Gastroenterology 2019;-:1–9

1 61
2 62
3 63
4
5
Dexlansoprazole and Esomeprazole Do Not Affect Bone 64
65
6 Homeostasis in Healthy Postmenopausal Women 66
7 67
8 Q7 Karen E. Hansen,1,* Jeri W. Nieves,2,* Sai Nudurupati,3 David C. Metz,4 and 68
9 Maria Claudia Perez3 69
10 70
11 1
School of Medicine and Public Health, University of Wisconsin, Madison, Wisconsin; 2Mailman School of Public Health, 71

CLINICAL AT
12 Columbia University, New York, New York; 3Takeda Development Center Americas, Inc., Deerfield, Illinois; and 4Perelman 72
13 School of Medicine, University of Pennsylvania, Philadelphia, Pennsylvania 73
14 74
15 BACKGROUND & AIMS: Epidemiological studies have associ- or antacids.2 As a result, several PPIs are available over the 75
16 ated proton pump inhibitor (PPI) therapy with osteoporotic counter. However, observational studies suggest that long- 76
17 fractures, but it is not clear if PPIs directly cause osteoporosis. term PPI therapy is associated with osteoporotic fractures, 77
18 We evaluated the effect of dexlansoprazole and esomeprazole hypomagnesemia, and vitamin B12 deficiency.2,3 78
19 on bone turnover, bone mineral density (BMD), true fractional A large body of epidemiological data has detected an 79
20 calcium absorption (TFCA), serum and urine levels of minerals, association between PPI use and fracture risk.4,5 The clinical 80
21 and levels of parathyroid hormone (PTH) in healthy post- importance of this association is unclear because odds ratios 81
22 menopausal women. METHODS: We performed a prospective, are low, there is no consistent dose-response relationship, 82
23 multicenter, double-blind study of 115 healthy, post- and confounding factors exist in many studies.4 Although a 83
24 menopausal women (45 to 75 years of age) from November 4, 84
plausible mechanism explaining a causal relationship be-
25 2010, through August 7, 2014. Women were randomized to 85
tween PPI use and fracture has not been definitively
26 groups given dexlansoprazole (60 mg), esomeprazole (40 mg), 86
established,4 hypotheses include decreased calcium ab-
or placebo daily for 26 weeks. We measured plasma levels of
27 sorption due to hypochlorhydria,6,7 osteoclast V-ATPase 87
procollagen type 1 N-terminal propeptide (P1NP) and C-ter-
28 inhibition, increased activity of parathyroid hormone (PTH) 88
minal telopeptide of type 1 collagen (CTX) at 0 (baseline), 13,
29 induced by hypergastrinemia,8,9 and decreased magnesium 89
and 26 weeks. Primary outcomes were percent change in P1NP
30 absorption.4,10–13 Mixed results also have been reported for 90
and CTX between weeks 0 and 26. We also measured changes
31 in serum and urine levels of mineral, BMD, PTH (all subjects), bone mineral density (BMD) in relation to PPI use.14–17 91
32 and TFCA (n ¼ 30). RESULTS: Between baseline and week 26, Some studies evaluating PPI effects on vitamin B12 and 92
33 there were no significant within-group differences in markers calcium absorption found a decrease associated with PPI 93
34 of bone turnover; there was a nonsignificant increase in CTX use,18–20 whereas others showed no effect.7,21–23 94
35 levels in the dexlansoprazole group (0.12 ng/mL). The eso- We evaluated the effects of PPIs on bone homeostasis in 95
36 meprazole and dexlansoprazole groups had significantly postmenopausal women, a population with the highest 96
37 increased levels of P1NP (18.2% and 19.2%, respectively) and prevalence of osteoporosis and related fractures.24,25 We 97
38 CTX (22.0% and 27.4%, respectively) at week 26 compared measured bone turnover, BMD, true fractional calcium ab- 98
39 with the placebo group, although these values remained within sorption (TFCA), PTH, and serum and urine mineral levels in 99
40 normal ranges. There were no statistically significant differ- women at baseline and 26 weeks after randomization to 100
41 ences between groups in serum or urine levels of minerals, placebo or 2 potent PPIs, dexlansoprazole and esomepra- 101
42 BMD, or PTH at week 26. PPI therapy did not reduce TFCA. zole. Bone turnover markers were the primary study 102
43 CONCLUSIONS: In a prospective study of postmenopausal outcome, as they are consensus standard markers for 103
44 women, we found significant increases in markers of bone assessment of bone resorption (C-terminal telopeptide of 104
45 turnover in women given PPI therapy compared with women
type 1 collagen [CTX]) and formation (procollagen type 1 N- 105
46 given placebo, but levels remained within the normal reference 106
terminal propeptide [P1NP]) related to both fracture risk
47 range. We found no significant differences among groups in 107
and monitoring osteoporosis medications.26,27 We also
48 changes in BMD, PTH, serum or urine levels of minerals, or 108
measured bone-specific alkaline phosphatase (BsAP) and
TFCA. Our findings indicate that 26 weeks of treatment with a
49 urinary N-terminal telopeptide (NTx), which have shown 109
PPI has no clinically meaningful effects on bone homeostasis.
50 110
Clinicaltrials.gov no: NCT01216293
51 111
52 *Authors share co-first authorship 112
53 Keywords: Clinical Trial; Proton Pump Inhibitor; Osteoporosis; 113
Abbreviations used in this paper: AE, adverse event; BMD, bone mineral
54 Intestinal Calcium Absorption. density; BsAP, bone-specific alkaline phosphatase; CI, confidence inter- 114
55 val; CTX, C-terminal telopeptide of type 1 collagen; NTx, N-terminal telo- 115
peptide; P1NP, procollagen type 1 N-terminal propeptide; PPI, proton
56 pump inhibitor; PTH, parathyroid hormone; TFCA, true fractional calcium 116
57 117
58
59
P roton pump inhibitors (PPIs) have been widely used
for decades to treat acid-related diseases.1 PPIs
exhibit a well-established safety profile and reduce gastric
absorption; 25(OH)D, 25-hydroxyvitamin D.

© 2018 by the AGA Institute

0016-5085/$36.00
118
119
60 acid more effectively than histamine-2 receptor antagonists https://doi.org/10.1053/j.gastro.2018.11.023 120

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 2 Hansen et al Gastroenterology Vol. -, No. -

vitamin D3 intake. Participants were given vitamin D3 to

121 WHAT YOU NEED TO KNOW 181
maintain 25-hydroxyvitamin D (25[OH]D) levels 32 ng/mL
122 182
BACKGROUND AND CONTEXT and calcium carbonate 600 mg/d to maintain a total (diet plus
123 183
supplement) intake of 1200 mg/d.
124 Placeholder text  Placeholder text  Placeholder text  184
Placeholder text  Placeholder text  Placeholder text  Participants were randomly assigned to receive placebo,
125 dexlansoprazole 60 mg once daily, or esomeprazole 40 mg once 185
126 Placeholder text  Placeholder text  Placeholder text  186
Placeholder text  daily in a 1:1:1 ratio for 26 weeks, with study medication
127 dispensed in bottles in a double-blind fashion. P1NP, CTX, urine 187
128 NEW FINDINGS NTx, and BsAP were measured at 0, 13, and 26 weeks. Serum 188
129 Placeholder text  Placeholder text  Placeholder text  samples were analyzed for P1NP using an electro- 189
130 Placeholder text  Placeholder text  Placeholder text  chemiluminescence immunoassay on the Elecsys 2010 auto- 190
131 Placeholder text  Placeholder text  Placeholder text  mated analyzer (Roche Diagnostics, Indianapolis, IN) and for 191
CLINICAL AT

