Bone Mineral Density and Fracture Rate in Response to Intravenous and Oral Bisphosphonates in Adult Osteogenesis Imperfecta

Transcription

1 Calcif Tissue Int (2010) 87: DOI /s y Bone Mineral Density and Fracture Rate in Response to Intravenous and Oral Bisphosphonates in Adult Osteogenesis Imperfecta Jay R. Shapiro Carol B. Thompson Yimei Wu Martin Nunes Carolynn Gillen Received: 12 November 2009 / Accepted: 13 May 2010 / Published online: 11 June 2010 Ó Springer Science+Business Media, LLC 2010 Abstract The effect of bisphosphonate treatment on bone mineral density (BMD) and fracture rates was assessed in adults with osteogenesis imperfecta (OI). This observational nonrandomized study included 90 OI adults treated with intravenous pamidronate (n = 28), oral alendronate (n = 10), or oral residronate (n = 17) or not treated (n = 35). There were 63 type I, 15 type III, and 12 type IV OI patients. BMD results were observed for up to 161 months and an average of 52 months of treatment. For type I and grouped type III/IV patients, treatment with pamidronate showed an increasing rate in L1 L4 BMD from baseline (0.006 [P = 0.03] and [P \ 0.001] gm/cm 2 /year, respectively); oral bisphosphonate treatment showed a significant increasing rate in L1 L4 BMD (0.004 gm/cm 2 /year [P = 0.047]) for type I patients. Pamidronate-treated type III/ IV and oral bisphosphonate-treated type I patients showed significant increases in total-hip BMD (0.006 [P = 0.003] The authors have stated that they have no conflict of interest. J. R. Shapiro (&) Y. Wu M. Nunes Osteogenesis Imperfecta Program, Kennedy Krieger Institute, 707 North Broadway, Baltimore, MD 21205, USA shapiroj@kennedykrieger.org Y. Wu Wu@kennedyKrieger.org M. Nunes Nunes@kennedyKrieger.org C. B. Thompson Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St., Rm. E-3150, Baltimore, MD , USA C. Gillen Department of Nursing, Kennedy Krieger Institute, 707 North Broadway, Baltimore, MD 21205, USA and [P = 0.046] gm/cm 2 /year, respectively). Bisphosphonate effect on fracture rate was assessed for 5-year periods before and after treatment in 51 treated and 22 nontreated individuals matched for age at which bisphosphonate was first administered to the treated group. Bisphosphonate treatment did not decrease fracture rate in type I OI patients. Fracture rate decreased in type III/IV patients following pamidronate but not following oral bisphosphonate treatment. These results underscore a need to consider whether bisphosphonate treatment is appropriate for all adults with OI. Keywords Osteogenesis imperfecta Bisphosphonate Bone mineral density Fracture Pamidronate Osteogenesis imperfecta (OI) is a heritable disorder of connective tissue estimated to affect approximately 25,000 individuals in the United States [1]. Recent interest in OI has focused on two aspects of the disorder. First, 90% of the cases studied to date have been found to result from mutations affecting type I collagen genes [2]. More recently, mutations involving the CRTAP, LEPRE 1, or PPIB genes which alter proline hydroxylation of type I collagen have been reported in severe and recessively inherited OI types [3, 4]. Second, multiple observational studies using different bisphosphonates, but most frequently pamidronate, have reported that bisphosphonates increase bone density, decrease musculoskeletal pain, increase vertebral height, and decrease fracture rate in OI children [5, 6]. However, two recent analyses of published studies highlight the limited data with which to support a uniformly positive bisphosphonate effect on fracture rate in children [7, 8]. These generally favorable outcomes in children have proven more difficult to demonstrate in adults for several reasons. First, the measurement of bone mineral density

2 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas 121 (BMD) by dual-energy X-ray absorptiometry (DXA) is more difficult in many OI adults because of the progression of their bone deformity over time. OI adults frequently have surgical hardware in their limbs or spine which affects the ability of these sites to be measured by DXA. Second, unlike the relatively high fracture rate in children, the incidence of fractures decreases dramatically after puberty. Thus, assessing fracture incidence in response to treatment during midlife is more difficult. Finally, compliance with treatment and testing schedules is less rigorous with adults compared with children, where parents oversee adherence to treatment schedules and clinic visits. Although bisphosphonates have been widely used in the treatment of OI adults, their effectiveness, in terms of bone density or fracture rate, has not been fully evaluated. This report focuses on the effects of treatment with intravenous pamidronate and the oral bisphosphonates alendronate and residronate with respect to two major outcome measures: annualized linear rates of change in BMD and fracture incidence before and after treatment. Subjects and Methods The analysis of patient data for this study was approved by the Institutional Review Board of the Johns Hopkins Hospital. This is a retrospective observational treatment study conducted at the Osteogenesis Imperfecta Clinic of the Kennedy Krieger Institute (Baltimore, MD). The designation of OI type was based on the Sillence classification [9]. Patients classified as type I OI had blue sclerae, had absent or minimal limb deformity, and were fully ambulatory. Subjects classified as type III or IV OI had a short stature, white or blue sclerae, scoliosis, and limb deformities which were severe enough to impair normal ambulation and required constant use of a wheelchair (type III) or the use of a cane or walker (type IV). Included in the physical examination was an evaluation