Risedronate for the prevention and treatment of postmenopausal osteoporosis.
Abstract
Background
Postmenopausal osteoporosis results in an increased susceptibility to low‐trauma fractures due to reduced bone volume and microarchitectural deterioration. Risedronate, a third generation bisphosphonate, has been shown in multiple clinical trials to reduce fracture risk and improve bone mineral density in postmenopausal women with osteoporosis. First and second generation bisphosphonates are known to have gastrointestinal side‐effects and risedronate may be better tolerated.
Objectives
To systematically review the efficacy of risedronate on bone density, and fracture reduction in postmenopausal women.
Search methods
The Cochrane Controlled Trials Registry Medline, and Current Contents were searched from 1990 ‐ 2001. The electronic search was supplemented by handsearching four osteoporosis journals and their conference proceedings, as well as contacting content experts and industry sources for unpublished data.
Selection criteria
We included eight trials that randomised women to risedronate or an alternative (placebo or calcium and /or vitamin D) and measured bone mineral density for at least one year.
Data collection and analysis
For each trial three independent reviewers assessed the methodological quality and abstracted data. Data was extracted for outcomes of fracture, bone mineral density and adverse events. The more conservative random effects model was used to pool data. The quality of trials was assessed according to the Jadad five‐point scale.
Main results
Both vertebral and non‐vertebral fractures were statistically and clinically reduced with risedronate. Eleven out of one hundred women who received risedronate had a vertebral fracture compared to 17 out of one hundred of those who received calcium and vitamin D (pooled relative risk for vertebral fractures of 0.64 (95% CI 0.52 ‐ 0.77). Three percent of participants who received risedronate had a non‐vertebral fracture compared to 4.6% of those who received calcium and vitamin D (pooled relative risk for nonvertebral fractures of 0.73 (95% CI 0.61 ‐ 0.87). The weighted mean difference for the percent change from baseline for bone mineral density with 5 mg daily for lumbar spine, femoral neck and trochanter was 4.54% (95%CI 4.12 ‐ 4.97), p<0.01; 2.75% (95% CI 2.32 ‐ 3.17), p<0.01; and 4.38% (95% CI 3.51 ‐ 5.25), p<0.01 respectively.
Authors' conclusions
There is good evidence for the efficacy of risedronate in the reduction of both vertebral and non‐vertebral fractures. In addition, there is evidence from randomized trials that risedronate is able to achieve this without increasing risk for overall withdrawals due to adverse effects.
PICOs
Plain language summary
HOW WELL DOES RISEDRONATE WORK TO TREAT AND PREVENT OSTEOPOROSIS IN WOMEN AFTER MENOPAUSE?
To answer this question, scientists found and analysed 8 high quality studies. These studies tested over 14 500 women after menopause who had mild or severe bone loss or a break. Women received a placebo (sugar pill) with or without calcium with vitamin D, or 5mg of risedronate daily. These studies provide the best evidence we have today.
What is osteoporosis and how can risedronate help?
Osteoporosis is a condition of weak brittle bones that break easily. In osteoporosis, breaks or fractures of the spine and hip or wrist (non‐spinal fractures) may occur and often without a fall. Risedronate is a bisphosphonate ‐ a drug that is often prescribed to women after menopause to decrease fractures (or breaks) by slowing the loss of bone. There is some debate about whether risedronate decreases all types of fractures and whether it causes few side effects.
How well did risedronate decrease fractures and increase bone density?
After 3 years, fewer women after menopause have spine or non‐spinal (such as wrist or hip) fractures when receiving risedronate than a placebo.
‐11 out of 100 women taking risedronate versus 17 out of 100 with a placebo had a spine fracture
‐3 out of 100 women taking risedronate versus 5 out of 100 with a placebo had a non‐spinal fracture.
Over 3 years, bone mineral density in the spine and hip increased more in women taking 5 mg of risedronate daily than in women taking a placebo or calcium and vitamin D.
Were there any side effects?
Side effects such as nausea, stomach upset or pain, and diarrhea may occur. However, the number of people who stopped taking risedronate due to side effects was about equal to the number of people who stopped taking a placebo.
What is the bottom line?
There is "platinum" level evidence that in women after menopause, risedronate at 5mg daily over 3 years decreases spine and non‐spinal fractures.
After 3 years, risedronate 5 mg increases bone density in the spine and hip.
Most women do not appear to have side effects that would cause them to stop taking risedronate.