132 Placeholder text  BsAP using an enzyme immunoassay with the Metra BsAP 192
133 LIMITATIONS enzyme immunoassay kit (Quidel, San Diego, CA). Plasma 193
134 samples were analyzed for CTX using an electro- 194
Placeholder text  Placeholder text  Placeholder text 
135 Placeholder text  Placeholder text  Placeholder text  chemiluminescence immunoassay with the beta-crosslaps 195
136 Placeholder text  Placeholder text  Placeholder text  immunoassay kit on the Elecsys 2010 automated analyzer. 196
137 Placeholder text  Urine samples were analyzed for NTx using an enzyme-linked 197
138 immunosorbent assay with the Osteomark kit (Wampole Lab- 198
IMPACT
139 oratories, Waltham, MA) and read on a SpectraMax microplate 199
140 Placeholder text  Placeholder text  Placeholder text  spectrophotometer. Subjects were instructed to fast without 200
Placeholder text  Placeholder text  Placeholder text  taking study medication, calcium, or vitamin D3 supplements
141 201
Placeholder text  Placeholder text  Placeholder text  for a minimum of 8 hours before all laboratory tests (excluding
142 Placeholder text  202
first screening visit). BMD was measured in each subject using
143 203
the same Hologic bone densitometer at baseline (‒12 weeks),
144 204
significant associations with BMD response to osteoporosis 26, and 52 weeks. Calcium and vitamin D3 supplements were to
145 205
therapies.28 be taken with the first meal of the day, approximately 1 hour
146 206
after study drug administration. A comprehensive schedule of
147 study procedures is reported in Supplementary Table 2. 207
148 Our primary outcomes were changes in bone formation and 208
149 Methods 209
bone resorption. Changes in bone turnover precede changes in
150 Study Design BMD and predict harmful (or beneficial) effects of an interven- 210
151 We conducted a phase 4, randomized, double-blind, pla- tion on BMD.29–31 We selected P1NP and CTX as our primary 211
152 cebo-controlled study in healthy postmenopausal women from measures of bone formation and resorption, as they change more 212
153 November 4, 2010 (first participant in) to August 7, 2014 (last quickly with pharmacologic interventions.29–32 We had no pilot 213
154 participant out), at 12 US centers. Institutional review board data on PPI-mediated changes in bone turnover by which to 214
155 approval and participant consent were obtained before the calculate a potential sample size for the study. However, within- 215
156 study. This study was funded by Takeda Pharmaceuticals In- subject changes in the bone resorption marker CTX can be as 216
157 ternational, Inc. Takeda Pharmaceuticals International, Inc., was high as 36%.33 We therefore selected a sample size of 240 217
158 responsible for and sponsored the study design, data collection, subjects so that the width of the 95% confidence interval (CI) for 218
159 data interpretation, and writing of this article. All authors had the difference in the percent change from baseline in each of the 219
160 access to the study data and reviewed and approved the final markers between PPI and placebo was restricted to no more 220
161 manuscript. than 30%. Finally, we chose a study duration of 6 months, as this 221
162 Eligible women were between 45 and 75 years of age, with is the duration of therapy for dexlansoprazole approved by the 222
163 body mass index between 18 and 30 kg/m2 and with no Food and Drug Administration. In addition, a 6-month exposure 223
menses for at least 5 years. Eligibility requirements included would minimize risks to participants while still providing sen-
164 224
bone turnover markers and clinical laboratory evaluations sitive information about bone turnover and alterations in min-
165 225
within normal postmenopausal ranges (postmenopausal eral homeostasis resulting from PPI therapy.
166 226
follicle-stimulating hormone levels 23 IU/L; creatinine clear- The primary outcome was percent change from baseline to
167 227
ance 50 mL/min; BMD T-score > ‒2.0 at the total hip, femoral week 26 in the bone formation marker P1NP and the bone
168 neck, and lumbar spine by dual-energy x-ray absorptiometry); 228
resorption marker CTX. The secondary endpoint was percent
169 and willingness to take daily supplements of vitamin D and change from baseline to week 26 in urine NTx (resorption 229
170 calcium carbonate (CaCO3). Participants could not use medi- marker) and BsAP (formation marker). Additional endpoints 230
171 cations and substances with a potential effect on bone ho- included the 13-week percent change in P1NP and CTX; the 26- 231
172 meostasis during the study (Supplementary Table 1). Further and 52-week percent change in femoral neck, total hip, and 232
173 eligibility criteria and BMD screening methods are reported in lumbar spine BMD, and the incidence of fractures, including 233
174 the Supplementary Methods. clinical vertebral fractures. We also measured 26-week changes 234
175 The study consisted of a 12-week screening period, a 26- in 24-hour urinary calcium and magnesium excretion, serum 235
176 week treatment period, and a follow-up visit at week 52 for PTH, calcium, phosphorus, magnesium, and 25(OH)D. A TFCA 236
177 BMD assessment. Dietary assessment was performed between study was performed at baseline and 6 months in a subset 237
178 the screening period and week 13 to assess calcium and of participants using dual stable 44Ca (oral) and 42Ca 238
179 239
180 240

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 - 2019 Dexlansoprazole, Esomeprazole in Bone Homeostasis 3

(intravenous) isotopes34 and an inpatient 24-hour urine (Supplementary Figure 1). No substantial differences in
241 301
collection. The inpatient diet provided during the 24-hour TFCA baseline demographics were observed between treatment
242 302
study period matched the participants’ outpatient diet based on groups (Table 1). Most participants (>90%) were white,
243 303
244
4-day food diaries (additional information regarding the TFCA with mean age of 62 ± 6 years, and mean body mass index 304
245
substudy is provided in the Supplementary Methods section). of 25.2 ± 2.8 kg/m2. Subjects’ serum 25(OH)D levels were 305
Adverse events (AEs) were classified according to Medical 39 ± 10 ng/mL at randomization. The calcium absorption
246 Dictionary for Regulatory Activities and by event severity. 306
247 substudy enrolled 34 participants (30 completing) with 307
248 demographics and baseline characteristics representative of 308
249 Statistical Analyses the overall study population. Mean study drug compliance 309
250 The safety data set included all participants who received at was >95% in each treatment group. The incidence of the 310
251 least 1 dose of the study drug. The pharmacodynamic data set most common concurrent medical conditions relevant to 311

CLINICAL AT
252 included all participants who had baseline and post-baseline study outcomes (osteoarthritis, osteopenia, and back pain) 312
253 values for any primary or secondary endpoint. Assuming a was comparable across treatment groups. The proportion of 313
254 standard deviation of percent change from baseline levels to be subjects with protocol deviations was also comparable 314
255 no more than 40% for each bone turnover marker, a sample across treatment groups (placebo 76%, dexlansoprazole 315
256 size of 80 participants per arm provided a 95% CI of the esti- 68%, and esomeprazole 74%) and primarily attributed to 316
257 mated difference between PPI and placebo that extended no deviations in visit windows, which did not affect the overall 317
more than 15% in each direction, with an allowance for up to outcome of the study.
258 318
30% dropouts.
259 319
After all study visits were completed, one of the trial sites
260 Bone Biomarkers 320
(site 6019) was disqualified by the Food and Drug Adminis-
261 tration, and all data collected at site 6019 were excluded from
321
In all treatment arms, bone turnover remained within
262 these analyses. Even without this site, the 95% CI for the 322
the normal reference range for postmenopausal women at
263 between-arm percent change in all bone turnover markers was 323
baseline, 13, and 26 weeks. Levels of P1NP, BsAP, CTX, and
264 33% wide, which is still comparable to the original assump- 324
NTx showed little change from baseline, and final values
265 tion (95% CI to have a width not to exceed 30%). All differ- 325
remained within the normal range for this population
266 ences in percent change from baseline endpoints between 326
(Figure 1).
267 dexlansoprazole or esomeprazole and placebo were estimated 327
Bone formation, as assessed by P1NP, showed a statisti-
268 using the nonparametric Hodges-Lehmann estimator and the 328
cally significant increase in the dexlansoprazole group at
269 Moses method for the 95% CI. Comparisons were considered 329
week 26 (19.2%; 95% CI 7.0%–30.2%) and the esomepra-
270 statistically significant if the 95% CI excluded zero. 330
zole group at week 26 (18.2%; 95% CI 6.7%–30.4%) relative
271 331
to placebo (Table 2). Levels of the formation marker BsAP
272 332
also increased significantly in the dexlansoprazole group at
273 Results week 26 (6.8%; 95% CI 0.4%–12.8%) and at week 26 in the
333
274 Subjects 334
esomeprazole group (7.2%; 95% CI 0.6%–12.9%) relative to
275 335
Of the 556 women screened at 11 centers, 115 were placebo. Week 13 changes in P1NP and BsAP values in the
276 336
enrolled, and 93 participants completed the study esomeprazole group were not statistically significant.
277 337
278 338
279 Table 1.Baseline Demographics 339
280 340
Placebo, Dexlansoprazole, Esomeprazole, All subjects,
281 n ¼ 41 60 mg QD, n ¼ 38 40 mg QD, n ¼ 35 N ¼ 114
341
282 342
283 Mean age (SD), ya 61.2 (5.6) 62.4 (6.7) 63.2 (5.1) 62.2 (5.9) 343
284 Range 5173 5275 5474 5175 344
285 Race, n (%) 345
286 White 37 (90) 34 (90) 33 (94) 104 (91) 346
Black/African American 2 (5) 2 (5) 2 (6) 6 (5)
287 Otherb 2 (5) 2 (5) 0 4 (4)
347
288 Mean BMI (SD), kg/m2 25.6 (2.6) 24.5 (3.0) 25.7 (2.6) 25.2 (2.8) 348
289 Range 21.330.9 18.231.3 19.630.7 18.231.3 349
290 Current drinker, n (%) 30 (73) 31 (82) 27 (77) 88 (77) 350
291 Serum calcium at randomization, mean (SD), mg/dL 9.5 (0.4) 9.5 (0.3) 9.6 (0.3) 9.5 (0.3) 351
292 Range 8.610.1 9.010.1 9.110.2 8.610.3 352
Serum vitamin D at randomization, mean (SD), ng/mL 40 (9) 40 (12) 36 (9) 39 (10)
293 353
Range 2969 2687 1955 1987
294 354
295 355
296 BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); QD, once daily; SD, standard 356
297 deviation. 357
298
a
n ¼ 40 for placebo for age, because age was calculated at first dose and 1 participant was not dosed. 358
b
299 Other races included American Indian/Alaska Native, Asian, or multiracial. 359
300 360

