IN this issue of Anesthesiology, Paul et al.  report results of their meta-analysis on femoral nerve block (FNB) analgesia and outcome after total knee arthroplasty.1The authors conclude that FNB (plus patient-controlled analgesia) is a good alternative to patient-controlled analgesia alone or epidural for postoperative analgesia in patients having total knee arthroplasty and that the current evidence does not support using either a sciatic nerve block or continuous FNB (CFNB) in addition to a single-injection FNB in this patient population.

The report in this issue of Anesthesiology by Paul et al.  is likely to create controversy because their conclusions are at odds with current practice in many centers where CFNB is an apparent “gold standard” for postoperative analgesia in patients having total knee arthroplasty. Regardless, with clearly defined inclusion criteria, the authors identified 23 randomized controlled trials (1,016 patients) that compared FNB with opioid-based patient-controlled analgesia or epidural analgesia, but suprisingly only two randomized controlled trials that directly compared CFNB with single-injection FNB. To examine theoretical differences between the single FNB and CFNB, the authors used a Bayesian random-effects model that allows estimation of comparative treatment effects for interventions that were not directly compared in the original studies (sometimes called “indirect comparisons”). This sophisticated analytical approach relies on several assumptions, including that the groups are similar in respect to other possible confounding factors; yet even when this is the case, indirect comparisons are not randomized comparisons and, as Paul et al. , suggest, carry with them the limitations of observational data. Regardless, the authors cannot be criticized for being forced to impute data (variability and ranges) because the available original studies were simply not available or forthcoming with this information.

Although there is a plethora of literature on the efficacy of various FNB and CFNB techniques and their modifications on quality of analgesia, there are only a few randomized controlled trials that compare a single-injection FNB (or its equivalent) with CFNB.2–4The lack of randomized controlled trials on FNB versus  CFNB is particularly surprising, given their widespread practice in patients having total knee arthroplasty. With regard to analgesic outcome, the two relevant studies reported significant improvements in analgesia at rest and movement as well as the opioid-sparing benefits of CFNB.3,4Ilfeld et al.  reported shorter time to discharge readiness in patients receiving CFNB.3However, actual “length of stay” was not shorter as this is a difficult outcome variable to assess, and studies may not have been designed to look at early discharge. For instance, in Ilfeld et al. , mean (±SD) duration of hospitalization was 3.6 (±0.6) days in the CFNB group and 3.5 (±0.6) days in the single-injection FNB (placebo) group (P = 0.74). However, patients treated with 4 days of perineural ropivacaine attained three major discharge criteria in a median (25th–75th percentiles) of 25 (21–47) h compared with 71 (46–89) h for those in the placebo group (estimated ratio = 0.47, 95% confidence interval: 0.32–0.67; P < 0.001). Unfortunately, their study protocol required patients to remain hospitalized until at least postoperative day 3, even if they were discharge-ready before then.3Salinas et al. , also included in the meta-analysis, did not find a difference between treatment groups in actual discharge because their physical therapists and surgeons were simply not prepared to discharge patients early.4In addition, the analgesic benefits of the CFNB were likely lessened by their study protocol where the CFNB infusion (48 h) was provided after a single injection of ropivacaine 0.5% which often lasts more than 24 h.