Authors' conclusions
Background
Osteoporosis is defined by the World Health Organization as a reduction in bone mineral density (BMD) of more than 2.5 standard deviations below the mean value in young women and it is estimated to affect approximately 30% of postmenopausal women. (Kanis 1994) The cumulative lifetime osteoporosis related fracture risk for a 50 year old woman may be as high as 40% (Cooper 1993). Low BMD, increasing age and a history of fragility fractures are notable risks for future fractures (Brown 2002). For every 1 standard deviation (SD) reduction in BMD, there is an increase in relative risk of fracture of approximately 1.5 to 2.6 (Marshall 1996). Seventy percent of hip fractures have been estimated to be attributable to osteoporosis and the mortality in the first year after a hip fracture ranges from 12 ‐ 20% (Hawker 1998; Cooper 1993). In addition, only 1/3 of women who sustain hip fractures are able to return to their previous living arrangements and functional abilities (Jagal 1996; Jagal 1998). Various treatment guidelines have been developed to outline a framework for the management of postmenopausal osteoporosis (Brown 2002, Royal College of Phy).
Risedronate, a third generation bisphosphonate, inhibits osteoclast‐mediated bone resorption by reducing activity of osteoclasts, and hastening osteoclast apoptosis. Recent large randomised controlled trials in postmenopausal osteoporosis have demonstrated a positive impact of risedronate on BMD and a reduction in the risk of vertebral and non‐vertebral fractures.
Objectives
To assess the efficacy of risedronate versus control in the prevention and treatment of postmenopausal osteoporosis. The major endpoints were: (1) reduction in vertebral and non‐vertebral fractures, (2) change in BMD and (3) toxicity of risedronate.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials of at least one year in duration comparing risedronate to placebo (or calcium and/or vitamin D combinations) were included. The studies included at least one of the primary outcomes of bone mineral density or fractures.
Types of participants
Trials including postmenopausal women were included. The two major groups of women that were included were those with mild bone loss (prevention) or those with more severe bone loss as defined by a T score < ‐2.0 SD or prevalent vertebral fracture (treatment).
Types of interventions
Interventions included all doses of risedronate for durations of at least one year. Calcium and/or vitamin D were allowed as concurrent therapy, if given to both treatment groups in equal doses.
Types of outcome measures
(1) Number of women with incident vertebral and non‐vertebral fractures in risedronate versus control groups.
(2) Percent change in BMD as measured by Dual Energy X‐ray Absorptiometry (DXA) at baseline and yearly intervals.
(3) Percent of patients with withdrawals due to adverse events and withdrawals due to gastro‐intestinal side‐effects.
Results for fractures and withdrawals were calculated to provide an indication of the number needed to treat (NNT).
Search methods for identification of studies
The electronic search strategy consisted of the following key and text words: risedronate, actonel, osteoporosis, postmenopausal, bisphosphonates, and randomised control trial. The search strategy as outlined by Haynes et al. (Haynes 1994) was used and modified by for the Cochrane Musculoskeletal group.
We searched the Cochrane Controlled Trials Registry Medline, EMBASE, and Current Contents for the years 1990 ‐ December 2000. Risedronate is a relatively new therapy so we did not expect to find any publications earlier than 1990. There was no restriction made on language.
1 Osteoporosis, Postmenopausal/
2 Osteoporosis/
3 Osteoporosis.tw
4 Exp bone density/
5 Bone loss$.tw
6 (Bone adj2 densit$).tw
7 menopause/
8 post‐menopause$.tw
9 or/2‐6
10 or/7‐8
11and/9‐10
12 1 or 11
13 exp diphosphonates/
15 diphosphonates$.tw,rn.
16 bisphosphonates$.tw,rn.
17 risedronate.tw,rn.
18 actonel.tw,rn
19 or/14‐18
20 13 and 19
21 clinical trial.pt.
22 randomised controlled trial.pt.
23 tu.fs.
24 dt.fs
25 random$.tw
26 (double adj blind$).tw.
27 placebo$.tw.
28 or/21‐27
29 20 and 28
We hand searched four primary musculoskeletal journals (Osteoporosis International, Journal of Bone and Mineral Research, Calcified Tissue International and Arthritis and Rheumatism) and the conference proceedings from 1998 ‐ 2001. We also hand searched Food and Drug Administration proceedings related to risedronate. We contacted content experts and industry to ensure completeness of our search.
Data collection and analysis
The abstracts that were returned from the search strategy were assessed by two independent reviewers (AC, NZ) for potential eligibility based on the inclusion criteria. The full publication of all abstracts that appeared to meet the inclusion criteria were retrieved. Each publication was reviewed and the data was independently extracted by two reviewers (NZ, AP or AC). Details of the study population, duration of intervention, baseline demographic data, and the outcome measures previously noted were extracted. Differences in data extraction were resolved by referring to the original publication and establishing consensus.
Methodological quality was assessed by three independent reviewers (LW, AC, NZ) using a validated scale (Jadad 1996) This scale includes three items pertaining to descriptions of randomisation, double‐blinding, and the inclusion of data for dropouts and withdrawals, with a total score of five.
For fractures and toxicity outcomes, a weighted relative risk was determined, using Review Manager 4.1 (Fleiss 1993). Heterogeneity of the treatment effect was calculated using a chi‐square test with n‐1 degrees of freedom, where n is the number of studies (Fleiss 1993). Where possible the analyses were based on intention‐to‐treat data from the individual clinical trials. For BMD a weighted mean difference of the percent change between treatment and control groups for different BMD sites was calculated. Data was extracted for lumbar spine, femoral neck, trochanter, and distal radius. Meta‐analysis was conducted with the more conservative random effects model.
