Parathyroid Hormone for Treatment of Osteoporosis

ORIGINAL INVESTIGATION Parathyroid Hormone for Treatment of Osteoporosis

Carolyn Crandall, MD

Background: Osteoporosis is a common condition as- pathic osteoporosis. Parathyroid hormone may protect sociated with multiple deleterious consequences. No against gonadotropin-releasing hormone agonist–related therapy entirely abolishes fracture risk. bone loss. Effects are less clear at nonspine sites when PTH is used as part of combination or sequential therapies or Methods: A MEDLINE database (1966 to the present) for treatment of glucocorticoid-induced osteoporosis. Para- search was performed for randomized controlled trials thyroid hormone decreased the incidence of radiographi- in humans using the keywords osteoporosis and parathy- cally detected spinal fractures. The numbers of nonverte- roid hormone (PTH) or parathyroid hormone and frac- bral fractures were too low to be broken down by individual ture. The Cochrane database was searched using the search site. Parathyroid hormone injections were difficult for some terms osteoporosis and parathyroid hormone. patients to comply with. Occasionally, PTH-associated hypercalcemia may be dose-dependent, often manifesting Results: Parathyroid hormone (usually subcutaneous) early in treatment. An increase in cancer risk from PTH is dosages varied markedly across the 20 randomized con- not reported in humans. trolled trial studies retrieved. In the range of 50 to 100 µg/d, effects may be dose-related. Results of larger trials (up to Conclusions: Parathyroid hormone decreases verte- 1637 patients) were conflicting as to whether effects were bral fractures and increases spinal bone density in post- limited to the spine and suggested detrimental effects on menopausal osteoporosis and glucocorticoid-induced os- radius bone mineral density. Little data analyzed the ef- teoporosis, but at the expense of a decrease in radius bone fects of PTH in older vs younger subjects or directly com- density. The long-term safety and nonvertebral fracture pared the effects by sex. Increases in spine bone mineral efficacy are unknown. density are induced by PTH in postmenopausal osteoporosis, glucocorticoid-induced osteoporosis, and idio- Arch Intern Med. 2002;162:2297-2309

STEOPOROSIS IS a com- has intensified surrounding the potential mon condition. An esti- role of bone formation inducers alone or mated 17 to 23 million in combination with antiresorptive agents. women and 8 to 15 mil- Several of the bone formation induc- lion men older than 50 ers (also called anabolic agents) have been yearsO in the United States have osteopo- investigated in animal models of osteoporosis or low bone density, with conse- rosis, but trial results in humans were ini- quential elevated fracture risk.1 Osteopo- tially disappointing. In humans, fluoride rosis is associated with multiple deleterious treatment unexpectedly increased the risk consequences, including deformity, pain, of fracture, presumably due to abnormal loss of ambulation, and death.2-6 bone quality, despite impressive increases To date, all approved osteoporosis in bone density.11-13 Fluoride has not been treatments are of the antiresorptive class. approved for management of osteoporosis Antiresorptive agents block resorption of in the United States. More recent efforts bone, but they do not induce bone forma- have focused on different fluoride dosages tion. They induce a significant decrease in and preparations, such as slow-release fluo- From the Department of fracture risk, but they do not decrease the ride preparations, as well as other bone for- Medicine, David Geffen School 7-9 of Medicine, and Iris risk completely. The increased bone den- mation inducers, such as parathyroid hor- Cantor–UCLA Women’s Health sity induced by antiresorptive agents is re- mone (PTH). 14,15 Center, University of lated to more complete secondary bone Preliminary studies of PTH have California, Los Angeles, UCLA mineralization, rather than increased syn- been overshadowed by concerns about ad- School of Medicine. thesis of new bone.10 Therefore, interest verse effects at nonspine areas, similar to

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2297

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 1. Randomized Controlled Trials of Parathyroid Hormone With Bone Density Outcomea

Source Subjects Regimenb Calciumc or Vitamin D Hodsman and 20 Patients aged 50-78 y, the men having hPTH–(1-38) 400 IU/d SC for 14 d alone or followed None mentioned Fraher,16 1990 idiopathic osteoporosis and women by CT 75-100 U/dd SC for 56 d, then 30 d Hodsman et al,17 having postmenopausal osteoporosis, treatment-free, randomized, 1 cycle in 1990 1991 all with symptomatic vertebral fractures report, 2 cycles (total 200 days) in 1991 Finkelstein et al,18 40 Women aged 20-44 y with Nafarelin acetate alone (200 µg intranasally bid) or If dietary urinary calcium excretion 1994 endometriosis with hPTH–(1-34) 40 µg SC qd for 6 mo, not was above 350 mg/24h, dietary double-blinde calcium intake was decreased by 50%. If serum calcium was Finkelstein et al,19 43 Women aged 21-45 y with Nafarelin alone (200 µg intranasally bid) or with Ͼ10.5 mg/dL, hPTH dosage or 1998 symptomatic endometriosis hPTH–(1-34) 40 µg SC qd for 12 mo, not dietary calcium intake was double-blindf decreased by 30%-50%. Hodsman et al,20 30 Women, mean age 67 y, with hPTH–(1-34) 50 µg/d SC for 28 days followed by Calcium supplement 500 mg 1997 osteoporosisg placebo vs hPTH for 28 days followed by CT 75 U elemental calcium per day, no Hodsman et al,21 29 Osteoporotich women, mean age 67±7 SC qd for 42 days, followed by 20 d vitamin D supplement. 2000 years, vs 15 untreated osteoporotic treatment-free in cycles, for 2 y, blinding not controls with severe osteoporosis and discussed multiple vertebral fractures Lane et al,22 1998 51 Postmenopausal women aged 50-82 y hPTH–(1-34) 40 µg/d SC+estrogen or estrogen Calcium carbonate 1000 mg/d and Lane et al,23 2000 with osteoporosis, taking HRT and alone, 12-mo treatment, followed by 12-mo vitamin D 800 IU/d as 2 multiple glucocorticoidsi observation in the 2 later studies for total duration vitamins. 24 Lane et al, 2000 of 24 mo, placebo injections were not given Cosman et al,25 52 Women aged 58-63 with osteoporosis hPTH–(1-34) 400 IU/d SC+HRT vs HRT alone for Calcium intake maintained at 1500 2001 taking HRTj 3 y, followed by1yofHRTalonek mg/d total. Vitamin D 400 IU/d. Dempster et al,26 Subgroup of 8 men and 8 hPTH–(1-34) 400 IU/d SC+HRT vs HRT alone for Calcium intake maintained at 1500 2001 postmenopausal women from the 2001 3 y, followed by1yofHRTalonek mg/d total. Vitamin D 400 IU/d. Cosman et al25 trial, mean age 54 y Cosman et al,27 10 Women with osteoporosis already hPTH–(1-34) 400 IU/d SC + alendronate vs Calcium intake maintained at 1500 1998 taking alendronate sodium, aged alendronate continuation alone for 6 wk open mg/d total from diet and 59-71 yl assignment supplements. No vitamin D. Neer et al,28 2001 1637 Postmenopausal women aged hPTH–(1-34) 20 or 40 µg vs placebo SC qd, median Calcium 1000 mg/24 h and vitamin 69-70 y with prior vertebral fracturesm duration of treatment 16-17 mo, follow-up 21 mo D 400-1200 IU daily supplements. Lindsay et al,29 34 Postmenopausaln women with hPTH–(1-34) 25 µg/d SC with HRTp vs HRT alone Calcium supplementation to bring 1997 osteoporosiso for 3 y, placebo injections were not given total intake to 1500 mg/d, no vitamin D mentioned. Fujita et al,30 220 Women aged 45-95 y with hPTH–(1-34) 50, 100, or 200 Ur SC weekly for None mentioned. 1999 osteoporosisq 48 wk, double-blind Rittmaster et al,31 66 Postmenopausal osteoporotic women hPTH–(1-84) 50, 75, or 100 µg/d SC for 1 y, Calcium carbonate 500 mg/d and 2000 aged 50-75 ys followed by alendronate 10 mg/d alone for 1 y vitamin D 400 IU/d. (sequential therapy), first year was double-blind, second year open label Kurland et al,32 23 Men aged 30-68 y with idiopathic hPTH–(1-34) 400 IU vs vehicle SC qd for 18 mo, All received calcium to maintain total 2000 osteoporosist double-blind intake of 1500 mg/d, all took vitamin D 400 IU/d.

