Oral Bisphosphonate Therapy for Osteogenesis Imperfecta: A Systematic Review and Meta‐Analysis of Six Randomized Placebo‐Controlled Trials

Objective To assess the effectiveness and safety of oral bisphosphonates in increasing bone mineral density (BMD), reducing fractures, and improving clinical function in patients with osteogenesis imperfecta (OI). Methods Studies were eligible for inclusion if they were randomized controlled trials of directly comparing oral bisphosphonate therapy with placebo‐group in OI patients. Data synthesis regarding to bone mineral density as measured by dual‐energy X‐ray absorptiometry (DEXA), decreased fracture incidence, change in biochemical markers of bone and mineral metabolism, bone histology, growth, bone pain, quality of life, and others were assessed, and meta‐analysis done when possible. Results From 98 potential references and six randomized controlled studies a total of 263 participants receiving oral bisphosphonates and 143 placebo treatments contributed data to meta‐analysis. Pooled meta‐analysis of three studies suggested that there was significant difference between bisphosphonate treated group and placebo in number of patients with at least one fracture (mean difference 0.53, 95% confidence interval 0.32–0.89, P = 0.02). Pooled meta‐analysis of two studies suggested that significant difference was noted between bisphosphonate treated group and placebo in mean percentage change in spine BMD (T‐score) (mean difference 28.43, 95% confidence interval 7.09‐49.77, P = 0.009). The similar effect was shown in the term of mean change (Z‐score) in spine BMD. Conclusions Significant improvement in lumbar areal BMD in patients affected with OI has been shown when treated with oral bisphosphonates, even though only a small population was enrolled. We cannot draw a definite conclusion that the increase in BMD can be translated into fracture reduction and clinical functional improvement. The optimal method, dose, type, initiation, and duration of oral bisphosphonates therapy still remains unclear. Well‐designed, adequately‐powered, placebo‐controlled RCTs investigating the effects of oral bisphosphonates on fractures reduction and improvement in quality of life in both children and adults are studied here.


Introduction
O steogenesis imperfecta (OI) is an inherited disorder of connective tissue, caused by mutations in the genes that encode type I collagen [1][2][3] . OI is characterized by increased bone fragility of varying severity and low bone mass 4,5 . It commonly presents with joint hypermobility, blue or grey-blue scleral color, dentinogenesis imperfecta, and premature hearing loss. This genetic rare disease is almost always caused by mutations in one of the two genes encoded with type I αchains (COL1A1 and COL1A2) 6,7 . The incidence of osteogenesis imperfecta is estimated to be between one and two per 20000 people [8][9][10] . Normal bone matrix is composed of 10% non-collagenous proteins and 90% type I collagen fibers which provide bone resilience. Individuals with OI have less and/or poorer quality type I collagen than unaffected people, causing their bones to deform and/or fracture. Further refinement of these classifications was made with molecular genetic analyses 3 . Recently, the addition of OI types V, VI and VII have been proposed 9 . It is unclear whether OI types V, VI and VII will be classified with OI in the future as these individuals do not have evidence of type I collagen mutations.
Intravenous and oral bisphosphonates are first-line established therapy for the treatment of most patients with osteoporosis, with proven efficacy to reduce fracture risk at the spine, hip, and other nonvertebral skeletal sites 11 . Bisphosphonates act by inactivating osteoclasts, the cells that break down bone tissue, thereby inhibiting bone resorption 43 . There are two different types of bisphosphonates, nitrogenous and non-nitrogenous. Nitrogenous bisphosphonates including alendronate, eridronate, olpadronate, pamidronate, risedronate and zolendronate disrupt osteoclast formation, survival and cytoskeletal dynamics. Nonnitrogenous bisphosphonates including clodronate and etidronate and tiludronate initiate osteoclast apoptosis. The bisphosphonates vary in their efficacy and absorption when taken orally, making direct comparison challenging. They increase the areal bone mineral density (BMD) and reduce the incidence of osteogenesis imperfecta. Beneficial effects have also been reported in children with osteogenesis imperfecta [12][13][14][15][16][17][18]29 .
Several important quasi-randomized or randomized controlled trials presented increase of BMD and decrease of reported fracture rates in children or adults with osteogenesis imperfecta by cyclic intravenous neridronate and pamidronate [19][20][21] . But treatment with intravenous bisphosphonate needs medication at regular intervals at home or during hospital stays. This invasive and inconvenient medication administration unavoidably affects schooling, distracting parents from their work commitments. The intravenous administration is also unfriendly to the patients especially to the children. Oral treatment provides merits in terms of convenience, cost, and reduced individual distress. Therefore, there is still no definite conclusion as to whether oral administration of bisphosphonate can improve quality of life in adults and children with osteogenesis imperfecta, especially regarding the reduced fracture rate.
The purpose of this study was to conduct a systematic review and meta-analysis of recent studies to address if oral bisphosphonate therapy for osteogenesis imperfecta was equivalent to the placebo group regarding bone mineral density as measured by dual-energy X-ray absorptiometry (DEXA), decreased fracture incidence, lessening of deformity, reduced pain, and improved growth and mobility. Our hypothesis was that the oral bisphosphonate therapy would offer the patients with improved quality of life and reduced fracture rate than placebo group.

Inclusion and Exclusion Criteria
Studies were eligible for inclusion if they were randomized controlled trials that directly compared oral bisphosphonate therapy with placebo-group in OI patients, and measured primary outcomes of bone mineral density as measured by dual-energy X-ray absorptiometry (DEXA), decreased fracture incidence, lessening of deformity, reduced pain, and improved growth and mobility (Table 1). These variables were selected because at least half of the studies included each of these measures. Children (defined as age 0 to 18 years) and adults with OI diagnosed using accepted diagnostic criteria, based on clinical or laboratory findings, or both, were eligible. Individuals affected with all types of OI are included in this review. We excluded randomized controlled trials in which any enrolled patients received the bisphosphonate therapy using IV administration. Although pathology features harvested from the operated site would be the ideal outcome measure, this would require following large number of participants for decades. No   to 2019 using MESH terms, "osteogenesis imperfecta", "randomized controlled trial", and "Randomised Controlled Trials," were also carried out by the authors.

Data Collection
Two reviewers (Zhi-min Ying, Bin Hu) independently reviewed each abstract and extracted data. The references for each of these studies also were examined for other relevant studies. If there were any discrepancies, a third individual reviewed the articles. We extracted change between the two different therapies in bone mineral density as measured by dual-energy X-ray absorptiometry (DEXA), decreased fracture incidence, change in biochemical markers of bone and mineral metabolism, bone histology, growth, bone pain, quality of life. Two reviewers independently assessed each trial's risk of bias, assessing factors such as randomization, allocation concealment, blinding, completeness of outcome assessment, and selective reporting. Where necessary we contacted authors to obtain information on primary outcome factors.

Assessment of Risk of Bias in Included Studies
We assessed the methodological quality of the trials based on the method described by Jüni 25 . In addition, we described the generation of allocation sequence and concealment of allocation sequence as adequate, inadequate, or unclear. We also evaluated each trial for the degree of blinding and whether an intention-to-treat analysis was undertaken. The methodological quality ratings, and details as to why they assigned these ratings for each criterion was presented ( Table 2). The authors have presented methodological quality ratings, and details as to why they assigned these ratings for each criterion. However, studies were not weighted on the basis of their assigned methodologic quality.

Data Analysis
We converted all the different outcomes to standardized mean differences, calculating a standardized mean difference of hanged from base-line in treatment and control groups.
Where clinically useful, we estimated a benefit in units of percentage change since baseline from the standardized mean differences by estimating the pooled standard deviation from the means of the standard deviation of the outcomes in double-row and single-row groups for each study, and multiplying the standardized mean differences by this. We calculated statistical heterogeneity using a χ 2 test on N-1 degrees of freedom, with significance conservatively set at 0.10. We also assessed inconsistency I 2 using the formula [(Q-df)/Q] × 100%, where Q is the χ 2 statistic and df is its degrees of freedom, to describe the percentage of the variability in effect estimates due to heterogeneity. We considered a value greater than 50% as denoting substantial heterogeneity. For each study, relative risks (RRs) with 95% confidence intervals (CIs) and standardized or weighted mean differences with 95% CIs were calculated for dichotomous outcomes and continuous outcomes, respectively.
A fixed-effects or random-effects model was applied dependent on the heterogeneity of the studies. Quality appraisal was performed according to the CONSORT 2010 checklist and Cochrane scale was used to assess the risk of bias 26 . When heterogeneity was considered substantial, we explored its causes by carrying out pre-specified subgroup analyses where data were available; that is, subgroup by sex, tear size, base profiles, compliance, and adequacy of allocation concealment. Where there weren't clear clinical reasons or study methodology reasons for substantial heterogeneity between studies, we proceeded to meta-analysis using random effects models. All analyses were carried out in Review Manager 5 (Computer program. Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014.). When possible we used intention to treat data in analyses, but if these were not available we used, in order of preference, data from available data or per protocol analyses. Assessment of publication bias was by funnel plot.

Description of Studies
Of the 13 studies were identified and reviewed, six randomized placebo-controlled trials met the inclusion criteria and were included in this comparison 22,27-31 (see Fig. 1). Of these six randomized controlled studies, a total of 263 participants receiving oral bisphosphonates and 143 placebo treatments contributed data to meta-analysis. Further details of the characteristics of included studies can be obtained in Tables 3  and 4. Trials were excluded if they are not RCTs or they did not evaluate effect of medicine bisphosphonate on OI patients or they did not assess the clinical outcome as in the form of bone density or fracture reduction. Five trials are investigating the effect of oral bisphosphonate clinical results in adults with osteogenesis imperfecta while only one trial studied the effect of oral bisphosphonate in children with osteogenesis imperfecta. They all have two studies that evaluated alendronate, risedronate and olpadronate separately.

Risk of Bias in Included Studies
Despite differences in methodological quality, the results of each study were considered equally and were not weighted in the analysis. Further details of the methodological quality of the included studies can be found in Table 2.
Sakkers and Chevrel trials 27,28 were described as randomized, by computer-generated random numbers and were deemed adequate. The Chevrel trial was described as double-blinded (study personnel and participants). For all the trials, an intention-to-treat analysis was undertaken. Two participants (one placebo and one treatment) withdrew from the trial but were accounted for in the final analysis. It was also reported that an intention-to-treat analysis were performed in the Chevrel trial.

Number of Patients with at Least One Fracture
Each of the five trials reported on this outcome including three studies in children 27,30,31 and one trial in adults 28 . The Sakkers trial reported a 31% reduction in relative risk for fracture after treatment with oral olpadronate, and, when analyzed in the review, this produced a hazard ratio of 0.69 (95% CI 0.52 to 0.91) and a significantly decreased fracture number and relative rate (RR) 0.40 (95% CI 0.24 to 0.69) 27 . The Chevrel trial showed that the incidence of vertebral and peripheral fractures was not significantly different between the alendronate and placebo groups. Two vertebral and 17 peripheral fractures occurred in 11 patients in the placebo group versus no vertebral and 17 peripheral fractures in 10 patients in the alendronate group 28 . Similar results are further confirmed by Ward et al. 30 . They reported that 83% of the ALN patients and 92% of placebo patients sustained at least one investigator-reported fracture (P = 0.070).
Pooled meta-analysis of three studies 27,30,31 suggested that there was significant difference between bisphosphonate treated group and placebo group in number of patients with at least one fracture (mean difference 0.53, 95% confidence interval 0.32-0.89, P = 0.02) (Fig. 2).  The study was coordinated and organized under the control of an independent steering committee, whose members were not involved in the study as investigators. Intention-to-treat.

Bishop 2013 (Risedronate)
Method not stated Patients were stratified by age (4-9, 10-15 years) and randomly assigned to receive treatment for 1 year with risedronate tablets or placebo in a 2:1 ratio by a telephone-based interactive voice response system in several permuted blocks of ten to 12 Pooled meta-analysis of two studies 29,30 suggested that significant difference was noted between bisphosphonate treated group and placebo in mean percentage change in spine BMD (T-score) (mean difference 28.43, 95% confidence interval 7.09-49.77, P = 0.009). The similar effect was shown in the term of mean change (Z-score) in spine BMD. Meta-analysis of three studies [29][30][31] showed that the bisphosphonate treated group presented significantly higher Zscore compared with placebo group (mean difference 5.70, 95% confidence interval 1.30-10.11, P = 0.01) (Fig. 3).

Mean Percentage Change in Total Femur BMD
Only two trials reported on this outcome 28,29 . Chevrel et al. reported that the mean observed change in total femur BMD in the alendronate treated groups and placebo groups was +0.024 AE 0.004 and − 0.002 AE 0.005 g/cm 2 , respectively. This increase in the alendronate group was significantly greater compared with placebo groups (P = 0.001). The increase in the alendronate group was significantly greater than that in the placebo group; total femur BMD, MD 3.00 (95% CI 2.73 to 3.27) 28 . However, another study conducted by Rauch et al. showed that there was no significant difference between the risedronate treated groups and placebo groups for changes in DXA parameters for hip [Mean (SD), Risedronate: 12.4 (10.8); Placebo: 6.5 (5.9), P = 0.11] 29 . As these two trials involve two different populations (children and adults), the meta-analysis cannot be performed.

Change in Biochemical Markers of Bone and Mineral Metabolism
There were five clinical trials that reported on this outcome [27][28][29][30][31] . There were different markers selected for study in the included trials so direct comparison could not be performed. Sakkers et al. reported no significant change was found in urine or serum markers (serum concentrations of creatinine,γ-glutamyl transpeptidase, aspartate and alanine aminotransferases) between the olpadronate and placebo groups 27  aminotransferase, and complete blood count 29 . A significant rise in PTH levels was observed only at month 3 in the ALN treated groups compared with the placebo group (P = 0.049). Also statistically significant difference was found in the ALN treated groups in 1, 25-dihydroxy vitamin D at month 24 compared with the placebo group (P = 0.048). uNTx levels decreased significantly in the groups treated with ALN than in the placebo groups (P = 0.001). However, there was no significant treatment difference between the two groups with regard to changes in serum total alkaline phosphatase activity 30 . Significant mean percentage decreases were both observed in urine NTx/creatinine and serum bone specific alkaline phosphatase concentration at 3 and 6 months in the risedronate group, while significant difference was noted between the risedronate and placebo groups at months 6 and 12 for both markers 31 .

Bone Histology
There were two studies 29,30 that showed histology results. Histomorphometric analysis of these samples by iliac bone biopsies showed that bone formation and resorption parameters, bone size, cortical width, or the amount of trabecular bone were similar between the risedronate and placebo groups 29 . Transiliac histomorphometric safety data showed the ALN group had a significant reduction in osteoid volume per bone volume after treatment. This difference between treatment groups was highly significant in the semiquantitative evaluation of iliac bone samples which were obtained at the end of the treatment period revealing the presence of at least one large osteoclast (>50 mm). However, no significant treatment difference was found between the two groups regarding the presence of calcified cartilage 30 . Type I, Autosomal dominant Fractures with little or no limb deformity, blue sclera, normal stature, hearing loss, dentogenesis imperfecta rare; Type II Autosomal dominant Lethal perinatal type: undermineralized skull, micromelic bones,"beaded" ribson x-ray, bone deformity, platyspondyly; Type III Progressive deformingtype: limb deformities, sclera huevaries, very short stature, dentogenesis imperfect common; Type IV Sclerae blue, grey, grey/blue, or white, mild/ moderate limb deformity with fracture, variable short stature, dentogenesis imperfect common, some hearing loss.  metacarpal bone 29 . The mean midline vertebral height was also similar between the ALN and placebo groups (P = 0.444) 30 .
reported that the pain score was similar in the two groups from 0 to 30 months and reported end-of-trial data that showed an increase at final follow-up with alendronate treated groups (mean 1.30, 95% CI 0.14 to 2.46) 28 . Also, another study confirmed the results that there was no significant difference between these two groups 29 . However, one trial completed by Ward et al. reported that significantly fewer ALN patients experienced bone pain at month 24 than at baseline (P = 0.001). But there was no significant difference between the ALN and placebo groups in the percentage of patients who experienced bone pain (P = 0.065) 30 . As these two trials involving different populations (children and adults), the meta-analysis cannot be performed.

Quality of Life
Five included trials evaluated at least one quality of life outcome 22,[27][28][29][30] . Sakkers reported that no changes were found in the changes of grip or hip flexor strength or mobility when comparing olpadronate to placebo, and the difference in changes was found to be similar between the two groups 27 . Hearing was assessed in the study by Chevrel et al., which found no difference between the alendronate-treated group and placebo group (mean: -0.10, 95% CI -2.88 to 2.68) 28 . No statistically significant difference was observed by two studies in self-care or mobility functional skills scaled scores and in grip force 29,30 . We found only slight differences in quality of life, which was measured using self-perception profile for children (SPPC) and health-utility index (HUI) in favor of the bisphosphonate group. A small but not significant decrease in pain was detected in the bisphosphonate group 22 .

Discussion
S ix randomized controlled studies were included in our systematic and meta-analysis 22,[27][28][29][30][31] . Two studies conducted by Sakkers et al. and Kok et al. 22,27 reported on the same patient groups with olpadronate but involving different results: the former focuses on skeletal effects and functional outcomes while the latter one focuses on quality of life. Five trials are investigating the effect of oral bisphosphonate clinical results in children with osteogenesis imperfecta while only one trial studies the effect of oral bisphosphonate in adults with osteogenesis imperfecta 28 . Five studies enrolled children 22,27,[29][30][31] and one enrolled adults 28 .
All the included studies investigated changes to varying degrees in BMD, fracture rate, markers of bone turnover, growth, pain, and quality of life with bisphosphonate therapy. For a summary of outcomes reported by different trials, see Table 4. All studies assessing BMD independently reported significant increases after treatment with either oral or IV bisphosphonate and at separate sites (spine, hip, femur) [27][28][29][30][31] . However, it is difficult to compare all these trials directly as different populations were included (adults versus children; for accurate comparisons children cannot be compared to adults due to high bone turnover during childhood and adolescence and open epiphyses). Additionally, different reporting indices were used (Z-score versus total BMD). But  data from some of these studies can contribute to the analysis when the same population and reporting indices were employed as well as the same locations were measured. Pooled meta-analysis of two studies 29,30 suggested that significant difference was noted between bisphosphonate treated group and placebo in mean percentage change in spine BMD (mean difference 28.43, 95% confidence interval 7.09-49.77, P = 0.009). A similar effect was shown in the term of mean change (Zscore) in spine BMD by three studies 29,31,32 . Both randomized 29,31,32 and non-randomized 33 studies all reported that oral bisphosphonate therapy has shown increases in BMD. The majority of the patients enrolled by Bishop et al. and Sakkers et al. were children with mild forms of the disease. They concluded that the oral bisphosphonate treatment is not suited to severe types of OI, but is suited to mild types. The effect of oral bisphosphonate in increasing and reducing fracture rates in BMD are consistent with those treated by intravenous bisphosphonate in uncontrolled, observational studies 12,14 and randomized controlled trial 19 . Each of the four trials reported on this outcome including three studies on children 27,30,31 and one trial on adults 28 . As mentioned before, an increase in the growth and BMD in children and adolescents with OI, together with the trend for decreased fractures, makes data comparison from all the trials extremely difficulty. The metaanalysis was performed when the same indices were reported among the studies. Pooled meta-analysis of three studies 27,30,31 suggested that there was significant difference between bisphosphonate treated group and placebo group in the number of patients with at least one fracture (mean difference 0.53, 95% confidence interval 0.32-0.89, P = 0.02). Improvement in the height of vertebral bodies and cortical width has been shown among the children with OI treated with IV bisphosphonate 19,[35][36][37][38] . Similar increases were observed by DiMeglio et al. in terms of the height of vertebral bodies between high-dose oral bisphosphonate and IV treatment 39 . In contrast, the oral bisphosphonate (alendronate, risedronate and olpadronate) therapy was successful in increasing the lumbar BMD but was not accompanied by improvement in other skeletal parts such as metacarpal and iliac cortical width 27,29,30,32 . One explanation for this phenomenon is that primary and secondary spongiosa are the main component of the vertebral body. More and more primary spongiosa are converted to secondary spongiosa during the child growing phase 34 .
Mild and hardly noticeable change in the bone metabolism has been reported among the patients with OI treated with oral bisphosphonate even though the definite treatment effect on lumbar spine areal BMD has been shown [27][28][29][30]32 . Sakkers et al. reported that no significant change was found in urine or serum markers (serum concentrations of creatinine,γglutamyl transpeptidase, aspartate and alanine aminotransferases) between the olpadronate and placebo groups 27 . Rauch et al. also showed that there was no significant differences between the risedronate-treated and placebo groups in terms of the serum levels of phosphorus, creatinine, 25-hydroxy vitamin D, 1,25-dihydroxy vitamin D, PTH, urinary calcium/ creatinine ratio, alanine aminotransferase, aspartate aminotransferase, and complete blood count 29 . No commonly accepted biochemical markers of bone turnover are used, but it is a public assumption that they can act as a proxy monitoring the efficacy of therapy. The varied markers selected for study in the trials made the direct comparison impossible. These different biological bone turnover markers can help doctors to assess the drug therapeutic effect of the bisphosphonates, drug dosing, and participant compliance.
Bone pain and mobility are the two main key clinical indices that have direct influence on the quality of life of the participants. Improvements in terms of these parameters have been shown in several studies 19,34,40,41 . However, no obvious change in mobility was noted after 2 years of oral bisphosphonates 27,30 and no improved to bone pain was detected in the trial by Bishop et al. using oral risedronate 32 . One of the possible reasons for this phenomenon is that oral bisphosphonates provided less benefit than IV treatment due to its lower therapeutic effect by oral administration. However, one trial conducted by Chevrel et al. found no difference in self-reported pain scores, with the exception of increased pain with bisphosphonates at 36 months. They also assessed the hearing and found no change in Rinne testing 28 . Taken together, no consistent improvements in these quality-of-life indicators with bisphosphonate administration was demonstrated by the current available literature.
Importantly, oral bisphosphonate administration was associated with few adverse effects. Ward et al. reported that mean bisphosphonate oral bioavailability is similar among the child and adult participants and the individual oral bioavailability varies as much as ten-fold 42 . Chevrel et al. reported that gastrointestinal symptoms were more common in the group of patients treated with alendronate than in those placebo ones, although these symptoms were not responsible for treatment withdrawal. Their results are in contrast with these other four studies 27,[29][30][31] , which indicated that gastrointestinal symptoms were not more common in ALN groups than in placebo groups. Based on current evidence, oral bisphosphonate administration treated for 1-3 years appears to be safe and well tolerated in the patients affected with OI. Most of the adverse effects reported in the trials are few and minor including gastrointestinal complaints, fever, headache, nausea, arthralgia, and others, and the drugs are generally well tolerated.

Conclusion
Significant improvement in lumbar areal BMD in patients affected with OI have been shown when treated with oral bisphosphonates, even though only a small population was enrolled. We cannot draw a definite conclusion that the increase in BMD can translate into fracture reduction and clinical functional improvement. The optimal method, dose, type, initiation and duration of oral bisphosphonates therapy still remains unclear. Well designed, adequately-powered, placebo-controlled RCTs investigating the effects of oral bisphosphonates on fracture reduction are important to determine improvements to quality of life in both children and adults.