Regression analysis was conducted that allowed us to determine which doses and years we could pool across (Cranney 2002 Endocrine Reviews 2002; 23(4):517‐23). The model allowed us to pool across all years but did not permit pooling across all doses for bone mineral density. The model allowed us to pool across doses of 2.5 mg cyclical, 2.5 mg daily, and 5.0 mg cyclical for BMD outcomes. For fractures, we pooled across all doses and years.
A Priori Hypotheses Regarding Heterogeneity:
We developed prior to the pooling of studies, hypotheses that might account for heterogeneity of study results. We compared groups according to 1) dose, 2) treatment duration and 3) prevention versus treatment populations. Dosages for risedronate included 2.5mg cyclical, 2.5 mg daily and 5 mg cyclical and 5 mg daily 4) low versus high quality trials. Treatment was defined as a population of women with prevalent vertebral fractures or BMD T score < ‐2.0 SD. Several authors of the included studies and industry (Procter and Gamble) were contacted to obtain additional data.
Allocation concealment was evaluated using the criteria outlined in the Cochrane Hanbook. The four categories being: A) adequate (i.e. central randomization; numbered or coded bottles or containers; drugs prepared by the pharmacy; serially numbered, opaque, sealed envelopes; or other description that contained elements convincing of concealment), B) unclear (i.e. authors either did not report an allocation concealment approach at all or reported an approach that was neither adequate nor inadequate), C) inadequate (i.e. alternation or reference to case record numbers or to dates of birth), and D) not used.
Grading the strength of the evidence
The common system of grading the strength of scientific evidence for a therapeutic agent that is described in the CMSG module scope and in the Evidence‐based Rheumatology BMJ book (Tugwell 2003) was used to rank the evidence included in this systematic review. Four categories are used to rank the evidence from research studies from highest to lowest quality: Platinum, Gold, Silver, and Bronze. The ranking is included in the synopsis and abstract of this review.
Results
Description of studies
Twenty‐one studies were returned from the electronic and hand search strategy outlined above. Thirteen were excluded for reasons that included lacking a control arm, non‐randomised nature, duplication of data or wrong population (Delmas 1997; Eastell 1999; Goa 1998; Harris 1999 (2); Kendler 1999; Licata 1997; Miller 1999; Rezda 1999; Ribot 1999; Roux 1999; Singer 1995; Watts 1998; Watts 1999). The remaining eight were included (Clemmesen 1997; Fogelman 2000; Harris 1999; Hooper 1999; McClung 1998; McClung 2001; Mortensen 1998; Reginster 2000). The details of the included publications are appended in the tables.
Risk of bias in included studies
The quality of the included publications was assessed using the Jadad scale (Jadad 1996). The quality scores and lost to follow‐up are appended in the table of included studies. The mean quality of the included studies was 3.1/5, and the range was 1‐5/5. All eight trials used an intention to treat analysis. Despite the high quality scores, five trials had lost to follow‐up greater than 25%.
Effects of interventions
FRACTURES:
Vertebral Fractures:
A total of 5 trials (n=2604) assessed the efficacy of risedronate on vertebral fractures. The pooled estimate of the relative risk, using a random effects model, for all doses and years was 0.64 (95% CI 0.52 ‐ 0.77); p<0.01. The treatment effect was similar across trials (heterogeneity p=0.89). All doses (2.5 mg and 5.0 mg/day) were pooled together. We used the estimate that represented the longest duration available from each trial. If we restricted our analyses to 5 mg dose alone, the pooled RR was 0.62 (95% CI 0.51, 0.76). For trials with the 5 mg dose of risedronate, the absolute fracture rate was 10.3 % compared to 16.7% for controls, resulting in an absolute treatment benefit of 6.4% and an NNT of 16 ( See Clinical Benefit ‐ Additional tables).
Non‐Vertebral Fractures:
A total of 7 trials (n=12985) assessed the efficacy of risedronate on non‐vertebral fractures. The weighted relative risk, using a random effects model, for all doses and years was 0.73 (95% CI 0.61 ‐ 0.87); p<0.01; heterogeneity p=0.81. The doses and durations were pooled in the same manner as vertebral fractures. The absolute fracture rate with risedronate was 2.9 % in comparison to 4.6% for controls yielding an absolute treatment benefit of 1.7% and an NNT of 58.
When we pooled only trials that used 5 mg dose the pooled estimate for the RR from 5 trials was 0.66 (95 % CI 0.51, 0. 86), which is consistent with a 34 % relative risk reduction. The absolute treatment benefit was 2.8% (absolute fracture rate with risedronate 5.4% compared to 8.2% for controls), yielding an NNT of 35.
BMD:
We were not able to obtain BMD data from the risedronate hip fracture trial (McClung 2001).