(continued )

concerns raised surrounding fluoride administration. Re- RESULTS cent studies have attempted to address these questions and have sometimes involved combination therapy with antiresorptive agents to protect nonspine areas. Results are shown in Tables 1, 2, 3, 4, and 5. The Parathyroid hormone is under review by the US Food MEDLINE and Cochrane literature searches retrieved 191 and Drug Administration for use in injection form. It references, 20 of which described randomized con- would be the first bone formation inducer integrated into trolled trials involving PTH and osteoporosis (Table 1). clinical use. This article will discuss evidence for the ef- Outcomes of the same trial were sometimes reported in ficacy and safety of PTH for treatment of osteoporosis. separate articles, and some articles were continuations of results of earlier trials.16-26 The number of subjects per METHODS trial ranged from 935 to 1637.28 Treatment duration ranged from 6 weeks27 to 3 A MEDLINE database (1966 to the present) keyword search years.29 Most studies administered PTH subcutaneously was performed for randomized controlled trials in humans us- 18,19 ing the keywords osteoporosis and parathyroid hormone or para- (SC). Two studies investigated intranasal PTH. Para- thyroid hormone and fracture. The Cochrane database was also thyroid hormone has been combined with alendronate searched using the search terms osteoporosis and parathyroid sodium, hormone replacement therapy (HRT), calcito- hormone. Reference lists of retrieved articles were reviewed for nin, or calcitriol (Table 1). Parathyroid hormone has additional pertinent articles. also been compared with placebo and has been studied

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2298

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 1. Randomized Controlled Trials of Parathyroid Hormone With Bone Density Outcomea (cont)

Trial Subjects Regimenb Calciumc or Vitamin D Leder et al,33 11 Men aged 50-80 y with locally advanced hPTH 22.5 mg IV infusion before leuprolide acetate None mentioned. 2001 node-positive or biochemically recurrent (GnRH agonist) therapy, and again after 6 mo of prostate cancer without bone metastasesu confirmed GnRH-induced hypogonadism Neer et al,34 1993 30 Osteoporotic postmenopausal womenv hPTH–(1-34) 400-500 U/d + calcitriol 0.25 µg/d vs Calcium supplementation calciumw for 12 mo 1000-2000 mg/d. Reeve et al,35 2 Separate trials. Trial 1: 9 women, Trial 2: Trial 1: hPTH–(1-34) 500 U SC qd with HRTy vs HRT None mentioned. 2001 40 women, all with vertebral fractures and alone for 1 y open study. Trial 2: hPTH–(1-34) primary or postmenopausal osteoporosisx 400 U with HRT vs HRT alone randomized for 1 y. HRT continued after hPTH cessation.

aCT indicates salmon calcitonin; hPTH, human parathyroid hormone; HRT, hormone replacement therapy; SC, subcutaneously; bid, twice daily; qd, every day; and GnRH, gonadotropin-releasing hormone. bAll treatments are oral unless otherwise noted. cTo convert serum calcium from micrograms per deciliter to millimoles per liter, multiply micrograms per deciliter by 0.25. dDue to substantial nausea and flushing, the 100 U daily used during the first cycle was lowered to 75 U/d for the second cycle. eNeither subjects nor investigators were blinded to treatment, but readers of bone density tests were blinded. f Twenty five of the subjects were continuing the 6-month 1994 trial by the same authors (Finkelstein et al18). The remainder of the patients from the 1994 trial completed their therapy before the trial was extended to 12 months. gOsteoporosis was based on radiologic evidence of Ն1 vertebral compression fracture. hContinuation of the 1997 trial by the same authors (Hodsman et al20). i T score Ͻ−2.5 at lumbar spine or femoral neck, menopausal Ն3 years, taking hormone replacement therapy for Ն1 year, prednisone 5.0-20 mg/d or its equivalent average daily dose for prior 12 months. Excluded secondary osteoporosis other than rheumatic diseases and glucocorticoid use, and renal or hepatic dysfunction. Details of type and dose of hormone replacement not given. jContinuation of the 2001 study by Lindsay et al.29 khPTH dosage is equivalent to 25 µg/d, but HRT regimen not specified. lNinety percent of subjects had history of osteoporotic fractures at the spine, hip, forearm, or other site, and all were diagnosed as having osteoporosis, ie, had T score Ͻ−2.5 at the spine and/or hip. Women were already taking alendronate, 10 mg/d, for at least 4 months at entry. Mean (± SD) age for those in the hPTH plus alendronate group was 69.0 ± 1.8 years; alendronate alone, 64.0 ± 4.8 years. mWomen with Ͻ2 moderate fractures had to have hip or lumbar BMD T score of Ͻ−1. nMean age of hPTH group was 59.5 years, estrogen group, 64.2 years. oOsteoporosis was defined as TϽ−2.5 or atraumatic fractures. pConjugated equine estrogen 0.625 mg/d, or transdermal estradiol, 50 µg/d. Women with a uterus also received medroxyprogesterone acetate, 5-10 mg/d, for Ն10 d/mo or 2.5 mg/d continuously. Patients had taken hormone replacement for Ͼ1 year before study entry. qOsteoporosis was defined by a radiographic scoring system. Mean lumbar BMD was 0.694 g/cm2. rCorresponding amounts of the peptide were 15, 30, and 60 µg, respectively. sWomen were at least 5 years postmenopausal with vertebral T score Ͻ−2.5. tZ score Ͻ−2.0 or T score Ͻ−2.5 at lumbar spine or femoral neck, using male reference database. Seventy-eight percent had sustained fractures of spine, hip, or appendicular skeleton, and 22% had presented with back pain. Hypogonadism was an exclusion criterion, but 2 men in the placebo group were taking stable doses of testosterone replacement by injection. uAll subjects had normal testosterone at baseline and none had received prior androgen deprivation therapy. vOsteoporosis was not defined in detail. Patients were stratified by age and spinal trabecular bone density. wCalcium given as carbonate, 1000 to 2000 mg/d. Follow-up duration was at least 12 months, total duration not detailed. xStudy 1: 10 women and 1 man with Ն2 vertebral fractures with primary or postmenopausal osteoporosis, and 1 woman with 1 vertebral fracture and low bone density. Study 2: subject characteristics not described. yHRT consisted of conjugated equine estrogen, 0.625 mg/d, with norethindrone acetate, 1 mg/d, for days 19-28.

for differing dose effects. The efficacy of PTH has been U/wk SC for 48 weeks. Corresponding amounts of the pep- tested for prevention of gonadotropin-releasing hor- tide were 15, 30, and 50 µg, respectively. Increases in lum- mone (GnRH)–associated bone loss in both sexes. bar bone mineral density (BMD) were dose-related (range, Mechanisms of the other medications are briefly sum- 0.6%-8.1%), but there were no changes at the femoral neck marized in Table 3. with any PTH dosage. The second trial31 compared 50, 75, Results are discussed in this section according to dos- or 100 µg/d SC for 1 year, followed by alendronate alone age and duration of therapy, study size, age, sex, baseline for 1 year. Although all dosages increased spine and fem- disease, anatomical site, histomorphometric effects, bone oral neck BMD compared with placebo, specific dose- marker changes, fracture outcome, adverse effects, and com- effect analysis was not presented. Increases in bone den- pliance. (Single-dose studies are not discussed.) sity ranged from 4.3% to 9.2% at the lumbar spine. Although there was no change at the femoral neck with EFFECT BY DOSAGE, SCHEDULE, PTH compared with placebo, the group treated with PTH AND DURATION OF THERAPY followed by alendronate had better femoral neck bone density than the group treated with placebo followed by alen- Dosages of PTH ranged markedly across trials. They were dronate. The third trial28 compared 20 or 40 µg/d SC with variously reported in micrograms or units per day in the placebo for 16 to 17 months. Both dosages increased spine different articles. Furthermore, because of different com- and hip (total hip, femoral neck, and trochanter) BMD combinations across trials and different comparison groups, pared with placebo, but dose-effect analysis per se was not conclusions regarding dose effects are difficult. Three tri- provided. Bone density increases with 20-µg dosaging were als28,30,31 directly compared different dosages of PTH with 9.7%, 2.6%, 2.8%, and 3.5% at the lumbar spine, total hip, each other. The first30 of the 3 compared 50, 100, and 200 femoral neck, and femoral trochanter, respectively. Cor-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2299

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 2. Bone Mineral Density Response to Parathyroid Hormone*