of dental status and hearing. Patients were physician- or self-referred to the OI Clinic Program and were not randomly assigned to receive a type of treatment. The decision to treat was based on several factors: the frequency of postpubertal fractures and a baseline DXA Z score or T score at the lumbar spine or total hip of -1.5 or lower (with the added risk factor of OI) as recommended by the American Association of Clinical Endocrinologists [10]. Patients had not previously received bisphosphonates. Once having elected to receive treatment, the patient selected pamidronate or oral bisphosphonate. To facilitate analysis where patient treatment numbers were small, we elected to combine type III/IV subjects as one group. Similarly, we combined alendronate and residronate into a single category of oral bisphosphonates. The intravenous dose of pamidronate (Novartis, Basel, Switzerland) was 1.5 mg/kg body weight to a maximum of 60 mg/infusion administered as a single dose over 4 hours in normal saline. Patients were treated every 3 4 months depending on their adherence to schedules. Alendronate (Merck, West Point, PA) and residronate (Ely Lilly, Indianapolis, IN) were administered in standard weekly oral doses of 70 and 35 mg, respectively, with instructions as provided by the manufacturers. Diets were supplemented with 800 1,200 mg of calcium/day and with 400 IU of vitamin D/day. Fracture incidence was determined for a 5-year period preceding initial bisphosphonate treatment and for a 5-year period after treatment was started. In order to compare fracture rates for the nontreated group with fracture rates during the pre- and posttreatment periods in the treated groups, treated subjects were stratified by ages at which treatment was administered. These average ages were applied to the nontreated group by concordant age groupings to establish a comparable pre- and posttreatment time frame for those not treated. Since self-reporting fractures may be inconsistent in OI patients, each patient s medical record was examined for X-ray or orthopedic reports of the event. When required, patients or their local facilities were contacted for X-ray confirmation of a fracture. DXA scans were obtained at the Johns Hopkins Hospital Department of Nuclear Medicine on a Hologic (Waltham, MA) 4500 instrument. Although most patients had DXA scans performed on this machine, several patients had DXA scans performed at their local facility. Only scans performed on the same machine are reported. BMD measurements were not included from sites where surgical hardware would alter the readings or where bone deformity or severe scoliosis would produce an artifact in measurement. DXA scans were obtained at approximately 12-month intervals. Statistical Analysis The distribution of patients with DXA examinations across OI types and treatments was reviewed to determine which, if any, categories would need to be excluded or combined with other categories because of sample sizes too small for analysis. Three sites from the DXA scans were considered for analysis: L1 L4, total hip, and femoral neck. For each OI category, a linear random coefficient growth curve analysis [11] was performed with all available DXA scan data to compare annualized rates of BMD change between treatment groups, adjusting for gender and age at baseline DXA scan. BMD analysis included 55 treated and 35 nontreated patients. Annualized rates of change for each treatment group and between groups within an OI category were determined from these analyses. For each OI category, a

3 122 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas negative binomial analysis using generalized estimating equations was performed with the 5-year pre- and posttreatment counts of fractures to compare the relative fracture rates between treatment groups and between treated and nontreated groups, adjusting for gender and age at baseline DXA scan [11]. Relative fracture rates by treatment group within OI categories were determined from these analyses. Statistical analyses were performed with STATA v10 (StataCorp, College Station, TX). Results Demographic and Treatment Characteristics The 94 patients tracked in this cohort had the following distribution of OI types: type I (n = 64), type III (n = 15), type IV (n = 12), and type V (n = 3). Because of the small number (n = 3) of patients with type V OI and one type I patient with no baseline BMD measurement, the study population was restricted to 90 patients for analysis. Of these 90 patients, 55 were treated with intravenous pamidronate or oral alendronate or residronate and 35 were not treated. The distribution of patients in each clinical and treatment group is presented in Table 1 along with summary statistics on age, gender, and length of treatment. Overall, 67% of the patients were male and the mean (SD) age of the patients was 39.4 (11) years. Dentinogenesis imperfecta and hearing loss occurred in each OI type: 28.6% of the patients had dentinogenesis imperfecta and 24.2% reported hearing loss [12, 13]. Thirty-five of the patients were not treated. These individuals had initial baseline studies but follow-up DXA studies at longer intervals than for the treated patients because of compliance with scheduling. Nineteen of the nontreated patients had only baseline DXA measurements. For the other 16 nontreated patients, the period of observation varied from 3 to 118 months. For treated patients, the period of observation varied from 13 to 161 months. The type I treated group was older than the nontreated group, 42.1 and 37.3 years, respectively. For type III/IV patients, mean ages for the nontreated and treated groups were 40.9 and 36.2 years, respectively. Table 2 compares treated to nontreated patients at baseline