Ipriflavone in the Treatment of Postmenopausal Osteoporosis a Randomized Controlled Trial

ORIGINAL CONTRIBUTION

Ipriflavone in the Treatment of Postmenopausal Osteoporosis A Randomized Controlled Trial

Peter Alexandersen, MD Context Data on the efficacy and safety of ipriflavone for prevention of postmeno- Anne Toussaint, MD pausal bone loss are conflicting. Claus Christiansen, MD, PhD Objectives To investigate the effect of oral ipriflavone on prevention of postmenopausal bone loss and to assess the safety profile of long-term treatment with iprifla- Jean-Pierre Devogelaer, MD, PhD vone in postmenopausal osteoporotic women. Christian Roux, MD, PhD Design and Setting Prospective, randomized, double-blind, placebo-controlled, 4-year Jacques Fechtenbaum, MD, PhD study conducted in 4 centers in Belgium, Denmark, and Italy from August 1994 to July 1998. Carlo Gennari, MD, PhD Participants Four hundred seventy-four postmenopausal white women, aged 45 Jean Yves Reginster, MD, PhD to 75 years, with bone mineral densities (BMDs) of less than 0.86 g/cm2. for the Ipriflavone Multicenter Interventions Patients were randomly assigned to receive ipriflavone, 200 mg 3 times European Fracture Study per day (n = 234), or placebo (n = 240); all received 500 mg/d of calcium.

TUDIES OF IPRIFLAVONE, A SYN- Main Outcome Measures Efficacy measures included spine, hip, and forearm BMD thetic isoflavone derivative, have and biochemical markers of bone resorption (urinary hydroxyproline corrected for creatinine and urinary CrossLaps [Osteometer Biotech, Herlev, Denmark] corrected for suggested that it inhibits bone re- creatinine), assessed every 6 months. Laboratory safety measures and adverse events sorption and stimulates osteo- were recorded every 3 months. Sblast activity in vitro in cell cultures1,2 and Results Based on intent-to-treat analysis, after 36 months of treatment, the annual in vivo in experimental models of osteo- 3 percentage change from baseline in BMD of the lumbar spine for ipriflavone vs pla- porosis. For example, ipriflavone was cebo (0.1% [95% confidence interval {CI}, −7.9% to 8.1%] vs 0.8% [95% CI, 45 demonstrated to inhibit Ca release from −9.1% to 10.7%]; P = .14), or in any of the other sites measured, did not differ sig- fetal long-bone cultures, both spontane- nificantly between groups. The response in biochemical markers was also similar ously and after stimulation with para- between groups (eg, for hydroxyproline corrected for creatinine, 20.13 mg/g [95% thyroid hormone,1 and to inhibit resorp- CI, 18.85-21.41 mg/g] vs 20.67 mg/g [95% CI, 19.41-21.92 mg/g]; P = .96); uri- tion pits induced by osteoclast activity.2 nary CrossLaps corrected for creatinine, 268 mg/mol (95% CI, 249-288 mg/mol) vs Incubation of rat osteosarcoma cells (cell- 268 mg/mol (95% CI, 254-282 mg/mol); P = .81. The number of women with new line UMR 106-a) with ipriflavone re- vertebral fracture was identical or nearly so in the 2 groups at all time points. Lym- phocyte concentrations decreased significantly (500/µL (0.5ϫ109/L]) in women sulted in increased release of alkaline treated with ipriflavone. Thirty-one women (13.2%) in the ipriflavone group devel- 4 phosphatase into the media. Further- oped subclinical lymphocytopenia, of whom 29 developed it during ipriflavone treat- more, ipriflavone has been shown to ment. Of these, 15 (52%) of 29 had recovered spontaneously by 1 year and 22 inhibit bone loss in osteoporotic rats (in- (81%) of 29 by 2 years. 3 duced by corticosteroids). These en- Conclusions Our data indicate that ipriflavone does not prevent bone loss or affect couraging results led to a number of clini- biochemical markers of bone metabolism. Additionally, ipriflavone induces lympho- cal trials to test the efficacy on bone mass cytopenia in a significant number of women. in various populations. This inhibition JAMA. 2001;285:1482-1488 www.jama.com of bone loss in these populations was typically mirrored by a reduction in the Author Affiliations are listed at the end of this article. Corresponding Author and Reprints: Peter A complete list of the members of the lpriflavone Mul- Alexandersen, MD, Center for Clinical and Basic Re- concentration of biochemical markers of ticenter European Fracture Study Group was published search, Ballerup Byvej 222, 2750 Ballerup, Denmark bone metabolism.5 In postmenopausal previously (Calcif Tissue Int. 1997;61:528-532). (e-mail: [email protected]).