Source Lumbar Total Hip Femoral Neck Trochanter 1/3† Radius Hodsman et ↑ 13% vs baseline (PϽ.01) ...... ↓ by 11% vs baseline al,17 1991 with PTH alone vs no (PϽ.05) with PTH change with PTH followed alone vs no change by CT, but no difference in with PTH followed change between groups, by CT, but no by DPA difference in change between groups, used absorptiometry Finkelstein et ↑ 2.1% PϽ.001 vs baseline . . . ↓ slightly during first 6 mo, ↓ slightly during first No significant change al,19 1998, then returned to baseline 6 mo, then returned in either group, Cosman et after 1 y, PϽ.001 nafarelin to baseline after P not significant, al,27 1998§ acetate+PTH vs nafarelin 1y,P = .04 nafarelin+PTH nafarelin+PTH vs nafarelin vs nafarelin Hodsman et ↑ 10.2% with PTH vs 7.8% . . . No significant difference ...... al,20 1997 with PTH+CT, no compared with baseline or significant difference between groups with PTH between groups or PTH+CT Lane et al,22 ↑ in PTH+estrogen group No significant No significant difference No significant No significant 1998 9.8% vs estrogen group difference in in change between difference in difference in (no change), PϽ.001 change between PTH+estrogen and change between change between PTH+estrogen and estrogen alone PTH+estrogen and PTH+estrogen and estrogen alone estrogen alone estrogen alone Lane et al,24 12 mo after cessation of 3.4% ↑ with Nonsignificant ↑ vs baseline Nonsignificant ↑ Nonsignificant ↓ in 2000 PTH ↑ was 11.9% with PTH+estrogen vs with PTH+estrogen, vs baseline with both groups 1.5%, estrogen+PTH vs estrogen estrogen alone, no significant difference PTH+estrogen, no significant alone, PϽ.001 between PϽ.01, 12 mo after between groups 12 mo no significant difference between groups cessation of PTH after cessation of PTH difference between groups 12 mo after groups12 mo after cessation of PTH cessation of PTH Cosman et ↑13.4±1.4% with PTH+HRT, 4.4±1.0% with ...... al,25 2001 vs no significant change PTH+HRT,vsno with HRT‡ significant change with HRT Neer et al,28 ↑ 9.7±7.4% with 20 µg and ↑ 2.6±4.9% with 20 ↑ 2.8±5.7% with 20 µg and ↑ 3.5±6.8% with 20 ↓ 2.1±4.2% with 20 µg 2001 13.7±9.7% with 40 µg, µg and 3.6±5.4% 5.1±6.7% with 40 µg, µg and 4.4±7.5% (P = .09 vs placebo) PϽ.001 vs placebo with 40 µg, PϽ.001 vs placebo with 40 µg, PϽ.001 and 3.2±4.5% with for both PϽ.001 vs placebo for both vs placebo for both 40 µg (PϽ.001) for both vs placebo) Lindsay et ↑ 13% with PTH+estrogen, ↑ 2.7% with ... ↑ 2.6% with ↑ by 1% with al,29 1997 PϽ.02 PTH+estrogen vs PTH+estrogen, PTH+estrogen, PTH+estrogen vs estrogen alone࿣ PϽ.001 P = 0.11 estrogen alone PTH+estrogen vs PTH+estrogen vs estrogen alone estrogen alone

(continued )

responding values for the 40-µg/d dosage were 13.7%, 3.6%, ing intranasal PTH. The trial was designed with the goal 5.1%, and 4.4%. Both dosages were associated with of preventing bone loss induced by GnRH agonist therapy. decreases in radius BMD (−2.1% and −3.2% for 20- and Therefore, the intranasal route of PTH has not been tested 40-µg/d dosages, respectively) compared with placebo in the context of glucocorticoid-induced osteoporosis or (Table 2). postmenopausal osteoporosis, nor has it been tested in Therefore, on the basis of the little existing data, it men. The trial found that intranasal PTH for 12 months is possible that dose-related effects exist throughout the increased lumbar BMD 2.1% and protected against fem- dosage range of 50 to 100 µg/d SC. More dose-effect in- oral neck and trochanter bone loss. Neither nafarelin ac- formation is needed. etate nor PTH plus nafarelin affected radius BMD. Because almost all trials used human PTH–(1-34) Therefore, although a few studies support the in- (hPTH–[1-34]), conclusions about any relative advan- tranasal administration of PTH to prevent bone loss in tage of hPTH–(1-38) or hPTH–(1-84) are not possible. women receiving specific endometriosis therapies, most The 2 trials (described in 3 articles) that did not involve data available refer to the SC route. hPTH–(1-34) were small and involved postmenopausal Study duration varied substantially across trials, rang- osteoporosis16,17,31 or idiopathic osteoporosis.16,17 ing from 6 weeks27 to 3 years29; the longest trial involv- All trials except 2 involved administration by the SC ing SC administration was 3 years (Table 1). The more route. The first exception was 1 trial (2 articles18,19) test- recent trials had access to more advanced bone density

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2300

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 2. Bone Mineral Density Response to Parathyroid Hormone* (cont)

Source Lumbar Total Hip Femoral Neck Trochanter 1/3† Radius Fujita et al,30 ↑ 0.6%, 3.6%, and 8.1% with 50, ...... 1999 100, and 200 U, respectively, PϽ.05 for each of l00 U vs 50 U,200Uvs50U,200Uvs 100 U Rittmaster et During PTH alone for 1 y ↑ of . . . During PTH alone for 1 y ...... al,31 2000 1.3%, 4.3%, 6.9%, or 9.2% no change from baseline with placebo, 50, 75, or 100 in any group and no µg, respectively, no significant difference between further change during the groups. Over 2 years, second year (alendronate ↑ in PTH + alendronate sodium only). Over 2 years, ↑ vs placebo + alendronate in PTH-alendronate vs P = .09, P = .009, P = .02, placebo-alendronate P = .003, and P = .12 vs baseline PϽ.001, and P = .08 for for placebo, 50, 75, 100-, 75-, and 50-µg groups and 100 µg vs placebo Kurland et ↑ 13.5±3% vs no change in No significant ↑ 2.9±1.5%, PϽ.05 . . . Nonsignificant ↓ with al,32 2000 untreated controls, PϽ.001 change in between groups PTH vs increase in between groups either PTH untreated controls, or untreated PϽ.05 between control group groups Neer et al,34 ↑ 12% by DPA with PTH, ...... 1993 no change with calcium¶ Reeve et al,35 ↑ more with PHT+HRT (12.2%) . . . No significant difference No significant By pQCT, ↑ at distal 2001 than with HRT alone (4.8%), between ↑ with difference between radius 1 y after P = .03 PTH+HRT (1.5%) vs HRT ↑ with PTH+HRT treatment was 4.8% alone (−0.2%), P = .70 (2.0%) vs HRT for PTH+HRT vs alone (2.4%), 1.9% for HRT alone, P = .66 P = .70

*PTH indicates parathyroid hormone; HRT, hormone replacement therapy; CT, salmon calcitonin; DPA, dual-photon absorptiometry; and pQCT, peripheral quantitative computed tomography. †Bone mineral density by dual x-ray absorptiometry (DEXA), unless otherwise noted. ‡The study documented that continuation of HRT for an additional year maintained the increases in bone density observed with PTH/HRT combination therapy. §The earlier trial by the same authors is not included because the 1998 trial is the continuation. ࿣During 1-year observation on HRT alone, after combination therapy, mineral insignificant decreases in spine and total hip BMD were observed.25 ¶Cortical bone mass decreased 1.7% with PTH plus calcitriol vs 5.7% with calcium alone. Trabecular bone mass increased 32% with PTH plus calcitriol but did not change with calcium alone.

measurement technology, such as dual-energy x-ray absorptiometry, and measured a greater number of sites than Table 3. Mechanisms of Action of Medications Included in Osteoporosis Trials of Parathyroid Hormone did the older studies (Table 2). Correspondingly, results are sometimes expressed in the context of differ- Inhibitors of bone resorption ent bone density measurement techniques. Alendronate sodium First, the longer (1-year and 3-year) trials will be de- Calcitonin scribed. A 1-year trial24 reported a greater increase in spine Estrogen replacement therapy (11.9% higher) and total hip (3.4% higher) BMD 12 months Mechanisms incompletely understood and necessary for optimal after treatment with PTH plus estrogen compared with es- intestinal calcium absorption trogen alone. However, changes were not significant with Calcitriol Gonadotropin releasing hormone agonists (GnRH agonists) inducing PTH plus estrogen or estrogen alone at the femoral neck, hypogonadism trochanter, or radius. The estrogen preparation and dos- Leuprolide acetate age were not specified. One 18-month trial32 revealed in- Nafarelin acetate creases in lumbar (13.5% higher; PϽ.001) and femoral neck (2.9% higher; PϽ.05) BMD in subjects receiving PTH vs untreated control subjects (Table 2). However, total hip BMD did not change with PTH treatment, and radius BMD nate administration. Increases were less marked at the decreased insignificantly with PTH, compared with an in- femoral neck (Table 2). Subjects received PTH or pla- crease in the untreated controls (PϽ.05 between groups). cebo for 1 year, followed by a second year of open-label (The possible decrease in radius BMD emphasizes the im- alendronate, 10 mg/d. portance of inclusion of control groups in trials of osteo- The single 3-year study,29 which compared PTH plus porosis therapies.) In the other 18-month trial,31 lumbar HRT with PTH alone, showed that combination therapy BMD increased (range, 4.3%-9.2%, depending on the dos- with PTH plus HRT induced higher increases in spine and age) with 1 year of PTH administration compared with pla- total hip BMD (13% and 2.7%, respectively) compared with cebo, but did not change in the following year of alendro- HRT alone. Differences between groups were less marked

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2301

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 4. Effect of Parathyroid Hormone Administration on Fractures*

