Osseous Manifestations of Adult Gaucher Disease in the Era of Enzyme Replacement Therapy

Abstract

Enzyme replacement therapy (ERT) for Gaucher disease with mannose-terminated glucocerebrosidase has proved its therapeutic position with salutary effects on hematologic abnormalities, visceral infiltration, and quality of life. The frequency of new bone complications is reduced but not eliminated. Established osteonecrosis is beyond salvage. A systematic description of the burden of bone manifestations, persisting despite ERT, should inform future remedial strategies. Thus, we conducted this study to quantify the burden of residual skeletal disease and to explore putative relationships between clinical, radiologic, and biochemical factors and bone sequelae associated with disability.

Consecutive adult patients attending 3 referral centers in the United Kingdom were invited to participate. A representative group of 100 patients agreed to a structured interview, clinical examination, radiologic review, and completion of questionnaires. Osteonecrosis was evident in 43%, Erlenmeyer flask deformity in 59%, fragility fracture in 28%, osteomyelitis in 6%, and lytic lesions in 4%. Mobility was impaired in 32% of patients, while 15% experienced significant pain. The EuroQol 5D (EQ5D) quality of life summary measure was reduced and was associated with osteonecrosis and fragility fracture. Eight patients experienced new osteonecrosis after the start of ERT, though the presentation and evolution were often atypical. Nine patients had been treated from childhood and had an excellent outcome. Osteonecrosis was associated with age of presentation and with splenectomy-indeed, we observed a strong temporal association between splenectomy and incidence of osteonecrosis.

The biomarkers PARC/CCL18 and chitotriosidase were associated with prevalent osteonecrosis, and, in particular, with osteonecrosis occurring despite treatment. This study documents significant residual skeletal pathology and disability in patients in the mature phase of their treatment in a developed region. The temporal association between splenectomy and osteonecrosis implies causation. The relationship between clinical and biochemical markers and existing bone complications sets the scene for future prospective studies that will focus on management strategies informed by credible assessment of risk.

Abbreviations: ELISA = enzyme-linked immunosorbent assay, EQ5D = EuroQol 5D, ERT = enzyme replacement therapy, ICGG = International Collaborative Gaucher Group, MRI = magnetic resonance imaging, PARC/CCLI8 = pulmonary and activation-regulated chemokine/chemokine (C-C motif) ligand 18, WHO = World Health Organisation.

INTRODUCTION

With manifestations principally affecting the macrophage system, Gaucher disease (OMIM [Online Mendelian Inheritance in Man; http://www.ncbi.nlm.nih.gov/sites/entrez?db=omim] #230800, #231000) was the first lysosomal disorder for which effective enzymatic complementation was developed.³ Mannose-terminated human glucocerebrosidase (alglucerase, Ceredase and imiglucerase, Cerezyme [Genzyme, Cambridge, MA]) is preferentially taken up by macrophages and corrects the inborn error of glycosphingolipid metabolism in Gaucher disease outside the central nervous system. Enzyme replacement thus has salutary effects on the abnormal blood counts and pathologic infiltration of viscera by storage cells-as well as quality of life.^14,35 In a registry-based study, Weinreb and colleagues³⁵ demonstrated the clear and dramatic improvement in platelet count, hemoglobin concentration, spleen and liver size and in a number of bone-related outcomes, particularly bone pain and episodes of bone crisis.

Over the 2 decades since it was first introduced, nearly 5000 patients with Gaucher disease in more than 60 countries have received enzyme replacement therapy (ERT). Understanding and perceptions of the disease have changed during this therapeutic era. Experienced physicians no longer view it as a disorder with cytopenias causing bleeding, anemia, and susceptibility to infection due to bone marrow infiltration and splenomegaly, but rather as a principally skeletal disorder with disabling effects on bone structure and metabolism that, once established, are refractory to therapy. The skeletal manifestations of Gaucher disease include osteoporosis, osteolytic lesions, and osteonecrosis.³⁶ Bone injury causes pain and disability, often with the need for surgical intervention. The frequency of new bone complications is reduced, but not eliminated by enzyme therapy.^5,35,36 Although these episodes may reflect the effects of nearby disease in marrow tissue, their pathogenesis is not well understood,⁶ and, as a consequence, neither definitive treatment nor predictive testing for their occurrence is possible.

Modeling deformities, particularly the Erlenmeyer flask abnormality in the distal femur, are often-described radiologic features of Gaucher disease, although they appear to be asymptomatic.³⁶ The Erlenmeyer flask deformity has been estimated to occur in 30%-45% of adults with the condition, but hitherto no relationship between this radiologic finding and other skeletal complications of Gaucher disease has been reported.

