3-year-old boy judged unaffected by D N A analysis, it will be important to establish long2term clinical follow-up if further NF-1 manifestations appear, although the D N A test result excludes NF-1 with 98% certainty.

The majority of requests for D N A testing were to exclude a diagnosis of NF-1. The D N A analysis predicted that 21 (88%) of the 24 persons requesting confirmatory testing were unaffected. With the increasing availability of D N A analysis, it may be tempting for primary care practitioners to send material for a laboratory test rather than to evaluate a patient's condition by means of clinical criteria, In at least one third of the patients referred to us, evaluation by N I H criteria was either incomplete or not noted at the time that D N A testing was performed, even though the evaluations were performed by physicians at N F centers. The most frequent reason for incompleteness was the lack of a slit-lamp examination, a procedure that is simple, relatively inexpensive, and noninvasive. 16 Lisch nodules usually develop with time; in one study they were present in 40% of 3- to 4-year-old patients with NF-1, in 55% of those 5 to 6 years of age, and in 100% of those older than 20 years. 16 When present, Lisch nodules are a useful diagnostic criterion, especially in younger patients with NF-1, in whom they are found more commonly than neurofibromas. 16 Although reasonable for children less than 6 years of age, in whom sufficient clinical criteria may not be present, D N A testing may not be the most cost-effective option for those older than 6 years. The D N A linkage studies for NF-1 cost approximately $1000 per family, whereas the estimated cost of a clinical genetic and ophthalmologic evaluation is $500. In younger patients with a family history of NF-1, with few or no clinical features, D N A analysis may be a

Volume 120 Number 3

more attractive option than the anxiety generated by delay while waiting for a clinical diagnosis and the financial cost of additional examinations. A drawback of DNA analysis not encountered in diagnosis by clinical examination is the requirement of participation by key family members. Participation may not be possible for personal reasons, such as denial, confidentiality, autonomy, finances, or geography. Any reason that prevents the necessary thorough clinical examination of participating family members may limit the availability of DNA testing. Two families who called our center directly to find out about DNA testing could not pursue it because of limited finances, which prohibited their obtaining the requisite clinical evaluations. One other family withdrew from testing after key family members decided not to participate. Another potential disadvantage of DNA testing relates to the variable expression of the NF-1 gene and our consequent inability to predict which individuals will have severe NF-1 manifestations. Our sample is too small to generalize about patient attitudes with regard to prenatal diagnosis and decisions to continue or terminate a pregnancy. Even though parents may choose to know either antenatally or during infancy whether their child has NF-I, the impact of this diagnosis on child rearing has not been fully explored. Adverse psychologic effects of labeling persons have been noted for other conditions in which morbidity and life expectancy are variable and unpredictable, such as sickle cell anemia, 17 benign cardiac murmurs, 18 and hypertension.19 Although an unknown number of inquiries were made about DNA testing to the NNFF, during a 12-month period only 64 requests for DNA testing were received by the NNFF. The precise denominator is unknown, but it is at least equal to those on the N N F F mailing list and to patients served by the N F centers. This response is lower than expected, given the projected level of interest shown by NF1-affected individuals before the availability of DNA testing. Before the availability of DNA linkage for NF-I, the NIH conducted a survey among households in which a family member either had NF-1 or was related to a patient with NF- 1.2~Among 300 respondents the potential demand for confirmatory testing was projected to be 87%, the demand for potential prenatal testing 68%, and that for terminating an affected pregnancy 19%. in a similar survey done at Baylor College of Medicine, 62% of NF- 1 clinic respondents indicated that they would use prenatal testing, 44% said they would use the test to prepare emotionally for an affected child, and 19% of'the total group said that they would abort an affected fetus. 21 We have yet to explore the reasons for the lower-than-expected use of the test, but our experience is similar to that with Huntington disease,

Familial neurofibromatosis and DNA testing

397

another autosomal dominant condition, in which few affected families came forward when the test finally became available.22 The issue of DNA testing for sporadic NF-1, which can be a greater diagnostic dilemma, has not been addressed. In one study, clinical confirmation of sporadic cases of NF-1 was possible in only 32% of affected children less than 6 years of age. 14 The detection of numerous potential new mutations may be difficult because of the relatively large size of the gene, although technical advances may facilitate detection. In the future, the availability of a protein biochemical test would be useful and perhaps cheaper than either DNA linkage or mutation analysis for familial NF- 1, and especially for sporadic cases of NF- 1 in which the diagnosis cannot be made on clinical grounds. This kind of test would also be useful in the small percentage of patients with familial NF-1 whose pedigree structure makes them unsuitable for DNA linkage testing, as was the case in eight families referred to our center. It is our impression that the use of NF-1 DNA linkage testing by the population at risk and by health care providers is lower than expected. By comparing DNA results with clinical examination, we have shown that the NIH clinical criteria for NF-1 are highly accurate in patients 6 years of age and older. DNA testing in familial NF-1 may therefore be most useful and cost-effective for (1) identifying new cases in young patients who can then be monitored for known complications, with the potential for early intervention, (2) introducing the possibility of reproductive choice in couples with an NF-l-affected partner, and (3) establishing the diagnosis in an ambiguous case to relieve the anxiety of uncertainty. We thank Haig H. Kazazian, MD, and Abram Gabriel, MD, for valuable discussion. REFERENCES