Source Fracture Outcome Hodsman et al,20 1997 No hip or other appendicular fracture. Vertebral fracture incidence no different between PTH+CT and PTH alone. Very low incidence of radiographically detected vertebral fracture rates over 2 y. Fewer fractures with PTH group vs PTH+CT group, but small sample size yielded no difference between groups (P = .08). No nonspinal fractures. Lane et al,22 1998 No patient in PTH group and 1 patient in estrogen-only group had new vertebral fractures. Nonvertebral fractures occurred in 2 patients in each group. Cosman et al,25 2001 During 3 y, radiographically detected vertebral fractures occurred less often with PTH+HRT vs HRT alone (P = .001). Combination therapy decreased the percentage of women with incident vertebral fractures from 37.5% vs 8.3% (PϽ.02). Radiographs not routinely performed 1 y after therapy, but no clinical fractures occurred. Neer et al,28 2001 20-µg and 40-µg doses decreased risk of Ն1 new vertebral fracture by 65% and 69%, respectively. Risk of Ն2 fractures was reduced by 77% and 86%, respectively, and risk of Ն1 moderate or severe vertebral fracture was reduced by 90% and 78% respectively. PTH also decreased the total number of vertebral fractures. The 20-µg and 40-µg groups were 35% and 40% respectively less likely to have Ն1 new nonvertebral fracture compared with the placebo group, and 53% and 54% respectively, less likely to have Ն1 new nonvertebral fragility fracture.† Lindsay et al,29 1997 Fewer vertebral fractures with estrogen+PTH vs estrogen alone, P = .09. Number of women with vertebral fractures not significantly different between groups, P = .12. Fujita et al,30 1999 Radiographically detected spinal fractures in 3 subjects in 50-U, 5 subjects in 100-U, and 0 in 200-U dosing. Difference between groups not significant. Kurland et al,32 2000 Numbers of radiographically detected vertebral fractures too small to establish significance. Reeve et al,35 2001 P = .28 for number of patients having new vertebral deformity for HRT vs HRT+PTH.‡

*PTH indicates parathyroid hormone; HRT, hormone replacement therapy; and CT, salmon calcitonin. †Absolute risk of Ն1 nonvertebral fractures was 10% with placebo vs 6% with PTH. This translated to a relative risk of 0.47. The numbers of women with new nonvertebral fractures were too small to estimate the incidence of each type of fracture. ‡Study was not powered to measure effect of treatment on vertebral fractures, although new vertebral fractures were observed radiologically.

at the trochanter and radius. One year after cessation of PTH, radius BMD. In contrast, although smaller studies re- with continuation of HRT alone, gains in BMD were main- ported positive hip and lumbar effects, they generally sug- tained.25 The specific hormone regimen used in this trial gested only minor bone loss at the radius (Table 2). was conjugated equine estrogen, 0.625 mg/d orally, or estradiol, 50 µg/d transdermally, combined with sequential EFFECT BY AGE, SEX, AND BASELINE DISEASE or continuous medroxyprogesterone acetate. In contrast, the 1-year trial24 already described did not specify exact hor- Only one study30 was designed to formally compare bone mone dosage and preparation. density changes according to age and sex. Parathyroid A briefer duration of therapy also results in in- hormone benefits on bone density were similar with sub- creased BMD30 (Table 2). Therefore, overall effects of 48 jects 65 years and older vs younger than 65, at a weight weeks to 3 years of therapy with PTH appear to be more of at least 50 kg vs less than 50 kg, whether baseline ver- marked at the spine than at other sites, may be detri- tebral fracture was present or not, and regardless of time mental at the radius, may be enhanced by combination since menopause. with estrogen, and may be advantageous when com- Most studies pertained to older men and women with bined sequentially with alendronate vs alendronate alone. established osteoporosis, with 2 exceptions.19,33 In the first The longest duration of therapy involved 3 years of PTH study,19 PTH was added in an effort to prevent bone loss combined with HRT and showed benefit at the spine and induced by GnRH agonists in young women with endo- total hip. It may take 12 months after treatment cessa- metriosis (Table 1). Although lumbar BMD was in- tion for the maximal anabolic effect of PTH to manifest creased compared with baseline in subjects receiving PTH at the hip.24 (Combinations of PTH and other medica- along with the usual GnRH agonist treatment, femoral tions are the focus of the “PTH in Combination With neck and trochanter BMD decreased with PTH and ra- Other Agents” subsection of the “Results” section.) dius BMD did not change (with or without inclusion of PTH) compared with baseline. The second study33 EFFECT BY STUDY SIZE investigated bone marker changes in older men with advanced prostate cancer. Augmentation of bone resorp- The largest studies involved 22030 and 1637 patients.28 tion associated with GnRH agonist–induced hypogonad- Both trials made use of several different dosages of PTH. ism was blunted by PTH administration. All other stud- The first30 found that each of the dosages tested (15, 30, ies included older men and women already affected by and 60 µg/d SC) increased spine BMD during 48 weeks osteoporosis (Table 1). and that the increases were dose-related. The other study28 Specifically, osteoporosis was postmenopausal os- described BMD increases at the spine, femoral neck, total teoporosis,16,17,20,21,27-31,34,35 idiopathic male osteoporo- hip, and trochanter, but decreases at the radius, with dos- sis,16,17,32 or glucocorticoid-induced osteoporosis.22-24 These ages of 20 and 40 µg/d SC vs placebo. Therefore, results 3 clinical scenarios are individually reviewed herein. of larger trials were conflicting as to whether effects were Postmenopausal osteoporosis is the setting of most limited to the spine and suggested detrimental effects on studies. In this setting, PTH consistently increased lum-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2302

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 5. Adverse Effects of Parathyroid Hormone Therapy*

Source Histomorphometric Effects Hypercalcemia Other Hodsman and ...... Calciuresis (persistent) Fraher,16 1990 Finkelstein et al,18 . . . Calcium rose but remained within normal Hypercalciuria in 2 women, resolved with reduced 1994 limits calcium intake Finkelstein et al,19 . . . Calcium increased with nafarelin vs None serious. Mild nausea and arthralgias reported 1998, Kurland nafarelin+hPTH, and calcium remained more commonly with hPTH than without, no et al,32 2000† in normal range in both groups difference between groups for other AES. Local discomfort or erythema at injection site. Hodsman et al,20 . . . No significant change from baseline over Local inflammatory reactions at injection sites. 1997 2 y with PTH treatment§ Three patients withdrew for failure to learn injection technique. Increased urinary calcium but by study end no different from baseline. Increase in serum creatinine at month 4, still in normal range. Hodsman et al,21 Cyclical PTH increased trabecular Rose initially but not at 2 y . . . 2000 bone turnover and induced positive remodeling balance without detrimental change to cortical bone Lane et al,22 1998 . . . Increased with PTH but stayed in normal Two subjects developed hypercalciuria that range. Three women developed returned to normal when calcium intake was hypercalcemia after 3 mo of PTH. reduced. Transient mild headaches and injection Responded to decreased calcium intake site tenderness with PTH injection. and reduced PTH dosage. Lane et al,24 2000 . . . Increased slightly but remained in normal ... range Cosman et al,25 . . . None observed No nausea. High incidence of injection site 2001 reactions, including subcutaneous nodules at injection site in 2 subjects.

(continued )

bar BMD. This was observed when PTH was given alone data specifically analyzed the effects of PTH in older vs or with sequential calcitonin, HRT, calcitriol, or alen- younger subjects or directly compared effects according dronate. Effects were less consistently seen at the total to sex. hip, femoral neck, trochanter, and radius (Table 2). Sec- ond, glucocorticoid-induced osteoporosis was studied in EFFECT BY 1 trial.22 Combined PTH and HRT had no advantage over ANATOMICAL SITE estrogen alone at the hip or radius. Furthermore, the benefit of PTH at the lumbar spine was only apparent after Parathyroid hormone increased lumbar spine BMD in all 1 year, and the benefit at the total hip was only apparent studies, at several dosages, for any duration, in different 1 year after cessation of PTH. The benefit at the femoral clinical situations, and in combination with multiple neck or trochanter was not statistically significant. Ef- agents (Table 2). However, results at the different hip sub- fects of combination therapy on the radius showed an sites were conflicting. Femoral neck bone loss induced insignificant decrease compared with estrogen alone.22,24 by nafarelin was only partly prevented by PTH.19 Para- Last, idiopathic osteoporosis in men was investigated in thyroid hormone was protective against bone loss at the 1 trial.17 Interestingly, increased lumbar spine BMD as- radius. In other settings, no changes in femoral neck BMD sociated with PTH was counteracted by the addition of were attributable to PTH therapy.20,22,24,30,31,35 In other stud- calcitonin, although calcitonin may have protected against ies,28,32 femoral neck BMD increased with PTH therapy. PTH-induced bone loss at the radius. Bone turnover ef- Femoral trochanter BMD loss caused by nafarelin fects of PTH in idiopathic male osteoporosis are dis- was not prevented by PTH, and significant BMD changes cussed in the “Effects on Bone Formation Markers” sub- at this site were not found in several trials.22,24,35 Fem- section of the “Results” section. oral trochanter BMD increased in other PTH trials.28,29 In summary, increases in lumbar BMD are induced Therefore, PTH effects at the hip vary across studies and by PTH (alone and possibly combined with other medi- clinical diseases (see the “Effect by Age, Sex, and Base- cations, as shown in the “PTH in Combination With Other line Disease” subsection of the “Results”section). Agents” subsection of the “Results” section) in post- A single study28 documented increases at all of the menopausal osteoporosis, glucocorticoid-induced osteo- multiple sites (ie, spine and all hip sites). It was also the porosis, and idiopathic osteoporosis. Effects at other sites only trial comparing the effect of different dosages of PTH are insignificant or conflicting. Parathyroid hormone may vs placebo at multiple sites for treatment of postmeno- protect against GnRH agonist–related bone loss. Al- pausal osteoporosis. This study was the largest, most re- though BMD end points were beneficially affected by PTH cent, well-designed trial and was the only trial compar- in all studies, with some variability between sites, little ing different dosages of PTH with placebo, as opposed

