Fifteen articles (n=1,062) were included: 1 cohort study (n=79), 11 prospective case series (n=435) and 3 retrospective case series (n=548).
The quality of evidence was generally low as nearly all of the studies were case series, with high potential for confounding. Many studies (9 of the 15) administered (unspecified) cointerventions and several studies (5 of the 15) had high (37 to 50%) or undisclosed drop-out rates. Some used questionnaires of limited relevance and some were poorly reported. Some publication bias might have occurred as positive studies were more likely to have been published. Moreover four potentially relevant studies were excluded by the language restriction. There was a high degree of clinical heterogeneity between the studies.
Pain.
All studies apparently reported this outcome.
A cohort study (n=79) reported a statistically significant decrease in pain score in the intervention group, from 19 on a 25-point scale at baseline to 9 at 24 hours post-procedure (-52.6%), while the score in the control group, who received conservative treatment, changed from 20 to 19 (-5%). In 4 case series (n=284), there was a statistically significant decrease in VAS score from 7.8 on a 10-point scale at baseline to 3.1 within 24 hours of the procedure (-60.3%). Another case series (n=16) reported a decrease of over 50% for 44% of patients at 3 days post-procedure.
Most studies reported a statistically significant improvement in pain at long-term follow-up (mean 17 months, range 6: to 65), though 5 studies reported a non significant increase in pain score during follow-up.
Quality of life.
Eight studies reported this outcome.
The cohort study reported a 29% improvement in physical functioning (by the Barthel index) on day 1, with no improvement in the conservatively treated group. However, there was no difference between the groups at 6 or 12 months.
In 7 case series, the results for quality of life were mixed. Overall, most of the studies reported an improvement, though it was unclear in some cases whether this was statistically significant.
Safety.
All studies reported this outcome.
The authors stated that over all 15 studies a mean of 2.4% of patients experienced clinical complications (range: 0.4 to 15.4%). However, this estimate did not include cement leakage, which was reported in 3% to 76% of patients (13 studies). New fractures were reported in a mean of 17% of patients (range: 5.5 to 52) over long-term follow up (9 studies). One study reported no statistically significant evidence that the risk of vertebral fracture was higher in the vicinity of a cemented vertebra than elsewhere. Another study reported a significant increase in new fractures near the cemented vertebra.