1. Littler M, Morton NE. Segregation analysis of peripheral neurofibromatosis(NF 1). J Mcd Genet 1990;27:307-10. 2. Riccardi VM, Lewis RA. Penetrance of yon Recklinghausen neurofibromatosis:a distinction between predecessors and descendants. Am J Hum Genet 1988;42:284-9. 3. SergeyevAS. On the mutation rate ofneurofibromatosis.Hum Genet 1975;28:129-38. 4. Fountain JW, Wallace MR, Bruce MA, et al. Physical mapping of a translocation breakpoint in neurofibromatosis.Science 1989;244:1085-7. 5. O'Connell PO, Leach R, Cawthon RM, et al. Two NF 1 translocations map within a 600-kilobasesegment of 17q11.2. Science 1989;244:1087-8. 6. Cawthon RM, Weiss R, Xu G, et al. A major segment of the neurofibromatosis type 1 gene: cDNA sequence, genomic structure, and point mutations. Cell 1990;62:193-201. 7. Wallace MR, Marehuk DA, Anderson LB, et al. Type 1 neu-

398

8.

9. 10.

11.

12.

13.

14.

Hofman and Boehm

The Journal of Pediatrics March 1992

rofibromatosis gene: identification of a large transcript disrupted in three NF1 patients. Science 1990;249:182-6. Ward K, O'Connell P, Carey JC, et al. Diagnosis of neurofibromatosis 1 by using tightly linked flanking DNA markers. Am J Hum Genet 1990;46:943-9. Stumpf DA, Alksne JF, Annegers JF, et al. Neurofibromatosis conference statement. Arch Neurol 1988;45:575-8. Fain PR, Goldgar DE, Wallace MR, et al. Refined physical and genetic mapping of the NF 1 region on chromosome 17. Am J Hum Genet 1989;45:721-8. O'Connell PO, Leach R J, Ledbetter DH, et al. Fine structure DNA mapping studies of the chromosomal region harbouring the genetic defect in neurofibromatosis type 1. Am J Hum Genet 1989;44:51-7. Barker DF, Wright EC, Andersen WL, Willard HF, van Tuinen P, Ledbetter DH. Isolation and physical localization of chromosome 17 RFLP markers [Abstract]. Cytogenet Cell Genet (HMG 10) 1989;51:957. Goldgar DE, Green P, Parry DM, Mulvihill JJ. Multipoint linkage analysis in neurofibromatosis 1: an international collaboration. Am J Hum Genet 1989;44:6-12. Obringer AC, Meadows AT, Zackai EH. The diagnosis of neurofibromatosis-1 in the child under the age of 6 years. Am J Dis Child 1989;143:717-9.

15. Huson SM, Compston DAS, Harper PS. A genetic study of von Reeklinghausen neurofibromatosis in southeast Wales. II. Guidelines for genetic counselling. J Med Genet 1989;26:71221. 16. Lubs ML, Bauer MS, Formas ME, Djokic B. Lisch nodules in neurofibromatosis type 1. N Engl J Med 1991;324:1264-6. 17. Hampton ML, Anderson J, Lavizzo BS, Bergman AB. Sickle cell "nondisease": a potentially serious public health problem. Am J Dis Child 1974;128:58-61. 18. Bergman AB, Stamm SJ. The morbidity of cardiac nondisease in schoolchildren. N Engl J Med 1967;276:1008-13. 19. MacDonald LA, Sackett DL, Haynes RB, Taylor DW. Labeling in hypertension: a review of the behavioral and psychological consequences. J Chronic Dis 1984;37:933-42. 20. Watson B, Eldridge R, Stewart K, Schlesinger S, Parry DM, Mulvihill JJ. Neurofibromatosis 1: attitudes toward predictive genetic testing [Abstract]. Am J Hum Genet 1989; 45:A282. 21. Crandall KA, Edwards JG, Riccardi VM. Attitudes of individuals affected with neurofibromatosis toward prenatal diagnosis [Abstract]. Am J Hum Genet 1988;43:A165. 22. Quaid KA, Brandt JA, Folstein SE. The decision to be tested for Huntington's disease [Letter]. JAMA 1987;257:3362.

FELLOWSHIPS

Beginning in January 1992, fellowships available in pediatric subspecialties and those for general academic pediatric training will be listed once a year, in January, in THE JOURNAL OF PEDIATRICS. Each July, forms for listing fellowships available for the academic year beginning 18 months after publication are sent to the Chairman of the Department of Pediatrics at most major hospitals in the United States and Canada. In addition, a copy of the application form appears in the August and September issues of THE JOURNAL (please use the current form). Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to M o s b y - Y e a r Book, Inc., by October 15 preceding the listing year to ensure publication. Additional forms wilf be supplied on request from the Journal Editing Department, M o s b y - Y e a r Book, Inc., 11830 Westline Industrial Drive, St. Louis, M O 63146-3318/ 800-325-4177, ext, 4317, or 314-453-4317.