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

women, data on the efficacy of iprifla- medication known to affect bone me- 3 years. An internal quality assurance vone on the prevention of bone loss are tabolism. Women participating in the control was set up at the Danish center conflicting. Nevertheless, most studies study were identified by advertise- for the BMD measurements, as previ- have shown that ipriflavone (typical dos- ments and via the national registration ously described.14 For women who age, 600 mg/d) is able to prevent bone office (Denmark only). dropped out of the study, the last BMD loss,5-9 and some data have even sug- observation was carried forward. gested that ipriflavone may increase bone Ethical Aspects Biochemical Markers of Bone Turn- mass in postmenopausal women.10-13 The study was approved by the local eth- over. Bone formation was determined by However, reports of lymphocytopenia in ics committees and health authorities in serum alkaline phosphatase automatic women taking ipriflavone have gener- all 3 countries recruiting participants to analyzer (Cobas Mira Plus, Roche Diag- ated some concerns regarding the safety the study (Belgium, Denmark, and Italy). nostic Systems, Basel, Switzerland). Bone of ipriflavone.7 The study was conducted in accor- resorption was evaluated by fasting uri- Ipriflavone is marketed and easily dance with the Helsinki Declaration, and nary hydroxyproline corrected by cre- available as an over-the-counter prod- all participants were informed about the atinine by spectrophotometry (UV-160 uct in several countries (ie, Ostovone in study and gave written informed con- A), as described previously.15 In addi- the United States, Osten in Japan, Os- sent before entering the study. tion, bone metabolism was evaluated by teochin in Hungary, and Osteofix and serum calcium, serum phosphorus, and Iprosten in Italy). We report the results Settings urinary excretion of calcium corrected from a large, randomized, double- The study was conducted at 4 Euro- for creatinine (Cobas Mira Plus). In the blind, placebo-controlled, 3-year clini- pean centers. The 2 Belgian centers re- Danish subpopulation, we also mea- cal study,14 designed to investigate the ef- cruited 205 and 52 subjects, the Dan- sured urinary CrossLaps (Osteometer ficacy and safety of ipriflavone on bone ish center recruited 197, and the Italian Biotech A/S, Herlev, Denmark) cor- density, biochemical markers of bone center recruited 20 subjects. Details rected for urinary creatinine as deter- turnover, and fracture rate in postmeno- about the study design have been pub- mined by enzyme-linked immunosor- pausal women with osteoporosis. lished elsewhere.14 bent assay.15 Biochemical markers were assessed at baseline and every 6 months METHODS Study Treatment Groups throughout the study. All analyses were Subjects Women were randomly assigned in performed when the study was com- Four hundred seventy-four white women blocks (assigned to each center) to ei- pleted. between the ages of 45 and 75 years, with ther ipriflavone (200 mg 3 times a day) Incident Nontraumatic Vertebral a natural menopause at least 1 year be- or placebo administered orally in con- Fractures. The incidence of nontrau- fore entering the study, with low bone nection with meals in a double-blind matic vertebral fractures was evaluated mass defined as a bone mineral density fashion. Tablets (ipriflavone or pla- as a secondary end point. Lateral radi- (BMD) of the lumbar spine (L2-L4) be- cebo) were all identical in appearance ography of the thoracic and lumbar spine low 0.86 g/cm2, as determined by the (white, round), smell, taste, and weight. was performed according to a standard- QDR 1000 (Hologic Inc, Waltham, All participants received a concomitant ized acquisition procedure, and as- Mass), corresponding to at least 2 SDs calcium supplementation of 500 mg/d. sessed in a central facility.16 The x-ray below the premenopausal mean value Individual participant treatment code film examination was performed at base- were included in the study. No women envelopes were provided to the investi- line (unless this had been done less than with a body mass index lower than 30 gator by the sponsor prior to allocation. a year prior to entry in the study) and kg/m2 were enrolled. Protocol exclu- The lead investigator kept the treat- again after years 1, 2, and 3. The x-ray sion criteria were (1) any x-ray film that ment code envelopes in a locked, secure films were evaluated by a radiologist documented previous vertebral frac- storage facility. Unblinding of the indi- blinded to treatment. A fracture was de- ture, substantial scoliosis, osteophyto- vidual treatment codes occurred upon fined as a 20% or greater reduction of the sis, or spinal secondary osteoporosis, or completion of the study by all subjects. anterior, middle, or posterior height of bone-related diseases; (2) significant con- a vertebra at the level of T4-L4.16 The to- comitant disease or medical history that End Points tal number of incident spinal fractures could interfere with the study; (3) alco- Bone Mineral Density. Lumbar spine and the number of women with inci- hol abuse; (4) medication such as sex ste- (L2-L4), total hip, and distal radius BMD dent fractures were then calculated for roids, bisphosphonates, calcitonin, fluo- was determined by dual-energy radio- each group.14 ride, glucocorticoids within 12 months graph absorptiometry (QDR 1000). Cali- Measurement of Ipriflavone- prior to randomization, or any iprifla- bration was performed with a phantom Circulating Metabolites in the Plasma. vone intake in the month prior to ran- before measurement at each skeletal site Ipriflavone and its active plasma domization. In addition, at the time of on a daily basis. The BMD was deter- metabolites M-III and M-V were ana- inclusion, participants were taking no mined every 6 months throughout the lyzed by hydrolosis followed by a spe-