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2303

Downloaded From: https://jamanetwork.com/ on 09/27/2021 Table 5. Adverse Effects of Parathyroid Hormone Therapy* (cont)

Source Histiomorphometric Effects Hypercalcemia Other Dempster et al,26 Cancellous bone area was ...... 2001 maintained in women and increased in men P = .07. Wall width of trabecular packets was maintained in women and significantly increased in men. Cortical width increased slightly in women and significantly in men, PϽ.02. Most patients had increased trabecular connectivity by 3-dimensional scan. Neer et al,28 2001 . . . Mild hypercalcemia (Ͼ10.6 mg/dL) in 2%, No significant differences among the 3 groups in 11%, and 28% of placebo, 20 µg, and serious AEs. Cancer incidence was higher in the 40-µg PTH groups, respectively. Of the placebo than PTH groups. More withdrawals high serum calcium values, 95% were with PTH than placebo (11% for 40 µg vs 6% Ͻ11.2 mg/dL. Only one third had each for placebo and 20 µg). Nausea and persistence of hypercalcemia on headache were more common with PTH confirmatory retesting. Women who did 40 µg than placebo (PϽ.001 and P = .01, not have hypercalcemia during the first respectively), but incidence was similar with 6 months of treatment rarely developed 20 µg compared with placebo. hypercalcemia later on. Lindsay et al,29 . . . No significant change Pain and redness at injection site. One PTH 1997 withdrawal because of back pain, another because of a subcutaneous nodule at the injection site. No medication-related systemic effects, no serious AEs. No hypercalciuria. Fujita et al,30 1999 . . . Calcium decreased vs baseline with all None serious. AEs increased with increasing dosages of PTH. Remained Ն9 mg/dL in dosage of PTH. Dropout of 3, 10, and 16 all dosage groups. patients in 50-U, 100-U, and 200-U groups because of AEs; AEs in 19%, 19%, and 19%, respectively. Urinary calcium fell below basal level.‡ Kurland et al,32 . . . No significant change. Two patients had No difference from baseline or between groups 2000 hypercalcemia (Ͼ10.5 mg/dL) in PTH in urinary calcium. Redness at injection site. group, leading to dose titration. Those No significant differences between groups in patients still had increase in spinal BMD. systemic, musculoskeletal, or neurological reactions. Leder et al,33 2001 . . . No significant change. (Checked just before ... infusion at 6 mo.) (P = .24). Reeve et al,35 2001 ...... No significant difference from baseline in change in urinary calcium, but difference between groups was significant (P = .03).

*PTH indicates parathyroid hormone; hPTH, human PTH; AE, adverse effect; and BMD, bone mineral density. To convert serum calcium from micrograms per deciliter to millimoles per liter; multiply micrograms per deciliter by 0.25. †The earlier trials by the same authors are not included because the 1998 trial is the continuation. ‡Reactions to PTH included subcutaneous hemorrhage, whole-body reddening, facial flush, eczema, itching, lumbago, headache, dizziness, nausea, vomiting, abdominal pain, belching, yawning, thirst, anorexia, febrile sensation, febrile episode, sensation of weakness, general malaise, chills, and sleepiness. No comment regarding significant difference between groups or compared with placebo, nor whether reactions were thought to be medication-related. §PTH and PTH + CT groups analyzed together.

to use in sequential or combination regimens. This trial HISTOMORPHOMETRIC EFFECTS confirmed the detrimental effect of PTH on radius BMD suggested in other studies17,32 of postmenopausal osteo- Bone quality must be assessed by histomorphometry to porosis. However, 1 trial29 reported a small BMD in- prevent unintentional augmentation in fracture risk, even crease with PTH plus estrogen compared with estrogen if BMD increases occur. As an example, fluoride in- alone at the radius. Because of the paucity of data, there creased fracture risk, despite inducing impressive in- is as yet no proof that the decrement in radius BMD trans- creases in BMD.11-13 Histomorphometric findings have lates into increased radius fractures. been used as an indicator of bone quality. In summary, when used for postmenopausal osteo- Histomorphometric consequences of PTH use were porosis treatment, PTH compared with placebo in- investigated in 1 trial21 (Table 5). In women with post- creases BMD at the spine and at multiple sites of the hip. menopausal osteoporosis, cyclical PTH increased trabec- Effects are less clear at nonspine sites when PTH is used ular bone turnover and induced positive remodeling bal- as part of combination or sequential therapy or for treat- ance, without detrimentally affecting cortical bone.21 The ment of glucocorticoid-induced osteoporosis. improvement of cortical and cancellous microarchitec-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2304

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ture with daily PTH administrations has been reported Results of combination therapy trials confirm what in humans.26 would be predicted according to medication mecha- In a sex-comparative analysis after 3 years of com- nism. Neither alendronate nor HRT increased forma- bined PTH and HRT, cancellous bone area was main- tion markers when administered alone.25,27 Rather, only tained in women and increased in men26 (Table 5). Wall combination PTH with alendronate or HRT increased for- width of trabecular packets was maintained in women mation markers. Parathyroid hormone–induced in- and significantly increased in men. Cortical width increased bone resorption and formation peaked at 6 creases slightly in women and significantly in men. Most months.25 They returned to baseline by 30 months. Sub- patients had increased trabecular connectivity by three- sequently, when patients were maintained on HRT alone dimensional scan. for another year, no significant changes in turnover markers occurred.25 EFFECTS ON It may be possible to make use of serial bone marker BONE FORMATION MARKERS measurements to identify skeletal responders to anabolic therapy in estrogen-replete women with glucocor- Bone turnover markers are sometimes used as surrogates ticoid-induced osteoporosis.23 One study23 suggests that to assess therapeutic effects of osteoporosis medications, the response of bone markers predicted a gain in BMD although whether such application is clinically relevant resulting from PTH therapy with high diagnostic accu- is a matter of active controversy. Bone markers can give racy, although it did not predict the magnitude of the gain. useful insights into medication mechanisms. Markers of In summary, PTH increases bone formation mark- bone formation include serum alkaline phosphatase, se- ers. The increase in bone formation is probably larger in rum osteocalcin, and C-terminal propeptide of type I pro- magnitude and may have a different time of onset com- collagen. Parathyroid hormone would be expected to in- pared with the increase in resorption. crease alkaline phosphatase on the basis of its mechanism as an inducer of bone formation. Correspondingly, in some ANTIFRACTURE EFFICACY trials,18-20,22,25,30,31 PTH increased serum alkaline phosphatase, although 1 brief trial27 reported no increase during Antifracture efficacy is the single most critical efficacy 6 weeks. Some investigators found an increase with outcome for osteoporosis treatment. As it is difficult to PTH administered alone, but not when followed by anti- include adequate numbers of subjects to yield substan- resorptive treatment with calcitonin.16,17 Similarly, some tial numbers of fracture events, fracture outcome data trials18-20,22,27,29,31,32,35 reported an increase in serum osteo- for PTH are sparse. Although not a satisfactory state of calcin resulting from PTH administration. One trial16 found affairs, this has historically been the case with other os- an increase in serum osteocalcin with PTH alone and PTH teoporosis therapies as well. Often, fractures are only followed by calcitonin treatment. Two studies27,32 re- monitored as part of adverse outcome assessment, as ported an increase in serum C-terminal propeptide of type opposed to constituting a primary outcome. Table 4 I procollagen with PTH plus alendronate vs alendronate summarizes fracture outcome data for PTH. Rates of ra- alone, or with PTH vs placebo. diographically detected fractures were generally low. Fracture rates were sometimes so low as to prevent sta- EFFECTS ON tistical analysis.20,22,32,35 BONE RESORPTION MARKERS Rates of radiographically detected spinal fractures were not different among different doses of PTH over a Markers of bone resorption that increase during PTH 50- to 200-U dose range.30 The incidence of radiographi- therapy include urinary pyridinoline,18,30,32 deoxypyridin- cally detected vertebral fractures was no different with oline,19,22,30 and N-telopeptide20,29,31,32 excretion. sequential PTH plus calcitonin compared with PTH alone, Induction of hypogonadism in older men may cause although differences between groups were not signifi- changes in skeletal sensitivity to PTH.33 The increase in cant and the number of events was small.20 When HRT urinary N-telopeptide excretion (but not deoxypyridi- was compared with HRT plus PTH, differences between noline) during PTH infusion was greater after 6 months groups in vertebral fracture incidence were not signifi- of leuprolide acetate therapy than before induction of hy- cant.35 The vertebral fracture incidence was lower with pogonadism caused by leuprolide. PTH plus estrogen vs estrogen alone, although differ- Data regarding urinary hydroxyproline are conflict- ences between groups were not significant.29 Rates of non- ing. A decrease was reported in 1 trial.30 Other trials18-20 vertebral fracture were lower, making analysis of re- reported increases in urinary hydroxyproline in women duced fracture rates difficult or impossible20,22 (Table 4). with endometriosis. One trial16 found no change from The largest trial28 reported an impressive reduction baseline with PTH therapy alone, although levels at the in the rate of new vertebral fractures with PTH dosages study end were lower after sequential PTH plus calcito- of 20 and 40 µg/d SC for 17 months. The risk of 2 or more nin vs PTH alone. vertebral fractures was reduced, as was the risk of at least Some data suggest that the increase in bone forma- 1 moderate or severe vertebral fracture, regardless of which tion and resorption is disproportionate, suggesting that PTH dosage was used. Nonvertebral fractures were also PTH could be uncoupling bone formation and resorp- decreased by PTH. Parathyroid hormone recipients were tion.22 This idea may be supported by the observation that 35% to 40% less likely to have at least 1 new nonverte- PTH plus HRT (compared with HRT alone) did not in- bral fracture compared with placebo and 53% to 54% less crease resorption as much as formation.29 likely to have at least 1 new nonvertebral fragility frac-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2305

