A total of 5 RCTs (n=270), 8 prospective observational studies (n=390) and 10 retrospective observational studies (n=779) were included in the review.
Two RCTs (n=142) evaluated intra-articular facet joint injections; 3 prospective observational studies (n=189) and 2 retrospective observational studies (n=133) evaluated lumbar facet joint injections.
One RCT (n=73) and one prospective observational study (n=100) evaluated medial branch blocks.
Two RCTs (n=55), and 4 prospective (n=101) and 8 retrospective observational studies (n=646) evaluated medial branch neurotomy.
Two hundred and forty-six articles reported complications (no details were reported).
Facet joint injections.
The evidence for lumbar intra-articular facet joint injection was moderate in the short term (the only RCT reported negative findings but the reviewers re-classified the findings as positive; 4 of 5 observational studies reported positive findings) and limited in the long term (the only RCT reported negative findings; 2 of 5 observational studies reported positive findings).
There was negative evidence for cervical intra-articular facet joint injections (the only RCT reported negative findings).
Medial branch blocks.
The evidence for cervical medial branch blocks using local anaesthetics plus steroids was moderate (the only RCT reported positive short- and long-term findings).
The evidence for lumbar medial branch blocks using local anaesthetics plus steroids was moderate (the only observational study reported positive short- and long-term findings).
Facet joint neurotomy.
The evidence for radiofrequency neurotomy of medial branch nerves was moderate to strong: positive short- and long-term findings were reported for studies of cervical branch neurotomy (the single RCT plus 2 of 4 observational studies), lumbar spine neurotomy (the single RCT), lumbar branch neurotomy (3 of 6 observational studies) and thoracic cervical branch neurotomy (1 of 2 studies).
Complications.
The most common complications were related to the placement of needles and drugs. Potential complications were listed. The authors of the review stated that quantitative data were not generally reported.