Lumbar Spine BMD:
A total of 6 trials (n=2138) assessed the efficacy of risedronate 5mg on lumbar spine BMD. The weighted mean difference between treatment and control of the percent change from baseline, using a random effects model, was 4.54% (95% CI 4.12 ‐ 4.97); p<0.01.
We pooled across doses 2.5 mg cyclical, 5.0 mg cyclical and 2.5 mg daily for 7 trials (n=1842) and the weighted mean difference with final year results was 2.94 (95% CI 1.56 ‐ 4.36), p<0.01.
Femoral Neck BMD:
A total of 6 trials (n=2337) assessed the efficacy of risedronate 5 mg on femoral neck BMD. The weighted mean difference of the percent change from baseline, using a random effects model and final year data, was 2.75% (95% CI 2.32 ‐ 3.17); p<0.01. The treatment effect was consistent across trials (heterogeneity p=0.77).
For treatment after one year with 5 mg of risedronate the pooled WMD from 5 trials (n=1509) for femoral neck was 1.55 (95% CI 1.08, 2.02).
The WMD for femoral neck pooled across doses 2.5 mg cyclical, 5.0 mg cyclical and 2.5 mg daily for 1 year from 6 trials (n=1606) was 0.63 (95% CI 0.21 ‐1.06) p<0.01. The WMD for final year for pooled 2.5 mg cyclical, 5.0 mg cyclical and 2.5 mg daily from 6 trials was 1.72 (95% CI 1.19, 2.25).
Trochanteric BMD:
A total of 6 trials (n=2337) assessed the efficacy of risedronate 5 mg on trochanter BMD. The weighted mean difference of the percent change from baseline, using a random effects model, was 4.38% (95% CI 3.51 ‐ 5.25); p<0.01; heterogeneity p<0.05.
Distal Radius BMD:
Only one trial reported data for distal radius BMD and the weighted mean difference was 0. 90 (95% CI ‐0. 76 to 2. 56), p<0.29. For mid radius the same trial reported a WMD of 0.9 (95% CI 0.35 to 1.45), p =0.001
ADVERSE EVENTS:
Withdrawals due to Side Effects:
A total of 8 trials (n=13,998) reported the safety of risedronate through withdrawals due to side effects. The pooled relative risk for withdrawals due to side effects, using a random effects model, for all doses was 0.94 (95% CI 0.84 ‐ 1.06); p=0.2. These results suggest that side‐effects due to treatment with risedronate did not effect the risk of discontinuing treatment. For the 5 mg dose, the pooled relative risk for withdrawals due to side effects was 0.94 (95% CI 0.84 ‐1.06). (16.7 % withdrew from the treatment group compare with 17.2 % of controls with and absolute reduction of ‐0.50. For all doses combined, the absolute reduction was 0.66 %.
Gastrointestinal Side Effects:
A total of 4 trials (n=12,313) reported the safety of risedronate based on gastro‐intestinal side effects. The pooled estimate of relative risk for discontinuation due to GI side effects, using a random effects model, for all doses pooled was 0.97 (95% CI 0.91 ‐ 1.04); p=0.5; heterogeneity p=0.7. Gastritis occurred in 25% of risedronate treated patients versus 32% of controls with an absolute risk reduction of 7%.
Abdominal Pain:
Five trials (N=12,835) reported abdominal pain and the pooled relative risk was 0.93 (95% CI 0.83, 1.05) using a random effects model. Abdominal pain occurred in 9.3 percent of risedronate treated individuals versus 9.8% of controls, yielding an absolute risk reduction of 0.5 %.
Esophagitis:
Five trials ( N=12856) reported on the outcome of esophagitis and the pooled relative risk for all doses was 0.91 (95% CI 0.70, 1.19).
SUBGROUP ANALYSIS:
Treatment versus Prevention:
Subgroup analysis was done based on prevention or treatment of osteoporosis. Six publications were classified as treatment studies (Clemmesen 1997; Fogelman 2000; Harris 1999; McClung 1998; McClung 2001; Reginster 2000) and two publications were classified as prevention studies (Hooper 1999; Mortensen 1998). The doses and durations were pooled in the same manner as the primary analyses.
Vertebral Fractures:
There was only one publication available that assessed the efficacy of risedronate in a prevention population for vertebral fractures so we were unable to estimate a pooled relative risk (RR). The RR for one trial (n=111) was 2.43 (95% CI 0.12 ‐ 49.43). There were a total of four treatment trials (n=2,493) available and the pooled estimate of the RR, using a random effects model, was 0.63 (95% CI 0.52 ‐ 0.77); p<0.01; heterogeneity p=0.98.
Non‐Vertebral Fractures:
There was only one prevention trial for risedronate that evaluated non‐vertebral fractures so we were unable to estimate a pooled relative risk. The RR from one trial (n = 111) was 0.48 (95% CI 0.10 ‐ 2.26).