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 4 Hansen et al Gastroenterology Vol. -, No. -

361 Figure 1. Bone turnover 421

362 markers. Top and bottom 422
363 of boxes represent first
423
and third quartiles; the
364 interior line represents the 424
365 median. Whiskers termi- 425
366 nate at 1.5 times (inter- 426
367 quartile range) below the 427
368 first quartile or above the 428
369 third quartile. Horizontal 429
dotted lines indicate upper
370 430
and lower bound of normal
371 range for women aged 45 431
CLINICAL AT

372 to 75 years (A); upper and 432

373 lower bound of normal 433
374 range for women aged 45 434
375 to 110 years (B); upper 435
376 bound of normal range for 436
all women (lower bound ¼
377 0) (C); and upper and lower 437
378 bound of normal range for 438
379 all women (D). BCE, bone 439
380 collagen equivalent; Cr, 440
381 creatinine. 441
382 442
383 443
384 Resorption (CTX levels) significantly increased in the urinary calcium or magnesium levels (Supplementary 444
385 dexlansoprazole group at weeks 13 (15.9%; 95% CI 4.9%– Table 3), and all values remained stable and normal. As 445
386 29.6%) and 26 (27.4%; 95% CI 12.7%–43.0%) and in the expected, serum gastrin levels increased in participants 446
387 esomeprazole group at weeks 13 (16.1%; 95% CI 5.0%– randomized to PPI therapy, with median week 26 differ- 447
388 27.0%) and 26 (22.0%; 95% CI 8.4%–35.7%) relative to ences of 80 ng/mL (dexlansoprazole) and 85 ng/mL (eso- 448
389 placebo. For NTx, the only statistically significant difference meprazole) relative to placebo (Supplementary 449
390 from placebo occurred at week 26 in the dexlansoprazole Table 3).10,35 No statistically significant between-group 450
391 group (20.1%; 95% CI 4.0%–34.4%; Table 2). Within-group changes in PTH were found (Supplementary Table 3). 451
392 26-week changes in bone turnover markers did not show Likewise, there was no correlation between PPI-related 452
393 statistically significant differences from baseline, except for changes in gastrin and PTH during the study. 453
394 a small increase in CTX levels in the dexlansoprazole group In the substudy population, there were no decreases in 454
395 (0.12 ng/mL; 95% CI 0.03–0.23 ng/mL). TFCA in participants randomized to PPI therapy 455
396 (Supplementary Table 4). The only statistically significant 456
397 BMD and Fracture increase in TFCA relative to placebo occurred in the eso- 457
398 We detected no statistically significant between-group meprazole group (þ0.06; 95% CI 0.02–0.11). 458
399 changes in lumbar spine, total hip, or femoral neck BMD 459
400 at 26 and 52 weeks (Table 3). No subject experienced a 26- 460
401
Safety 461
week decrease in lumbar spine, femoral neck, or total hip The median duration of treatment was comparable
402 BMD >0.021 g/cm2. A 73-year-old woman randomized to 462
403 among groups. AEs were reported for 39% (16/41), 45% 463
dexlansoprazole experienced a 3% decrease in spine BMD, (17/38), and 37% (13/35) of participants in the placebo,
404 but her T-score at the spine was þ0.6 even after the 464
405 dexlansoprazole, and esomeprazole groups, respectively. 465
decrease. No fractures occurred during the treatment Most AEs (93%) were classified as being mild or moder-
406 period. One woman randomized to dexlansoprazole sus- 466
407 ate. The most frequently reported AEs were diarrhea and 467
tained a traumatic foot fracture at week 52 because of a car upper respiratory tract infection with placebo; diarrhea
408 accident. A humerus fracture (circumstance unknown) 468
409 with dexlansoprazole; and urinary tract infection, dyspepsia, 469
occurred at week 23 in one woman randomized to esome- and upper respiratory tract infection with esomeprazole.
410 prazole; she withdrew from the study after 17 weeks of 470
411 The frequency of drug-related AEs within treatment groups 471
treatment because of worsening fatigue that was considered was 9% (3/34; placebo), 30% (9/30; dexlansoprazole), and
412 to be drug related. 472
413 48% (13/27; esomeprazole). AEs resulting in discontinua- 473
414 tion were reported for 1 individual in the placebo group 474
415
Calcium Absorption, Mineral, and Hormone Levels (diarrhea), 4 individuals in the dexlansoprazole group 475
416 Serum calcium and phosphorus concentrations (diarrhea, 3; leukopenia, 1), and 2 individuals in the eso- 476
417 remained stable and normal, with no between-arm changes. meprazole group (nausea and fatigue). 477
418 Likewise, we detected no statistically significant between- The only drug-related serious AE, nephrolithiasis, 478
419 group changes in serum magnesium, 25(OH)D, or 24-hour occurred in a dexlansoprazole participant at day 142 and 479
420 480

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 540
539
538
537
536
535
534
533
532
531
530
529
528
527
526
525
524
523
522
521
520
519
518
517
516
515
514
513
512
511
510
509
508
507
506
505
504
503
502
501
500
499
498
497
496
495
494
493
492
491
490
489
488
487
486
485
484
483
482
481
Table 2.Percent Change From Baseline to Weeks 13 and 26 for Biomarkers of Bone Homeostasis

-
2019
Median (IQR) of percent change Difference vs placebo (95% CI)a

Dexlansoprazole Esomeprazole
Placebo 60 mg QD 40 mg QD Dexlansoprazole vs placebo Esomeprazole vs placebo

Week 13, Week 26, Week 13, Week 26, Week 13, Week 26,
n ¼ 38 n ¼ 32 n ¼ 36 n ¼ 31 n ¼ 34 n ¼ 26 Week 13 Week 26 Week 13 Week 26

Formation
FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

P1NP, ng/mL 7.2 (21.9) ‒0.4 (24.1) 21.4 (31.1) 19.3 (41.0) 11.5 (28.8) 16.9 (37.4) 12.3 (3.1 to 22.8) 19.2 (7.0 to 30.2) 5.8 (‒4.0 to 14.7) 18.2 (6.7 to 30.4)
BsAP, U/L ‒1.44 (15.0) 0.5 (14.1) 7.7 (12.0) 8.7 (20.1) 4.4 (15.0) 6.2 (17.8) 6.5 (0.9 to 11.4) 6.8 (0.4 to 12.8) 3.1 (‒1.7 to 8.7) 7.2 (0.6 to 12.9)
Resorption
CTX, ng/mL 2.7 (20.0) 2.4 (29.2) 18.8 (48.8) 29.5 (39.8) 18.7 (32.7) 24.0 (30.3) 15.9 (4.9 to 29.6) 27.4 (12.7 to 43.0) 16.1 (5.0 to 27.0) 22.0 (8.4 to 35.7)
NTx (corrected) 5.5 (27.3) ‒6.1 (36.2) 14.6 (36.7) 16.9 (39.7) 13.4 (26.9) 6.2 (40.3) 6.6 (‒6.6 to 19.2) 20.1 (4.0 to 34.4) 7.6 (‒2.4 to 18.2) 11.6 (‒2.7 to 28.3)

IQR, interquartile range; QD, once daily.

a
Difference estimated with Hodges-Lehmann estimate; 95% CIs of the differences calculated by the Moses method. Percent change values were considered statistically
significant if the 95% CI excluded zero (reported in bold text).

Table 3.Changes From Baseline in BMD

Placebo, Dexlansoprazole 60 mg QD, Esomeprazole 40 mg QD, Difference in percent change vs

Dexlansoprazole, Esomeprazole in Bone Homeostasis

n ¼ 38a n ¼ 36a n ¼ 34a placebo (95% CIb)

Median percent Median percent Median percent Dexlansoprazole vs Esomeprazole vs

n change (IQR)b n change (IQR)b n change (IQR)c placebo placebo

Lumbar spine, g/cm2

Week 26 19 0.104 (5.029) 19 ‒2.236 (4.345) 20 ‒1.182 (3.024) ‒0.888 (‒3.263 to 1.084) 0.840 (‒2.755 to 1.061)
Week 52 17 ‒0.392 (2.362) 18 ‒0.537 (3.469) 18 0.793 (2.913) ‒0.112 (‒1.657 to 1.846) 1.344 (‒0.202 to 3.171)
Total hip, g/cm2
Week 26 26 ‒0.181 (1.707) 27 ‒1.045 (2.402) 24 ‒0.774 (2.336) ‒0.664 (‒1.482 to 0.240) ‒0.358 (‒1.272 to 0.501)
Week 52 23 0.109 (1.849) 25 ‒0.441 (1.983) 23 ‒0.109 (1.721) ‒0.344 (‒1.168 to 0.587) ‒0.004 (‒0.682 to 0.972)
Femoral neck, g/cm2
Week 26 26 ‒0.688 (2.823) 27 ‒1.861 (3.002) 24 ‒1.515 (3.802) ‒0.966 (‒2.360 to 0.482) ‒0.434 (‒1.879 to 1.056)
Week 52 23 0.000 (3.892) 25 ‒0.437 (2.598) 23 0.000 (4.647) ‒0.237 (‒1.565 to 1.221) 0.285 (‒1.250 to 2.165)

IQR, interquartile range; QD, once daily.

a
Some scans were not evaluable and were excluded from the BMD analyses, which led to fewer participants than for other parameters.
b
Difference estimated with Hodges-Lehmann estimate. 95% CIs of the differences calculated by the Moses method.
c
Percent change calculated either from baseline to week 26 or from week 26 to week 52.

