Article Text

Download PDFPDF

Bone health and fracture rate in individuals with neurofibromatosis 1 (NF1)
  1. T Tucker1,
  2. C Schnabel2,
  3. M Hartmann3,
  4. R E Friedrich3,
  5. I Frieling4,
  6. H-P Kruse4,
  7. V-F Mautner3,
  8. J M Friedman1
  1. 1
    Department of Medical Genetics, University of British Columbia, Vancouver, BC, Canada
  2. 2
    Institute for Clinical Chemistry, University Hospital Eppendorf, Hamburg, Germany
  3. 3
    Department of Maxillofacial Surgery, University Hospital Eppendorf, Hamburg, Germany
  4. 4
    Center for Osteoporosis Hamburg-Neuer Wall, Hamburg, Germany
  1. Dr T Tucker, Department of Medical Genetics, Medical Genetics Research Unit, Children’s & Women’s Hospital, Box 153, 4500 Oak Street, Vancouver, BC V6H 3N1, Canada; tbtucker{at}interchange.ubc.ca

Abstract

Background: Patients with neurofibromatosis 1 (NF1) are shorter than expected and often have low bone mineral density (BMD), but the pathogenesis of these bony problems is poorly understood.

Methods: We performed an exploratory study of BMD, 18 laboratory measures of bone metabolism, and fracture history in 72 adult NF1 patients.

Results: Eight of the 18 clinical biochemical measures of bone health had at least 10% of NF1 patients outside the standard reference range. Serum 25-hydroxy-vitamin D concentrations were low in 56% of the NF1 patients, serum parathyroid hormone (PTH) concentrations were high in 34%, and urine deoxypyridinoline cross-link concentrations were high in 50%. Mean serum 25-hydroxy-vitamin D concentrations were significantly lower in people with NF1 than in season matched controls in both summer (p = 0.008) and winter (p<0.001). 36 (50%) of the 72 people with NF1 studied had BMD consistent with osteopenia, and 14 (19%) had BMD consistent with osteoporosis. High serum PTH concentration, high serum bone tartrate resistant acid phosphatase concentration, and high serum calcium concentration were associated with lower BMD among the NF1 patients. Males were more likely than females to have low BMD. The reported frequency of fractures in individuals with NF1 was much higher than in their unaffected siblings and spouses (p<0.001), and pathological fractures were reported only in NF1 patients.

Conclusion: People with NF1 often have a generalised abnormality of bone metabolism. Further studies are needed to determine the biochemical and molecular basis of this abnormality.

Statistics from Altmetric.com

Neurofibromatosis 1 (NF1) is one of the most common autosomal disorders, affecting 1 in 3500 individuals.1 The protein product of the NF1 gene, neurofibromin, is ubiquitously expressed. The penetrance of NF1 is complete, but the manifestations are extremely variable.2 The most frequent clinical features of NF1 are café-au-lait spots, axillary freckling, iris Lisch nodules, and neurofibromas.

Some patients with NF1 have characteristic focal bony problems, which include dystrophic scoliosis, tibial bowing or pseudarthrosis, and sphenoid wing dysplasia. Individuals with NF1 are also shorter than expected, but their body proportions are usually not altered, suggesting that there is a generalised decrease in bone growth over the whole skeleton.3 Recent studies have shown that bone mineral density (BMD) is often much lower than expected in individuals with NF1.48 However, it is not known if the low BMD seen in people with NF1 is associated with an increased risk of fractures, as occurs in other people with low BMD.

The pathogenesis of bony problems that occur in NF1 patients is poorly understood and appears to be complex. Alterations in the proliferation and function of osteoclasts in NF1 individuals,9 and alterations of both osteoclast and osteoblast proliferation and function, have been observed in mouse models of NF1.912

In addition, the average serum concentration of 25-hydroxy-vitamin D, which plays a key role in bone metabolism and modulates the absorption of dietary calcium and phosphorus,13 is lower in people with NF1 than in individuals without NF1.14 15 However, it is not known if there is any relationship between low serum vitamin D concentrations and low BMD in NF1.

To help define the pathogenesis and clinical consequences of the generalised bony abnormality in NF1, we performed an exploratory study of BMD, various clinical measures of bone metabolism, and fracture history in a series of adult NF1 patients.

METHODS

Adults with NF1 were recruited from the NF-Clinic, Hamburg, Germany; all fulfilled the NF1 diagnostic criteria.16 Individuals with chronic illnesses or treatments known to influence bone health were excluded. This study was approved by the ethics committee of the Medical Chamber in Hamburg.