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

cific high-performance liquid chroma- nitrogen, aspartate transaminase, ala- Compliance tography with UV detection. Calibration nine transaminase, ␥-glutamyltransfer- An account was made for each partici- curves were linear over the studied con- ase, lactic dehydrogenase, alkaline phos- pant concerning study drugs dis- centration range of 20 to 4000 ng/mL. phatase, total bilirubin, total cholesterol, pensed and returned at each visit. The lower limit of the calibration curves triglycerides, total amount of protein, Plasma levels of ipriflavone and its me- was taken as the limit of quantitation albumin, creatinine, sodium, potas- tabolites were determined after the first, of the method. The extraction recov- sium, chloride, thyrotropin [base- second, and third year. ery determined at 3 concentration lev- line], and vitamin D3 [baseline]) were els was higher than 80%, the precision determined using the Cobas Mira Plus.14 Statistical Analysis of the method ranged between 2% and Urinary analysis (glucose, protein, The data presented are based on intent- 9%, and the accuracy between 90% blood, chetonic bodies) was per- to-treat analysis. Baseline parameters and 113%.17 formed by urine dipstick.14 Micro- and changes in end points at various scopic examination was performed if time points (during the 3-year treat- Safety Determinations abnormal dipstick results were obtained. ment period) were compared by using Laboratory Safety Parameters. Rou- Physical Examination. A complete the t test. Changes over time in the ver- tine blood samples were collected after physical examination of each partici- tebral morphometry parameters (height 12 hours of fasting at baseline (prior to pant was performed every 3 months.14 ratios) were analyzed in the time in- randomization) and semianually dur- Adverse Events. Every 3 months, ad- terval baseline (3 years) by repeated ing the course of the study. Hematol- verse events were recorded. An ad- measures analysis of variance with base- ogy (including erythrocytes, leuko- verse event was defined as any adverse line values and center included as co- cytes, differential count, hematocrit, change from the baseline clinical or variates. The number of fractures after hemoglobin, and platelets) was deter- laboratory condition and classified by 1, 2, and 3 years of treatment was evalu- mined using the Sysmex (Toa Medical body system or preferred term. Rela- ated using 1-sided Fisher exact test. Electronics, Surrey, England), biochem- tionship of the adverse events to the Similarly, changes over time in BMD istry (including glucose, serum urea study drug was evaluated.14 and biochemical markers of bone turnover were analyzed in the same time interval by repeated measures of analy- Table 1. Demographic Data* sis of variance. The SAS statistical Ipriflavone Placebo software (SAS Institute Inc, Cary, NC) Characteristic (n = 234) (n = 240) was used for the statistical analysis, and Mean (SD) PϽ.05 was regarded as statistically sig- Age, y 63.2 (6.2) 63.4 (6.2) nificant. Height, cm 159.2 (6.5) 160.0 (6.6) Sample size calculation was based on Body weight, kg 62.8 (8.7) 63.2 (9.0) a mean (SD) annual decrease in spinal Bone mineral density, g/cm2 Spine† 0.76 (0.08) 0.76 (0.07) BMD of approximately 1% (6%) in pla- Hip‡ 0.74 (0.09) 0.74 (0.10) cebo-treated women, which gives a Follicle-stimulating hormone, mIU/mL 87 (28) 83 (36) power of just over 97% (240 placebo- Lymphocyte count, ϫ106/L 1.92 (0.56) 1.91 (0.59) treated women and P=.05). The ac- tual SD in the study was 3% per year No. (%) for both groups. Conversely, to detect Surgical menopause 35 (15) 48 (20) a 1% difference in spinal BMD with at Family history of osteoporosis 44 (19) 50 (21) least 90% power would require a study Smokers 50 (21.4) 63 (26.4) Daily calcium intake size of about 1300 subjects. Low (Ͻ800 mg/d) 101 (43.2) 106 (44.2) Normal (800-1500 mg/d) 100 (42.9) 106 (44.2) RESULTS High (Ͼ1500 mg/d) 30 (12.8) 24 (10.0) Population None 3 (1.3) 4 (1.7) TABLE 1 depicts the baseline charac- Physical exercise (walking) teristics for the 2 groups. None had a Low (Ͻ1 h/d) 124 (53.0) 124 (51.7) prevalent spine fracture in either group Moderate (1 h/d) 90 (38.5) 104 (43.3) at the time of randomization. At base- Intense (Ͼ1 h/d or other 19 (8.1) 11 (4.6) physical activity) line, 339 of the total population (71.6%) None 1 (0.4) 1 (0.4) had a spinal T score of less than −2.5 *There was no significant difference between the groups for any variable tested. SD. FIGURE 1 illustrates the progress of †Mean T score for ipriflavone is −2.89; placebo, −2.86. patients throughout the study. A total ‡Mean T score for ipriflavone is −1.97; placebo, −1.99. of 292 women completed the study: 132