5
CLINICAL AT
600
599
598
597
596
595
594
593
592
591
590
589
588
587
586
585
584
583
582
581
580
579
578
577
576
575
574
573
572
571
570
569
568
567
566
565
564
563
562
561
560
559
558
557
556
555
554
553
552
551
550
549
548
547
546
545
544
543
542
541
 6 Hansen et al Gastroenterology Vol. -, No. -

601 lasted for 37 days and then resolved. Analysis of the passed group. Likewise, serum calcium, urine calcium, and serum 661
602 stone indicated it was 91% calcium oxalate, and although phosphorus levels remained stable and normal during the 662
603 classified as drug related, it could also have been caused by study, with no between-arm changes. Together, these data 663
604 concomitant calcium or vitamin D supplementation. Three suggest that hypergastrinemia had no effect on PTH, cal- 664
605 serious AEs deemed unrelated to study drugs were breast cium, or phosphorus concentrations. 665
606 cancer (n ¼ 1, dexlansoprazole), aortic dissection (n ¼ 1, Another proposed mechanism by which PPIs might in- 666
607 dexlansoprazole), and hip arthroplasty (n ¼ 1, placebo). fluence bone homeostasis is through reduced calcium ab- 667
608 sorption due to increased gastric pH.7 TFCA, the 668
609 proportion of a given dose of calcium actually absorbed, is 669
610
Discussion approximately 25% in adult men and nonpregnant women 670
611 Recent population-based epidemiological studies have across a wide age range44 and declines on average by 0.2% 671
CLINICAL AT

612 generated conflicting results about a relationship between per year after age 40.45 Calcium is absorbed in the intes- 672
613 chronic PPI therapy and fracture risk.4,5,11–13,36 Low odds tinal tract by passive transport and by active transport that 673
614 ratios (<2), lack of dose response, biological implausi- requires adequate vitamin D status.46 The pH of the small 674
615 bility, and uncontrolled potential confounders limit any intestine is nearly neutral, which suggests that gastric pH 675
616 firm conclusions about the causal nature between PPI would not affect calcium absorption.7 Our results support 676
617 therapy and osteoporosis.37,38 Although some studies this theory. Serum 25(OH)D levels were stable throughout 677
618 have detected an association between PPI use and frac- the study, which suggests that vitamin D metabolism did 678
619 tures in postmenopausal women, prospective studies have not influence the TFCA results. Median TFCA values ranged 679
620 found no differences in BMD between PPI users and from 16% to 20% across the 3 treatment groups at the 680
621 nonusers.14,25,39 baseline and week 26 assessments, as expected in this 681
622 We undertook this prospective, double-blind, ran- population.44,45 Consistent with another study of 30-day 682
623 domized, placebo-controlled trial to evaluate the effects of administration of omeprazole,7 26-week PPI administra- 683
624 PPIs on bone homeostasis in healthy postmenopausal tion was not associated with a decreased TFCA in post- 684
625 women. Overall, we found no evidence of bone loss among menopausal women. 685
626 healthy postmenopausal women who received dexlanso- Although rare, hypomagnesemia has been reported in 686
627 prazole or esomeprazole for 26 weeks compared with patients treated with PPIs for at least 3 months (in most 687
628 those who received placebo. PPI therapy was associated cases after a year of therapy).47,48 In this study, neither PPI 688
629 with increases in both bone formation (P1NP increased reduced serum or urine magnesium, and no clinically sig- 689
630 17% to 19%; BsAP increased 6% to 9%) and bone nificant effects on 24-hour urine magnesium levels were 690
631 resorption (CTX increased 24% to 30%; corrected NTx observed for subjects in the TFCA substudy. 691
632 increased 6% to 17%) compared with placebo. However, There were no significant safety concerns identified in 692
633 absolute changes within treatment groups were small and this study. The proportion of subjects with at least 1 AE was 693
634 largely remained within the range expected for post- comparable in the 3 treatment groups, and the incidence of 694
635 menopausal women. individual AEs in each of the treatment groups was low and 695
636 In adults, uncoupled bone resorption and formation with consistent with previously reported AEs in randomized, 696
637 a net increase in resorption results in bone loss and placebo-controlled PPI studies.49–51 697
638 decreased BMD, leading to osteoporosis.40,41 In the current No fractures were reported within 30 days after the last 698
639 trial, resorption and formation appeared to remain coupled dose of study drug, although 1 traumatic foot fracture 699
640 over 26 weeks, with increases in both formation and (dexlansoprazole) and 1 humerus fracture (esomeprazole) 700
641 resorption. Furthermore, there were no significant between- occurred at day 359 and day 161, respectively, both after 701
642 arm differences in BMD during 26 weeks of PPI therapy or the patients took the last dose of study drug. Most of the AEs 702
643 at 52 weeks from baseline. The very small within-subject that resulted in discontinuation of the active study drug 703
644 changes in BMD observed in this study were consistent were consistent with the known AE profiles for these drugs. 704
645 with expected changes in postmenopausal women (1% to The serious AE reported during the study (nephrolithiasis in 705
646 2% bone loss per year during the first 5 to 8 years after the dexlansoprazole group) was possibly consistent with the 706
647 menopause).42 Our data are consistent with a Canadian supplementation of calcium and vitamin D throughout the 707
648 longitudinal study that found no association between PPI study. No other clinically important findings were noted in 708
649 use and BMD loss over 10 years, and with the Study of the postdose clinical laboratory safety measures, physical 709
650 Women’s Health Across the Nation (SWAN), a longitudinal examination, vital signs, or electrocardiogram results. 710
651 study of 2068 premenopausal or early premenopausal Strengths of this study include the prospective, ran- 711
652 women, that found no difference in BMD change in PPI users domized, placebo-controlled study design, use of multiple 712
653 compared with nonusers.14,39 bone turnover markers, use of standard methods to mea- 713
654 Gastrin-induced PTH production leading to excess bone sure BMD, and use of a gold standard method to measure 714
655 resorption has been proposed as a mechanism to explain TFCA. In addition, we standardized vitamin D status and 715
656 PPI effects on bone homeostasis.8,9,43 Gastrin levels were calcium intake in all subjects, so that PPI or placebo therapy 716
657 increased with PPI administration in this study (an increase was the only intervention during the trial. We also 717
658 from baseline of 80–85 pg/mL), but there were no corre- acknowledge limitations of this study. The first limitation is 718
659 sponding increases in PTH levels in either PPI treatment a relatively small sample size. We initially recruited 235 719
660 720

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 - 2019 Dexlansoprazole, Esomeprazole in Bone Homeostasis 7

721 women into the trial, only to learn after completion of the References 781
722 trial that one site’s data were unreliable. Had all 235 1. Klok RM, Postma MJ, van Hout BA, et al. Meta-analysis: 782
723 women’s data been available for analysis, we likely would comparing the efficacy of proton pump inhibitors in 783
724 have observed smaller CIs around the study outcomes of short-term use. Aliment Pharmacol Ther 2003;17:1237– 784
725 bone turnover, BMD, and TFCA. Even with a smaller sample 1245. 785
726 size, ours is the only study in which participants were 2. Gill JM, Player MS, Metz DC. Balancing the risks and 786
727 randomized to PPI or placebo with the specific objective of benefits of proton pump inhibitors. Ann Fam Med 2011; 787
728 measuring multiple parameters of skeletal health. The ulti- 9:200–202. 788
729 mate goal was to identify a potential mechanism whereby 3. Moayyedi P, Leontiadis GI. The risks of PPI therapy. Nat 789
730 PPI therapy might increase the risk of fracture. Thus, we feel Rev Gastroenterol Hepatol 2012;9:132–139. 790
731 that our study adds important information to the literature 4. Leontiadis GI, Moayyedi P. Proton pump inhibitors and 791

