Fifty-seven (n = 2,071) studies were included in the review: 15 RCTs (n = 647); six comparative studies (n = 640); 20 case series (n = 766); and 16 case reports (n = 18). The studies were generally considered to be of average quality. Most studies reported no losses to follow-up, but power calculations and intention-to-treat analyses were not reported in most studies. Nineteen studies (n=989; six RCTs, n = 306) were of anaesthesia and 38 studies (n=1,082; nine RCTs, n=341) were of analgesia.
Five of the six RCTs examined anaesthesia for breast surgery and one assessed herniorrhaphy.
Safety: Two RCTs assessed perioperative mortality and reported no deaths. Five RCTs assessed postoperative nausea and vomiting; the pooled relative risk showed significantly lower incidence in the paravertebral block groups (relative risk 0.25, 95% CI: 0.13 to 0.50; three RCTs). Results were also reported for a wide range of other safety outcomes, including: pleural puncture; pneumothorax; epidural spread; blood loss; urinary retention; local anaesthetic toxicity; bradycardia and hypotension; and other complications. None of these displayed a significant difference between groups reported by an RCT.
Efficacy: Four RCTs and eight other studies of breast surgery reported rates of effective block ranging from 80% to 100% (range in RCTs was 93.2% to 100%). One small RCT in herniorrhaphy found significantly higher rates of effective block in the paravertebral block group compared to the field block group (p<0.01). Eight studies, including two RCTs, reported rates of conversion to general anaesthesia. These ranged from 0 to 33% (range in RCTs was 0 to 13.3%). One RCT reported that 10% of paravertebral block patients required intraoperative re-block. Hospital stay was comparable between groups in one RCT and significantly shorter in the paravertebral block group in another. High levels of patient satisfaction were reported in two RCTs. This was significantly higher than in the general anaesthesia group in one RCT (p = 0.008). There was evidence that both operative (five RCTs) and postoperative (five RCTs) use of a range of analgesic medication was equivalent or lower in the paravertebral block groups than in the comparator groups.
Other efficacy outcomes reported were: number of dermatomes blocked, bilateral block, time to effective block, duration of surgery and duration of anaesthesia. In none of these was a significant difference between groups reported by an RCT.
Of the nine RCTs, two assessed each of breast surgery, cardiovascular surgery and cholecystectomy and thoracotomy; one assessed herniorrhaphy
Safety: Three RCTs assessed mortality and reported no deaths. No significant differences in postoperative nausea and vomiting were reported (four RCTs). One RCT recorded an incidence of vascular puncture of 8.7%. No RCTs reported incidences of pleural puncture or pneumothorax. One RCT reported significantly lower arterial blood pressure in the paravertebral block group compared to the thoracic epidural group (p<0.001); no other significant differences in blood pressure or heart rate were found (two RCTs). There was little or no RCT evidence of the following complications: epidural spread, urinary retention, local anaesthetic toxicity and other complications.
Efficacy: Rates of effective block in RCTs ranged from 80% to 92%. Two RCTs reported equivalent numbers of dermatomes blocked in paravertebral block and comparator groups. One RCT reported a lower rate of unintended bilateral block in the paravertebral block group (30% compared to 100% of epidural patients). Time to effective block did not differ between paravertebral block and thoracic epidural groups in one RCT.
Conflicting results for use of supplemental medication were found: one RCT reported significantly higher use in the paravertebral block group and one reported no differences, while four reported lower use in the paravertebral block group. A similarly conflicting pattern was observed for reported postoperative pain. No significant differences were found for duration of surgery or duration of analgesia, or patient satisfaction.