Patients with Gaucher disease frequently have a lower bone mass than expected for their age and sex. Diffuse bone loss is exacerbated with increasing age,^4,26 but the risk of fragility fractures due to osteoporosis and its association with biochemical markers and other clinical parameters have not been formally quantified in Gaucher disease, to our knowledge.

Osteonecrosis and episodes of bone infarction are characteristic features; they may proclaim the diagnosis and frequently dominate the clinical course of Gaucher disease; osteonecrosis causes pain, disability, and a poor quality of life¹⁴ in patients of all ages.⁴ Typically the epiphyseal regions of the long bones in the hip, knee, and shoulder are affected, but osteonecrosis may also affect other sites. Once established, osteonecrosis is beyond rescue with ERT. Definitive diagnosis usually depends on magnetic resonance imaging (MRI), although plain radiography may detect late and destructive complications including collapse and secondary arthritis of joints.^22,23 The association between splenectomy (often required to manage life-threatening cytopenia in the era before specific treatment) and osteonecrosis has been known for many years,²⁴ but the nature of the link has been debated.²⁰ Mistry and coworkers²⁴ have also explored the link between osteonecrosis and the interval between diagnosis and initiation of specific treatment.

Osteolytic foci appear to be closely related to regions where Gaucher tissue is packed tightly with pathological macrophages in the intramedullary space. Osteolysis is associated with expanding lesions that typically cause cortical thinning and pathological fractures of the shaft of long bones.

The perspective of Gaucher disease as a reversible hematologic complaint with visceral involvement has been refined; it is now more appropriately considered as a longstanding metabolic disorder best suited to individualized treatment goals-and management as a chronic disease such as diabetes mellitus.^27,34 However, in the absence of rigorously defined therapeutic targets-including surrogate biomarkers analogous to hemoglobin A1C for diabetes with strong predictive value for disease outcome-the goals introduced for the long-term management of Gaucher disease lack scientific justification and prognostic validity. A comprehensive description of the burden of manifestations in late-stage Gaucher disease that persist despite treatment should inform remedial strategies for management of the condition.

As part of an effort to determine how Gaucher disease affects contemporary patients with access to enzyme therapy, and to what extent the disorder influences well-being and other clinically relevant features in the long-term, we documented the osseous manifestations of the condition in a national cohort of adult patients. Accordingly, we sought to quantify the burden of residual disease; to explore putative relationships between clinical factors (such as splenectomy and age of clinical onset) and sequelae in bone; and to identify other factors associated with disability and expression of the disease in the skeleton.

PATIENTS AND METHODS

Patients

The study was conducted with the approval of a national medical ethics review board for the United Kingdom. At the time of interview (September 2003 to September 2006), 2 specialist referral centers (Cambridge and London) were designated and funded for the management of adult patients with Gaucher disease. A third center (Manchester) was a pediatric unit that followed adult patients as part of a clinical partnership with the Cambridge center. As the centers were funded centrally by the United Kingdom Department of Health to provide these services, it is unlikely that more than a handful of treated adult patients attended other specialist clinics. Approximately 200 adult patients were known to the patient organization and the treatment centers at that time.

Inclusion criteria were as follows: a confirmed enzymologic diagnosis of Gaucher disease (Types I and III), age 18 years or over, and regular attendance at 1 of the participating centers. There were no specific exclusion criteria. Consecutive adult patients attending Gaucher disease outpatient clinics at the 3 study sites were invited to participate by letter and directly at scheduled clinic visits. At the time of recruitment, 120 patients attended the Cambridge clinic, 70 patients attended the London clinic, and 16 patients attended the Manchester clinic. Seventy-two patients were recruited at the Cambridge site, 16 patients at the London site, and 12 patients at the Manchester site. All patients gave signed informed consent. One hundred patients in total were recruited and completed the study; 4 patients with poor records of compliance or attendance were specifically not approached, and 5 declined to participate. Blood samples were taken on the day of the clinical assessment. Eleven children with Gaucher disease were recruited at 2 sites and will be reported elsewhere.

Structured Assessment

Clinical Bone Registry

A database was established for recording predefined clinical information obtained through interview, physical examination, and questionnaires. This was an observational, cross-sectional study of the prevalence of bony complications of Gaucher disease, putative clinical risk factors, radiologic imaging, quality of life data, and biochemical risk markers. Osteonecrosis was determined by a single clinician (PBD), based on a combination of clinical and radiologic evidence. Osteonecrosis, for the purposes of this study synonymous with the less accurate term "avascular necrosis," was defined on the basis of the clinical history of bone crises with supportive radiologic imaging, or the presence of conclusive evidence on MRI or plain radiologic imaging.³⁶

Bone mineral density was assessed by dual-energy X-ray absorptiometry (DEXA) at the lumbar vertebrae, hip, and wrist. Only the single most recent assessment was recorded. Values were reported as Z-scores and T-scores, in relation to the deviation in units of standard deviation from the median age-specific and young adult reference values, respectively. Data from anatomical sites affected by radiologically evident osteonecrosis were excluded.