CLINICAL AT
732 on PPI and fracture. The second limitation of our study is risk of bone fractures. Curr Treat Options Gastroenterol 792
733 the short duration of exposure to PPI therapy. We designed 2014;12:414–423. 793
734 this study to evaluate the direct effects of PPIs on skeletal 5. Targownik LE, Lix LM, Metge CJ, et al. Use of proton 794
735 health. We chose bone turnover as our primary outcome pump inhibitors and risk of osteoporosis-related frac- 795
736 because changes in bone turnover precede changes in BMD tures. CMAJ 2008;179:319–326. 796
737 and serve as an early marker of adverse (or beneficial) in- 6. Corleto VD, Festa S, Di Giulio E, et al. Proton pump in- 797
738 terventions. We had no pilot data regarding the effects of hibitor therapy and potential long-term harm. Curr Opin 798
739 PPIs on bone turnover by which to calculate a potential Endocrinol Diabetes Obes 2014;21:3–8. 799
740 sample size for the study. However, within-subject changes 7. Hansen KE, Jones AN, Lindstrom MJ, et al. Do proton 800
741 in the bone resorption marker CTX can be as high as 36%.33 pump inhibitors decrease calcium absorption? J Bone 801
742 Thus, we selected a sample size of 240 subjects, so that the Miner Res 2010;25:2786–2795. 802
743 width of the 95% CI for the difference in the percent change 8. Grimelius L, Johansson H, Lundqvist G, et al. The para- 803
744 from baseline in each of the markers between dexlanso- thyroid glands in experimentally induced hyper- 804
745 prazole or esomeprazole and placebo was restricted to no gastrinemia in the rat. Scand J Gastroenterol 1977;
805
more than 30%. Despite having only 115 subjects with valid 12:739–744.
746 806
747 data, the width of the 95% CI for the between-arm changes 9. Gagnemo-Persson R, Hakanson R, Sundler F, et al.
807
748 in bone turnover markers were still 33%. Finally, we Growth of the parathyroid glands in omeprazole-treated
808
studied healthy postmenopausal women and cannot state chickens. Scand J Gastroenterol 1994;29:493–497.
749 809
whether study outcomes might differ in other populations. 10. Yang YX. Proton pump inhibitor therapy and osteopo-
750 810
rosis. Curr Drug Saf 2008;3:204–209.
751 811
11. Vestergaard P, Rejnmark L, Mosekilde L. Proton pump
752 812
Conclusions inhibitors, histamine H2 receptor antagonists, and other
753 813
antacid medications and the risk of fracture. Calcif Tissue
754 Overall, this prospective, randomized, double-blind 814
Int 2006;79:76–83.
755 clinical trial detected minimal changes in bone homeosta- 815
12. Yang YX, Lewis JD, Epstein S, et al. Long-term proton
756 sis related to 26 weeks of PPI therapy. Increases in bone 816
pump inhibitor therapy and risk of hip fracture. JAMA
757 turnover were detected, but formation and resorption both
2006;296:2947–2953. 817
758 increased and remained coupled. There were no PPI- 818
13. Andersen BN, Johansen PB, Abrahamsen B. Proton
759 associated declines in BMD, serum or urine mineral levels,
pump inhibitors and osteoporosis. Curr Opin Rheumatol 819
760 PTH, or TFCA. Our study provides strong evidence against
2016;28:420–425. 820
761 PPI-mediated alterations in calcium absorption or mineral
14. Solomon DH, Diem SJ, Ruppert K, et al. Bone mineral 821
762 homeostasis. If there is a causal relationship between PPI
density changes among women initiating proton pump 822
763 and fracture, that relationship is not mediated by the most
inhibitors or H2 receptor antagonists: a SWAN cohort 823
764 common metabolic pathways predisposing to fracture. study. J Bone Miner Res 2015;30:232–239. 824
765 The findings reported here suggest that PPI use alone 825
15. Yu EW, Blackwell T, Ensrud KE, et al. Acid-suppressive
766 may not be a reason for postmenopausal women to have medications and risk of bone loss and fracture in older 826
767 additional bone density evaluations. The US Preventive adults. Calcif Tissue Int 2008;83:251–259. 827
768 Services Task Force recommends screening for osteoporosis 16. Arj A, Razavi Zade M, Yavari M, et al. Proton pump in- 828
769 in women aged 65 years and older, as well as in younger hibitors use and change in bone mineral density. Int J 829
770 women with comparable or greater fracture risk.52 Rheum Dis 2016;19:864–868. 830
771 17. Amoako AO, Jafilan L, Nasiri P, et al. Correlation of bone 831
772 mineral density scores and proton pump inhibitors use in 832
773 Supplementary Material the elderly. Curr Rheumatol Rev 2016;12:162–166. 833
774 Note: To access the supplementary material accompanying 18. O’Connell MB, Madden DM, Murray AM, et al. Effects of 834
775 this article, visit the online version of Gastroenterology at proton pump inhibitors on calcium carbonate absorption 835
776 www.gastrojournal.org, and at https://doi.org/10.1053/j. in women: a randomized crossover trial. Am J Med 2005; 836
777 gastro.2018.11.023. 118:778–781. 837
778 838
779 839
780 840

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 8 Hansen et al Gastroenterology Vol. -, No. -

19. Lam JR, Schneider JL, Zhao W, et al. Proton pump in- handling and patient preparation to reduce analytical
841 901
hibitor and histamine 2 receptor antagonist use and variability. Osteoporos Int 2017;28:2541–2556.
842 902
vitamin B12 deficiency. JAMA 2013;310:2435–2442. 34. Hansen KE, Johnson RE, Chambers KR, et al. Treatment
843 903
20. Dharmarajan TS, Kanagala MR, Murakonda P, et al. Do of vitamin D insufficiency in postmenopausal women: a
844 904
acid-lowering agents affect vitamin B12 status in older randomized clinical trial. JAMA Intern Med 2015;
845 905
adults? J Am Med Dir Assoc 2008;9:162–167. 175:1612–1621.
846 906
21. Serfaty-Lacrosniere C, Wood RJ, Voytko D, et al. 35. Larson GM, Sullivan HW, Rayford P. Omeprazole-
847 907
Hypochlorhydria from short-term omeprazole treatment induced hypergastrinemia: role of gastric acidity. J Surg
848 908
does not inhibit intestinal absorption of calcium, phos- Res 1986;40:504–509.
849 phorus, magnesium or zinc from food in humans. J Am 909
36. Kaye JA, Jick H. Proton pump inhibitor use and risk of
850 Coll Nutr 1995;14:364–368. 910
hip fractures in patients without major risk factors.
851 22. Koop H, Bachem MG. Serum iron, ferritin, and vitamin Pharmacotherapy 2008;28:951–959. 911
CLINICAL AT