by OI type for the three DXA scan sites. BMD results were observed for up to 161 months and an average of 52.2 months for the 55 treated patients and 39.5 months for all nontreated patients including those with only baseline DXA. The type I treated group had significantly lower BMD measurements, about 9 and 7 gm/cm 2, than the nontreated group at the L1 L4 and total-hip sites, respectively. However, significant differences were not evident for the more severely affected type III/IV patients. The treated vs. nontreated groups did not differ on baseline BMD at the femoral neck site in either OI type. Table 1 Demographic characteristics of study patients by treatment and OI type Treatment Type I Type III/IV Total Age (years) Mean (SD) Males (%) Months observed (25th a, 75th a ) N Age (years) Mean (SD) Males (%) Months observed (25th a, 75th a ) n Age (years) Mean (SD) Males (%) Months observed (25th a, 75th a ) n Not treated All 37.3 (12.2) Baseline only Baseline plus Treated 41.1 (14.6) 32.8 (6.8) All 42.1 (11.3) Pamidronate 45.0 (13.3) Residronate 39.4 (10.2) Alendronate 40.2 (6.4) 65.4 (0, 27) (10.0) 50.0 (0, 0) (10.5) 83.3 (19, 52) (10.9) 75.7 (27, 72) (9.5) 76.5 (26, 73) (8.9) 92.3 (41, 60) (4.3) 42.9 (27, 69) (2.3) 55.6 (0, 27) (11.6) 60.0 (0, 0) (13.3) 50 (21, 72) (8.7) 55.6 (21, 65) (11.3) 54.6 (32, 65) (12.1) 50.0 (26, 112) (9.1) 66.7 (13, 19) (9.9) 62.9 (0, 27) (0, 0) (20, 52) (23, 72) (27, 72) (31, 109) (19, 49) 10 a 25th indicates the 25th percentile; i.e., 25% of the patients were observed for less than or equal to this number of months. This is similar for 75th

4 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas Table 2 Comparison of treated and nontreated patients at baseline visit by OI type for three DXA sites Site OI type Group n BMD Mean (SD) gm/cm 2 L1 L4 Total hip Femoral neck I 0.02 Not treated (0.14) Treated (0.11) III/IV 0.3 Not treated (0.16) Treated (0.19) I 0.04 Not treated (0.11) Treated (0.12) III/IV 0.62 Not treated (0.35) Treated (0.20) I 0.15 Not treated (0.12) Treated (0.17) III/IV 0.85 Not treated (0.22) Treated (0.19) BMD Response to Pamidronate Annualized linear rates of BMD change (gm/cm 2 /year) from baseline to an average of 39.5 months of observation were adjusted for age at baseline DXA scan and gender. The rates were estimated for each treatment group within the two OI types (I and III/IV) as shown in Table 3. Subsequent references to BMD change describe these annualized linear rates. For type I patients in the nontreated group, there was a nonsignificant change in L1 L4 BMD of gm/cm 2 / year (P = 0.6). Treatment with intravenous pamidronate showed a significant annualized increase in L1 L4 BMD of gm/cm 2 /year (P = 0.03). Pamidronate treatment did not show a significant rate of BMD change at the total hip or at the femoral neck in type I patients. For type III/IV patients, pamidronate treatment showed significantly increasing rates of L1 L4 BMD change of gm/cm 2 /year (P \ 0.001) and of total-hip BMD of gm/cm 2 /year (P = 0.046). As with type I patients, type III/IV OI patients treated with pamidronate did not show significant rates of BMD change at the femoral neck. Nontreated patients in this group showed no significant change in BMD at any of the three sites. P BMD Response to Oral Bisphosphonates (Alendronate and Residronate) Patients administered alendronate or residronate were combined for the analysis of response to oral bisphosphonates because of small numbers in each treatment group (Table 3). In type I patients, treatment with oral biphosphonates showed significant increasing rates of change for L1 L4 BMD (0.004 gm/cm 2 /year, P = 0.047) and for totalhip BMD (0.006 gm/cm 2 /year, P = 0.003). No significant rate of change was found at the femoral neck site. Administration of alendronate/residronate to type III/IV patients did not show a significant rate of BMD change at any of the three sites. Comparison of BMD Response Between Treatments Figures 1 and 2 show the rates of BMD change projected over 5 years by treatment regimes within each type of OI patient by BMD site. Type I patients with pamidronate treatment had marginally greater rates of BMD change than those with no treatment (Table 3). At L1 L4 the difference was gm/cm 2 /year (P = 0.08), and at the total hip the difference was gm/cm 2 /year (P = 0.08). Patients treated with oral bisphosphonates, however, showed a significantly greater rate of BMD change (0.011 gm/cm 2 / year, P = 0.01) than the nontreated group at the total hip. The one difference between rates of BMD change for type III/IV patients was at the L1 L4 site, where patients treated with oral bisphosphonates had a lower rate of BMD change ( gm/cm 2 /year, P = 0.03) than patients treated with pamidronate. Table 4 shows the percent change in L1 L4 BMD for type I and type III/IV patients by type of treatment: pamidronate and oral (alendronate and residronate), where treatment extended to between 18 and 36 months from baseline. Pamidronate-treated type III/IV patients had a 15.9% mean (7.2% median) change in L1 L4 BMD over a treatment period of 31 months, whereas type I patients showed a mean (median) change of 4.3% (3.6%). Type III/ IV patients also had a greater mean response to oral bisphosphonate (6.2%) compared to type I patients (3.1%). Bisphosphonate Effect on Fracture Rate Table 5 summarizes the results of the fracture incidence comparisons. Fracture data were available for 72 of the 90 patients, 51 type I and 21 type III/IV. For the 72 patients, the average observed/treatment period was 41.1 (38.9) months. For type I and type III/IV nontreatment patients the ratios of fractures for the 5-year posttreatment to 5-year pretreatment intervals were 0.65 (P = 0.18) and 0.48 (P = 0.26), respectively.

