IN this issue of Anesthesiology, Hadzic et al. 1report on hand surgery outpatients randomly assigned to receive general anesthesia (GA) or infraclavicular nerve block. The specific aim was to compare discharge times (a 97-min difference favoring infraclavicular); the secondary aim was to compare bypass of the postanesthesia care unit (PACU; 76% infraclavicular vs. 24% GA). This study promotes the use of regional anesthesia (RA) to shorten the outpatient stay.
Based on these findings, two questions are raised: (1) What is the recent evidence indicating that RA use will reduce discharge times, and (2) what criteria should be used to delineate the PACU bypass milestone?
In clinical practice, many orthopedic procedures have been converted from same-day admit to same-day discharge. Converting admissions into same-day discharges is one of two important strategies in reducing duration of stay. Duration of stay reduction via same-day discharge has recently been demonstrated by White et al. 2; this randomized trial compared a popliteal block with placebo infusion against the same block with a continuous infusion; both treatments were combined with GA. All placebo patients (10 of 10) were admitted for pain control, versus 4 of 10 of treatment group patients (P = 0.03). A multidisciplinary approach is necessary to convert same-day admissions into same-day discharges in clinical practice.3
Duration of stay reductions on the day of surgery are more challenging because staffing patterns and management styles vary across institutions. However, discharge time values within institutions are useful, especially in studies that involve large data sets of multiple procedures and capture clinical practice patterns. In a prospective case series (n = 479), Pavlin et al. 4noted a 90-min reduction of discharge times when peripheral nerve block (n = 72) was used versus GA (n = 230). This time savings is consistent with that reported by Hadzic et al. 1In a follow-up study (n = 175), Pavlin et al. 5showed that (1) maximum pain score predicted recovery time, (2) intraoperative fentanyl use predicted postoperative fentanyl use (in women), and (3) cumulative fentanyl dose predicted postoperative nausea and vomiting. Our group’s work has shown that each postoperative nursing intervention is associated with a 27- to 45-min delay in discharge,3a notion corroborated by Kitz et al. ,6who showed that postoperative administration of controlled analgesics is more costly than the administration of noncontrolled analgesics. These studies refute the intuitive notion that simply administering more intraoperative opioids via conventional bolus methods will prevent discharge time increases. To my knowledge, advanced opioid dosing techniques (e.g. , target-controlled infusions) have not been compared with RA techniques with respect to analgesic outcome differences.
Recently developed target-controlled infusion technologies7could in theory obviate the need for RA techniques to produce an ideal analgesic recovery profile after invasive outpatient orthopedic surgery. This may be especially true if target-controlled opioid dosing was associated with lower rates of postoperative nausea and/or vomiting (PONV) and other opioid-related side effects. However, limited evidence available shows high incidences of PONV after both target-controlled remifentanil (27% nausea, 10% vomiting)8and alfentanil (30% nausea) infusions.9The alfentanil target-controlled study group was compared to patients receiving conventional perioperative morphine boluses, and both treatment groups had a postoperative nausea rate of 30%.9Separating the therapeutic index of opioid analgesia and the “toxic index” of opioid-induced PONV (and other undesirable side effects for outpatients) may prove to be a daunting task. If the following scenarios were to occur, a significant clinical advance would be made, perhaps obviating the pressing need for labor-intensive RA analgesic procedures in complex orthopedic outpatient procedures: (1) a synthetic parenteral opioid that does not produce PONV in therapeutic analgesic doses; (2) appropriate supporting technology for a target-controlled infusion with this opioid; and (3) a reliable oral opioid with low diversion potential (i.e. , low street value) that produces therapeutic analgesic levels free from nausea/vomiting, somnolence, respiratory depression, pruritus, urinary retention, and constipation. The scenario described above does not even account for the PONV risk associated with volatile agent use when compared with the use of total intravenous propofol GA.
Opioids are not the sole analgesic option for outpatient surgery; multimodal analgesia is an important concept. Pavlin et al. 5showed that intraoperative parenteral ketorolac predicted (1) lower use/consumption of perioperative fentanyl, (2) less PONV, and (3) faster discharge time. Whether parecoxib10,11has a similar impact as does ketorolac in reducing PONV and providing faster discharge time remains to be seen. In either case, there is likely a procedural pain threshold above which it seems unlikely that nonsteroidal antiinflammatory drugs would be sufficient, rendering nerve blocks potentially valuable in reducing discharge times and avoiding unplanned admissions. These “procedural pain thresholds” require further research.