852 B12 during prolonged omeprazole therapy. J Clin Gas- 912

37. Moayyedi P, Cranney A. Hip fracture and proton pump
853 troenterol 1992;14:288–292. inhibitor therapy: balancing the evidence for benefit and 913
854 23. Wright MJ, Sullivan RR, Gaffney-Stomberg E, et al. harm. Am J Gastroenterol 2008;103:2428–2431. 914
855 Inhibiting gastric acid production does not affect intes- 38. Moayyedi P, Yuan Y, Leontiadis G. on behalf of the CAG 915
856 tinal calcium absorption in young, healthy individuals: a Clinical Affairs. Canadian Association of Gastroenter- 916
857 randomized, crossover, controlled clinical trial. J Bone ology position statement: hip fracture and proton pump 917
858 Miner Res 2010;25:2205–2211. inhibitor therapy: a 2013 update. Can J Gastroenterol 918
859 24. World Health Organization. Assessment of fracture risk and 2013;27:593–595. 919
860 its application to screening for postmenopausal osteopo- 39. Targownik LE, Leslie WD, Davison KS, et al. The 920
861 rosis: report of a WHO study group. Geneva, Switzerland: relationship between proton pump inhibitor use and 921
862 World Health Organization, 1994. WHO Technical Report longitudinal change in bone mineral density: a 922
863 Series, No. 843. Available at: http://apps.who.int/iris/ population-based study [corrected] from the Canadian 923
864 handle/10665/39142. Accessed May 31, 2018. Multicentre Osteoporosis Study (CaMos). Am J Gastro- 924
865 25. Gray SL, LaCroix AZ, Larson J, et al. Proton pump in- enterol 2012;107:1361–1369. 925
866 hibitor use, hip fracture, and change in bone mineral 40. Camacho P, Kleerekoper M. Biochemical markers of 926
867 density in postmenopausal women: results from the bone turnover. In: Favus MJ, ed. Primer on the metabolic 927
868 Women’s Health Initiative. Arch Intern Med 2010; bone diseases and disorders of mineral metabolism. 6th 928
869 170:765–771. ed. Washington, DC: American Society for Bone and 929
870 26. Morris HA, Eastell R, Jorgesen NR, et al. Clinical use- Mineral Research, 2006. Q4 930
871 fulness of bone turnover marker concentrations in oste- 41. Watts NB. Clinical utility of biochemical markers of bone 931
872 oporosis. Clin Chim Acta 2016;467:34–41. remodeling. Clin Chem 1999;45:1359–1368. 932
873 27. Vasikaran S, Eastell R, Bruyère O, et al. Markers of bone 42. Manolagas SC. Birth and death of bone cells: basic 933
874 turnover for the prediction of fracture risk and monitoring regulatory mechanisms and implications for the patho- 934
875 of osteoporosis treatment: a need for international genesis and treatment of osteoporosis. Endocr Rev 935
876 reference standards. Osteoporos Int 2011;22:391–420. 2000;21:115–137. 936
877 28. Wheater G, Elshahaly M, Tuck SP, et al. The clinical utility 43. Raggatt LJ, Partridge NC. Cellular and molecular mech- 937
878 of bone marker measurements in osteoporosis. J Transl anisms of bone remodeling. J Biol Chem 2010; 938
879 Med 2013;11:201. 285:25103–25108. 939
880 29. Black DM, Greenspan SL, Ensrud KE, et al. The effects of 44. Hunt CD, Johnson LK. Calcium requirements: new esti- 940
881 parathyroid hormone and alendronate alone or in com- mations for men and women by cross-sectional statisti- 941
882 bination in postmenopausal osteoporosis. N Engl J Med cal analyses of calcium balance data from metabolic 942
2003;349:1207–1215. studies. Am J Clin Nutr 2007;86:1054–1063.
883 943
884 30. Chen P, Satterwhite JH, Licata AA, et al. Early changes in 45. Heaney RP, Recker RR, Stegman MR, et al. Calcium 944
biochemical markers of bone formation predict BMD absorption in women: relationships to calcium intake,
885 945
response to teriparatide in postmenopausal women with estrogen status, and age. J Bone Miner Res 1989;
886 946
osteoporosis. J Bone Miner Res 2005;20:962–970. 4:469–475.
887 947
31. Lindsay R, Nieves JW, Formica M, et al. Randomized 46. Van Cromphaut SJ, Dewerchin M, Hoenderop JG, et al.
888 948
controlled study of effect of parathyroid hormone on Duodenal calcium absorption in vitamin D receptor-
889 949
vertebral-bone mass and fracture incidence among knockout mice: functional and molecular aspects. Proc
890 950
postmenopausal women on oestrogen with osteopo- Natl Acad Sci U S A 2001;98:13324–13329.
891 rosis. Lancet 1997;350:550–555. 951
47. Assadi F. Hypomagnesemia: an evidence-based
892 952
32. Garnero P, Hausherr E, Chapuy MC, et al. Markers of approach to clinical cases. Iran J Kidney Dis 2010;
893 bone resorption predict hip fracture in elderly women: the 953
4:13–19.
894 EPIDOS Prospective Study. J Bone Miner Res 1996; 954
48. Regolisti G, Cabassi A, Parenti E, et al. Severe hypo-
895 11:1531–1538. 955
magnesemia during long-term treatment with a proton
896 33. Szulc P, Naylor K, Hoyle NR, et al. Use of CTX-1 and 956
pump inhibitor. Am J Kidney Dis 2010;56:168–174.
897 P1NP as bone turnover markers: National Bone Health 957
49. Kuipers EJ, Sung JJ, Barkun A, et al. Safety and tolera-
898 Alliance recommendations to standardize sample 958
bility of high-dose intravenous esomeprazole for
899 959
900 960

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 - 2019 Dexlansoprazole, Esomeprazole in Bone Homeostasis 9

prevention of peptic ulcer rebleeding. Adv Ther 2011; Health, Room 4124 MFCB, 1685 Highland Avenue, Madison, WI 53705.
961 e-mail: keh@medicine.wisc.edu; fax: 608-263-7353. Q1 1021
28:150–159.
962 1022
963 50. Peura DA, Metz DC, Dabholkar AH, et al. Safety profile of Acknowledgments
1023
This study was supported by Takeda Development Center Americas, Inc, Q6
dexlansoprazole MR, a proton pump inhibitor with a novel
964 Deerfield, Illinois. Medical writing assistance was provided by Nikhilesh 1024
dual delayed release formulation: global clinical trial
965 Sanyal, PhD, and Jacob Edelstein, PhD, of inVentiv Medical
Communications, LLC, a Syneos Health group company, and supported by
1025
experience. Aliment Pharmacol Ther 2009;30:1010–1021.
966 Takeda Development Center Americas, Inc. 1026
51. Metz DC, Howden CW, Perez MC, et al. Clinical trial:
967 Sai Nudurupati’s current affiliation: AbbVie Inc, North Chicago, Illinois. 1027
dexlansoprazole MR, a proton pump inhibitor with dual Author contributions: Drs Hansen, Nudurupati, Metz, and Perez contributed
968 to the study design. Drs Hansen and Nieves contributed to the data collection. 1028
delayed-release technology, effectively controls symp-
969 Drs Hansen, Nieves, Nudurupati, Metz, and Perez contributed to data 1029
toms and prevents relapse in patients with healed erosive interpretation and writing the manuscript.
970 oesophagitis. Aliment Pharmacol Ther 2009;29:742–754.
1030
971 Conflicts of interest 1031

CLINICAL AT
52. US Preventive Services Task Force. Screening for osteo- Dr Hansen was paid for her work as a consultant in the design and conduct of
972 the study and received a grant to conduct the study. Dr Nieves received a 1032
porosis: US Preventive Services Task Force recommen-
973 dation statement. Ann Intern Med 2011;154:356–364. research grant from Takeda to conduct this research. Dr Nudurupati was an 1033
974 employee of Takeda Pharmaceuticals at the time this study was conducted.
Dr Metz received administrative and editorial support for study design and
1034
975 analysis of the data. Dr Perez is employed by Takeda Pharmaceuticals. Q2 1035
976 1036
Received August 23, 2018. Accepted November 8, 2018. Funding
977 This study was funded by Takeda Pharmaceuticals International, Inc. Takeda 1037
978 Reprint requests Pharmaceuticals International, Inc., was responsible for and sponsored the 1038
Address requests for reprints to: Karen E. Hansen, MD, MS, Associate study design, data collection, data interpretation, and writing of this
979 Professor of Medicine, University of Wisconsin School of Medicine & Public manuscript. Q3 1039
980 1040
981 1041
982 1042
983 1043
984 1044
985 1045
986 1046
987 1047
988 1048
989 1049
990 1050
991 1051
992 1052
993 1053
994 1054
995 1055
996 1056
997 1057
998 1058
999 1059
1000 1060
1001 1061
1002 1062
1003 1063
1004 1064
1005 1065
1006 1066
1007 1067
1008 1068
1009 1069
1010 1070
1011 1071
1012 1072
1013 1073
1014 1074
1015 1075
1016 1076
1017 1077
1018 1078
1019 1079
1020 1080

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 9.e1 Hansen et al Gastroenterology Vol. -, No. -

1081 Supplementary Methods participants received a 600-mg calcium supplement, along 1141
1082 with 2 stable calcium isotopes (8 mg of 44Ca by mouth in 1142
1083 calcium-fortified apple juice and 3 mg of 42Ca intrave- 1143
Study Randomization and Blinding nously), and consumed food for the next 24 hours that
1084 1144
Randomization of patients in a 1:1:1 ratio to dexlanso- replicated the macronutrient and micronutrient content of
1085 1145
prazole 60-mg capsules, esomeprazole 40-mg capsules, or their regular diet, including kilocalories, calcium, carbohy-
1086 1146
1087
placebo was scheduled and maintained using an interactive drates, protein, fat, fiber, vitamin D3, sodium, magnesium, 1147
voice-activated response system. Patients were randomized caffeine, and oxalate. Uneaten food was bagged and weighed
1088 1148
using a blocked design stratified by vitamin D3 supple- to be assessed by a nutritionist for calcium intake. Research
1089 1149
mentation (low or high) required by the subject at the nurses collected all urine for 24 hours during an inpatient
1090 1150
screening visit (week 12) after the result of 25(OH)D was stay. Urine was frozen at ‒70 Celsius and analyzed in
1091 1151
1092
received, and the entire block of randomization numbers batches to determine the concentrations of 42Ca and 44Ca 1152
1093
was assigned to each site. The investigational drug blind relative to 43Ca. The inpatient diet provided during the 24- 1153
(double-blind) was maintained using the interactive voice- hour TFCA study period matched the participants’ outpa-
1094 1154
activated response system and was not broken except for tient intake of kilocalories, calcium, carbohydrates, protein,
1095 1155
1096
reasons of medical necessity. fat, fiber, vitamin D3, sodium, magnesium, caffeine (servings 1156
1097 per day), and oxalate (servings per day). TFCA was calcu- 1157
1098 Additional Inclusion and Exclusion Criteria lated using the formula of Eastell and colleagues1: 1158
1099 Participants could not take over-the-counter PPIs within 1159
9 months of study start or during the study, apart from % excess 44Ca ðoralÞ
1100 TFCA ¼ 1160
1101 study medication. Participants were ineligible if they had % excess 42Ca ðintravenousÞ 1161
1102 PTH or thyroid-stimulating hormone levels outside of the natural abundance 44Ca dose 42Ca 1162
 