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

in the ipriflavone group, and 160 in was similar at all time points. FIGURE 3 ture after 1, 2, or 3 years of treatment the placebo group. shows the changes in serum alkaline (TABLE 2). In the ipriflavone group, the (Figure 3A), in urinary hydroxypro- plasma ipriflavone concentration (mean Efficacy Parameters line corrected for creatinine (Figure 3B), [SEM]) was 93 (10) ng/mL after 1 year, The data on efficacy presented are for and in urinary CrossLaps corrected for 107 (20) ng/mL after 2 years, and 64 (16) intent-to-treat. Data for validated com- creatinine (Figure 3C). We found no sta- ng/mL after 3 years of treatment. The pleters were essentially identical to the tistically significant difference between concentration of the active metabolite intent-to-treat analysis (data available the 2 treatment groups for any of the bio- M-V was 1001 (58) ng/mL after 1 year, upon request). FIGURE 2 shows the chemical markers of either bone forma- 1115 (95) ng/mL after 2 years, and 886 change in BMD (mean [90% confi- tion or bone resportion. There was no (95) ng/mL after 3 years and similar val- dence interval {CI}]) in the 2 groups statistically significant difference be- ues were found for the other metabo- during the course of the study. tween the 2 groups regarding inciden- lite M-III (916 [425], 3900 [1695], and After 36 months of treatment, there tal vertebral fractures or the number of 1233 [671] ng/mL, respectively). The was no statistically significant differ- subjects with an incident vertebral frac- values for the placebo group regarding ence between annual percentage change from baseline in BMD of the lumbar Figure 1. Flow Diagram spine (ipriflavone vs placebo, 0.1% [95%

CI, −7.9% to 8.1%] vs 0.8% [95% CI, 680 Patients Screened −9.1% to 10.7%]; P=.14); or in any of the other sites between the 2 groups. The 206 Not Randomized 87 Not Willing to Give Informed Consent, Lack of response in the excretion of hydroxy- Time, Other Personal Reason proline corrected for creatinine was also 119 Ineligible 46 Significant Background Disease similar between groups (ipriflavone vs 27 Prevalent Vertebral Fracture 40 Bone Mineral Density ≥ 0.86 g/cm2 placebo, 20.13 mg/g [95% CI, 18.85- 3 Body Mass Index >30 kg/m2 21.41 mg/g] vs 20.67 mg/g [95% CI, 3 Receiving Medication 19.41-21.92 mg/g]; P=.96) and in the excretion of CrossLaps corrected for cre- 474 Randomized atinine was 268 mg/mol (95% CI, 249- 288 mg/mol) vs 268 mg/mol (95% CI, 234 Assigned to Receive Ipriflavone Plus 240 Assigned to Receive Placebo Plus 254-282 mg/mol); P=.81. The lympho- Calcium, 500 mg/d Calcium, 500 mg/d cyte concentration decreased significantly in women treated with ipri- Cumulative Withdrawals Cumulative Withdrawals flavone and 29 women developed 33 at 6 mo 34 at 6 mo 66 at 12 mo 52 at 12 mo subclinical lymphocytopenia (Ͻ500/µL 94 at 24 mo 78 at 24 mo Ͻ ϫ 9 102 at 36 mo 80 at 36 mo [ 0.5 10 /L]) during ipriflavone treat- 49 Adverse Events 35 Adverse Events ment. Of these, 52% had recovered spon- 9 Protocol Violations 10 Protocol Violations 4 Preexisting Diseases 7 Preexisting Diseases taneously by 1 year and 81% by 2 years. 40 Personal Reasons 28 Personal Reasons The development in spine BMD (Fig- ure 2A), hip BMD (Figure 2B), or arm 132 Completed Trial 160 Completed Trial BMD (Figure 2C) between the 2 groups

Figure 2. Time-Related Change in Bone Mineral Density (BMD)

A Lumbar Spine B Femur C Radius

2 0.78 0.75 0.45

0.76 0.74 0.44

0.74 0.73 0.43 Placebo Ipriflavone Change in BMD, g/cm 0.72 0.72 0.42 0612 18 24 30 36 0612 18 24 30 36 0612 18 24 30 36 Month Month Month

Values are expressed as changes in mean (90% confidence interval).