There was a total of six treatment trials that evaluated nonvertebral fractures (n=12,847). The pooled estimate of the relative risk, using a random effects model, was 0.73 (95% CI 0.61 ‐ 0.88); p<0.01; heterogeneity p=0.97.
Lumbar Spine BMD: (5 mg dose, final year data)
The weighted mean difference of the percent change from baseline from two prevention trials (N=230) with 5 mg was 4.70% (95% CI 3.74 ‐ 5.67); p<0.001; heterogeneity p=0.28.
The weighted mean difference of the percent change from baseline from four treatment trials (n=1908), using a random effects model, was 4.53% (95% CI 3.95 ‐ 5.11); p<0.01.
Femoral Neck BMD:
The weighted mean difference of the percent change from baseline from two prevention trials (n=230), using a random effects model, was 3.22% (95% CI 2.35 ‐ 4.08); p<0.01; heterogeneity, p=0.88.
The weighted mean difference of the percent change from baseline from four treatment trials (n=1908), using a random effects model, was 2.60% (95% CI 2.12‐ 3.09); p<0.01; heterogeneity p=0.9.
Discussion
Prevention of fractures is the primary outcome used to assess the efficacy of osteoporosis treatments. There is good evidence to support the use of risedronate to reduce the risk of both vertebral and non‐vertebral fractures based on our results. The pooled estimate of the relative risk for both vertebral and non‐vertebral fractures was 0.64 and 0.73 respectively. This corresponds to a relative risk reduction of 36% for vertebral fractures (absolute treatment benefit of 6.4%) and 27% for non‐vertebral fractures. Both estimates were statistically significant and there was little heterogeneity of treatment effect. The estimate for non‐vertebral fractures was affected by the addition of one trial that only included hip fractures in their publication (McClung 2001). This trial represented 45.1% of the weight in treatment trials and 44.5% of the weight when all trials were pooled. It may create a slightly more conservative estimate of effect in both cases. Using the 5 mg/day estimate, a more clinically relevant measure, the pooled relative risks are 0.62 for vertebral fractures and 0.66 for non‐vertebral fractures, which are slightly more favourable than the pooled estimate for combined doses.
BMD is a surrogate outcome for fractures that is used when assessing the efficacy of osteoporosis treatments. There is adequate evidence available to support the use of risedronate for increasing BMD at multiple sites. There was a significant increase in lumbar spine BMD with the 5 mg daily dose and with pooled cyclical and 2.5 mg daily dose. Femoral neck BMD showed statistically significant improvement in the percent change from baseline, and there was no heterogeneity of treatment effect with similar results for treatment and prevention groups. There was a trend to a larger treatment effect in the prevention population but the confidence intervals overlapped those of the prevention population.
There was no statistically significant increase in the relative risk for adverse events measured by withdrawals due to side effects and discontinuations due to gastro‐intestinal side effects. Gastro‐intestinal side effects were grouped into gastritis and abdominal pain. Neither demonstrated an increased risk with risedronate.
The quality of the included trials was high based on the Jadad score. There is evidence from previous reviews that suggests publications which score poorly on the Jadad scale are more likely to bias the treatment estimate away from the null. The Jadad scale, however, does not consider the quality of the drop‐outs in a particular trial, just whether dropouts are reported. An important methodological limitation of these trials was that 5 of the 8 included trials had losses to follow‐up of greater than 25%. Loss to follow up can threaten the validity of the trial since the event rate may be very different in those lost to follow‐up versus those who complete the trial. However, in this case it appears unlikely that loss to follow‐up has created a bias in favour of risedronate given that the patients lost to follow‐up in the placebo arm were a particularly high risk group as reflected by the larger number of patients who had a vertebral fracture at baseline.
All of the 8 trials did conceal allocation. This was clarified for a couple of the trials by checking with industry.
Comparison 1 VERTEBRAL FRACTURES, Outcome 1 2.5 mg / day.
Comparison 1 VERTEBRAL FRACTURES, Outcome 2 5 mg / day.
Comparison 1 VERTEBRAL FRACTURES, Outcome 3 pooled doses.
Comparison 2 NON‐VERTEBRAL FRACTURES, Outcome 1 2.5 mg / day.
Comparison 2 NON‐VERTEBRAL FRACTURES, Outcome 2 5 mg / day.
Comparison 2 NON‐VERTEBRAL FRACTURES, Outcome 3 pooled doses.
Comparison 3 BMD: LUMBAR SPINE, Outcome 1 2.5 mg / day.
Comparison 3 BMD: LUMBAR SPINE, Outcome 2 5 mg / day.
Comparison 3 BMD: LUMBAR SPINE, Outcome 3 pooled doses.
Comparison 4 BMD: FEMORAL NECK, Outcome 1 2.5 mg / day.
Comparison 4 BMD: FEMORAL NECK, Outcome 2 5 mg / day.
Comparison 4 BMD: FEMORAL NECK, Outcome 3 pooled doses, 2.5 mg daily or cyclically or 5 mg cyclically.