5 124 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas Table 3 Comparison of annualized linear rates of BMD change (gm/cm 2 /year) between treatments for type I and types III/IV OI patients OI type I Treatment/comparison a L1 L4 (n = 57) Rate b (95% CI) P Total hip (n = 51) Rate (95% CI) P Femoral neck (n = 53) Rate (95% CI) P Not treated (-0.009, 0.006) (-0.012, 0.003) (-0.018, ) Pamidronate (0.008, 0.012) (-0.003, 0.013) (-0.008, 0.008) Alendronate/residronate (0.0006, 0.008) (0.002, 0.011) (-0.007, 0.005) Pamidronate vs. not treated ( , 0.018) ( , 0.02) (-0.002, 0.02) Alendronate/residronate vs. not treated (-0.002, 0.014) (0.003, 0.019) (-0.002, 0.019) Alendronate/residronate vs. pamidronate (-0.009, 0.004) (-0.007, 0.01) (-0.011, 0.009) III/IV L1 L4 (n = 27) Rate (95% CI) P Total hip (n = 17) Rate (95% CI) P Femoral neck (n = 18) Rate (95% CI) P Not treated (-0.01, 0.02) (-0.04, 0.06) (-0.05, 0.07) Pamidronate \ (0.008, 0.023) (0.0002, 0.02) (-0.017, 0.01) Alendroante/residroante (-0.006, 0.012) (-0.014, 0.02) (-0.013, 0.019) Pamidronate vs. not treated (-0.005, 0.027) (-0.47, 0.05) (-0.073, 0.045) Alendronate/residronate vs. not treated (-0.018, 0.016) (-0.058, 0.044) (-0.066, 0.055) Alendronate/residronate vs. pamidronate (-0.024, ) (-0.029, 0.012) (-0.012, 0.028) a Analysis of linear combinations of treatment estimates b Adjusted for age at baseline DXA scan, gender In type I patients, the ratios of post- to pretreatment fractures were 1.67 (P = 0.19) and 1.44 (P = 0.33) for the pamidronate- and oral bisphosphonate-treated, respectively, neither of which was a statistically significant change. However, the post- to pretreatment ratio of fractures for pamidronate-treated patients was marginally greater than the ratio for nontreated patients (P = 0.06). In type III/IV pamidronate-treated patients, the ratio of fracture incidence post- to pretreatment was 0.42 (P = 0.05), indicating a statistically significant decrease in fracture rate in this group. By contrast, treatment with oral bisphosphonates was not associated with a decrease in pre- to posttreatment fracture incidence ratio (P = 0.67). Table 6 summarizes the change in the number of fractures over 5-year periods pre- and posttreatment by fracture location, OI type, and treatment type. Fractures are grouped as vertebral, hip, and nonvertebral fractures, which includes appendicular fractures and peripheral sites (ankle, digits). The numbers of fractures pre- and post- 5-year time intervals in the nontreated groups (type I and III/IV) were small and larger for the treated categories as noted above. For treated type I patients, the distribution of vertebral, hip, and nonvertebral fractures was similar. For treated type III/IV patients, the small numbers of fractures at each site limited potential comparisons. Discussion The use of bisphosphonates, either orally or intravenously, represents a major advance in the treatment of osteoporotic disorders. Several observational studies have reported increases in BMD and decreases in fracture rate in OI children treated with pamidronate or alendronate [5, 6, 14].