A total of 72 individuals with NF1 had blood drawn and urine collected in the winter to assess various clinical measures of bone health. Serum 25-hydroxy-vitamin D concentration was also measured on a blood sample drawn in the summer. Reference values (2.5 and 97.5 centile cut-offs) for both summer and winter 25-hydroxy-vitamin D were determined from 312 patients (56 for summer, 256 for winter) without NF1 and any known diseases that affect 25-hydroxy-vitamin D who were assessed by the same laboratory during the same time frame as the NF1 patients. Standard clinical reference ranges for all other analytes were provided by the manufacturer of each test based on the 2.5 and 97.5 centile cut-offs of the distribution observed in a normal population. The results for each analyte in each subject were classified as low (<2.5%), normal (2.5–97.5%) or high (>97.5%), using age and sex specific reference ranges when appropriate.

BMD was determined for all individuals with NF1 by dual energy x ray absorptiometry (DXA) scans of the total hip and lumbar spine using a Hologic QDR 1000 Plus Bone Densitometer. We defined normal BMD as a T score above −1 (that is, higher than 1 standard deviation (SD) below the mean for young adult women) at the total hip or lumbar spine site, osteopenia as a T score between −1 and −2.5 SD, and osteoporosis as a T score −2.5 SD or more below the mean (www.dv-osteologie.org). Control values for each region were provided by the Hologic QDR manufacturer (Hologic Inc, Bedford, Massachusetts, USA).

A questionnaire was administered to all subjects with NF1 to assess height, weight, history of bone fractures, amount of sun exposure, usual clothing coverage, occupation, exercise history, vitamin D intake, medication and dietary supplement use, and reproductive history (appendix 1). The levels of sun exposure, exercise and vitamin D intake were each classified using a point system that combines the answers from several related questions from the questionnaire, as previously described (appendix 2).17 Medications were grouped into therapeutic categories as multivitamin supplements, cardiovascular medications, central nervous system medications, gastrointestinal medications, thyroid medications, oral contraceptives, and other medications.

The history of bone fractures for each NF1 patient was obtained from the questionnaire. For purposes of comparison, a follow-up telephone call was made to each NF1 patient to obtain fracture histories for her or his spouse and/or siblings using the same questions as on the questionnaire. The circumstances surrounding each fracture were used to classify it as pathological (associated with trauma not considered to be sufficient to cause the fracture) or non-pathological (associated with trauma sufficient to cause the fracture). Pathological fractures were diagnosed by the physicians who reviewed the patient’s history during the study and from the reports of radiologists who reviewed the radiographs of the fractures.

Statistical analysis

Statistical analysis was performed using SPSS version 11.0 (SPSS, Inc, Chicago, Illinois). A Mann–Whitney U or Kruskal–Wallis H test was used to compare the means of each of the measures of bone health with respect to sex, medication use (none, any), sun exposure (sufficient >5 points, insufficient ⩽5 points), vitamin D intake (high >8 points, medium 5–8 points, poor <5 points) and amount of exercise (large 3–4 points, moderate 1–2 points, none 0 points). Between group comparisons of categorical variables were performed using a χ2 test. A value of p<0.05 was considered to be significant.

Ordinal logistic regression analysis was used to determine if overall BMD (classified as normal, osteopenic or osteoporotic) was associated with each of the measures of bone health, height or weight. First, a univariate model that used BMD as the dependent variable was tested for each measure of bone health, height or weight, with age and sex as covariates. Variables that had a value of p⩽0.1 were included in a multivariable model. Variables with a value of p<0.05 in the multivariable model were considered to be significant.

Non-parametric correlations were performed to determine if there was an association between overall BMD and sun exposure, vitamin D intake or exercise. A value of p<0.05 was considered to be significant.

The frequency of reported fractures (classified as none versus one or more) was compared between individuals with NF1 and their unaffected siblings and spouses using a χ2 test. A value of p<0.05 was considered to be significant.

Binary logistic regression analysis was used to determine if the presence of one or more fractures was associated with each of the laboratory measures of bone health. First, a univariate model was tested for each measure of bone health that used BMD as the dependent variable, with age and sex as covariates. Variables that had a value of p⩽0.1 were included in a multivariable model. Variables with a value of p<0.05 in the multivariable model were considered to be significant. Binary logistic regression analysis was also used to determine if the occurrence of one or more fractures was associated with BMD in individuals with NF1.