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ture.28 The numbers of nonvertebral fractures were too groups. The longest (3-year) trial25 reported that nausea low to be broken down by individual site. It could not did not occur. be concluded from the study if the notable decrease in Hypercalciuria occurred infrequently in some tri- radius BMD in this trial may have led to increased frac- als,16,18,22 but by the study end returned to baseline in 1 ture rates. study.20 Generally, hypercalciuria resolved with a de- These findings are replicated in the longest and most crease in calcium intake.22 Other studies29,32,35 reported recent trial.25 Radiographically detected vertebral frac- no change from baseline in urinary calcium levels. One tures were significantly decreased with PTH plus HRT study30 reported that urinary calcium fell below the basal during 3 years compared with HRT alone. The percent- level. age of women with incident vertebral fractures de- There was sometimes difficulty with PTH injection creased from 37.5% to 8.3% by PTH plus HRT com- technique20 or compliance30 (Table 5). The largest trial28 pared with HRT alone. Again, the rate of nonvertebral found more subject withdrawal from the study with PTH fractures was low. than with placebo treatment. In summary, when data were robust enough to al- Rats that were given nearly lifetime daily injections low adequate statistical power, PTH decreased the inci- of PTH manifested increased osteosarcoma incidence in dence of radiographically detected spinal fractures, de- a dose-dependent fashion as a result of PTH administra- creased the incidence of nonvertebral fractures, and may tion.28 However, this effect is absent in primates (eg, mon- have added to the fracture reduction ability of estrogen key models). Furthermore, chronic hyperparathyroid- replacement therapy. ism is not associated with osteosarcoma in humans.28 None of the trials reviewed herein reported an increased cancer ADVERSE EFFECTS AND COMPLIANCE incidence associated with PTH, and the cancer incidence was higher with placebo vs PTH therapy.28 In summary, Adverse effects of PTH are summarized in Table 5. No PTH administration is associated with hypercalcemia in serious medication-related adverse effects were attrib- a small percentage of patients, possibly in a dose- uted to PTH. In some trials,19,22 PTH increased calcium dependent fashion, and early in treatment. Other minor levels, but they remained in the normal range. One trial30 adverse effects are also associated with PTH. No pub- reported minor decreases in calcium levels with PTH lished trial results have indicated that cancer risk is in- therapy. Other trials20,29,32,33 reported no change in cal- creased by PTH use in humans. cium levels. The largest trial28 reported mild hypercalcemia (Ͼ10.6 mg/dL [Ͼ2.65 mmol/L]) in 2%, 11%, and PTH IN COMBINATION 28% of the placebo, 20-µg, and 40-µg PTH groups, re- WITH OTHER AGENTS spectively. Of the high serum calcium values, 95% were less than 11.2 mg/dL (2.80 mmol/L). A 28-day “pulse” of daily hPTH–(1-34) injections at a Monitoring of serum calcium levels varied across tri- dosage sufficient to elevate serum calcium levels to the als. In some trials, subjects were monitored every 3 upper normal reference range was proposed as a way of months22,25 or every 6 months,29 whereas other sub- activating bone turnover and osteoblastic synthesis of new jects20,21,24,28,30,32 were monitored more frequently at the bone matrix.21 The underlying hypothesis is that, once beginning and less frequently later on. Women who did bone turnover had been activated, the anabolic phase of not manifest hypercalcemia during the first 6 months of bone remodeling would be left to complete itself during treatment rarely developed hypercalcemia later on.28 The the remainder of the 90-day treatment cycle before ini- trial allowed adjustment of calcium intake in patients ex- tiation of the next cycle. Calcitonin given sequentially periencing hypercalcemia. Examples of protocol adjust- after the PTH pulse would be designed to limit the de- ments triggered by hypercalcemia include a decrease in gree of PTH-induced bone resorption, thereby enhanc- PTH dosage for calcium levels greater than 10.5 mg/dL ing the net anabolic effect of PTH. (2.63 mmol/L), with subsequent documentation of nor- The idea of pairing PTH with calcitonin in such a malization of calcium levels,18,19 or a decrease in medi- manner did not yield the anticipated beneficial effects, cation dosage only for persistence of hypercalcemia, de- in that the 2 groups (PTH alone vs PTH with calcitonin) spite a decrease in dietary intake.28 In the latter trial, the had the same incremental gain in vertebral BMD at the numbers of subjects withdrawn for hypercalcemia were end of a 2-year trial.20 The corresponding histomorpho- 1 in the placebo group, 1 in the 20-µg PTH group, and 9 metric analyses for all the 29 treated patients vs a sepa- in the 40-µg group. rate group of biopsies from control patients with un- Besides hypercalcemia, other adverse effects asso- treated osteoporosis revealed an increase in trabecular ciated with PTH therapy include local injection site re- bone turnover with a positive remodeling balance, with- actions, transient mild headaches, nausea, and arthral- out deleterious effect on cortical bone.21 gia (Table 5). There was some suggestion that adverse Six other trials studied PTH given in combination effects increased with increasing dosage of PTH.30 How- with an antiresorptive (Tables 1 and 2). These com- ever, the largest trial,28 which compared placebo, 20-µg bined alendronate with PTH,27,31 estrogen replacement PTH, and 40-µg PTH, reported no significant difference with PTH,22-24,29,35 or PTH with calcitriol.34 These trials between groups in the incidence of serious adverse ef- reported a bone density benefit over individual therapy fects. Interestingly, the incidence of nausea and head- at the lumbar spine22-24 and had conflicting results re- ache was higher with the 40-µg/d dosage compared with garding different areas of the hip22-24,29,31,35 (Table 2). Dur- placebo, but were similar in the 20-µg/d and placebo ing a 1-year observation on maintenance HRT, follow-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2306