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

Figure 3. Changes in Markers of Bone Formation

A Serum Alkaline Phosphatase B Hydroxyproline Corrected for Creatinine C Urinary CrossLaps

190 27 320

25 170 280 23 150 240 U/L mg/g 21 mg/mol 130 200 Placebo 19 Ipriflavone 110 17 160 0612 18 24 30 36 0612 18 24 30 36 0612 18 24 30 36 Month Month Month

Values are expressed as changes in mean (90% confidence interval). The transient decrease in both groups in A and C after 6 months of treatment probably is due to the effect of calcium supplementation. the 2 groups. During the 4-year study * Table 2. Incident Vertebral Fractures period, the number compliant (Ն75%) Ipriflavone Placebo Ͻ (n = 233) (n = 239) and noncompliant ( 75%) were 130 No. RR (95% CI)† (55.6%) vs 18 (7.7%) for ipriflavone, New vertebral fractures respectively, and 153 (63.8%) vs 21 12 mo 0 2 (8.8%) for placebo, respectively. Data 24 mo 7 9 0.82 (0.36-1.89) were missing for 86 (36.8%) for ipri- 36 mo 11 11 1.07 (0.53-2.16) flavone and 66 (27.5%) for placebo. No. (%) P Value‡ Women with Յ1 vertebral fracture COMMENT 12 mo 0 (0) 2 (0.8) .26 Ipriflavone (7-isopropoxy-isoflavone) 24 mo 6 (2.6) 6 (2.5) .63 is a synthetic daidzein derivative of 36 mo 7 (3) 8 (3.3) .52 natural isoflavone,7 the natural com- *One woman in each group had a vertebral fracture that was considered traumatic and excluded from the analysis. †RR indicates relative risk; CI, confidence interval. pound daidzein is its metabolite M-II, ‡Calculated using 1-sided Fisher exact test. thought to have a positive effect on health similarly to other isoflavones, plasma concentrations of ipriflavone and discriminate between lymphocyte such as genistein. Several in vitro stud- its metabolites were not detectable. subpopulations). Of the 31 women with ies have suggested that ipriflavone (typi- low lymphocyte counts, 29 developed cally 200 mg orally 3 times per day) in- Safety Parameters them during the course of the study, hibits bone resorption1,5-9 and increases TABLE 3 and TABLE 4 summarize the while 2 had lymphocytopenia at 36 bone formation,4 believed to be the number of adverse events reported in the months. Of the 29 with low lympho- mechanism by which ipriflavone may study for the 2 groups. The most impor- cyte concentrations, 15 (52%) had recov- prevent bone loss in postmenopausal tant adverse event in this study was ered to normal values within 12 months women.6,11 However, the present study lymphocytopenia (total lymphocyte con- and 22 (81%) in 24 months after dis- did not confirm the previous findings centration Ͻ500/µL). The mean lym- continuation of ipriflavone. All cases of on bone metabolism in terms of bio- phocyte percentage fraction of lympho- lymphocytopenia were subclinical (ie, chemical markers of bone turnover, cytes relative to the total white blood cell clinically asymptomatic). There were no BMD, or fracture rates. count in the entire ipriflavone group significant changes between the groups There may be several potential expla- decreased significantly from about 33% in any other clinical or laboratory param- nations for the present lack of efficacy (1900/µL [1.9ϫ109/L]) to about 27% eter investigated (data not shown). One of ipriflavone observed in this study. (1400/µL [1.4ϫ109/L]) (PϽ.001) patient (3%) had lymphocytopenia once First, it could be speculated that the dos- (FIGURE 4). This decrease had occurred before the baseline counts. The cut-off age used in this study was not correct, after 6 months and remained stable point was when the absolute lympho- or at least suboptimal. Second, it is rel- throughout the treatment period. In 31 cyte count was below 500/µL. Five of the evant to question whether the statisti- of the subjects treated with ipriflavone lymphocytopenic women (16%) are still cal power of the study was sufficient to (13.2%), the concentration of circulat- being monitored. detect statistically significant differ- ing lymphocytes decreased signifi- No statistically significant differences between the ipriflavone group and cantly below 500/µL during the treat- ences in overall treatment and in cal- the placebo group for the end points ment period (this study did not cium compliance were found between evaluated. And third, it could be ques-