6 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas 125 Fig. 1 Projected levels of BMD change for type I OI patients by treatment group and site starting at a baseline measurement across a span of 5 years. The projected BMD changes are based on the following (annualized linear rates of change [95% CI] adjusted for age at baseline DXA scan and gender): for L1 L4, no treatment ( [ to 0.006]), pamidronate (0.006 [ ]), alendronate/residronate (0.004 [ ]); for total hip, no treatment ( [ to 0.003]), pamidronate (0.005 [-0.003, 0.013]), alendronate/residronate (0.006 [ ]); for femoral neck, no treatment ( [ to ]), pamidronate (0.001 [ to 0.008]), alendronate/residronate ( [ to 0.005]) Fig. 2 Projected levels of BMD change for type III/IV OI patients by treatment group and site starting at a baseline measurement across a span of 5 years. The projected BMD changes are based on the following (annualized linear rates of change [95% CI] adjusted for age at baseline DXA scan and gender): for L1 L4, no treatment ( [-0.01 to 0.02]), pamidronate (0.016 [ ]), alendronate/residronate (0.003 [-0.006, 0.012]); for total hip, no treatment (0.009 [-0.04 to 0.06]), pamidronate (0.011 [ ]), alendronate/residronate (0.003 [ to 0.02]); for femoral neck, no treatment (0.009 [-0.05 to 0.07]), pamidronate ( [ to 0.01]), alendronate/residronate (0.003 [ to 0.019])

7 126 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas Table 4 Percent change in L1 L4 BMD for type I and type III/IV adult OI patients by type of treatment Rx category/type Mean months Rx a (SE) Median months Rx a Mean % BMD change b (SE) Median % BMD change b Pamidronate/type I 28.2 (1.1) (2.5) 3.6 Pamidronate/type III/IV 30.7 (1.2) (8.0) 7.2 Oral/type I 26.5 (1.4) (2.8) 4.4 Oral type III/IV 25.1 (2.1) (3.6) 7.0 a Months Rx compares the time from baseline to the latest visit between 18 and 36 months b % BMD change compares the BMD measurements from baseline to the latest visit between 18 and 36 months Table 5 Comparison of post- to pretreatment 5-year fracture rate between treatment groups for type I and types III/IV OI patients OI type I OI type III/IV Treatment/comparison a 5-year fracture rate 5-year fracture rate n Ratio post to pre b (95% CI) P n Ratio post to pre (95% CI) P Not treated c ( ) ( ) Pamidronate ( ) ( ) Alendronate/residronate ( ) ( ) Pamidronate vs. not treated ( ) ( ) Alendronate/residronate vs. not treated ( ) ( ) Alendronate/residronate vs. pamidronate ( ) ( ) a Analysis of linear combinations of treatment estimates b Adjusted for age at baseline DXA scan, gender c Beginning of treatment for controls based on comparable ages from treatment group However, two recent meta-analyses of the published reports highlight the limited numbers of randomized controlled studies applicable to bisphosphonate treatment in children [7, 8]. Current experience is that bisphosphonates have been generally accepted as the standard of care for adult OI patients. However, to date, there is limited information about the effects of bisphosphonates on adult bone density and fracture rates. The results of this study demonstrate that intravenous pamidronate and the oral bisphosphonates alendronate and residronate have a positive effect on BMD when measured as the annualized linear rate of change in BMD after a minimum of 13 months of treatment. In type I patients, BMD was increased by both pamidronate and the oral bisphosphonates at L1 L4 and by the oral bisphosphonates at the total hip. In type III/IV patients pamidronate increased BMD at the L1 L4 site and at the total hip. The oral agents did not increase BMD at the hip in type III/IV patients. Analysis of the site-specific effects of bisphosphonates on BMD indicates that the femoral neck is unresponsive to bisphosphonate treatment. While the reason for this disparity is not apparent, experience in patients with age-related osteoporosis indicates that the BMD response to bisphosphonates at the femoral neck is frequently less than that observed at the lumbar spine [15]. For type I patients, those receiving treatment compared to the nontreated group were older and had significantly lower baseline BMD values at L1 L4 and the total hip, consistent with the decision to treat based on BMD and recent fracture history. Treated type III/IV patients were younger than the nontreated group. In the more severely affected younger population the decision to be treated may have also been influenced by the patient s awareness of the reported benefits children gained by taking bisphosphonates. Thus, while