Assessment of Erlenmeyer flask deformity was carried out by a single physician (PBD). Plain radiologic skeletal survey was also categorized to document joint deformities, the presence of vertebral fractures and nonvertebral fractures, lytic lesions, and replacement joint status. Fragility fractures were defined as fractures (including vertebral fractures) occurring in the context of little or no trauma. MRI was used to assess the presence and approximate onset of osteonecrosis, together with volumetric quantification of liver and spleen size.

Severity score index was calculated according to the method of Zimran et al,³⁸ on the basis of the clinical and laboratory features of the patient at presentation to the clinic and before definitive treatment.

EuroQol 5D (EQ5D) quality of life measures and Brief Pain Inventory determinations were recorded from questionnaires completed by the patients. In the EQ5D system, patients rate their health in 5 dimensions. The scores are summarized into a single "utility index score" or "health state score" according to a "time trade-off" weighting system derived from a representative survey of United Kingdom inhabitants,¹¹ where 0 represents death and 1 represents full health. In the Brief Pain Inventory, patients report their pain in the last 24 hours according to a scale of 0-10, with 0 indicating no pain and 10 being the most severe pain imaginable.

Laboratory Procedures

Chitotriosidase activity was estimated using the standard fluorimetric assay using 4MU-β-D-chitotrioside.¹⁷ The assay was carried out in several batches over the course of the study.

We estimated the chemokine (C-C motif) ligand 18 (CCL18)/pulmonary and activation-regulated chemokine (PARC) concentration using a flow cytometric bead array immunoassay system carried out on the Luminex system, using a modification of an enzyme-linked immunosorbent assay (ELISA) system,⁹ employing anti-CCL18 antibodies from R & D Systems (Minneapolis, MN) and recombinant CCL18 standards also from R & D Systems. The assay was carried out in a single batch on samples stored at −80°C.

Blood counts were determined in the routine laboratory using Coulter Counter technology (Beckman Coulter, Fullerton, CA). Glucocerebrosidase genotypes were determined for a limited panel of 5 frequent mutations.

Statistical Methods

We compiled descriptive statistics using Analyse-It software (version 2.2, 2008, Leeds, UK). For the purpose of testing associations between adverse bone outcome measures and putative clinical and biochemical risk factors, we used nonparametric statistical tests as reported in the Results section, also using Analyse-It software. Output variables (adverse bony outcomes) were established a priori and are listed here: osteonecrosis, fragility fracture, lytic lesions, pain, mobility, quality of life measures, and orthopedic surgical procedures. Input variables (putative risk factors) were established a priori and are listed here: glucocerebrosidase gene mutation, age at presentation and diagnosis, splenectomy status, severity score index at presentation, Erlenmeyer flask deformity, medication, and biomarkers of alternative macrophage activation.

RESULTS

Patient Group

Clinical characteristics of the 100 patients are shown in Table 1. Sixty patients were women, 4 had Type III (chronic neuronopathic) disease.

TABLE 1: Demographic Information for 100 Adult Patients With Gaucher Disease

Ninety-two patients were in receipt of ERT, and the mean duration of treatment was 8 years and 6 months. Only 9 patients had received ERT for less than 3 years. Nine patients had started ERT in childhood (aged ≤16 yr). Of the 8 patients not in receipt of ERT, 6 were on no specific treatment, 1 was in receipt of the oral substrate inhibitor miglustat, and 1 had undergone a successful bone marrow transplant at the age of 11 years. The latter patient, with complete engraftment, was excluded from analysis of biomarkers as these macrophage-derived proteins were of donor origin in this patient.

Thirty-two patients were receiving bisphosphonate drugs.

Mutations identified by molecular analysis of the human glucocerebrosidase gene are depicted in Table 2. Information was available for 89 patients and was categorized based on the detection of the N370S and L444P mutations. All other known mutations were categorized as "other." Patients in whom genotyping failed to detect 1 of 5 frequent mutations were categorized as "unknown." Information on genotype was not available for 11 patients. Thirteen patients (15%) were homozygous for the frequent mutation, N370S. Two patients were homozygous for the L444P mutation, commonly found in Type III disease. Twenty-one patients (24%) were compound heterozygotes, N370S/L444P.

TABLE 2: Frequency of Glucocerebrosidase Mutations in 89 Adult Patients With Gaucher Disease

No data were collected in this study in relation to anti-imiglucerase antibody status.