Two important trials did not meet the criteria for inclusion in the meta-analysis by Paul et al.  The first is a randomized, double-masked, placebo-controlled trial comparing a single-injection FNB (equivalent) to a single-injection FNB plus 4 days of perineural infusion3; the second is a comparison of epidural, CFNB, and intravenous patient-controlled analgesia on surgical outcome and duration of rehabilitation.5Because Paul et al.  included studies that “compared the analgesic effects of epidural or patient-controlled analgesia opioid analgesia versus  FNB (single-injection or continuous) on analgesia outcomes after total knee arthroplasty” (intervention), the aforementioned study of Ilfeld et al.  did not make it to the meta-analysis because the comparison group received an overnight perineural infusion in addition to the initial mepivacaine FNB. Likewise, the report by Capdevilla et al.  was disqualified probably because it included knee surgeries (arthrolysis) other than arthroplasty.5Of note, both Ilfeld et al.  and Capdevilla et al.  reported analgesic and/or outcome advantages of CFNB. The issue of single-injection FNB versus  CFNB should also be viewed from a logic of the clinical practice. A single injection of predetermined concentration and type of local anesthetic in FNB results in quadriceps muscle weakness and inability to ambulate for most of the analgesic duration. In contrast, the ability to control the concentration or local anesthetic dose, type of local anesthetic, and infusion regimen theoretically should allow better preservation of the quadriceps function with CFNB. Therefore, an injection of a long-acting local anesthetic may be well suited for the duration of surgery and postanesthesia care unit stay, but the use of shorter-acting local anesthetic, followed by adjustable CFNB, allows mobilization of the patients the afternoon of surgery. This is important because there is an increasing trend to decrease time to early ambulation (even the same day of surgery). As an example, the study by Hirst et al.  (included in the meta-analysis) used bupivacaine 0.5% with epinephrine for the initial surgical blocks, followed by only 48 h of perineural infusion.2Because the duration of nerve blockade with bupivacaine is very long, the difference between treatment groups was (predictably) small. Therefore, although the Hirst et al.  study met the criteria for inclusion in the meta-analysis, it does not really contribute clinically relevant data on whether CFNB provides patient benefits over a single-injection FNB in today's orthopedic environment.

In conclusion, the meta-analysis by Paul and colleagues is well conducted and presented and includes a well-designed systematic review of the literature. The authors appropriately conclude that analgesia after single-injection FNB is superior to intravenous patient-controlled analgesia in patients having total knee replacement, as well as that the available data are unclear as to whether CFNB and/or sciatic nerve block confers additional analgesic and rehabilitation benefits when added to a single-injection long-acting FNB. Paul et al.  should be commended for setting a nice standard for the reporting of Bayesian meta-analysis and identifying the need for randomized controlled trials examining CFNB and single-injection FNB.

*Department of Anesthesiology, College of Physicians and Surgeons, Columbia, University, New York, New York. †Department of Anesthesiology, Wake Forest University School of Medicine, Winston-Salem, North Carolina. ‡Department of Anesthesiology and Critical Care Medicine, Lapeyronie University Hospital and Montpellier School of Medicine, Montpellier, France. §Division of Regional Anesthesia and Acute Pain Medicine, Department of Anesthesiology, University of California, San Diego, San Diego, California.


Paul JE, Arya A, Hurlbert L, Cheng J, Thabane L, Tidy A, Murthy Y: Femoral nerve block improves analgesia outcomes after total knee arthroplasty: A meta-analysis of randomized controlled trials. Anesthesiology 2010; 113:1144–62
Hirst GC, Lang SA, Dust WN, Cassidy JD, Yip RW: Femoral nerve block. Single injection versus continuous infusion for total knee arthroplasty. Reg Anesth 1996; 21:292–7
Ilfeld BM, Le LT, Meyer RS, Mariano ER, Vandenborne K, Duncan PW, Sessler DI, Enneking FK, Shuster JJ, Theriaque DW, Berry LF, Spadoni EH, Gearen PF: Ambulatory continuous femoral nerve blocks decrease time to discharge readiness after tricompartment total knee arthroplasty: A randomized, triple-masked, placebo-controlled study. Anesthesiology 2008; 108:703–13
Salinas FV, Liu SS, Mulroy MF: The effect of single-injection femoral nerve block versus  continuous femoral nerve block after total knee arthroplasty on hospital length of stay and long-term functional recovery within an established clinical pathway. Anesth Analg 2006; 102:1234–9
Capdevilla X, Barthelet Y, Biboulet P, Ryckwaert Y, Rubenovitch J, d'Athis F: Effects of perioperative analgesic technique on the surgical outcome and duration of rehabilitation after major knee surgery. Anesthesiology 1999; 91:8–15