8 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas 127 Table 6 Change in number of fractures over 5-year period pre- and posttreatment by fracture location, OI type, and treatment type Type I No treatment Pamidronate Oral bisphosphonate b Decrease a Same a Increase a Decrease Same Increase Decrease Same Increase Type I Vertebral 4 (19.0%) 16 (76.2%) 1 (4.8%) 1 (7.1%) 10 (71.4%) 3 (21.4%) 3 (15.0%) 16 (80.0%) 1 (5.0%) Hip 2 (9.5%) 18 (85.7%) 1 (4.8%) 0 (0.0%) 13 (92.9%) 1 (7.1%) 2 (10.0%) 15 (75.0%) 3 (15.0%) Other NV 4 (19.0%) 14 (66.7%) 3 (14.3%) 2 (14.2%) 6 (42.9%) 6 (42.9%) 4 (20.0%) 8 (40.0%) 8 (40.0%) All V/NV 7 (33.3%) 12 (57.1%) 2 (9.5%) 3 (21.4%) 5 (35.7%) 6 (42.9%) 4 (20.0%) 7 (35.0%) 9 (45.0%) Type II/IV Vertebral 1 (16.7%) 5 (83.3%) 0 (0.0%) 4 (44.4%) 4 (44.4%) 1 (11.2%) 1 (16.7%) 5 (83.3%) 0 (0.0%) Hip 0 (0.0%) 6 (100.0%) 0 (0.0%) 0 (0.0%) 9 (100.0%) 0 (0.0%) 0 (0.0%) 4 (66.7%) 2 (33.3%) Other NV 2 (33.3%) 3 (50.0%) 1 (16.7%) 6 (66.6%) 2 (22.2%) 1 (11.2%) 3 (50.0%) 1 (33.3%) 2 (16.7%) All V/NV 3 (50.0%) 2 (33.3%) 1 (16.7%) 6 (66.7%) 1 (11.1%) 2 (22.2%) 2 (33.3%) 2 (33.3%) 2 (33.3%) a Number of patients where count of a patient s fractures in posttreatment period were less than, equal to, or more than the count of fractures in pretreatment period, respectively b Alendronate and residronate NV other nonvertebral locations (limb, ankle, peripheral, and mixed occurrence); V/NV all vertebral and nonvertebral locations noted above the allocation of patients to treatment groups in this study differs from that followed in a randomized clinical trial, it mirrors the decision process encountered in clinical practice. Frequently, documentation of fractures in adult OI patients is difficult because of inadequate X-ray confirmation and the tendency of adults to forget dates and fracture sites. Conversely, patients may assume they have, for example, injured a rib after experiencing chest pain but lack subsequent X-ray confirmation of the event. Although orthopedic reports, radiological reports, and X-rays were reviewed to confirm the occurrence of a fracture, the likelihood remains that fractures may have been missed in both the treated and nontreated groups. However, when assessed by fracture site, no trend for treatment effect on fracture number was observed for either OI type grouping. The study was not powered to permit a site-specific effect of treatment on vertebral, hip, or nonvertebral fractures. Pamidronate led to a significant decrease in fracture rate only in the type III/IV patients during the 5-year period of observation. These results related to fracture incidence illustrate two significant findings. First, the post- to pretreatment fracture ratio decreased, but not to a significant extent, in all groups including the nontreatment group except for type I patients with pamidronate or oral bisphosphonate treatment, where the ratio increased to a marginal extent. We suggest that this illustrates the influence that inclusion in a structured clinical program may have in determining treatment outcomes. For example, program affiliation could have resulted in improved vitamin D and calcium intake, or alterations in activity could have lessened fracture risk. The marginal increase (P = 0.06) in the postto pretreatment ratio of fractures for pamidronate-treated patients compared to nontreated patients requires a larger number of observations to confirm or refute the possibility that treatment in type I patients might have increased, rather than decreased, fracture incidence. Second, analyzing fracture incidence for a 5-year period pre- and posttreatment during an average treatment period of 44 months does provide an adequate time to evaluate the response to treatment. Although average treatment duration was less than the stated 5 years, 44 months of treatment should have provided fracture protection. Prolongation of the bisphosphonate effect on fracture prevention after a defined period of treatment has been demonstrated in the Fracture Intervention Trial Long-Term Extension (FLEX) [16]. The results in that trial suggested that after 5 years of alendronate treatment, discontinuation for up to 5 years did not appear to significantly increase fracture risk. We submit that a 5-year period of observation pre- and posttreatment provides an adequate interval with which to evaluate the treatment effects of the bisphosphonates on bone mass and fracture incidence. Two previous studies have addressed the effects of bisphosphonate treatment on BMD and fracture rates in OI adults. Chevrel et al. [17] studied the effects of oral alendronate on BMD in 64 adult patients in a 3-year randomized placebo-controlled trial. Mean increases in lumbar spine BMD were 10.1 ± 9.8%, and total-hip BMD increased by 3.3 ± 0.5%. However, they concluded that the sample size was not sufficient to determine an effect of alendronate on fracture rate. Adami et al. [18] treated 46 OI adults for 2 years at 3- month intervals with intravenous neridronate compared with a no-treatment group of 23 individuals. Spine and hip