Bone Disease and its Consequences

The bone complications experienced by these patients are listed in Table 3. Osteonecrosis (Figure 1) was detected in 43 patients; 28 suffered fragility fractures. Acute osteomyelitis had occurred in 6 patients and chronic osteomyelitis was established in 3; Erlenmeyer flask deformity was identified after examination of plain radiographs of the lower femur in 59% of patients examined. Lytic lesions (Figure 2) were found in radiographs of 4 patients (6%). Bone complications were the presenting feature of Gaucher disease in 4 of 84 patients. Joint replacement surgery had been carried out in 22 of 100 patients. In 13 of these, multiple joints had been replaced.

TABLE 3: Prevalence of Bone Manifestations in 100 Adult Patients With Gaucher Disease

FIGURE 1: Plain radiograph of the pelvis and hips of a 40-year-old man with Type 1 Gaucher disease showing bilateral deformity and flattening of the femoral heads as a consequence of childhood osteonecrosis. Surgical clips outline an orthotopic liver transplant.

FIGURE 2: Plain radiograph of the left femur of a 47-year-old woman with Type 1 Gaucher disease showing expanded lytic lesions and pathologic fractures of the upper femoral shaft and within the lytic region.

Bone mineral density: Median Z-score at the lumbar vertebrae (n = 80) was −0.22, and median lumbar T-score was −0.6. Four patients met the criterion for osteoporosis according to the World Health Organisation (WHO) criterion of a T-score of −2.5 or less. Median hip Z-score was +0.05; median T-score was −0.24 (n = 64). Two patients met the criterion for osteoporosis at this site at the time of assessment. Median forearm Z-score was −0.8 (n = 53); median T-score was −1.2. Eleven patients met the WHO criterion for osteoporosis at this site.

Dates for 87 episodes of osteonecrosis in 41 patients were obtained for the group as a whole (asplenic and spleen-intact), principally based on reported bone crises with predefined classical symptomatic features. Although 43 patients in total experienced osteonecrosis, data were obtainable in relation to dates of episodes in only 41 patients. The remaining 2 patients were unable to provide details of the timing of episodes.

Patients With Osteonecrosis Despite Therapy

Eight patients experienced osteonecrosis after the start of ERT; far more (35 patients) experienced osteonecrosis only before the start of treatment. The pattern of osteonecrosis in patients who were in receipt of therapy and under close medical observation was noteworthy in that despite regular clinical surveillance, the complication was often completely asymptomatic but was observed radiologically to progress, albeit slowly. Brief clinical summaries of the 8 patients with osteonecrosis despite therapy are presented below.

Patient A, a 39-year-old man with Type 1 disease, suffered a series of classical bone crises starting 3 years after ERT was initiated, immediately after splenectomy for a persistent splenic abscess. The bone crises, all subsequently confirmed as osteonecrosis on MRI, affected L4, L5, S1, and S2 vertebral bodies and both humeral heads. Bilateral shoulder replacement was required. The initial dose of ERT was 20 U/kg per month, but episodes of osteonecrosis continued despite an increase in dose to 60 U/kg per month. He had no episodes during the last 2 years on a dose of 120 U/kg per month.

Patient B, a 41-year-old woman with Type 1 disease who was splenectomized just before she started treatment, was noted to have slowly evolving but asymptomatic osteonecrosis in the acetabulum and ileum on regular MRI (Figure 3). This extended over a 10-year period from the start of ERT. Treatment was interrupted twice during the course of pregnancies. The dose was 28 U/kg per month (given initially in divided doses 3 times weekly), increasing in several increments to 120 U/kg per month as biweekly infusions.

FIGURE 3: Axial T2-weighted MRI with gadolinium enhancement of the pelvis in a 41-year-old asplenic woman treated with imiglucerase for Type 1 Gaucher disease. Geographic alteration in signal intensity enhancing with gadolinium is evidence of a zone of granulation at the edge of avascular bone and a sign of ongoing osteonecrosis.

Patient C, a 38-year-old woman with Type I disease and intact spleen, had suffered 2 bone crises with findings of osteonecrosis on MRI of the right femoral head and left femoral shaft. ERT was begun at a dose of 24 U/kg per month. During the course of treatment, which was interrupted many times to accommodate pregnancies and breast feeding, she suffered intermittent, subacute, widespread episodes of bone pain. Clear, but slowly evolving changes of osteonecrosis were noted on MRI of both femoral shafts over a 5-year period.

Patient D, a 47-year-old woman with Type I Gaucher disease, intact spleen, and coexisting celiac disease, experienced several classical bone crises affecting the left knee before ERT. After the start of ERT at 26 U/kg per month, pain continued intermittently, and initial MRI documented an established area of osteonecrosis of the left femoral condyle. Subsequent imaging showed progression of the lesion, which culminated in a collapse of the joint surface.