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

tioned if the study population studied study, the values of urinary excretion Table 3. Adverse Events* was too old, or had too little bone mass. of CrossLaps were in the same range as No. of Adverse Ipriflavone Placebo With respect to the dosage issue, the found in other postmenopausal women Events (n = 234) (n = 240) tablets used in this study each con- with low bone mass treated with cal- 0 15 (6.4) 13 (5.4) tained 200 mg of ipriflavone, and were cium.15 The different results when com- 1 17 (7.3) 21 (8.8) 2 16 (6.8) 23 (9.6) given 3 times a day with a meal. This regi- paring our study with previous stud- 3 17 (7.3) 13 (5.4) men has typically been used in previ- ies showing a bone-preserving effect of 4 11 (4.7) 24 (10.0) Ն ous clinical studies.5-14 Measurements of ipriflavone may in part be due to the 5 158 (67.5) 146 (60.8) the physiologically active ipriflavone me- fact that our study population was older *Values are expressed as number (percentage). tabolites M-III and M-V in plasma dur- than in most of these other controlled Table 4. Women With at Least 1 Adverse ing the course of the study showed val- studies. We also found no effects on the Effect Categorized by Body System ues that were comparable with those fracture incidence (primary end point) Ipriflavone Placebo 6 obtained in positive studies, indicating in the ipriflavone-treated women. How- System* (n = 234) (n = 240) that the dosage of ipriflavone presum- ever, our study did not have sufficient Gastrointestinal tract 200 (85.5) 152 (63.3) ably was clinically sufficient. There- power to detect an effect of iprifla- Metabolic and 25 (10.7) 27 (11.7) nutritional fore, it seems unlikely that the dosage of vone on fracture incidence. So far, no Musculoskeletal 153 (65.4) 193 (80.4) ipriflavone used in the present study studies on ipriflavone have been suffi- Neoplasms† 3 (1.3) 7 (2.9) Psychiatric/mental 53 (22.6) 40 (16.7) should have been insufficient. ciently powered to study fracture inci- Reproductive 9 (3.8) 3 (1.3) Concerning the second question of dence in relation to ipriflavone.10 Re- Infections 32 (13.7) 20 (8.3) Respiratory 79 (33.8) 93 (38.8) statistical power, the size of this study cent fracture studies performed (eg, of Dermal and appendage 35 (15.0) 36 (15.0) population was large enough to detect bisphosphonates or raloxifene)18-20 have Urinary 38 (16.2) 40 (16.7) a statistically significant change in both enrolled several hundred, or even thou- Vascular (extracardiac) 29 (12.4) 13 (5.4) Lymphocytopenia 29 (12.4) 1 (0.4) bone turnover parameters and in bone sands of patients. Thus, with regard to Other 42 (17.9) 39 (16.3) mass. Although only 292 individuals the surrogate end points bone mass and *Body system or preferred term developed during the study. were valid completers, we found no ef- the markers of bone resorption, of †One neoplasm in the placebo group was benign. fect on any of these parameters. The which CrossLaps is known to be both 15 methods used are all generally accepted specific and sensitive, we consis- Figure 4. Changes in Circulating and bone mass measurements were sub- tently found no difference from pla- Concentration of Lymphocytes ject to quality control.14 Furthermore, our cebo. data were consistent (ie, similar re- Therefore, it remains to be consid- 40 Placebo Ipriflavone sponses to treatment were found for the ered if we looked at the right study popu- various biochemical markers), and in lation. Studies of early and later post- turn these results were mirrored in the menopausal women6,11 and of women 30 change in bone mass. Therefore, it is with senile osteoporosis10 have re- most likely that the results are reliable ported positive results with ipriflavone % Lymphocytes, in terms of bone mass and biochemical (ie, ipriflavone may prevent postmeno- 20 0612 18 24 30 36 markers. pausal bone loss). The population con- Month One previous study7 found that in sidered in the current study was post- early postmenopausal women, uri- menopausal women with established Values are expressed as changes in mean (90% confidence interval). After 12 months, women treated with nary deoxypyridinoline did not de- osteoporosis but no prevalent vertebral ipriflavone had a significant decrease in their average crease more in a group receiving ipri- fractures, and therefore a positive result lymphocyte concentration (24%; PϽ.001) that re- flavone and calcium compared with a in terms of bone mass should be able to mained constant throughout the study. group receiving calcium only. Deoxy- be detected. Theoretically, our popula- pyridinoline has been demonstrated to tion could have been too osteoporotic ferences in behavioral, nutritional, and correlate highly and significantly to and this might explain the lack of effect environmental factors influencing bone CrossLaps,15 the resorption marker used of ipriflavone in these women, but we ob- mass, which were not controlled in this in a subpopulation of our study. De- served not even a tendency toward an ef- study. However, the study was random- spite this, in the study cited,7 iprifla- fect in ipriflavone-treated women com- ized and conducted in accordance with vone prevented bone loss compared pared with placebo. Furthermore, other Good Clinical Practice guidelines. Evalu- with calcium alone and the difference interventions to increase BMD are effec- ation of efficacy and safety parameters (approximately −2%) was signifi- tive in this population.21 Another factor was performed with the investigators be- cantly different. Other studies have may be that the women in this study ing unaware of treatment groups. There- shown that ipriflavone decreases bone came from various parts of Europe, and fore, we conclude that the reason for a resorption and prevents bone loss com- geographic differences within subpopu- lack of effect probably is not in the de- pared with placebo.5-11 In the current lations may be important because of dif- sign of the study, nor in the quality of