1103 reference range for the testing laboratory at week ‒12; a natural abundance 42Ca dose 44Ca 1163
1104 25(OH)D level <10 ng/mL at week ‒12 or <32 ng/mL at 1164
1105 week ‒2; a disorder strongly associated with osteoporosis; a 1165
1106 clinical history of fragility of wrist, hip, spine, or leg frac- Quality Control 1166
1107 tures; or a family history of genetic bone disorders. Partic- Quality control via 10 repeated scans of the Bona Fide 1167
1108 ipants who used any nicotine-containing products within 3 Phantom (BioClinica, Inc, Newtown, PA) within each center 1168
1109 months of study start also were excluded. Potential candi- resulted in a coefficient of variance <0.5% during the study. 1169
1110 dates also were excluded if they had <2 evaluable vertebrae There were no changes in dual-energy x-ray absorptiometry 1170
1111 or a condition that interfered with dual-energy x-ray ab- machines and no equipment issues noted. 1171
1112 sorptiometry measurement (ie, hip replacement, vertebral Monitoring visits were made to study sites periodically 1172
1113 deformity, or severe lumbar scoliosis). during the study to ensure that the protocol was followed. 1173
1114 Glucocorticoids, immunosuppressants, antiepileptics, Protocol deviations were made only to eliminate immediate 1174
1115 selective serotonin reuptake inhibitors, lithium, bisphosph- hazard to study participants and were documented in 1175
1116 onates, loop and thiazide diuretics, PPIs, histamine-2 re- source documents. Quality assurance audits were done by 1176
1117 ceptor antagonists, antacids, systemic fluoride, aromatase the sponsor or regulatory agencies, such as the US Food and 1177
1118 inhibitors, antidiabetic treatments, RANK ligand inhibitors, Drug Administration. The study was conducted according to 1178
1119 estrogen therapy, hormone replacement therapy, low- ethical principles based on the Declaration of Helsinki and 1179
1120 molecular-weight heparin, warfarin, clopidogrel, oral anti- the international Guideline for Good Clinical Practice. 1180
1121 fungals, and multivitamins could not be used for up to 1 1181
1122 year before the study. 1182
1123 1183
1124 Calcium Absorption Substudy Design Supplementary Reference 1184
1125 A 4-day food diary was completed by all substudy par- 1. Eastell R, Vieira NE, Yergey AL, et al. One-day test using 1185
1126 ticipants. On day ‒1 and week 26, participants arrived at the stable isotopes to measure true fractional calcium ab- 1186
1127 research ward at 7 AM while fasting. With breakfast, sorption. J Bone Miner Res 1989;4:463–468. 1187
1128 1188
1129 1189
1130 1190
1131 1191
1132 1192
1133 1193
1134 1194
1135 1195
1136 1196
1137 1197
1138 1198
1139 1199
1140 1200

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 - 2019 Dexlansoprazole, Esomeprazole in Bone Homeostasis 9.e2

1201 1261
1202 1262
1203 1263
1204 1264
1205 1265
1206 1266
1207 1267
1208 1268
1209 1269
1210 1270
1211 1271
1212 1272
1213 1273
1214 1274
1215 1275
1216 1276
1217 1277
1218 1278
1219 1279
1220 1280
1221 1281
1222 1282
1223 Supplementary Figure 1. Disposition of participants (CONSORT flowchart). QD, once daily. Volunteers screened: Numbers 1283
1224 exclude the subjects from one site that was disqualified by the US Food and Drug Administration. 1284
1225 1285
1226 1286
1227 1287
1228 1288
1229 1289
1230 1290
1231 1291
1232 1292
1233 1293
1234 1294
1235 1295
1236 1296
1237 1297
1238 1298
1239 1299
1240 1300
1241 1301
1242 1302
1243 1303
1244 1304
1245 1305
1246 1306
1247 1307
1248 1308
1249 1309
1250 1310
1251 1311
1252 1312
1253 1313
1254 1314
1255 1315
1256 1316
1257 1317
1258 1318
1259 1319
1260 1320

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

 1380
1379
1378
1377
1376
1375
1374
1373
1372
1371
1370
1369
1368
1367
1366
1365
1364
1363
1362
1361
1360
1359
1358
1357
1356
1355
1354
1353
1352
1351
1350
1349
1348
1347
1346
1345
1344
1343
1342
1341
1340
1339
1338
1337
1336
1335
1334
1333
1332
1331
1330
1329
1328
1327
1326
1325
1324
1323
1322
1321

9.e3
Hansen et al
Supplementary Table 1.Excluded Medications and Substances

Excluded medications Period of exclusion

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

Systemic glucocorticoids (eg, prednisone, cortisone, hydrocortisone, 6 mo before screening (visit 1) and through last dose of study drug
dexamethasone)
Topical glucocorticoids Topical glucocorticoids were allowed during the study as long as the use did not exceed
more than the equivalent of 5 mg/d prednisone for more than 14 d
Inhaled glucocorticoidsa Inhaled glucocorticoids were allowed during the study as long as the use did not exceed
21 d in a 3-mo period or 42 d in 1 y through last dose of study drug
Immunosuppressors 2 mo before screening (visit 1) and through last dose of study drug
Antiepileptic agents, SSRIs, and lithium 3 mo before screening (visit 1) and through last dose of study drug
Oral bisphosphonates Use within 1 y before screening (visit 1) and through last dose of study drug
Loop and thiazide diuretics Use within 1 y before screening (visit 1) and through last dose of study drug
PPIs, over-the-counter PPIs (more than 3 doses per month) 6 mo before screening (visit 1) and through last dose of study drug
H2RAs, antacids From screening (visit 1) through last dose of study drug
Intravenous bisphosphonates Ever taken
Systemic fluoride treatment, PTH analog Ever taken
Aromatase inhibitors, antidiabetic treatments 1 y before screening (visit 1) and through last dose of study drug
RANK ligand inhibitors 6 mo before screening (visit 1) and through last dose of study drug
Systemic estrogen therapy or hormone replacement therapyb 1 y before screening (visit 1) and through last dose of study drug
Selective estrogen-receptor modulator (tamoxifen and raloxifene) 3 mo before screening (visit 1) and through last dose of study drug
Low-molecular-weight heparin (dalteparin, enoxaparin, tinzaparin) 3 mo before screening (visit 1) and through last dose of study drug
Warfarin 3 mo before screening (visit 1) and through last dose of study drug
Clopidogrel 1 mo before screening (visit 1) and through last dose of study drug
Oral antifungals 2 mo before screening (visit 1) and through last dose of study drug
Multivitamins and other vitamins/supplements containing calcium or vitamin D3 other Screening (visit 1) and through last dose of study drug
than what was required per protocol

H2RA, histamine 2-receptor antagonist; RANK, receptor activator of nuclear factor kappa-B ligand; SSRI, selective serotonin-reuptake inhibitor.

Gastroenterology Vol.
a
Inhaled glucocorticoids were allowed during the study as long as the use did not exceed 21 consecutive days.
b
Local vaginal estrogen was permitted.

-,
No.
-
1440
1439
1438
1437
1436
1435
1434
1433
1432
1431
1430
1429
1428
1427
1426
1425
1424
1423
1422
1421
1420
1419
1418
1417
1416
1415
1414
1413
1412
1411
1410
1409
1408
1407
1406
1405
1404
1403
1402
1401
1400
1399
1398
1397
1396
1395
1394
1393
1392
1391
1390
1389
1388
1387
1386
1385
1384
1383
1382
1381
 1500
1499
1498
1497
1496
1495
1494
1493
1492
1491
1490
1489
1488
1487
1486
1485
1484
1483
1482
1481
1480
1479
1478
1477
1476
1475
1474
1473
1472
1471
1470
1469
1468
1467
1466
1465
1464
1463
1462
1461
1460
1459
1458
1457
1456
1455
1454
1453
1452
1451
1450
1449
1448
1447
1446
1445
1444
1443
1442
1441
Supplementary Table 2.Schedule of Study Procedures

-
2019
Screening Randomization Treatment Follow-up BMD

Visit 1 Visit 2 Visit 3 Visit 4 Visit 5

Wk 26 or early
Study procedure or assessmenta Wk ‒12 Wk ‒2 Day ‒1 Day 1 Wk 13 termination Wk 52

Medical history, demographics, and concurrent medical conditions X

Physical examination, vital signs, ECG X X X
FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

Height and BMI calculationb X X X

Clinical laboratory tests (chemistry, hematology, and urinalysis) X X X X
BMD by DXA X X X
24-hour urine calcium excretion, magnesium excretion and 24-hour urine X X
creatininec
Urine magnesium X X X
Estimated CrCl calculation X X X
Gastrin X X
FSH X
P1NP, CTX, urine NTx,c BsAP X X X
Serum phosphorus, calcium, albumin, and magnesium X X X
25(OH)D X X X X X
PTH X X X X
AE monitoring X X
Dispense double-blind treatments (dexlansoprazole 60 mg, esomeprazole X X
40 mg, or placebo)
Calcium absorption substudy X X

Dexlansoprazole, Esomeprazole in Bone Homeostasis

BMI, body mass index; CrCl, creatinine clearance; DXA, dual-energy x-ray absorptiometry; ECG, electrocardiogram; FSH, follicle-stimulating hormone.
a
All laboratory tests required an 8-hour fast except at week 12.
b
BMI calculation performed at visit 1 (week ‒12) only.
c
Urine NTx samples were collected in clinic with the second fasting void of the day during visit 2, visit 4, and visit 5. For visit 5, the second fasting void of the day was
defined as the first void of the day after completion of the 24-hour urine collection. For visit 2 and visit 4, the second fasting void of the day was defined as the first void after
6 AM.