Comparison 5 BMD: TROCHANTER, Outcome 1 2.5 mg / day.
Comparison 5 BMD: TROCHANTER, Outcome 2 5 mg / day.
Comparison 6 ADVERSE EVENTS, Outcome 1 Withdrawals due to Side Effects.
Comparison 6 ADVERSE EVENTS, Outcome 2 Gastro‐Intestinal Side Effects.
Comparison 6 ADVERSE EVENTS, Outcome 3 Abdominal Pain.
Comparison 6 ADVERSE EVENTS, Outcome 4 Gastritis.
Comparison 6 ADVERSE EVENTS, Outcome 5 Esophagitis.
Comparison 7 SUBGROUP ANALYSIS: TREATMENT VERSUS PREVENTION: POOLED DOSES, Outcome 1 Vertebral Fractures.
Comparison 7 SUBGROUP ANALYSIS: TREATMENT VERSUS PREVENTION: POOLED DOSES, Outcome 2 Non‐Vertebral Fractures.
Comparison 7 SUBGROUP ANALYSIS: TREATMENT VERSUS PREVENTION: POOLED DOSES, Outcome 3 Lumbar Spine BMD, 5 mg dose, Final year.
Comparison 7 SUBGROUP ANALYSIS: TREATMENT VERSUS PREVENTION: POOLED DOSES, Outcome 4 Femoral Neck BMD, 5mg dose Final year.
Comparison 8 SENSITIVITY ANALYSIS: QUALITY SCORE: 5 mg pooled, Outcome 3 Lumbar Spine BMD.
Comparison 8 SENSITIVITY ANALYSIS: QUALITY SCORE: 5 mg pooled, Outcome 4 Femoral Neck BMD.
Comparison 9 BMD: RADIUS, Outcome 1 Distal radius.
Comparison 9 BMD: RADIUS, Outcome 2 Mid radius.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 2.5 mg / day Show forest plot | 4 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.1 1 Year | 2 | 1941 | Risk Ratio (M‐H, Random, 95% CI) | 0.56 [0.40, 0.79] |
| 1.2 2 Years | 1 | 185 | Risk Ratio (M‐H, Random, 95% CI) | 0.98 [0.45, 2.14] |
| 1.3 3 Years | 1 | 88 | Risk Ratio (M‐H, Random, 95% CI) | 0.55 [0.22, 1.35] |
| 1.4 All Years Pooled | 4 | 2214 | Risk Ratio (M‐H, Random, 95% CI) | 0.61 [0.45, 0.82] |
| 2 5 mg / day Show forest plot | 4 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.1 1 Year | 2 | 1996 | Risk Ratio (M‐H, Random, 95% CI) | 0.40 [0.27, 0.59] |
| 2.2 2 Years | 1 | 237 | Risk Ratio (M‐H, Random, 95% CI) | 0.53 [0.24, 1.17] |
| 2.3 3 Years | 3 | 2138 | Risk Ratio (M‐H, Random, 95% CI) | 0.62 [0.50, 0.77] |
| 2.4 All Years Pooled | 4 | 2375 | Risk Ratio (M‐H, Random, 95% CI) | 0.62 [0.50, 0.76] |
| 3 pooled doses Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 3.1 1 Year | 2 | 2943 | Risk Ratio (M‐H, Random, 95% CI) | 0.48 [0.36, 0.64] |
| 3.2 2 Years | 1 | 297 | Risk Ratio (M‐H, Random, 95% CI) | 0.68 [0.36, 1.30] |
| 3.3 3 Years | 4 | 2307 | Risk Ratio (M‐H, Random, 95% CI) | 0.63 [0.51, 0.78] |
| 3.4 All Years Pooled | 5 | 2604 | Risk Ratio (M‐H, Random, 95% CI) | 0.64 [0.52, 0.77] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 2.5 mg / day Show forest plot | 3 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.1 1 Year | 0 | 0 | Risk Ratio (M‐H, Random, 95% CI) | 0.0 [0.0, 0.0] |
| 1.2 2 Years | 2 | 617 | Risk Ratio (M‐H, Random, 95% CI) | 0.69 [0.36, 1.33] |
| 1.3 3 Years | 1 | 88 | Risk Ratio (M‐H, Random, 95% CI) | 1.0 [0.27, 3.75] |
| 1.4 All Years Pooled | 3 | 705 | Risk Ratio (M‐H, Random, 95% CI) | 0.74 [0.41, 1.34] |
| 2 5 mg / day Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.1 1 Year | 0 | 0 | Risk Ratio (M‐H, Random, 95% CI) | 0.0 [0.0, 0.0] |
| 2.2 2 Years | 2 | 673 | Risk Ratio (M‐H, Random, 95% CI) | 0.65 [0.36, 1.20] |
| 2.3 3 Years | 3 | 2510 | Risk Ratio (M‐H, Random, 95% CI) | 0.66 [0.49, 0.89] |
| 2.4 All Years Pooled | 5 | 3183 | Risk Ratio (M‐H, Random, 95% CI) | 0.66 [0.51, 0.86] |
| 3 pooled doses Show forest plot | 7 | Risk Ratio (M‐H, Fixed, 95% CI) | Subtotals only | |
| 3.1 1 Year | 0 | 0 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.0 [0.0, 0.0] |
| 3.2 2 Years | 2 | 945 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.67 [0.40, 1.12] |