RESULTS

Seventy-two adults with NF1 were assessed for various measures of bone health. There were 29 males and 43 females. For males, the mean (SD) age was 43.4 (11.8) years (range 24–68 years), mean height was 172.0 (8.4) cm, mean weight was 72.9 (13.2) kg, and mean body mass index (BMI) was 24.8 (4.1) kg/m2. For females, the mean age was 42.1 (12.5) years (range 18–72 years), mean height was 161.8 (7.9) cm, mean weight was 61.5 (10.0) kg, and mean BMI was 23.4 (3.2) kg/m2. Sixteen of the 43 females had entered menopause at the time of the study. The average age of menopause among these women was only 41.8 (6.7) years (range 32–52 years).

We assessed 18 biochemical measures of bone health. The average values for each of the measurements for males and females are listed in table 1. For all measurements except serum 25-hydroxy-vitamin D concentration, the manufacturer’s reference range was used to classify values as low, normal or high. Eight measurements had at least 10% of the individuals outside the 2.5–97.5% reference range. Fifty-six per cent of the NF1 patients had low 25-hydroxy-vitamin D concentration (29/52 in the winter and 38/68 in the summer), 34% (24/71 patients) had high serum parathyroid hormone (PTH) concentrations, 24% (17/72 patients) had low serum bone specific alkaline phosphatase activity (AP), 13% (9/72 patients) had high serum osteocalcin concentrations, and 13% (9/71 patients) had high serum bone tartrate resistant acid phosphatase (bone TRAP5b) concentrations. In addition, 51% (36/71 patients) had high urine deoxypyridinoline cross-link concentrations, 31% (22/70 patients) had high 24 h urine calcium excretion, and 21% (15/70 patients) had high 24 h urine phosphate excretion.

Table 1 Mean (SD) values for markers of bone health among men and women with neurofibromatosis 1 (NF1); p values are given for the difference observed between the sexes

Seven males and 15 females reported taking multivitamin supplements; five males and six females, cardiovascular medications; three males and six females, central nervous system medications (for example, antidepressants); two males and three females, gastrointestinal medications; three males and three females, thyroid supplements; seven females, oral contraceptives; and four males and nine females other medications. Each group of medications was assessed to determine if there was a relationship to any of the 18 measures of bone health. With the exception of women on oral contraceptives being more likely to have a higher mean serum parathyroid hormone (PTH) concentration (p = 0.048), there was no association between the use of any of these medications and the measures of bone health.

Serum 25-hydroxy-vitamin D concentrations vary with the season and were, therefore, measured in both summer and winter. Mean serum 25-hydroxy-vitamin D concentrations were significantly lower in individuals with NF1 than in controls in both seasons (summer, p = 0.008; winter, p<0.001).

People with NF1 often have disfiguring skin lesions, but there was no difference in the mean serum 25-hydroxy-vitamin D concentration among the NF1 patients who did not cover their skin and either the 22 NF1 patients who covered most of their skin in the winter or the 13 NF1 patients who covered most of their skin in the summer.

Dietary vitamin D intake among the people with NF1 was estimated from a questionnaire. Sixteen of the NF1 patients reported poor vitamin D intake, 33 reported moderate intake and 23 reported high intake of vitamin D for both summer and winter. Reported vitamin D intake was not associated with serum 25-hydroxy-vitamin D concentrations in either winter (p = 0.398) or summer (p = 0.375).

BMD was measured by DXA at the lumbar spine and total hip sites. There were 23 NF1 patients (seven males and 16 females) with normal BMD (T score above −1.0), 34 (13 males and 21 females) with osteopenia (T score between −2.5 and −1.0) and 14 (nine males and five females, four of whom were post-menopausal) with osteoporosis (T score lower than −2.5). As shown in table 2, low BMD among the NF1 patients was associated with higher serum PTH concentration, higher serum bone TRAP5b concentration and higher serum calcium concentration, as well as with lower serum and 24 h urine creatinine concentrations. Sex was also significantly associated with BMD, with lower values in males than in females. BMD was not associated with height (p = 0.565) or weight (p = 0.513).

Table 2 Multivariable regression analysis of markers of bone health on bone mineral density (classified as normal, osteopenic, or osteoporotic) in individuals with neurofibromatosis 1 (NF1)

There was no association between BMD and physical activity (p = 0.399). Among nine individuals who reported no physical activity, three were females: one had entered menopause in her mid 30s and had osteoporosis in her mid 50s, one had osteopenia and one had normal BMD. Among the six males who reported no physical activity, two had osteoporosis, two had osteopenia, and two had normal BMD. There were 12 individuals who reported high physical activity. Among the nine females, three had osteopenia and six had normal BMD. The three males who reported high physical activity included one with osteoporosis, one with osteopenia, and one with normal BMD. There was no association of BMD with sun exposure (p = 0.510) or vitamin D intake (p = 0.173).

