NEU12-0212: Substitution Monotherapy with Levetiracetam Versus Older AEDs: A RCT. Revised
Journal: Calcified Tissue International
Authors: Tahir Hakami, Terence J O’Brien, Sandra J Petty, Mary Sakellarides, Jemma Christie, Susan Kantor, Marian Todaro, Alexandra Gorelik, Markus J Seibel, Raju Yerra, John D Wark.
Corresponding author: John D. Wark MB.BS, Ph.D., FRACP, Department of Medicine, The Royal Melbourne Hospital, The University of Melbourne, 4th Floor Clinical Sciences Building Royal Parade, The Royal Melbourne Hospital, Parkville 3050, Victoria, AUSTRALIA, Email: .
Supplemental material: details of pQCT assessment.
Figure e1: Stratec XCT 3000.
Figure e2: pQCT Tibia Position. A: 4% site; and B: 38% site.
A. B.
Patients were resting comfortably on an adjustable examination couch, adjacent to the pQCT. The manufacturers’ limb stabilisation devices were utilized. A scout view was taken; the anatomical reference line was positioned at the distal end of tibia. Tibial lengths (in millimetres) were measured. One 2.5 mm scan was taken at the 4% site and another at the 38% site of the tibia on the non-dominant limb. The in-plane pixel size was 300 μm and slice thickness was 2.5 mm.
Figure e3: pQCT Radius Position. A: 4% site; and B: 38% site.
A. B.
Patients were resting comfortably on an adjustable examination couch, adjacent to the pQCT. The manufacturers’ limb stabilisation devices were utilized. A scout view was taken; the anatomical reference line was positioned at the distal end of radius. Radial lengths (in millimetres) were measured. One 2.5 mm scan was taken at the 4% site and another at the 38% site of the radius on the non-dominant limb. The in-plane pixel size was 300 μm and slice thickness was 2.5
Supplemental material: Testing agreement between the results of the blood tests from the two laboratories: ANZAC Research Institute and Melbourne Health Shared Pathology Service (MHSPS).
For testing agreement between the results from the two laboratories, we have chosen serum samples of five patients that were analyzed at the ANZAC Research Institute to be reanalyzed at the MHSPS for all analytes (see the results from the two labs in Table e-2). The agreement was statistically evaluated using several methods: Pearson’s correlation, linear regression analysis, and Bland Altman analysis. Bland-Altman analysis is the plots of difference between paired values against the mean for each pair of values. Adjustment between the two paired values was performed if the bias is outside the 95% confidence intervals (CIs) for the linear plot.
The Bland-Altman analysis showed an excellent agreement for procollagen 1 N-terminal peptide (PΙNP), and a clinically unreliable agreement for 25-hydroxyvitamin-D (25OHD) and 1,25-dihydroxyvitamin-D (1,25(OH)2D (Figure e-1). The linear regression analysis for intact parathyroid hormone (iPTH) and C-terminal telopeptides of type I collagen (βCTX) revealed slopes significantly different from unity (Figure e-2).
Therefore, the results for 25OHD and iPTH from the ANZAC Research Institute were adjusted, using the generated linear regression equations, based on the results from the MHSPS. 1,25(OH)2D was adjusted based on the results from the ANZAC Research Institute as the results from the MHSPS were high compared to the published normal range for this analyte. The adjusted results had an excellent agreement when plotted against the other results from MHSPS (Figure e-3).
Table e-1 Assays and reference ranges for laboratory tests
Test / Melbourne Health (shared) Pathology Service (MHSPS), Melbourne / ANZAC Research Institute, SydneyAssay / Reference range / Unit / Assay / Reference range / Unit
βCTX / Elecsys® β-CrossLapsTM/Serum immunoassay (Roche Diagnostics, Mannheim, Germany) / Male: N/A
Premenopausal: <0.57
Postmenopausal:<1.01 / ng/ml / Elecsys® β-CrossLapsTM/Serum immunoassay (Roche Diagnostics, Mannheim, Germany) / Male: N/A
Premenopausal: <0.57
Postmenopausal:<1.01 / ng/ml
PΙNP / Elecsys® Total PΙNP assay (Roche Diagnostics, Mannheim, Germany) / Male: N/A