Patient E, a 37-year-old asplenic woman with Type I disease, had experienced several classical bone crises with established joint destruction before ERT. Eighteen months after the start of ERT at a dose of 20 U/kg per month, progressive asymptomatic osteonecrosis was noted in several vertebral bodies.

Patient F, a 69-year-old woman with Type I disease and an intact spleen, suffered a classical bone crisis with collapse of the femoral head before the start of ERT. One year after the onset of ERT at a dose of 22 U/kg per month, she suffered a further bone crisis with collapse of the right humeral head.

Patient G, a 55-year-old asplenic woman, had experienced 4 classical bone crises with documented radiologic collapse of the right femoral head before initiation of ERT. Eleven years later, while receiving 50 U/kg per month, she suffered chronic pain in the left tibia with evidence of fresh osteonecrosis on MRI.

Patient H was a 60-year-old man in whom Gaucher disease was diagnosed following splenectomy for suspected cancer. He suffered 3 classical bone crises with confirmed osteonecrosis on MRI, beginning 8 months postsplenectomy and 4 months after the start of treatment at an initial dose of 28 U/kg per month.

Bone Disease in Patients Treated From Childhood

Of the subgroup of 9 adults who were treated from childhood, 1 patient presented with established bone disease (osteonecrosis) and was subsequently treated. A second patient presented with splenomegaly at the age of 3 years, suffered a single episode of osteonecrosis at 5 years, and was treated from 9 years of age. Of the remaining 7 patients, none had osteonecrosis either before or after treatment. Of the total number of 9 patients, none experienced osteonecrosis after start of treatment. Of the group of 9 patients, 1 had been splenectomized and 1 had neuronopathic disease, defined only by long-standing lateral gaze palsy.

Quality of Life

The median health state score derived from the EQ5D was 0.727 (confidence interval, 0.691-0.796). Three patients had a health state score <0. Thirty-one of 99 patients assessed had a health state score of 0.6 or less. The United Kingdom population norm for this measure is 0.8. This measure was significantly lower (median, 0.679) in those patients with a history of osteonecrosis than in those who had no history of osteonecrosis (median, 0.796) (p < 0.01, Mann-Whitney test). The health state score was also significantly lower (median, 0.626) in those who had suffered a fragility fracture than in those who had not (median, 0.796) (p < 0.001, Mann-Whitney test).

On the Brief Pain Index rating scale of 0-10, 43 of 99 patients scored "pain on average in the last 24 hours" as 0. Thus, 56 patients experienced some pain in the preceding day, with 15 of 99 patients experiencing pain on average at 5/10 or greater. Twenty-six patients were receiving analgesic medication. Of these, 25 patients estimated the degree of relief from pain they achieved: median relief, 70%, 9 patients had 50% relief or less. Thirty-two of 99 patients had impaired gait, and 2 were wheelchair-bound.

Association Between Clinical Risk Factors and Outcome

The occurrence of osteonecrosis was associated with younger age at which symptomatic Gaucher disease presented (p < 0.05, Mann-Whitney test), younger age at diagnosis (p < 0.05, Mann-Whitney test), splenectomy (p < 0.0001, chi-squared test), and severity score index (p < 0.001, Mann-Whitney test). It is noteworthy that osteonecrosis was not associated with fragility fracture or the presence of the Erlenmeyer flask deformity, as determined independently by radiologic evaluation. Only 1 of 13 patients homozygous for the N370S mutation had a history of osteonecrosis (p < 0.05, chi-squared test). There was a distinct peak in the first occurrence of osteonecrosis between 10 and 15 years of age (Figure 4).

FIGURE 4: Frequency histogram of first episode of osteonecrosis in 41 patients with Gaucher disease related to age at which the episodes occurred.

To investigate a possible causal relationship between osteonecrosis and splenectomy, we examined the temporal association between these events. Among the 44 splenectomized patients, 29 patients had experienced osteonecrosis and were able to provide dates for episodes. We obtained dates for 71 episodes of osteonecrosis and related these to the date of splenectomy. The median onset of osteonecrosis was 4.4 years (95% confidence interval, 3.0-5.5 yr) after the date of splenectomy, with a large peak of incidence in the first 5 years after splenectomy (Figure 5).

FIGURE 5: Frequency histogram of 71 episodes of osteonecrosis in 28 splenectomised patients with Gaucher disease. Time is expressed in years; 0 represents the date of the individual patient's splenectomy, thus an episode associated with a positive number occurred postsplenectomy.