9 128 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas BMD rose by 3.0 ± 4.6% (SD) and by 4.3 ± 3.9%, respectively, within the first 12 months of treatment and an additional 3.91% and 1.49% during the second year. Following treatment, the relative risk in fracture rate was 0.14 (95% CI ). This study reports comparable increases (4.3% for pamidronate-treated type I patients) in L1 L4 BMD for treated compared to nontreated groups, with a larger increase in BMD (15.9%) in treated type III/IV patients (Table 4). This may reflect the slightly lower entry BMD in the more severely affected patients, where BMD response is typically greater. Additional factors must be considered in interpreting these outcomes. While the age range was years, mean (SD) age was 39 (11) years. The natural history of fractures in adult OI indicates that fractures are relatively frequent prepuberty but that fracture incidence falls off sharply after puberty, only to increase again when patients are in the age range [19]. One could assume that the fracture rate was low in the posttreatment period, reflecting the natural history of OI fractures, and that it was not likely to fall lower with the addition of bisphosphonate. An additional explanation may relate to differences between type I and type III/IV patients at the bone matrix level [19]. Bisphosphonates decrease osteoclastic bone resorption, permitting osteoblastic bone formation to complete remodeling cycles, leading to a net increase in BMD. However, osteoblastic bone formation is impaired in OI as a consequence of the defective syntheses of type I collagen as well as that of other matrix proteins [20]. Furthermore, although bone turnover rates are increased in children with OI, thus facilitating the bisphosphonate antiresorptive effect on osteoclasts, limited data in adults suggest that bone turnover is usually decreased in adults with type I OI compared to children [21]. It could be anticipated that the effect of bisphosphonates on bone turnover in adult type I patients would be less robust than that observed following treatment in children. Bone turnover rates based on bone biopsy analysis are not published for adults with types III/ IV OI. However, elevated levels of bone resorption markers (C-terminal procollagen propeptide and urinary pyridinoline crosslinks) in type III and IV adults indicate that bone resorption is increased [22, 23]. Here, we observed increased BMD in this group following pamidronate treatment at the L1 L4 site and total hip. We postulate that a higher rate of bone turnover in type III/IV patients would facilitate the positive pamidronate effect on both BMD and fracture rate. There are several limitations in this study. First, this is an observational nonrandomized study but one including several years of patient involvement. Patient selection was based on severity and DXA results. The value of nonrandomized clinical studies has been confirmed by Wilkes et al. [24], who reported that in a cross-design synthesis of bisphosphonate-treated osteoporotic women total fractures were similarly reduced by bisphosphonates among postmenopausal women in randomized trials (23.6%) and highly compliant patients in observational studies of large databases from routine practice (20.3%). The study indicates similar outcome events when observational studies are compared to randomized clinical trials where compliant patients are involved. Second, men and women were not considered separately, although the analyses for percent BMD change were adjusted for gender. Third, the limited numbers of III/IV patients required merging OI groups for analysis. Fourth, merging results for alendronate and residronate could obscure a positive effect of either of these agents. In each instance, larger numbers of participants would help to clarify these issues. Also, this study did not address secondary treatment outcomes such as the relief of musculoskeletal discomfort or improvements in the quality of daily activities, which have been studied following treatment in OI children [25]. In summary, this study provides information about the BMD and fracture response of OI adults to bisphosphonate treatment. It does not conclusively answer the question as to whether adult OI patients should be treated with bisphosphonates. Certainly, the response in adults differs from that in children and this should be incorporated in the clinical decision as to whether bisphosphonate treatment is appropriate for an individual patient. Acknowledgements This study was supported by The Charitable & Research Foundation in Punta Gorda, Florida. References 1. Martin E, Shapiro JR (2007) Osteogenesis imperfecta: epidemiology and pathophysiology. Curr Osteoporos Rep 5: Rauch F, Glorieux FH (2005) Osteogenesis imperfecta, current and future medical