9.e4
1560
1559
1558
1557
1556
1555
1554
1553
1552
1551
1550
1549
1548
1547
1546
1545
1544
1543
1542
1541
1540
1539
1538
1537
1536
1535
1534
1533
1532
1531
1530
1529
1528
1527
1526
1525
1524
1523
1522
1521
1520
1519
1518
1517
1516
1515
1514
1513
1512
1511
1510
1509
1508
1507
1506
1505
1504
1503
1502
1501
 1620
1619
1618
1617
1616
1615
1614
1613
1612
1611
1610
1609
1608
1607
1606
1605
1604
1603
1602
1601
1600
1599
1598
1597
1596
1595
1594
1593
1592
1591
1590
1589
1588
1587
1586
1585
1584
1583
1582
1581
1580
1579
1578
1577
1576
1575
1574
1573
1572
1571
1570
1569
1568
1567
1566
1565
1564
1563
1562
1561

9.e5
Hansen et al
Supplementary Table 3.Baseline and Week 26 Median Urinary Calcium, Serum Calcium, Phosphorus, Magnesium, Urine Magnesium, PTH, Serum Vitamin D, and Gastrin
Levels

Dexlansoprazole 60 mg QD,
Placebo, n ¼ 38 n ¼ 36 Esomeprazole 40 mg QD, n ¼ 34 Difference vs placebo (95% CI)a
FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

Difference from Difference

Difference from baseline from baseline Dexlansoprazole Esomeprazole
n Median (range) baseline (95% CI)a n Median (range) (95% CI)a n Median (range) (95% CI)a vs placebo vs placebo

Serum gastrin, pg/mL

Baseline 28 27.5 (13.0–325.0) – 30 28.5 (18.0–60.0) – 25 29.0 (11.0–73.0) – – –
Week 26 28 30.5 (15.0–703.0) 1.0 (‒5.0 to 7.0) 30 105.5 (32.0–684.0) 78.0 (56.0 to 99.0) 25 113.0 (30.0–892.0) 84.0 (64.0 to 108.0) 79.5 (55.0 to 106.0) 85.0 (66.0 to 110.0)
PTH, pg/mL
Baseline 36 28 (16- 53) – 30 27 (15–77) – 31 31 (15–48) – – –
Week 26 32 30 (15–70) 1 (‒4 to 7) 29 31 (16–82) 1 (‒4 to 8) 26 33 (20–57) 3 (‒2 to 9) 1 (‒3 to 4) 2 (‒2 to 7)
Urinary calcium, mg/dL
Baseline 29 145 (42–325) – 29 174 (59–346) – 24 121 (57–355) – – –
Week 26 29 119 (14–326) ‒26 (‒66 to 16) 29 112 (24–286) ‒50 (‒92 to ‒11) 24 103 (23–242) ‒34 (‒84 to 1) ‒17 (‒59 to 21) ‒8 (‒64 to 27)
Serum calcium, mg/dL
Baseline 36 9.5 (8.6–10.1) – 32 9.5 (9.0–10.1) – 30 9.6 (9.1–10.3) – – –
Week 26 32 9.5 (8.1–10.5) 0.0 (‒0.2 to 0.2) 31 9.5 (8.8–10.0) 0.0 (‒0.2 to 0.2) 26 9.6 (9.1–10.1) ‒0.1 (‒0.2 to 0.1) 0.0 (‒0.2 to 0.2) 0.0 (‒0.3 to 0.2)
Serum phosphorus, mg/dL
Baseline 36 3.8 (2.8–4.2) – 32 3.8 (3.2–4.9) – 30 3.9 (3.1–4.5) – – –
Week 26 32 3.8 (3.2–5.2) 0.0 (‒0.2,0.2) 31 3.9 (3.0–5.1) 0.1 (‒0.1 to 0.3) 26 4.0 (3.5–4.7) 0.0 (‒0.1 to 0.2) 0.0 (‒0.2 to 0.2) ‒0.1 (‒0.3 to 0.2)
Urine magnesium, mEq/L
Baseline 35 3.7 (0–8) – 32 3.6 (1–18) – 31 3.6 (1–13) – – –
Week 26 31 3.3 (1–8) ‒0.2 (‒1.4 to 1.0) 31 3.7 (1–8) ‒0.3 (‒1.5 to 1.0) 26 3.1 (1–12) ‒0.1 (‒1.5 to 0.9) ‒0.3 (‒1.6 to 0.9) ‒0.3 (‒1.6 to 1.0)
Serum magnesium, mEq/L
Baseline 36 1.7 (1.5–2.0) – 32 1.7 (1.4- 2.0) – 30 1.8 (1.6–2.0) – – –
Week 26 32 1.7 (1.4–2.0) 0.0 (‒0.1 to 0.1) 31 1.7 (1.5–1.9) 0.0 (‒0.1 to 0.0) 26 1.8 (1.5–2.1) 0.0 (‒0.1 to 0.0) 0.0 (‒0.1 to 0.0) 0.0 (‒0.1 to 0.0)
25(OH)D, ng/mL
Baseline 36 39 (29–69) – 32 38 (26–87) – 31 35 (19–55) – – –

Gastroenterology Vol.
Week 26 32 36 (25–56) ‒2 (‒6 to 2) 30 40 (27–50) 1 (‒3 to 5) 26 38 (22–49) 2 (‒2 to 6) 3 (‒2 to 8) 4 (‒1 to 9)

QD, once daily.

a
Difference from baseline or placebo estimated using Hodges-Lehmann estimate; 95% CIs of the differences in actual values calculated by the Moses method. Com-
parisons were considered statistically significant if the 95% CI excluded zero (reported in bold text). Q5

-,
No.
-
1680
1679
1678
1677
1676
1675
1674
1673
1672
1671
1670
1669
1668
1667
1666
1665
1664
1663
1662
1661
1660
1659
1658
1657
1656
1655
1654
1653
1652
1651
1650
1649
1648
1647
1646
1645
1644
1643
1642
1641
1640
1639
1638
1637
1636
1635
1634
1633
1632
1631
1630
1629
1628
1627
1626
1625
1624
1623
1622
1621
 1740
1739
1738
1737
1736
1735
1734
1733
1732
1731
1730
1729
1728
1727
1726
1725
1724
1723
1722
1721
1720
1719
1718
1717
1716
1715
1714
1713
1712
1711
1710
1709
1708
1707
1706
1705
1704
1703
1702
1701
1700
1699
1698
1697
1696
1695
1694
1693
1692
1691
1690
1689
1688
1687
1686
1685
1684
1683
1682
1681
-
Supplementary Table 4.True Fractional Calcium Absorption Change From Baseline to Week 26
2019

Placebo, n ¼ 12 Dexlansoprazole 60 mg QD, n ¼ 9 Esomeprazole 40 mg QD, n ¼ 10

Baseline, median (range) 0.16 (0.10 to 0.29) 0.18 (0.11 to 0.24) 0.19 (0.10 to 0.23)
Week 26, median (range) 0.17 (0.08 to 0.27) 0.19 (0.11 to 0.29) 0.20 (0.14 to 0.30)
Change, median (range) ‒0.02 (‒0.08 to 0.07) 0.02 (‒0.04 to 0.10) 0.07 (‒0.07 to 0.08)
Difference between treatment vs placebo (95% CI)a ‒ 0.03 (‒0.02 to 0.08) 0.06 (0.02 to 0.11)

QD, once daily.

a
Difference estimated with Hodges-Lehmann estimate. Change values were considered statistically significant if the 95% CI excluded zero (reported in bold text).

FLA 5.5.0 DTD
YGAST62254_proof
15 February 2019
7:37 pm
ce

Dexlansoprazole, Esomeprazole in Bone Homeostasis
9.e6

1800
1799
1798
1797
1796
1795
1794
1793
1792
1791
1790
1789
1788
1787
1786
1785
1784
1783
1782
1781
1780
1779
1778
1777
1776
1775
1774
1773
1772
1771
1770
1769
1768
1767
1766
1765
1764
1763
1762
1761
1760
1759
1758
1757
1756
1755
1754
1753
1752
1751
1750
1749
1748
1747
1746
1745
1744
1743
1742
1741