Fragility fracture was associated with younger age at presentation (p < 0.05, Mann-Whitney test), severity score index (p < 0.0001, Mann-Whitney test), and splenectomy (p < 0.05, chi-squared test), but not with current age or Erlenmeyer flask deformity. Only 1 of 13 N370S homozygotes had suffered a fragility fracture, a significantly lower rate than in the remaining patients (p < 0.05, chi-squared test). Fragility fracture was associated with forearm T-score (p < 0.05, Mann-Whitney), but was not significantly associated with bone mineral density at other sites. As expected, history of fragility fracture was associated with the use of bisphosphonate therapies (p < 0.0001, chi-squared test).

Treatment Factors and Outcome

Ninety-two patients were receiving ERT at the time of the study; median dose was 30 U/kg per month (range, 13-94 U/kg per mo). (Some patients received higher doses after the study closed, including Patient A and Patient B described above.) Median duration of ERT was 8.5 years. Historical dosing data were not collected. There was no association between dose of ERT and prevalence of osteonecrosis.

After the report of Mistry et al,²⁴ we examined whether there was a relationship between the prevalence of osteonecrosis and the interval between diagnosis and initiation of ERT. Data were available from 77 patients. We noted that there was a significantly longer interval between diagnosis and treatment in those patients with osteonecrosis (p < 0.05, Mann-Whitney test), although this was compounded by splenectomy status.

Dates of 86 episodes of osteonecrosis were obtained in 40 patients (asplenic and spleen-intact) and related to the date of start of ERT. Results are shown in Figure 6. Seventy-three episodes occurred before the start of treatment, and 13 occurred after the start of treatment. One patient who had undergone a successful bone marrow transplant and who was not receiving ERT was excluded.

FIGURE 6: Frequency histogram of 86 episodes of osteonecrosis in 40 patients (asplenic and spleen-intact) who eventually received ERT. Time is expressed in years; 0 represents the date of the start of ERT for each individual, thus an episode with a positive number occurred after the start of treatment.

Biomarkers of Gaucher Disease Activity and Outcome

The PARC/CCL18 concentration was significantly greater in patients with a history of osteonecrosis (p < 0.01, Mann-Whitney test). The number of anatomical sites involved with osteonecrosis also correlated with PARC/CCL18 concentration (Pearson, r = 0.37; p < 0.001). Plasma chitotriosidase activity was also associated with history of osteonecrosis (p < 0.01, Mann-Whitney test). Chitotriosidase activity correlated with the number of sites affected by osteonecrosis (Pearson, r = 0.33; p < 0.01).

Among the 92 patients in receipt of ERT, PARC/CCL18 concentration was significantly elevated (median, 767 ng/mL) in the group of 8 patients who experienced osteonecrosis while receiving ERT, when compared with all remaining 84 patients (median, 315 ng/mL; p < 0.01, Mann-Whitney) or when compared with the group of 34 who experienced osteonecrosis only before enzyme treatment (median, 296 ng/mL; p < 0.05, Mann-Whitney) (Figure 7). Plasma chitotriosidase activities were elevated in the same group of 8 patients (median, 2643 μmol/L per h), but this was significant only when compared with the remaining 84 patients (median, 1069 μmol/L per h; 0.05, Mann-Whitney) and not with the group of 34 who experienced osteonecrosis before ERT (median, 1427 μmol/L per h). Subsequent analysis27a demonstrated that when chitotriosidase activity was corrected for the presence of the common inactivating mutation in the chitotriosidase gene, chitotriosidase activity was significantly higher in the aforementioned group of 8 patients than in the 34 whose osteonecrosis occured only before treatment. Neither marker was found to be associated with fragility fractures.

FIGURE 7: Distribution of PARC/CCL18 concentration in 3 groups of patients with Gaucher disease in receipt of ERT: Group 0 (n = 50) never experienced osteonecrosis, Group 1 (n = 34) experienced osteonecrosis only before the start of ERT, while Group 2 (n = 8) suffered osteonecrosis after the start of ERT. Individual data points are shown as open circles. The box represents the 25th, 50th, and 75th centiles. The solid T-bars represent the range of the data, while the hatched T-bars represent the 95th centile.

DISCUSSION

Here we describe 100 adult patients with Gaucher disease from several specialist centers in the United Kingdom in whom we have characterized the osseous manifestations of the condition, with particular reference to its prevalence and clinical and biochemical associations, and its effects on operational measures of life quality.

This study documents significant residual pathology and disability resulting from skeletal manifestations in many patients with Gaucher disease: it is clear that in the current therapeutic era of ERT, the principal effects of Gaucher disease are in bone. The complex osseous effects of this disease on health-related quality of life, pain, and disability continue to pose a challenge to clinical management as well as to scientific investigators who seek an understanding of pathogenesis.