Downloaded From: https://jamanetwork.com/ on 09/30/2021 IPRIFLAVONE FOR POSTMENOPAUSAL OSTEOPOROSIS

the study, its conduction, and surro- Ipriflavone-treated women with lym- teoporotic drugs available. Its use in treat- gate end points, nor is it related to dos- phocytopenia withdrawn from the study ment is not supported by these data.

age or compliance issues of the drug, and also were regularly monitored even af- Author Affiliations: Center for Clinical and Basic Re- we exclude having considered the wrong ter the termination of the study, and search, Ballerup, Denmark (Drs Alexandersen and end points or the wrong population. The some (16%) are still being followed up Christiansen); Bone/Cartilage Metabolism Unit, CHU Brull, Lie` ge, Belgium (Drs Toussaint and Reginster); most obvious explanation is that iprifla- in the safety study until normalization. Rheumatology Unit, St-Luc University Hospital, Uni- vone does not have a significant effect on However, no statistically significant dif- versite´ Catholique de Louvrain, Brussels, Belgium (Dr Devogelaer); Centre d’Evaluation des Maladies Os- the factors evaluated in this population. ference in opportunistic infections, neo- seuses, Hoˆ pital Cochin, Paris, France (Drs Roux and We observed that women treated with plastic events, or other adverse effects Fechtenbaum); Department of Rheumatology, Uni- versity Rene´ Descartes, Paris, France (Dr Roux); and ipriflavone had a signficant decrease in was found between the ipriflavone and Institute of Internal Medicine, University of Siena, Siena, lymphocyte concentrations from about placebo groups and the lymphocytope- Italy (Dr Gennari). Author Contributions: Study concept and design: 33% to about 27%, whereas women nic and nonlymphocytopenic, iprifla- Alexandersen, Toussaint, Christiansen, Gennari, treated with placebo did not. The de- vone-treated women. Hence, all causes Reginster. Acquisition of data: Alexandersen, Toussaint, Chris- crease occurred after 6 months of treat- of lymphocytopenia observed were sub- tiansen, Devogelaer, Roux, Fechtenbaum, Gennari, ment and the lymphocyte concentra- clinical (ie, all subjects remained clini- Reginster. tion thereafter remained stable. A cally healthy). The importance of the Analysis and interpretation of data: Alexandersen, Toussaint, Christiansen, Roux, Gennari, Reginster. number of the ipriflavone-treated lymphocytopenia observed in associa- Drafting of the manuscript: Alexandersen, Tous- women developed lymphocytopenia tion with ipriflavone in terms of health saint, Christiansen, Reginster. Ͻ 22,23 Critical revision of the manuscript for important in- ( 500/µL). The lymphocyte subpopu- remains unknown. tellectual content: Alexandersen, Toussaint, Chris- lations CD4 and CD8 were not deter- In conclusion, the present large ran- tiansen, Devogelaer, Roux, Fechtenbaum, Gennari, Reginster. mined in this study, and thus it is un- domized, double-blind, placebo- Obtained funding: Toussaint, Reginster. known if these subpopulations were controlled study failed to find any ef- Administrative, technical, or material support: Alex- andersen, Toussaint, Christiansen, Roux, Fechten- affected equally. However, this effect of fect of ipriflavone on calcium metabolism baum, Reginster. ipriflavone has previously been re- in women with postmenopausal osteo- Study supervision: Alexandersen, Toussaint, Chris- ported, albeit in a smaller sample.10 In porosis, but indicated that some women tiansen, Roux, Reginster. Other: Devogelaer. the current trial, this adverse labora- treated with ipriflavone may develop sub- Funding/Support: Cheisi Farmaceutici SpA (Parma, tory condition returned to normal within clinical lymphocytopenia that may take Italy), which manufactures ipriflavone, sponsored the study. 24 months in 81% of the patients, more than 24 months to resolve. On the Acknowledgment: We thank Cheisi Farmaceutici, SpA, whereas the remaining 19% were fol- basis of our results, the relative benefit- for sponsoring the study and the participants for their participation. The statistical evaluation and analyses lowed up at regular intervals until the risk ratio of ipriflavone appears low when were assisted by PPD Development Inc (Cambridge, lymphocyte count returned to normal. compared with the alternative antios- England).