Downloaded From: https://jamanetwork.com/ on 09/27/2021 ing 3 years of combined HRT plus PTH, the increases in clinical corollary—pain. Preliminary findings indicate that BMD persisted.25 there was improvement in backache associated with PTH These studies of combination therapy can be consid- treatment in one trial30 and that new or worsening back ered according to whether the components are given con- pain was reported less often with PTH than placebo in comitantly or sequentially. Concomitant therapy will be another trial.28 The protective effects of PTH against new considered first. In the only study25 of combination therapy nonvertebral fractures and nonvertebral fragility frac- that reported fracture outcome, PTH was given simulta- tures were rapid; effects were evident after 9 to 12 months neously with HRT to treat postmenopausal osteoporosis. of therapy.28 Although direct comparisons have not been The concomitant use resulted in a decrease in radiographi- performed, approximately the same degree of fracture re- cally detected vertebral fractures and in the percentage of duction results from PTH treatment as from risedronate women with incident vertebral fractures. Unfortunately, sodium, alendronate, or raloxifene hydrochloride treat- most studies of concomitant therapy had too few frac- ment in other trials. tures to assess22,29,35 or were not designed to deter- When combined with standard oral or transdermal mine23,24,27 fracture outcome. These latter studies23,24,27 with estrogen alone in postmenopausal osteoporosis, PTH in- bone density outcomes suggested that concurrent PTH and duced increased BMD at the lumbar spine and femoral estrogen use had an advantage over estrogen alone at the neck, but not at the femoral trochanter, compared with spine in women with postmenopausal osteoporosis. The estrogen alone.29 However long-term studies are lack- advantage may24,29 or may not35 exist at the hip. ing, and different routes of administration and dosages In considering sequential therapy, studies again had may have different effects. Moreover, estrogen and pro- too few fractures to allow fracture outcome assess- gesterone components were not investigated sepa- ment16,20,32 or were not designed for assessment of frac- rately.29,35 Each hormone component probably has dif- ture outcome.17,31 It seemed that the use of calcitonin fol- ferent effects on bone metabolism. It is also conceivable lowing PTH therapy had no benefit in postmenopausal that adding PTH to ongoing HRT may have effects dif- or idiopathic male osteoporosis,17,20 but that the use of ferent from those of initiating PTH and HRT concur- alendronate following PTH protected the bone density rently.29 Based on the data reviewed, because of differ- gains incurred during PTH therapy of postmenopausal ing regimens and subject characteristics across trials, it osteoporosis.25,31 is difficult to say if calcium and vitamin D administra- Therefore, in comparing concurrent vs sequential PTH tion affected BMD or fracture outcome effects of PTH. regimens, the data overall would support concurrent therapy Cortical and trabecular bone are the 2 main types with PTH plus estrogen to increase bone density and de- of bone. Cortical bone constitutes the shell around can- crease the incidence of vertebral fractures, and sequential cellous bone. The peripheral skeleton is predominantly therapy with alendronate following PTH therapy to pro- cortical bone, but the axial skeleton (eg, spine) is a com- tect bone density gains. Data specifically do not support bination of cortical and trabecular bone.36 Metabolism sequential therapy with PTH followed by calcitonin. These and rates of bone loss differ in the 2 types of bone, and comments refer only to postmenopausal osteoporosis; data diseases often differentially affect the 2 types of bone.36 in other patient subgroups are lacking. The theoretical possibility of unintentionally causing a In summary, regarding the use of PTH in combina- fracture (or decreased bone density as a possible surro- tion regimens for postmenopausal osteoporosis, the ad- gate for fracture) at one type of bone while simulta- dition of PTH to combination HRT or alendronate had neously benefiting another type of bone is the reason for advantages over HRT alone at the lumbar spine, but ef- strict attention to differentiating cortical and trabecular fects at the radius, trochanter, and hip are conflicting. effects. In 1 trial,34 cortical bone mass decreased 1.7% with For glucocorticoid-induced osteoporosis, combined PTH PTH plus calcitriol vs 5.7% with calcium alone, but tra- and estrogen replacement may be of added benefit com- becular bone mass increased 32% with PTH plus cal- pared with estrogen replacement alone at the spine, but citriol. The varying effects according to bone composi- not necessarily at the total hip or radius. Because only tion (ie, trabecular vs cortical) may relate to the varying one trial tested each regimen, conclusions cannot be made results seen according to measurement technique. For regarding any particular combination involving PTH or example, spine BMD differences between estrogen and about relative advantages of sequential vs concomitant estrogen plus PTH groups were 9.8% by dual-energy x- therapy. Effects of combining one HRT regimen vs an- ray absorptiometry and 33.5% by calcitonin treat- other HRT regimen with PTH are not known. ment.22 There may be subtle differences between men and women in the degree of histomorphometric response to 26 17,24,32,34 COMMENT PTH. The suggestion in several studies of a decrease in radius BMD in association with PTH use par- The increase in BMD induced by PTH seems to be dose- allels the recognized differing effects of hyperparathy- dependent. Based on the results of 1 trial,30 BMD in- roidism on trabecular vs cortical bone. Even if BMD effects creases are similar for subjects 65 years and older vs at different sites become well established, BMD changes younger than 65, weighing at least 50 kg vs less than 50 during therapy are only a surrogate for fractures.37 We kg, with or without baseline vertebral fractures, and re- need to directly determine fracture rates site by site in gardless of time since menopause. prospective studies with adequate statistical power. Parathyroid hormone protects against vertebral frac- Strictly speaking, treatment with PTH should be lim- ture.28 It not only decreases the incidence of radiographi- ited to less than 2 years, given the lack of long-term safety cally detected vertebral fractures but also decreases the and efficacy data. Parathyroid hormone treatment should

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2307

Downloaded From: https://jamanetwork.com/ on 09/27/2021 be followed by antiresorptive therapy, with the goal of nonvertebral fractures. However, there were several draw- maintaining BMD gains. Evidence supports continua- backs to the lack of data: numbers of nonvertebral frac- tion of HRT after combination PTH plus HRT to pre- tures were too low to be broken down by individual site, serve increases in BMD.25 However, the latter medica- it is unknown whether the decrement in radius BMD tions must be continued long-term to preserve beneficial associated with PTH results in an increased risk of future bone density effects. Long-term PTH effects are not clear; radius fracture, and the only population definitely dem- hence, optimal duration of therapy is not clear. Effects onstrated to have vertebral fracture reduction associated of intermittent vs continuous therapy are different and with PTH is women with established postmenopausal os- require additional elucidation.38-40 In the future, differ- teoporosis. ences in response to PTH must be analyzed by sex. In Parathyroid hormone administration induces hy- addition, determination of a clinical algorithm regard- percalcemia in a small percentage of patients, possibly ing appropriate response to PTH-associated hypercalce- in a dose-dependent fashion, and early in treatment. Other mia will be necessary. minor adverse effects are also associated with PTH. Al- In placing future research into proper context, es- though concerns were raised from animal research, pub- pecially given that PTH will be studied in combination lished trial results indicate that cancer risk is not in- with other therapies, several potential pitfalls warrant creased by PTH use in humans. Because only one trial mentioning. Identical bone density changes between 2 tested each individual regimen, conclusions cannot be active treatment groups might signify that both regi- made regarding any particular combination regimen in- mens work to increase BMD or that neither regimen works volving PTH. to increase BMD. In addition, instead of being dis- missed as proof of lack of efficacy, a lack of BMD change Accepted for publication May 8, 2002. compared with baseline may signify that BMD is being Corresponding author and reprints: Carolyn Crandall, preserved (ie, that bone loss is being prevented). Such MD, Department of Medicine, David Geffen School of Medi- considerations must determine appropriate comparison cine at University of California, Iris Cantor–UCLA Women’s groups in the design and proper interpretation of re- Health Center, UCLA School of Medicine, 100 UCLA Medi- sults of randomized controlled trials. cal Plaza, Suite 250, Los Angeles, CA 90095-7023 (e-mail: [email protected]). CONCLUSIONS REFERENCES Antifracture efficacy is the single most critical efficacy outcome for osteoporosis treatment, and fracture reduc- 1. Looker AC, Orwoll ES, Johnston CC Jr, et al. Prevalence of low femoral bone den- tion data are not robust for PTH, especially at nonver- sity in older U.S. adults from NHANES III. J Bone Miner Res. 1997;12:1761-1768. 2. Cauley JA, Thompson DE, Ensrud KC, Scott JC, Black D. Risk of mortality fol- tebral sites. Additional fracture data are anticipated and lowing clinical fractures. Osteoporos Int. 2000;11:556-561. will be the ultimate proof of efficacy. Bone density ef- 3. Ensrud KE, Black DM, Harris F, Ettinger B, Cummings SR, for the Fracture In- fects of 48 weeks to 3 years of therapy with PTH appear tervention Trial Research Group. Correlates of kyphosis in older women. JAm more marked at the spine than at other sites, may be det- Geriatr Soc. 1997;45:682-687. 4. Kado DM, Browner WS, Palermo L, Nevitt MC, Genant HK, Cummings SR, for the rimental at the radius, may be enhanced by combina- Study of Osteoporotic Fractures Research Group. Vertebral fractures and mortal- tion with estrogen, and may enhance effects of alendro- ity in older women: a prospective study. Arch Intern Med. 1999;159:1215-1220. nate. Maximal anabolic effect of PTH may require 12 5. Marottoli RA, Berkman LF, Leo-Summers L, Cooney LM Jr, for the Established months after treatment cessation before manifesting at Populations for Epidemiologic Studies of the Elderly. Predictors of mortality the hip. Dose-related bone density effects may exist in and institutionalization after hip fracture: the New Haven EPESE cohort. Am J Public Health. 1994;84:1807-1812. the dosage range of 50 to 100 µg/d SC. Although a few 6. Nevitt MC, Ettinger B, Black DM, et al. The association of radiographically de- studies support the intranasal route of PTH administra- tected vertebral fractures with back pain and function: a prospective study. Ann tion to prevent bone loss in women receiving specific en- Intern Med. 1998;128:793-800. dometriosis therapies, most data available refer to the SC 7. McClung MR, Geusens P, Miller PD, et al, for the Hip Intervention Program Study Group. Effect of risedronate on the risk of hip fracture in elderly women. N Engl route. However, the SC route was associated with de- J Med. 2001;344:333-340. creased compliance. Subjects in the trials reviewed had 8. Black DM, Thompson DE, Bauer DC, et al, for the FIT Research Group. Fracture the support of clinical trial nurses. Older patients en- risk reduction with alendronate in women with osteoporosis: the Fracture countered in community practice may have difficulty with Intervention Trial. J Clin Endocrinol Metab. 2000;85:4118-4124. injections. 9. Ettinger B, Black DM, Mitlak BH, et al, for the Multiple Outcomes of Raloxifene Evaluation (MORE) Investigators. Reduction of vertebral fracture risk in Results of larger trials were conflicting as to whether postmenopausal women with osteoporosis treated with raloxifene: results from effects were limited to the spine and suggested detrimen- a 3-year randomized clinical trial. JAMA. 1999;282:637-645. tal effects on radius BMD. Little data analyzed the effects 10. Seeman E, Delmas PD. Reconstructing the skeleton with intermittent parathy- of PTH in older vs younger subjects or directly compared roid hormone. Trends Endocrinol Metab. 2001;12:281-283. 11. Hedlund LR, Gallagher JC. Increased incidence of hip fracture in osteoporotic effects according to sex. Increases in lumbar BMD are in- women treated with sodium fluoride. J Bone Miner Res. 1989;4:223-225. duced by PTH in postmenopausal osteoporosis, glucocor- 12. Riggs BL, Hodgson SF, O’Fallon WM, et al. Effect of fluoride treatment on the ticoid-induced osteoporosis, and idiopathic osteoporo- fracture rate in postmenopausal women with osteoporosis. N Engl J Med. 1990; sis. Effects at other sites are insignificant or conflicting. 322:802-809. When data were robust enough to achieve adequate sta- 13. Kleerekoper M, Peterson EL, Nelson DA, et al. A randomized trial of sodium fluoride as a treatment for postmenopausal osteoporosis. Osteoporos Int. 1991;1: tistical power, PTH decreased the risk of new spinal frac- 155-161. tures in women with postmenopausal osteoporosis by 65% 14. Reeve J, Arlot M, Bernat M, et al. Calcium-47 kinetic measurements of bone turn- to 70% and may have decreased the overall incidence of over compared to bone histomorphometry in osteoporosis: the influence of hu-