Premenopausal: 15-60
Postmenopausal: 20-76 / μg/l / Orion Diagnostica UniQ PΙNP RIA radioimmunoassay (Uniq™ PΙNP RIA, Orion Diagnostica, Espoo, Finland) / Male: 22-87
Premenopausal: 16-96
Postmenopausal: 19-83 / μg//l
iPTH / DPC® IMMULITE 2000 intact PTH chemiluminescent immunoassay (DPC, Los Angeles, CA, USA) / 1.2-6.5 / pmol/l / DSL-8000 ACTIVE Intact PTH immunoradiometric (IRMA) assay (Diagnostic Systems Laboratories Inc., Webster, TX, USA). / 0.99-6.05 / pmol/l
25OHD / DiaSorin® LIAISON 25 (OH)D Total chemiluminescent immunoassay (DiaSorin, Stillwater, Minnesota USA) / Replete 55-108
Sub-opt 26-54
Deficient < 25 / nmol/l / DiaSorin® 25-OH- D RIA assay (DiaSorin, Stillwater, Minnesota USA). This assay shows 100% cross-reactivity between 25 (OH) D2 and 25 (OH) D3. / Deficiency 0-12.5
Insufficiency12.5.50
Hypovitaminosis D 50-100
Sufficiency 100-250
Toxicity >250 / nmol/l
1,25 (OH)2D / IDS® 1,25-OH-D RIA assay (Immunodiagnostic Systems Inc, Fountain Hills, AZ). / 78-190 / pmol/l / DiaSorin® 1,25-OH- D2 RIA assay (DiaSorin, Stillwater, Minnesota USA). It measures both vitamin D2 and D3 with 100% cross-reactivity and has a detection limit of <2 pg/ml, an intra-assay precision of 8.6% and an inter-assay precision of 12.3%. / 36-120 / pmol/l
Table e-2 The results of the paired serum samples used for testing agreement between the two laboratories
Test / Sample1 / 2 / 3 / 4 / 5
βCTXa ng/ml / 0.81 / 0.40 / 0.07 / 0.21 / 0.63
βCTXb ng/ml / 0.90 / 0.44 / 0.09 / 0.23 / 0.69
P1NPa μg/l / 106.5 / 14.5 / 80.5 / 61.0 / 30
P1NPb μg/l / 109 / 14.1 / 78.2 / 58.8 / 26.4
iPTHa pmol/l / 8.72 / 4.1 / 10.0 / 2.6 / 6.0
iPTHb pmol/l / 10.5 / 4.6 / 13.1 / 1.9 / 10.3
25OHDa nmol/l / 78 / 60 / 40 / 16 / 117
25OHDb nmol/l / 74 / 56 / 29 / 8 / 109
1,25 (OH)2Da pmol/l / 151 / 61 / 100 / 80 / 37
1,25(OH)2Db pmol/l / 220 / 130 / 220 / 190 / 84
aTests were performed at ANZAC Research Institute, Sydney , Australia. bTests were performed at Melbourne Health (shared) Pathology Service (MHSPS), Melbourne, Australia.
Figure e-4 Bland-Altman agreement plots for results from ANZAC Research Institute and MHSPS
Figure e-5 Linear regression scatter plots of results from ANZAC Research Institute vs. those from
MHSPS
Figure e-6 Assessment of agreement between the adjusted values and the reference laboratory values for 1,25(OH)D, 25OHD and iPTH
Table e-3 pQCT measures at the non-dominant tibia
pQCT measure / Within levetiracetam group change / Within older AED group change / Between group difference, p-valueBaseline mean (SD) / Mean change (SD) / % Change / p-value / Baseline mean (SD) / Mean change (SD) / % Change / p-value
Total bone area Trabecular (mm2) / 1126.90 (260.64) / -1.48 (47.08) / - 0.12 / 0.871 / 1216.18 (143.90) / 10.15 (54.06) / + 6.05 / 0.423 / 0.446
Trabecular bone area (mm2) / 947.27 (222.09) / -3.75 (48.99) / - 0.37 / 0.669 / 1022.92 (122.17) / 6.54 (55.86) / + 0.62 / 0.616 / 0.515
Trabecular BMC (mg/mm) / 241.50 (69.04) / -2.83 (22.92) / + 1.08 / 0.535 / 261.44 (42.49) / -1.03 (23.43) / - 0.39 / 0.850 / 0.799
Trabecular BMD (mg/ccm) / 254.46 (20.50) / -1.69 (13.16) / - 0.65 / 0.519 / 254.96 (20.73) / -2.00 (12.47) / - 0.80 / 0.494 / 0.937
Total bone area Cortical (mm2) / 382.59 (79.80) / 1.60 (10.59) / + 0.40 / 0.448 / 409.24 (49.87) / 6.29 (12.77) / + 1.50 / 0.046* / 0.186
Cortical bone area (mm2) / 288.73 (55.08) / 0.24 (8.88) / + 0.08 / 0.890 / 307.04 (44.76) / 0.88 (9.74) / + 0.28 / 0.699 / 0.821
Cortical BMC (mg/mm) / 341.95 (63.57) / 0.61 (8.56) / + 0.17 / 0.719 / 362.65 (51.69) / 0.15 (8.93) / + 0.04 / 0.943 / 0.861
Cortical BMD (mg/ccm) / 1185.86 (9.11) / 1.96 (7.86) / + 0.17 / 0.216 / 1181.95 (8.29) / -2.23 (9.52) / - 0.19 / 0.321 / 0.114
Cortical thickness (mm) / 5.59 (0.52) / -0.02 (0.13) / - 0.28 / 0.547 / 5.74 (0.58) / -0.05 (0.14) / - 0.90 / 0.123 / 0.392
Periosteal circumference (mm) / 71.1 (3.8) / 0.10 (0.92) / + 0.1 / 0.58 / 72.3 (4.33) / 0.49 (1.1) / + 0.7 / 0.072 / 0.21