REFERENCES 1. Tsuda M, Kitazaki T, Ito T, Fujita T. The effect of lumbar bone mineral density and bone metabolic mark- healthy volunteers at steady-state: evaluation of two ipriflavone (TC-80) on bone resorption in tissue cul- ers in postmenopausal women with low bone mass. different dosage schemes. Eur J Drug Metab Pharma- ture. J Bone Miner Res. 1986;1:206-211. Horm Res. 1999;51:178-183. cokinet. 1998;23:172-177. 2. Notoya K, Yoshida K, Taketomi S, Yamazaki I, Ku- 10. Agnusdei D, Bufalino L. Efficacy of ipriflavone in 18. Liberman UA, Weiss SR, Broll J, et al, for the Alen- megawa M. Inhibitory effect of ipriflavone on pit for- established osteoporosis and long-term safety. Cal- dronate Phase III Osteoporosis Treatment Study Group. mation in mouse unfractionated bone cells. Calcif Tis- cif Tissue Int. 1997;61:23-27. Effect of oral alendronate on bone mineral density and sue Int. 1992;51(suppl 1):S1-S6. 11. Agnusdei D, Crepaldi G, Isaia G, et al. A double the incidence of fractures in postmenopausal osteo- 3. Yamazaki I, Shino A, Shimizu Y, Tsukuda R, Shi- blind, placebo-controlled trial of ipriflavone for pre- porosis. N Engl J Med. 1995;333:1437-1443. rakawa Y, Kinoshita M. Effect of ipriflavone on glu- vention of postmenopausal spinal bone loss. Calcif Tis- 19. Harris ST, Watts NB, Genant HK, et al, for the Ver- cocorticoid-induced osteoporosis in rats. Life Sci. 1986; sue Int. 1997;61:142-147. tebral Efficacy With Risedronate Therapy (VERT) Study 38:951-958. 12. Moscarini M, Patacchiola F, Spacca G, Palermo Group. Effects of risedronate treatment on vertebral 4. Benvenuti S, Tanini A, Frediani U, et al. Effects of P, Caserta D, Valenti M. New perspectives in the treat- and nonvertebral fractures in women with postmeno- ipriflavone and its metabolites on a clonal osteoblas- ment of postmenopausal osteoporosis: ipriflavone. Gy- pausal osteoporosis: a randomized controlled trial. tic cell line. J Bone Miner Res. 1991;6:987-996. necol Endocrinol. 1994;8:203-207. JAMA. 1999;282:1344-1352. 5. Gennari C, Adami S, Agnusdei D, et al. Effect of 13. Passeri M, Biondi M, Costi D, et al. Effect of ipri- 20. Ettinger B, Black DM, Mitlak BH, et al, for the Mul- chronic treatment with ipriflavone in postmeno- flavone on bone mass in elderly osteoporotic women. tiple Outcomes of Raloxifene Evaluation (MORE) In- pausal women with low bone mass. Calcif Tissue Int. Bone Miner. 1992;19(suppl 1):S57-S62. vestigators. Reduction of vertebral fracture risk in post- 1997;61(suppl 1):S19-S22. 14. Reginster JY, Bufalino L, Christiansen C, et al. De- menopausal women with osteoporosis treated with 6. Adami S, Buffalino L, Cervetti R, et al. Ipriflavone pre- sign for an ipriflavone multicenter European fracture raloxifene: results from a 3-year randomized clinical vents radial bone loss in postmenopausal with low bone study. Calcif Tissue Int. 1997;61(suppl 1):S28-S32. trial. JAMA. 1999;282:637-645. mass over 2 years. Osteoporos Int. 1997;7:119-125. 15. Bonde M, Qvist P, Fledelius C, Riis BJ, Christiansen 21. Meunier PJ, Delmas PD, Eastell R, et al, for the 7. Gennari C, Agnusdei D, Crepaldi G, et al. Effect of C. Applications of an enzyme immunoassay for a new International Committee for Osteoporosis Clinical ipriflavone—a synthetic derivative of natural isofla- marker of bone resorption (CrossLaps): follow-up on hor- Guidelines. Diagnosis and management of osteopo- vones—on bone mass loss in the early years after mone replacement therapy and osteoporosis risk as- rosis in postmenopausal women: clinical guidelines. menopause. Menopause. 1998;5:9-15. sessment. J Clin Endocrinol Metab. 1995;80:864-868. Clin Ther. 1999;21:1025-1044. 8. Agnusdei D, Zacchei F, Bigazzi S, et al. Metabolic and 16. Genant HK, Wu CY, Van Kujik C, Nevitt MC. Ver- 22. Gergely P. Drug-induced lymphopenia: focus on clinicaleffectsofipriflavoneinestablishedpost-menopausal tebral fracture assessment using a semiquantitative CD4+ and CD8+ cells. Drug Saf. 1999;21:91-100. osteoporosis. Drugs Exp Clin Res. 1989;15:97-104. technique. J Bone Miner Res. 1993;8:1137-1148. 23. Castelino DJ, McNair P, Kay TW. Lymphocyto- 9. Ohta H, Komukai S, Makita K, Masuzawa T, 17. Acerbi D, Poli G, Ventura P. Comparative bio- penia in a hospital population—what does it signify? Nozawa S. Effects of 1-year ipriflavone treatment on availability of two oral formulations of ipriflavone in AustNZJMed. 1997;27:170-174.

Downloaded From: https://jamanetwork.com/ on 09/30/2021