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2308

Downloaded From: https://jamanetwork.com/ on 09/27/2021 man parathyroid fragment (hPTH 1-34) therapy. Metab Bone Dis Relat Res. 27. Cosman F, Nieves J, Woelfert L, Gordon S, Shen V, Lindsay R. Parathyroid re- 1981;3:23-30. sponsivity in postmenopausal women with osteoporosis during treatment with 15. Hesp R, Hulme P, Williams D, Reeve J. The relationship between changes in fem- parathyroid hormone. J Clin Endocrinol Metab. 1998;83:788-790. oral bone density and calcium balance in patients with involutional osteoporosis 28. Neer RM, Arnaud CD, Zanchetta JR, et al. Effect of parathyroid hormone (1-34) treated with human parathyroid hormone fragment hPTH (1-34). Metab Bone on fractures and bone mineral density in postmenopausal women with osteo- Dis Relat Res. 1981;2:331-334. porosis. N Engl J Med. 2001;344:1434-1441. 16. Hodsman AB, Fraher LJ. Biochemical responses to sequential human parathy- 29. Lindsay R, Nieves J, Formica C, et al. Randomised controlled study of effect of roid hormone (1-38) and calcitonin in osteoporotic patients. Bone Miner. 1990; parathyroid hormone on vertebral-bone mass and fracture incidence among 9:137-152. postmenopausal women on oestrogen with osteoporosis. Lancet. 1997;350: 17. Hodsman AB, Steer BM, Fraher LJ, Drost DJ. Bone densitometric and histomor- 550-555. phometric responses to sequential human parathyroid hormone (1-38) and salmon 30. Fujita T, Inoue T, Morii H, et al. Effect of an intermittent weekly dose of human calcitonin in osteoporotic patients. Bone Miner. 1991;14:67-83. parathyroid hormone (1-34) on osteoporosis: a randomized double-masked pro- 18. Finkelstein JS, Klibanski A, Schaefer EH, Hornstein MD, Schiff I, Neer RM. Para- spective study using three dose levels. Osteoporos Int. 1999;9:296-306. thyroid hormone for the prevention of bone loss induced by estrogen defi- 31. Rittmaster RS, Bolognese M, Ettinger MP, et al. Enhancement of bone mass in ciency. N Engl J Med. 1994;331:1618-1623. osteoporotic women with parathyroid hormone followed by alendronate. J Clin 19. Finkelstein JS, Klibanski A, Arnold AL, Toth TL, Hornstein MD, Neer RM. Endocrinol Metab. 2000;85:2129-2134. Prevention of estrogen deficiency–related bone loss with human parathyroid hor- 32. Kurland ES, Cosman F, McMahon DJ, Rosen CJ, Lindsay R, Bilezikian JP. Para- mone–(1-34): a randomized controlled trial. JAMA. 1998;280:1067-1073. thyroid hormone as a therapy for idiopathic osteoporosis in men: effects on bone 20. Hodsman AB, Fraher LJ, Watson PH, et al. A randomized controlled trial to com- mineral density and bone markers. J Clin Endocrinol Metab. 2000;85:3069- pare the efficacy of cyclical parathyroid hormone versus cyclical parathyroid hor- 3076. mone and sequential calcitonin to improve bone mass in postmenopausal women 33. Leder BZ, Smith MR, Fallon MA, Lee ML, Finkelstein JS. Effects of gonadal ste- with osteoporosis. J Clin Endocrinol Metab. 1997;82:620-628. roid suppression on skeletal sensitivity to parathyroid hormone in men. J Clin 21. Hodsman AB, Kisiel M, Adachi JD, Fraher LJ, Watson PH. Histomorphometric Endocrinol Metab. 2001;86:511-516. evidence for increased bone turnover without change in cortical thickness or po- 34. Neer M, Slovik DM, Daly M, Potts T Jr, Nussbaum SR. Treatment of postmeno- rosity after 2 years of cyclical hPTH(1-34) therapy in women with severe osteo- pausal osteoporosis with daily parathyroid hormone plus calcitriol. Osteoporos porosis. Bone. 2000;27:311-318. Int. 1993;3(suppl 1):204-205. 22. Lane NE, Sanchez S, Modin GW, Genant HK, Pierini E, Arnaud CD. Parathyroid 35. Reeve J, Mitchell A, Tellez M, et al. Treatment with parathyroid peptides and es- hormone treatment can reverse corticosteroid-induced osteoporosis: results of trogen replacement for severe postmenopausal vertebral osteoporosis: predic- a randomized controlled clinical trial. J Clin Invest. 1998;102:1627-1633. tion of long-term responses in spine and femur. J Bone Miner Metab. 2001;19: 23. Lane NE, Sanchez S, Genant HK, Jenkins DK, Arnaud CD. Short-term increases 102-114. in bone turnover markers predict parathyroid hormone–induced spinal bone min- 36. Managing Osteoporosis. Chicago, Ill: American Medical Association; 1999:1-20. eral density gains in postmenopausal women with glucocorticoid-induced os- 37. Crandall C. The role of serial bone mineral density testing for osteoporosis. teoporosis. Osteoporos Int. 2000;11:434-442. J Womens Health Gend Based Med. 2001;10:887-895. 24. Lane NE, Sanchez S, Modin GW, Genant HK, Pierini E, Arnaud CD. Bone mass 38. Podbesek R, Edouard C, Meunier PJ, et al. Effects of two treatment regimes with continues to increase at the hip after parathyroid hormone treatment is discon- synthetic human parathyroid hormone fragment on bone formation and the tissue tinued in glucocorticoid-induced osteoporosis: results of a randomized con- balance of trabecular bone in greyhounds. Endocrinology. 1983;112:1000-1006. trolled clinical trial. J Bone Miner Res. 2000;15:944-951. 39. Hock JM, Gera I. Effects of continuous and intermittent administration and in- 25. Cosman F, Nieves J, Woelfert L, et al. Parathyroid hormone added to established hibition of resorption on the anabolic response of bone to parathyroid hormone. hormone therapy: effects on vertebral fracture and maintenance of bone mass af- J Bone Miner Res. 1992;7:65-72. ter parathyroid hormone withdrawal. J Bone Miner Res. 2001;16:925-931. 40. Tam CS, Heersche JN, Murray TM, Parsons JA. Parathyroid hormone stimu- 26. Dempster DW, Cosman F, Kurland ES, et al. Effects of daily treatment with para- lates the bone apposition rate independently of its resorptive action: differential thyroid hormone on bone microarchitecture and turnover in patients with os- effects of intermittent and continuous administration. Endocrinology. 1982;110: teoporosis: a paired biopsy study. J Bone Miner Res. 2001;16:1846-1853. 506-512.

(REPRINTED) ARCH INTERN MED/ VOL 162, NOV 11, 2002 WWW.ARCHINTERNMED.COM 2309

Downloaded From: https://jamanetwork.com/ on 09/27/2021