An association between splenectomy and osteonecrosis in Gaucher disease has long been recognized,^16,28 but the nature of this association has been highly controversial. One view is that splenectomy simply reflects severity in patients with Gaucher disease and that the association is indirect.^8,20 The other view is that splenectomy accelerates the bony manifestations.^1,13,29,30 There has been little recent information to help resolve this question. In this context, the current findings, which demonstrate a temporal association and a peak incidence of osteonecrosis in the 5 years immediately after splenectomy, indicate a strong causal link between splenectomy and the occurrence of osteonecrosis. Splenectomy itself is associated with arterial and venous thrombosis as well as pulmonary arterial hypertension (reviewed by Crary and Buchanan⁷). Abnormalities in thrombosis, thrombolysis, and rheologic parameters such as blood viscosity and red-cell deformability have been proposed as contributory factors.

Although much is known about the histopathologic changes in skeletal Gaucher disease,¹⁹ the disturbances of normal bone physiology and growth that cause the osseous complications are hardly understood. To our knowledge, the detailed microvascular pathology of bone necrosis in Gaucher disease has yet to be examined, but an analogy with microcirculatory disorders such as Caisson disease and sickle cell anemia is inescapable. In these disorders where microvascular obstruction occurs at epiphyses, infarction of the humeri, femora, and vertebrae is well recognized.¹⁸ Abnormalities in rheologic parameters including blood viscosity and red-cell deformability have been described in splenectomized Gaucher disease patients, but were attributable to the splenectomy status and not the Gaucher disease.² In a review of a mixed group of patients with osteonecrosis (only a few of whom had Gaucher disease), abnormalities of thrombosis and fibrinolysis were detected in 74%.¹⁵ In the current study, the lack of association between osteonecrosis, fragility fractures, and the Erlenmeyer flask deformity emphasizes the distinct pathophysiologic mechanisms involved in the osseous manifestations of Gaucher disease.

The current study was neither powered nor designed to assess the response of bone density to treatment with bisphosphonate drugs in Gaucher disease. The lack of association between fragility fracture and most markers of bone density may be related to the time at which bone density was assessed, since this was carried out during the mature phase of treatment; in many cases osteopenia and osteoporosis had already resolved as a result of exposure to ERT and bisphosphonate drugs. The use of bisphosphonate therapy was clearly associated with a history of fragility fracture in this group. Unpublished experience at the Cambridge site is that lumbar bone density responds favorably to ERT and bisphosphonate drugs, and in combination these agents demonstrate striking therapeutic synergy. Abnormal bone density at the wrist appears to show little change in response to enzyme therapy and bisphosphonates, and thus more closely reflects the pretreatment state (Deegan, unpublished observation). This may be why wrist bone mineral density, even in the mature phase of treatment, is associated with history of fracture.

Data on the long term outcome of primary prevention of bone complications in children with Gaucher disease are difficult to obtain. Here we report the adult phenotype of a group of Gaucher patients, whose treatment was initiated in childhood. It is encouraging that, as predicted, there is a reduced incidence of splenectomy and osteonecrosis when compared with the group as a whole; moreover the effect is apparent despite the clear association between early onset of disease and later disabling complications. This observation is in keeping with similar observations made by Zimran and coworkers,³⁷ although to our knowledge the details have not been published in a peer-reviewed journal.

Those patients who experienced osteonecrosis despite treatment are particularly noteworthy-not only because the treatment failed, but also because the complication occurred during surveillance, including regular radiologic review of MRI studies. The clinical presentation and evolution of osteonecrosis in this subgroup was often cryptic: the patients infrequently reported the expected acute episodes of pain, and their lesions developed slowly-often occurring at unusual sites in the pelvis, particularly the ischium and sacroiliac regions. Without serial radiologic monitoring, and especially serial MRI, these treatment failures would not have been detected; in most instances, progressive focal necrosis of bone was identified only during retrospective evaluation during the conduct of this study. An alternative interpretation is that the presentation and evolution of osteonecrosis is modified in the context of the altered natural course of Gaucher disease induced specifically by ERT. Sims and colleagues,³¹ in a prospective study of 33 treatment-naive patients with skeletal complications of Gaucher disease, described a reduction in the incidence of osteonecrosis after starting treatment with imiglucerase. Nevertheless a few patients experienced bone crises, early in their course of treatment. Additional asymptomatic bone lesions were detected on imaging. In a large registry-based study, Mistry and coworkers²⁴ demonstrated that splenectomy status and the interval between diagnosis of Gaucher disease and initiation of ERT were independently linked to the incidence of osteonecrosis, despite treatment. The 8 patients in the current study were either splenectomized or had suffered osteonecrosis before the onset of treatment. Initial doses of ERT were <30 U/kg per month, though doses were increased latterly in response to complications. In the light of this evidence, it may be argued that splenectomy or prior episodes of osteonecrosis are an indication for higher initial dosing of ERT.

Data on current, but not previous, dose of ERT were gathered in this study (other than in the 8 cases reported). The doses of imiglucerase used in United Kingdom centers are less than in a number of countries such as the United States and Germany, but equivalent to doses used in The Netherlands and Israel. Little can be concluded from the absence of a relationship between dose of ERT and the prevalence of osteonecrosis in the group as a whole. Most osteonecrosis occurred before initiation of treatment, and history of osteonecrosis may have influenced decisions about treatment. As a result of the provision of reimbursement from national sources of health funding in the United Kingdom, doses of enzyme therapy have increased steadily. Further study is needed to investigate any relationship of the initial dose with the incidence of osteonecrosis occurring despite treatment.

Commonly used biochemical markers in Gaucher disease, the chitinase, chitotriosidase, and the chemokine PARC/CCL18, which are stable and present in the blood, correlate with the degree of abnormal bone marrow signal on MRI¹⁰ but have not hitherto been clearly associated with bone complications.⁹ Although the response of chitotriosidase to enzyme therapy was not associated with progressive bone disease, the therapeutic response of a newly identified blood biomarker, the cytokine macrophage inhibitory protein 1-beta (MIP-1β), has been reported and appears to predict bone outcomes.³³

In the present study, bone disease was carefully and independently characterized, and the biochemical markers were determined at the time of the clinical assessment. The association between biochemical markers and bone disease at this late phase of treatment may reflect resistance to enzyme therapy, at least in terms of biomarker response, in individuals with established bone disease. However, the association of serum PARC/CCL18 concentration with osteonecrosis occurring despite enzyme therapy suggests that this marker may identify a group of patients with developing bone pathology that is resistant to treatment. We recognize that it is not possible to draw firm conclusions in associative studies of this type; clearly a prospective investigation, before and after the introduction of enzyme therapy, will be required to explore the predictive power of surrogate biomarkers in relation to the osseous complications of Gaucher disease.

Several review articles have summarized the literature on orthopedic and osseous aspects of Gaucher disease,^{12,21,25,32,36} but to our knowledge there have been few large single studies, with the exception of the detailed study by Stowens and colleagues³² in the era before the introduction of enzyme therapy. The International Collaborative Gaucher Group (ICGG) registry has provided additional information,⁴ but these registry data are characteristically incomplete for most participants. We conducted the present study to describe systematically the skeletal involvement in a well-characterized and representative group of patients for whom specific treatment had been available for 15 years. The prevalence of major bone complications is higher in this group than in the patients reported in the ICGG registry.⁴ The difference may relate to the distribution of genotypes, with fewer N370S alleles in this cohort than in the ICGG registry; splenectomy was also more prevalent in the group here reported.

To a considerable extent, this national study operationally defines the contemporary, somewhat unsatisfactory, status of Gaucher disease in the current therapeutic era in a developed economic region. Disabling bone complications persist and their prevalent manifestations are principally based on patterns of osseous disease that were long-established in the pretreatment period. These include associations with splenectomy (no longer common practice) and the onset of bone infarction during puberty. Nonetheless, recurrent episodes of overt and silent osteonecrosis persist in British patients despite enzyme therapy, albeit at modest doses. It is striking that the patients in whom treatment started in childhood were free of active disease in the skeleton and had an excellent quality of life as mature adults.

Episodes of osteonecrosis, which show an association with disease activity (reflected in frequently measured surrogate biomarkers in the blood) clearly represent a failure of treatment and immediately suggest prospective studies to improve disease management and to develop more specific means to identify patients at risk from these complications for early intervention. With the introduction of orphan drug legislation and wide acceptance of the therapeutic efficacy of enzyme replacement in Gaucher disease, as well as a willingness of health funding organizations to reimburse its costs, there are now greater opportunities to overcome painful and disabling osseous complications. The emergence of enzyme preparations other than imiglucerase also offers an opportunity for wide-ranging prospective studies of dose-response in the skeleton, as well as a competitive containment of costs.

Our findings provide strong support for the earliest possible introduction of specific therapy in Gaucher disease; this will require greater diagnostic awareness of the condition and ethically appropriate methods to enhance its detection in the cryptic presymptomatic phase. Any integrated program of management for the condition thus needs to expedite diagnosis and facilitate the introduction of therapy adequate to prevent ongoing manifestations of Gaucher disease that, as we show here, remain prevalent in the skeleton.

ACKNOWLEDGMENTS

The authors acknowledge the kind support of the members of the Gaucher's Association and the cooperation of the patients attending our centres.