| 3.3 3 Years | 5 | 12013 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.74 [0.61, 0.89] |
| 3.4 All Years Pooled | 7 | 12958 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.73 [0.61, 0.87] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 2.5 mg / day Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 1 Year | 5 | 1531 | Mean Difference (IV, Random, 95% CI) | 2.22 [1.46, 2.97] |
| 1.2 2 Years | 5 | 1442 | Mean Difference (IV, Random, 95% CI) | 3.11 [1.28, 4.93] |
| 1.3 3 years | 1 | 170 | Mean Difference (IV, Random, 95% CI) | 4.3 [2.02, 6.58] |
| 1.4 Final year pooled | 6 | 1754 | Mean Difference (IV, Random, 95% CI) | 3.00 [1.45, 4.55] |
| 2 5 mg / day Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 1 Year | 5 | 1478 | Mean Difference (IV, Random, 95% CI) | 3.49 [3.01, 3.98] |
| 2.2 2 Years | 6 | 2356 | Mean Difference (IV, Random, 95% CI) | 4.90 [4.33, 5.47] |
| 2.3 3 Years | 2 | 1115 | Mean Difference (IV, Random, 95% CI) | 4.94 [3.63, 6.26] |
| 2.4 Final Year Pooled | 6 | 2138 | Mean Difference (IV, Random, 95% CI) | 4.54 [4.12, 4.97] |
| 3 pooled doses Show forest plot | 7 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Final year cyclical doses | 2 | 132 | Mean Difference (IV, Random, 95% CI) | 1.79 [0.25, 3.34] |
| 3.2 Final Year Pooled 2.5 mg cyclical,or daily or 5 mg cyclically | 7 | 1842 | Mean Difference (IV, Random, 95% CI) | 2.95 [1.56, 4.35] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 2.5 mg / day Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 1 Year | 5 | 1509 | Mean Difference (IV, Random, 95% CI) | 0.64 [0.18, 1.10] |
| 1.2 2 Years | 5 | 1560 | Mean Difference (IV, Random, 95% CI) | 1.66 [1.17, 2.15] |
| 1.3 3 Years | 1 | 258 | Mean Difference (IV, Random, 95% CI) | 1.77 [0.00, 3.54] |
| 2 5 mg / day Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 1 Year | 5 | 1509 | Mean Difference (IV, Random, 95% CI) | 1.55 [1.08, 2.02] |
| 2.2 2 Years | 6 | 2569 | Mean Difference (IV, Random, 95% CI) | 2.40 [1.94, 2.87] |
| 2.3 3 Years | 2 | 1313 | Mean Difference (IV, Random, 95% CI) | 2.88 [2.09, 3.68] |
| 2.4 5 mg years 1.5 ‐ 3 pooled | 6 | 2337 | Mean Difference (IV, Random, 95% CI) | 2.75 [2.32, 3.17] |
| 3 pooled doses, 2.5 mg daily or cyclically or 5 mg cyclically Show forest plot | 7 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 1 Year | 6 | 1606 | Mean Difference (IV, Random, 95% CI) | 0.64 [0.22, 1.06] |
| 3.2 1.5 to 3 Years | 6 | 1380 | Mean Difference (IV, Random, 95% CI) | 1.72 [1.20, 2.25] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 2.5 mg / day Show forest plot | 5 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 1.1 1 Year | 4 | 1449 | Mean Difference (IV, Random, 95% CI) | 1.54 [1.00, 2.08] |
| 1.2 2 Years | 4 | 1500 | Mean Difference (IV, Random, 95% CI) | 2.76 [1.81, 3.70] |
| 1.3 3 Years | 1 | 258 | Mean Difference (IV, Random, 95% CI) | 3.13 [1.16, 5.10] |
| 1.4 All Years Pooled | 5 | 1768 | Mean Difference (IV, Random, 95% CI) | 2.12 [1.24, 3.01] |
| 2 5 mg / day Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 2.1 1 Year | 5 | 1509 | Mean Difference (IV, Random, 95% CI) | 2.41 [1.85, 2.96] |
| 2.2 2 Years | 6 | 2569 | Mean Difference (IV, Random, 95% CI) | 3.86 [3.14, 4.58] |
| 2.3 3 Years | 2 | 1313 | Mean Difference (IV, Random, 95% CI) | 5.19 [3.01, 7.37] |
| 2.4 All Years Pooled | 6 | 2337 | Mean Difference (IV, Random, 95% CI) | 4.38 [3.51, 5.25] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Withdrawals due to Side Effects Show forest plot | 8 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.1 2.5 mg | 4 | 7660 | Risk Ratio (M‐H, Random, 95% CI) | 0.93 [0.67, 1.28] |