Endosteal circumference (mm) / 35.0 (2.6) / 0.18 (0.82) / + 0.5 / 0.28 / 36.6 (2.2) / 0.78 (0.98) / + 2.1 / 0.003** / 0.029*
SSIp / 1668.49 (503.99) / -12.80 (77.29) / - 0.70 / 0.398 / 1825.47 (329.33) / 50.21 (94.20) / + 2.70 / 0.032* / 0.017*
The comparison of within-group mean change and percentage change and between-group difference over a 12 month period in the changes in pQCT measures at the non-dominant tibia. Significant within groups change or between groups difference is indicated by * for P < 0.05, ** for P < 0.01, and *** for P < 0.001. Abbreviations: mm2: millimetre square; mg/mm: milligram per millimetre, mg/ccm: milligrams per cubic centimetre; SD: standard deviation; -: decrease; +: increase; SSIp: polar strength-strain index
Table e-4: Restricted Maximum Likelihood Analysis– Variance Components. Analysis of Lumbar Spine Bone Mineral Density change
Factor or covariate / Coefficient / Standard Error / p -value / 95% CIAge (years) / 0.0003 / 0.0003 / 0.246 / -0.0002 / 0.0009
Height (m) / -0.0292 / 0.0313 / 0.350 / -0.0906 / 0.0321
Weight at follow-up (kg) / -0.0005 / 0.0004 / 0.130 / -0.0012 / 0.0002
Weight change (kg) / 0.0017 / 0.0009 / 0.071 / -0.0001 / 0.0036
Time interval (months) / -0.0011 / 0.0008 / 0.174 / -0.0026 / 0.0005
Calcium intake / 0.0000 / 0.0000 / 0.432 / 0.0000 / 0.0000
Life-time smoking / -0.0001 / 0.0003 / 0.641 / -0.0006 / 0.0004
Treatment
1= levetiracetam (effect set at 0)
2= older AEDs / -0.0093 / 0.0116 / 0.422 / -0.0321 / 0.0135
Sex and Menopausal status
1= male (set at 0)
2= female postmenopausal / -0.0325 / 0.0163 / 0.047* / -0.0645 / -0.0005
3= female premenopausal / -0.0042 / 0.0132 / 0.749 / -0.0300 / 0.0216
Treatment*postmenopause female interaction / 0.0265 / 0.0250 / 0.290 / -0.0225 / 0.0755
Treatment*premenopause female interaction / 0.0109 / 0.0214 / 0.610 / -0.0311 / 0.0529
Alcohol consumption
1= 1-20 units per week (set at 0)
2= 21-40 units per week / 0.0078 / 0.0161 / 0.627 / -0.0237 / 0.0393
3= never or occasional / 0.0044 / 0.0091 / 0.627 / -0.0134 / 0.0223
Mixed-effects Restricted Maximum Likelihood (REML) regression analysis was fitted to examine the effect of specific factors and covariates on areal bone mineral density (aBMD) at the lumbar spine. The response (dependent) variable (i.e. aBMD) was utilized unadjusted for age, height, weight, or time interval, with these variables included in the fixed model. Factors examined included treatment groups, sex and menopausal status, and alcohol consumption. Covariates included age, height, weight change, follow-up interval, calcium intake, and life-time smoking. The response (dependent) variable (i.e. bone mineral density) was utilized unadjusted for age, height, weight, or time interval, with these variables included in the fixed model. The fixed model fitted included (1) Factors: drug (coded 1= levetiracetam (set at 0), 2= older AEDs), sex/menopausal status (coded: 1= male) (set at 0), 2= female post-menopausal, 3= female pre-menopausal, alcohol consumption (coded 1= 1-20 units per week (set at 0), 2= 21-40 units per week, 3= never or occasional), and an interaction between drug and sex/menopausal status. (2) Covariates examined included age, height, weight at follow-up, change in weight, follow-up interval, calcium intake, and life-time smoking. “Patient.id” was specified as a random effect, to take account of two levels of variation: between-patients, and within-patients. The estimated variance components were calculated between- and within patients: the component was 0.00094; using the residual variance model for the term “patient.id”, the parameter was 0.00013. Therefore, 12.1% of the variance was attributed to within-patient differences and the remaining 87.9% of variance was attributed to between-patients differences [= 0.00013/ (0.00094 + .00013)*100]. This indicates that majority of the unexplained residual variation exists at the between-patient level. Using this model post-menopausal females were a borderline significant predictor for the change in the lumbar spine aBMD (p= 0.047, χ2 test).