Among the 72 NF1 patients with information on fractures, 24 individuals (11 males and 13 females) reported a total of 41 fractures; 24 fractures in 14 individuals (8 males and 6 females) were considered to be pathological. There was no significant difference in the mean age of NF1 patients with (44.8 (12.6) years) or without a fracture (40.8 (12.4) years). Four pathological fractures occurred in four post-menopausal women. No significant association was observed between the occurrence of fractures and osteopenia or osteoporosis as defined by DXA (p = 0.163) or with any other measure of bone health among the NF1 patients.

The reported frequency of fractures in individuals with NF1 was compared with the frequency reported for their 73 unaffected siblings and 22 spouses. The average age of individuals with NF1 (42.4 (12.3) years) was not significantly different from the average age of their siblings and spouses without NF1 (43.1 (11.7) years). Among the 73 unaffected siblings, six fractures were reported in six individuals. No fractures were reported among the 22 spouses for whom fracture histories were available. Individuals with NF1 were much more likely than their unaffected siblings and spouses to have one or more fractures (p<0.001). Pathological fractures were reported only in individuals with NF1 and not in their unaffected siblings or spouses.

DISCUSSION

The extensive series of biochemical measures, DXA analysis of BMD, and bone health questionnaire used in this study together provide the most comprehensive assessment of bone metabolism ever conducted in a large series of individuals with NF1. The population studied is typical of adults seen in an NF1 specialty clinic. There is no reason to suspect that the patients who attended this clinic are more severely affected than the general NF1 population since bone health is rarely, if ever, a reason for referral of an adult to an NF1 specialty clinic.

We found frequent elevations (>10% of study population) of serum concentrations of PTH, bone TRAP5b and urinary deoxypyridinoline cross-links, and 24 h urine calcium and phosphate excretion. Brunetti-Pierri et al studied several markers of calcium and phosphorous metabolism and bone turnover in 16 individuals with NF1 who were selected because they had low BMD on DXA and found increased serum PTH concentrations in eight.15 These findings suggest that people with NF1 have increased osteoclastic activity that leads to increased bone breakdown. This interpretation is consistent with the association we observed between higher serum concentrations of PTH and bone TRAP5b with lower BMD.

However, the alterations of bone metabolism that occur in NF1 are complex, and, in addition to the evidence of increased bone breakdown, we observed low serum 25-hydroxy-vitamin D concentrations, as previously reported.14 15 25-hydroxy-vitamin D is converted in the body to 1,25-dihydroxy-vitamin D, which increases intestinal absorption of calcium and phosphate, which becomes available for bone formation.13 Lower serum 25-hydroxy-vitamin D concentrations might, therefore, be expected to reduce calcium and phosphate deposition into bone in people with NF1.

The abnormal biochemical measures of bone health observed in this study are consistent with a study that found increased in vitro proliferation and lytic activity of osteoclasts from individuals with NF1 compared to osteoclasts from individuals without NF1.9 12 Moreover, osteoclasts from Nf1+/−mice, which have a genetic defect analogous to that found in NF1 patients, show increased proliferation, survival, migration and lytic activity in vitro compared to wild type osteoclasts.9 12 An increased serum TRAP5b concentration, which is indicative of increased osteoclast numbers or activity,18 was found in our NF1 patients.

Osteoblast precursor cells from Nf1 haploinsufficient mice show premature apoptosis and higher than expected rates of proliferation in vitro.10 Although the majority of our results suggest a defect in osteoclast function in our patients, we cannot rule out concomitant osteoblastic defects. Recent histomorphometric analysis of bone biopsies from 14 NF1 patients showed a fourfold increase in the number of osteoblasts as well as a 10-fold increase in the number of osteoclasts in comparison to age and sex matched controls.19 Our observation that osteocalcin, a substance that is produced by mature osteoblasts during bone formation, was elevated in 14% of our study population, is consistent with an increase in the number of osteoblasts in NF1 bone. Brunetti-Pierri and colleagues15 also found elevated serum osteocalcin concentrations in two of 16 NF1 patients with low BMD.