| 1.2 5 mg | 6 | 9367 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.89, 1.06] |
| 1.3 final year ‐ all doses | 8 | 13998 | Risk Ratio (M‐H, Random, 95% CI) | 0.94 [0.84, 1.06] |
| 2 Gastro‐Intestinal Side Effects Show forest plot | 4 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.1 2.5 mg | 3 | 7130 | Risk Ratio (M‐H, Random, 95% CI) | 1.01 [0.92, 1.10] |
| 2.2 5 mg | 4 | 8768 | Risk Ratio (M‐H, Random, 95% CI) | 0.96 [0.89, 1.04] |
| 2.3 All doses pooled | 4 | 12313 | Risk Ratio (M‐H, Random, 95% CI) | 0.97 [0.91, 1.04] |
| 3 Abdominal Pain Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 3.1 2.5 mg | 4 | 7479 | Risk Ratio (M‐H, Random, 95% CI) | 0.92 [0.79, 1.06] |
| 3.2 5 mg | 5 | 9112 | Risk Ratio (M‐H, Random, 95% CI) | 1.00 [0.82, 1.23] |
| 3.3 All doses pooled | 5 | 12835 | Risk Ratio (M‐H, Random, 95% CI) | 0.93 [0.83, 1.05] |
| 4 Gastritis Show forest plot | 3 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 4.1 2.5 mg | 2 | 1179 | Risk Ratio (M‐H, Random, 95% CI) | 1.10 [0.52, 2.32] |
| 4.2 5 mg | 3 | 2801 | Risk Ratio (M‐H, Random, 95% CI) | 0.76 [0.45, 1.30] |
| 4.3 All doses pooled | 3 | 3393 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.53, 1.21] |
| 5 Esophagitis Show forest plot | 5 | 21983 | Odds Ratio (M‐H, Random, 95% CI) | 0.93 [0.76, 1.14] |
| 5.1 5 mg | 5 | 9127 | Odds Ratio (M‐H, Random, 95% CI) | 0.96 [0.70, 1.30] |
| 5.2 pooled doses | 5 | 12856 | Odds Ratio (M‐H, Random, 95% CI) | 0.91 [0.70, 1.20] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Vertebral Fractures Show forest plot | 5 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 1.1 Prevention | 1 | 111 | Risk Ratio (M‐H, Random, 95% CI) | 2.43 [0.12, 49.43] |
| 1.2 Treatment | 4 | 2493 | Risk Ratio (M‐H, Random, 95% CI) | 0.63 [0.52, 0.77] |
| 2 Non‐Vertebral Fractures Show forest plot | 7 | Risk Ratio (M‐H, Random, 95% CI) | Subtotals only | |
| 2.1 Prevention | 1 | 111 | Risk Ratio (M‐H, Random, 95% CI) | 0.48 [0.10, 2.26] |
| 2.2 Treatment | 6 | 12847 | Risk Ratio (M‐H, Random, 95% CI) | 0.73 [0.61, 0.88] |
| 3 Lumbar Spine BMD, 5 mg dose, Final year Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 Prevention | 2 | 230 | Mean Difference (IV, Random, 95% CI) | 4.70 [3.74, 5.67] |
| 3.2 Treatment | 4 | 1908 | Mean Difference (IV, Random, 95% CI) | 4.53 [3.95, 5.11] |
| 4 Femoral Neck BMD, 5mg dose Final year Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.1 Prevention | 2 | 231 | Mean Difference (IV, Random, 95% CI) | 3.22 [2.35, 4.08] |
| 4.2 Treatment | 4 | 2106 | Mean Difference (IV, Random, 95% CI) | 2.60 [2.12, 3.09] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 3 Lumbar Spine BMD Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 3.1 => 3/5 | 5 | 1945 | Mean Difference (IV, Random, 95% CI) | 4.62 [4.05, 5.18] |
| 3.2 < 3/5 | 1 | 193 | Mean Difference (IV, Random, 95% CI) | 4.46 [3.60, 5.32] |
| 4 Femoral Neck BMD Show forest plot | 6 | Mean Difference (IV, Random, 95% CI) | Subtotals only | |
| 4.1 => 3/5 | 5 | 2143 | Mean Difference (IV, Random, 95% CI) | 2.62 [2.14, 3.09] |
| 4.2 < 3/5 | 1 | 194 | Mean Difference (IV, Random, 95% CI) | 3.24 [2.32, 4.16] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1 Distal radius Show forest plot | 1 | 436 | Mean Difference (IV, Random, 95% CI) | 0.9 [‐0.76, 2.56] |
| 2 Mid radius Show forest plot | 1 | 436 | Mean Difference (IV, Fixed, 95% CI) | 0.90 [0.35, 1.45] |