treatment. Am J Med Genet C Semin Med Genet 15: Baldridge D, Schwarze U, Morello R, Lennington J, Bertin TK, Pace JM, Pepin MG, Weis M, Eyre DR, Walsh J, Lambert D, Green A, Robinson H, Michelson M, Houge G, Lindman C, Martin J, Ward J, Lemyre E, Mitchell JJ, Krakow D, Rimoin DL, Cohn DH, Byers PH, Lee B (2008) CRTAP and LEPRE1 mutations in recessive osteogenesis imperfecta. Hum Mutat 29: van Dijk FS, Nesbitt IM, Zwikstra EH, Nikkels PG, Piersma SR, Fratantoni SA, Jimenez CR, Huizer M, Morsman AC, Cobben JM, van Roij MH, Elting MW, Verbeke JI, Wijnaendts LC, Shaw NJ, Högler W, McKeown C, Sistermans EA, Dalton A, Meijers- Heijboer H, Pals G (2009) PPIB mutations cause severe osteogenesis imperfecta. Am J Hum Genet 85: Glorieux FH (2007) Experience with bisphosphonates in osteogenesis imperfecta. Pediatrics 119:S163 S Aström E, Jorulf H, Söderhäll S (2007) Intravenous pamidronate treatment of infants with severe osteogenesis imperfecta. Arch Dis Child 92:

10 J. R. Shapiro et al.: Bisphosphonate Treatment in Adult Osteogenesis Imperfectas Castillo H, Samson-Fang L (2009) Effects of bisphosphonates in children with osteogenesis imperfecta: an AACPDM systematic review. American Academy for Cerebral Palsy and Developmental Medicine Treatment Outcomes Committee Review Panel. Dev Med Child Neurol 51: Phillipi CA, Remmington T, Steiner RD (2008) Bisphosphonate therapy for osteogenesis imperfecta. Cochrane Database Syst Rev 4:CD Sillence DO, Senn A, Danks DM (1979) Genetic heterogeneity in osteogenesis imperfecta. J Med Genet 16: Hodgson SF, Watts NB, Bilezikian JP, Clarke BL, Gray TK, Harris DW, Johnston CC Jr, Kleerekoper M, Lindsay R, Luckey MM, McClung MR, Nankin HR, Petak SM, Recker RR AACE, Force Osteoporosis Task (2003) American Association of Clinical Endocrinologists medical guidelines for clinical practice for the prevention and treatment of postmenopausal osteoporosis: 2001 edition, with selected updates for Endocr Pract 9: Diggie PJ, Heagerty P, Liang KY, Zieger SL (2002) Analysis of longitudinal data. Ozford University Press, New York 12. Saeves R, Lande Wekre L, Ambjørnsen E, Axelsson S, Nordgarden H, Storhaug K (2009) Oral findings in adults with osteogenesis imperfecta. Spec Care Dentist 29: Pillion JP, Shapiro JR (2008) Audiological findings in osteogenesis imperfecta. J Am Acad Audiol 19: DiMeglio LA, Peacock M (2006) Two-year clinical trial of oral alendronate versus intravenous pamidronate in children with osteogenesis imperfecta. J Bone Miner Res 21: Cranney A, Wells G, Willan A, Griffith L, Zytaruk N, Robinson V, Black D, Adachi J, Shea B, Tugwell P, Guyatt G (2002) Osteoporosis Methodology Group and the Osteoporosis Research Advisory Group. Meta-analyses of therapies for postmenopausal osteoporosis. II. Meta-analysis of alendronate for the treatment of postmenopausal women. Endocr Rev 23: Black DM, Schwartz AV, Ensrud KE, Cauley JA, Levis S, Quandt SA, Satterfield S, Wallace RB, Bauer DC, Palermo L, Wehren LE, Lombardi A, Santora AC, Cummings SR FLEX, Group Research (2006) Effects of continuing or stopping alendronate after 5 years of treatment: the Fracture Intervention Trial Long-term Extension (FLEX): a randomized trial. JAMA 296: Chevrel G, Schott AM, Fontanges E, Charrin JE, Lina-Granade G, Duboeuf F, Garnero P, Arlot M, Raynal C, Meunier PJ (2006) Effects of oral alendronate on BMD in adult patients with osteogenesis imperfecta: a 3-year randomized placebo-controlled trial. J Bone Miner Res 21: Adami S, Gatti D, Colapietro F, Fracassi E, Braga V, Rossini M, Tatò L (2003) Intravenous neridronate in adults with osteogenesis imperfecta. J Bone Miner Res 18: Paterson CR, McAllion S, Stellman JL (1984) Osteogenesis imperfecta after the menopause. N Engl J Med 310: Fedarko NS, Robey PG, Vetter U (1995) Extracellular matrix stoichiometry in osteoblasts from patients with osteogenesis imperfecta. J Bone Miner Res 10: McCarthy EF, Earnst K, Rossiter K, Shapiro J (1997) Bone histomorphometry in adults with type 1A osteogensis imperfecta. Clin Orthop Relat Res 336: Braga V, Gatti D, Rossini M, Colapietro F, Battaglia E, Viapiana O (2004) Bone turnover markers in patients with osteogenesis imperfecta. Bone 34: Garnero P, Schott AM, Prockop D, Chevrel G (2009) Bone turnover and type I collagen C-telopeptide isomerization in adult osteogenesis imperfecta: associations with collagen gene mutations. Bone 44: Wilkes MM, Navikis RJ, Chan WW, Lewiecki EM (2010) Bisphosphonates and osteoporotic fractures: a cross-design synthesis of results among compliant/persistent postmenopausal women in clinical practice versus randomized controlled trials. Osteoporos Int 21: Seikaly MG, Kopanati S, Salhab N, Waber P, Patterson D, Browne R, Herring JA (2005) Impact of alendronate on quality of life in children with osteogenesis imperfecta. J Pediatr Orthop 25: