Eight randomised controlled trials (RCT) (2,391 patients) were included in the review. Random sequence generation was described in four trials but none of the studies adequately reported allocation concealment. Although two studies were stated to be double-blind, the blinding process was not fully described and three trials did not describe any blinding of participants and personnel. Outcome assessors were blinded in four trials. All the trials were judged to be free of selective outcome reporting but losses to follow-up were described adequately in two RCTs.
There were statistically significant benefits observed with treatment with teriparatide compared with placebo in spine BMD (WMD 8.14%, 95% CI 6.72 to 9.55%; eight RCTs, 2,206 participants) and total hip BMD (WMD 2.48%, 95% CI 1.67 to 3.29%; seven RCTs, 1,303 participants).
Of the studies that evaluated teriparatide with hormone replacement therapy compared with hormone replacement therapy alone, combined treatment was associated with significant improvements in lumbar spine BMD (WMD 10.98, 95% CI 10.81 to 11.16%) and total hip BMD (WMD 3.65%, 95% CI 3.53 to 3.76%). Statistically significant benefits were also observed with teriparatide and alendronate in BMD at the lumbar spine (WMD 5.22%, 95% CI 2.58 to 7.87) but no significant differences between these treatments was observed in BMD at the total hip site.
Patients who received teriparatide with calcium supplementation had significantly greater improvements in BMD measured at the hip bone (WMD 2.48, 95% CI 1.67 to 3.29; seven trials, 1,303 participants; Ι²=89%). Furthermore, patients whose intake of calcium was more than 1,500mg had significant improvements in hip BMD (WMD 3.72, 95% CI 3.42 to 4.03; two studies, 908 participants; Ι²=26%). There were no differences observed between patients who received teriparatide with calcium supplementation of less than 1,500mg/day and patients who received calcium supplementation alone; there was significant heterogeneity for this outcome (Ι²=87%). There were significant benefits observed with total spine BMD with calcium supplementation (WMD 8.14%, 95% CI 6.72 to 9.55) with significant statistical heterogeneity (Ι²=94%). Significant results were also observed with calcium supplementation both greater than 1,500mg (WMD 9.83, 95% CI 8.43 to 11.23; three trials; Ι²=90%) and less than 1,500mg (WMD 6.69, 95% CI 3.11 to 10.28; five trials; Ι²=95%).
There were no significant differences observed between patients who received treatment for durations longer than 18 months and patients who received treatment for less than 18 months. No significant differences between patients who received treatment for 24 months compared with patients who received short-term treatment.
Treatment with teriparatide was associated with a 38% reduction in risk of non-vertebral fracture (RR 0.62, 95% CI 0.44 to 0.87; three RCTs; 1,842 participants), and a 70% risk reduction of vertebral fracture (RR 0.30, 95% CI 0.21 to 0.44; three RCTs, 1,452 participants), with no statistically significant heterogeneity observed for these outcomes.
Visual appraisal of the funnel plots showed some evidence of asymmetry, indicating there was some potential for publication bias.