Neurofibromin negatively regulates p21ras (Ras), a key cell signalling molecule. Elevated Ras can activate a number of signalling pathways including mitogen activated protein kinase (MAPK) pathway and the phosphatidylinositol-3-phosphate kinase (PI-3K) pathway, both of which have been linked to the NF1 phenotype. Although the exact role that elevated Ras-MAPK plays in osteoblast activity is still under debate, the evidence from Nf1 mouse models suggests that the Ras-MAPK pathway plays a key role in both osteoclast and osteoblast biology.912

Our observation of BMD T scores consistent with osteopenia in 49% of our NF1 patients and with osteoporosis in 19% is similar to that of other investigations that also found decreased BMD in people with NF1.48 15 20 However, the predominance of males with BMD values in the osteoporosis range was unexpected. Most previous studies of BMD in people with NF1 have not reported the sex of those found to have osteoporosis, and Lammert and associates found a larger proportion of females than males with T scores below −2.5.8 The reason that we found a larger proportion of males with very low BMD is not clear, and this observation requires independent confirmation.

We observed a higher incidence of fractures in individuals with NF1 than in their siblings and spouses without NF1. A history of fractures affecting an even larger fraction (52%) of NF1 patients was recently reported in another study.15 It is possible that our NF1 patients failed to recall or were unaware of some of the fractures that had occurred in their siblings or spouses, but the rate of fractures, and especially of fractures associated with minimal trauma, in the NF1 patients themselves appears to be unusually high. This observation is consistent with a generalised abnormality of bone that decreases its strength in people with NF1 and suggests that the presence of NF1 may be a risk factor for fractures. We did not find a statistically significant relationship of fractures to BMD or any other measure of bone health assessed in this study, but our study may have been too small to detect relationships that do exist.

Our study is cross-sectional and cannot be used to infer an individual NF patient’s risk for osteoporosis. A longitudinal study is required to define the natural history of osteopenia and osteoporosis and their relationship to fractures and other clinical manifestations in people with NF1.

Our findings clearly indicate that people with NF1 often have a generalised abnormality of bone metabolism. Further studies are needed to elucidate the precise nature of this abnormality and its relationship to haploinsufficiency of the NF1 gene. The role of neurofibromin in the pathogenesis of osteoporosis in the general population has not been explored, but our findings raise the possibility that neurofibromin may also play a role in osteoporosis in people who do not have NF1.

Appendix 1

Appendix 1

Appendix 1 Modified questionnaire from German MOD study 2002 (Erkal et al17). Questionnaire has been translated for the purposes of publication. Part 1.

Appendix 1

Appendix 1 Modified questionnaire from German MOD study. Part 2.

Appendix 1

Appendix 1 Modified questionnaire from German MOD study. Part 3.

Appendix 2

Points for physical activity, vitamin D intake, and sun exposure were assigned as follows: The physical activity score consists of questions 8 and 9. For the type of work, patients received 1 point if their occupation involved light physical activity and 2 points if it involved manual labour. In addition, a patient was given 1 point for participation in a sport for less than 1 h/week and 2 points for participation more than 1 h/week. From the total of these two questions, the patients were classified as having high physical activity scores if they had 3–4 points and low physical activity if they had 1–2 points.

The vitamin D intake and frequency of intake was assessed using question 11. The food items listed are either rich in vitamin D (for example, milk, diary products, fish) or counter the effect of vitamin D (alcohol, coffee etc). For each product rich in vitamin D eaten daily, the patient received 4 points; for each vitamin D-rich product eaten 3–4 times/week, 3 points; for each vitamin D-rich produce eaten once a week, 2 points; and for each vitamin D-rich product eaten less than once a week, 1 point. For each substance that countered the effect of vitamin D the patient consumed, 1 point was subtracted. If the patient had a total of more than 8 points, the diet was considered to be high in vitamin D; if the total was 5–8 points, the vitamin D intake was considered to be medium; and if the total was less than 5 points, the diet was considered to be poor in vitamin D.

Sun exposure was assessed from questions 14 and 15. Patients received 2 points if they spend their holiday in warm and sunny countries and 2 points if they like the sun. Patients were given 4 points if the individual went out when the weather was sunny at home regardless of the season, 3 points for every day exposure during the summer months at home, 2 points for 1–2 outdoor exposures per week during the summer months at home, and 1 point for 1–2 exposures per month in the summer months at home. Individuals were considered to have sufficient sun exposure if they received 5 or more points from questions 14 and 15 combined.

REFERENCES

Footnotes

  • Funding: This work was supported by Wealth Management Capital Holding.

  • Competing interests: None declared.

  • Patient consent: Not required.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.