Two recent retrospective studies of outpatient anesthesia for hand surgery showed RA patients to have faster discharge times. In one, intravenous regional (Bier) block patients had faster discharge times than did GA and axillary block patients (by 32 and 48 min, respectively)12; in another, wrist block patients for carpal tunnel release were discharged sooner versus Bier block and GA patients (by 23 and 65 min, respectively).13The GA groups in both studies used simple bolus dosing of opioids, as did the current study by Hadzic et al. 1
Discharge time improvements shown by the aforementioned hand surgery studies have not yet translated to other outpatient orthopedic procedures. While calling for potential change in their institution’s discharge processes, Collins et al. 14reported a 7-min improvement in hospital discharge time when an RA program was developed for outpatient foot surgery. In another randomized trial (n = 60, knee arthroscopy patients), Jankowski et al. 15provided spinal, psoas-compartment lumbar plexus block, or GA and found no differences in discharge times, but these authors required voiding in all patients, a requirement which may no longer be necessary in low-risk patients undergoing regional/neuraxial techniques.16
It is difficult to generalize meaningful hospital staffing/cost-saving benefits if RA techniques were used as opposed to GA-only techniques (via PACU bypass or same-day discharge time reductions or both). Generally speaking, if the conversion from GA to RA techniques encompassed only 25–100 cases/yr, it seems highly unlikely that the duration-of-stay benefits would translate to meaningful reductions in hospital expenditures (for staffing etc. ). However, if 500–3,000+ cases/yr were converted from GA to RA and the results (PACU bypass, reduced duration of stay) by Hadzic et al. 1were able to be generalized to all cases/case categories subject to the anesthesia process transformation, it seems more likely that actual cost savings would be tenable.17–20The threshold is unknown for the proportion of outpatients requiring GA-to-RA conversions to produce meaningful hospital cost advantages.
Our group studied 1,432 consecutive outpatients undergoing anterior cruciate ligament reconstruction and more invasive knee procedures.17–20In our institution, these procedures were associated with a 17% unplanned admission rate when GA was used (without nerve blocks); GA was used 90% of the time (and nerve blocks none of the time) before July 1996. Nerve blocks evolved as the centerpiece of analgesia during the next 2 yr, leading to routine same-day discharge of complex knee surgery patients by 1998, when all of these patients were previously admitted.4
In addition to achieving routine same-day discharge after invasive outpatient orthopedic surgery (associated with the routine use of nerve blocks), we designed a PACU bypass score to address specific needs of RA patients. Our scoring system was modeled after the 10-point modified Aldrete score,21with special provisions forbidding the bypass of patients with any pain, PONV, shivering, pruritus, orthostatic symptoms, or hypotension. We believed it was inefficient to use a PACU nurse (trained in critical care nursing) for the treatment of postoperative patients who met the stated criteria. Similarly, we did not want to create “loopholes” in which patients were inappropriately transferred from the PACU, because patients can have pain or PONV without reflecting adversely on the Aldrete score.
Hadzic et al. 1used the modified Aldrete score of 9 or greater for PACU bypass (if following these criteria to the letter, nerve block patients could never achieve a score of 10 because these patients cannot move their blocked upper extremity). Meanwhile, Jankowski et al. 15used a Mayo Modified Discharge Scoring System score of 8 or greater to qualify for PACU bypass; these criteria were based largely on the same modified Aldrete score criteria. Thankfully, Hadzic et al. 1did not bypass patients who had pain scores higher than 3 (out of 10); this is a loophole in both the Aldrete and the Mayo criteria (because pain is not evaluated in either of these). Other than our group’s criteria17–20and the Mayo criteria, the only other reported criteria are those by White and Song,22which were not designed for patients undergoing RA. The White-Song criteria allow PACU bypass for patients with (1) transient vomiting or retching or (2) moderate to severe pain controlled with intravenous analgesics, if all other parameter scores are perfect. The White-Song criteria do not address shivering or pruritus.18 Table 1lists my suggestion for merging clinically useful PACU bypass quality indicators.
In conclusion, Hadzic et al. 1demonstrate the possibility of routinely using RA for peripheral orthopedic surgery in an effort to minimize same-day discharge times. Much work is required, especially via randomized controlled trials, to determine whether discharge time savings are found whenever RA is used or whether emerging pharmacology and technology (e.g. , target-controlled opioid infusions) may obviate the need for routine RA procedures in invasive outpatient orthopedic surgery. Standardizing PACU bypass criteria is an important step to render discharge time studies comparable, because nursing qualifications and staffing ratios in the PACU versus phase II may be an important covariate in the study of discharge times. PACU bypass criteria should prohibit undue increases in phase II labor intensity by discharging (or bypassing) a patient without fully addressing common labor intensive symptoms (PONV and pain) that are not in the forefront of critical care symptomatology (as are airway management and hemodynamics).
University of Pittsburgh, Pittsburgh, Pennsylvania. williamsba@anes.upmc.edu