For Outpatient Rotator Cuff Surgery, Nerve Block Anesthesia Provides Superior Same-day Recovery over General Anesthesia

The study was approved by the Institutional Review Board of St. Luke’s-Roosevelt Hospital Center, New York, New York. Patients were eligible for participation if they were aged 18–70 yr, had an American Society of Anesthesiologists physical status of I–III, and were scheduled to undergo outpatient open repair of the rotator cuff. Patients were recruited on the day of surgery by a coinvestigator and a research assistant. After obtaining written informed consent, patients were randomized to receive either ISB or fast-track GA (specifically designed for rapid wake up and same-day discharge), using standard protocols.

Data were recorded with respect to anesthesia drugs given and several physiologic parameters during anesthesia (heart rate, blood pressure, and oxygen saturation). Patients were monitored during surgery and recovery according to standard guidelines published by the American Society of Anesthesiologists.††

Patients assigned to receive ISB were given midazolam (1–2 mg intravenous) and alfentanil (250–500 μg) in the operating room (OR) before block placement. These premedications were used to decrease anxiety and discomfort during block injection while maintaining meaningful patient contact. Blocks were performed by a senior trainee or fellow under the direction of an attending anesthesiologist with extensive experience in ISB.

Supplemental oxygen (5 l/min) was administered by facemask throughout. The ISB was performed using a 22-gauge, 50-mm Stimuplex® block needle (B. Braun Medical Inc., Bethlehem, PA) and a nerve stimulator (Tracer II®; LifeTech Inc., Stafford, TX). The block was performed using a standard technique with the patient supine.12After the brachial plexus was localized with a current of 0.2–0.4 mA (0.1 ms), 35–40 ml ropivacaine (0.75%) was injected in divided doses.

After injection, surgeons proceeded with surgical preparation without waiting for complete onset of surgical anesthesia. During surgery, ISB patients received an intravenous infusion of propofol (Diprivan®; AstraZeneca Pharmaceuticals LP, Wilmington, DE), titrated to light sleep with easy arousability. No other intraoperative sedatives or opioids were allowed. After surgery, propofol was stopped, and the patient was taken to the phase 1 postanesthesia care unit (PACU).

Interscalene block patients with inadequate surgical anesthesia, or those requiring intraoperative intravenous opioids, were given GA. The research team predetermined that all patients with failed blocks would be analyzed in the ISB treatment group to follow principles of intent to treat.

General anesthesia patients were given preoperative dolasetron (12.5 mg intravenous) for prophylaxis against PONV, midazolam (1–2 mg), and fentanyl (50–100 μg). GA was induced with propofol (1.5–2.0 mg/kg); one dose of rocuronium (1 mg/kg) was given to facilitate intubation. Anesthesia was maintained with desflurane in a 1:1 mixture of nitrous oxide and oxygen. The end-tidal concentration of desflurane was maintained at 3–6%, based on mass spectrometry (Capnomac Ultima ULT1; Datex-Ohmeda, Helsinki, Finland). Fentanyl boluses (25–50 μg intravenous) were administered as deemed necessary by the attending anesthesiologist.

Surgeons prepared the limb as soon as correct placement of the endotracheal tube was confirmed. At the end of surgery, the incision was infiltrated with 5–10 ml bupivacaine (0.25%) followed by an intraarticular injection of 10–15 ml bupivacaine (0.25%). Patients were awakened after a wound dressing and an arm sling had been applied.

After surgery, patients were taken to the phase 1 PACU. Phase 1 PACU nurses were blinded to the anesthetic technique used and had no access to the (automated) anesthesia record. Patients were evaluated using a modified Aldrete score13by the PACU nurse who made a decision regarding the patient’s eligibility to bypass phase 1 PACU going directly to the phase 2 PACU. Patients could bypass phase 1 PACU only with the following criteria: modified Aldrete score of 9 or greater, no treatment for pain (visual analog scale [VAS] score < 3), and no PONV. If a patient was admitted to phase 1 PACU, his or her vital signs were determined according to PACU policy, and the presence of symptoms (e.g. , PONV) was recorded.

In phase 2 PACU, patients were assessed at 15-min intervals by the nurses. They determined when patients met discharge-to-home criteria (a score of ≥ 9 on the postanesthesia discharge scoring system).14There was no minimum time required for patients to remain in phase 2 PACU. Voiding was not required for discharge from the hospital.15 

Daily pain scores and overall satisfaction with anesthesia were assessed as single VAS scores (1–10); these scores were then arbitrarily trichotomized as 0–2 (unacceptable), 3–7 (marginal), and 8–10 (acceptable).

The severity of postoperative pain was repeatedly assessed using the VAS at 15-min intervals. If patients reported pain in phase 1 PACU, morphine (1–2 mg intravenous) was administered every 5–10 min until the patient was comfortable (VAS score ≤ 2). The pain management protocol in phase 2 PACU and at home consisted of acetaminophen (325 mg) with codeine (30 mg) every 4 h as needed.

Hospital time intervals (e.g. , induction time, OR time, PACU time) were recorded using data from the automated record-keeping system. Data on discharge time were collected from the nursing documentation and verified by research assistants.

The research assistant, who was blinded to the type of anesthetic used, collected patient data by phone at 24 h, 48 h, 72 h, and 2 weeks after surgery. During the first 3 postoperative days, data included highest VAS pain score, daily pill counts, and other parameters of anesthesia recovery (e.g. , appetite, self-care, ambulation, interest in daily activities, anxiety). At 2 weeks after surgery, patients were asked about the occurrence of potential complications (e.g. , prolonged numbness, radiating pain in the distribution of the brachial plexus, motor weakness), overall satisfaction with anesthesia care, and willingness to have the same anesthetic for a subsequent surgery (if needed).

Sample size estimates were based on time to home readiness and discharge (in minutes) because this variable was of primary interest to the study. It was estimated that a sample size of 18 patients/group would provide 80% power to detect a clinically meaningful difference of 90 min (within-group SD, 60 min) at α= 0.001. The probability of a type I error was set low to accommodate the multiple comparisons that were planned, particularly for the targeted time measures (e.g. , time to ambulation, time to intake of fluids and solids). The final sample size was increased to 25 patients/group as an additional assurance that α would not be inflated when demographic and postoperative data were analyzed.

Discrete categoric data are presented as n (%); continuous data are given as mean ± SD. Confidence intervals are reported for the specific aims (PACU bypass and discharge times), and number-needed-to-treat analysis is reported for PACU bypass ineligibility and unplanned hospital admission. Differences in demographic, surgical, anesthetic, and postoperative data were tested by independent Student t  test (continuous data) or by chi-square (categoric data) and Fisher exact tests (where appropriate). For descriptive purposes, P  value differences less than 0.05 are noted in the tables. All analyses were conducted using the Statistical Package for the Social Sciences (SPSS for Windows, version 11.0.1; Chicago, IL).

Recruitment began in April 2000, and study follow-ups were completed by March 2002. Fifty-four patients were enrolled in the study, with the only refusals to participate occurring after randomization for 4 patients (3 ISB, 1 GA). No patient refused to participate before signing the study consent form. Fifty patients (25 in each group) completed the study on the day of surgery. There were no failed blocks, so intent-to-treat analysis was not applicable. There were no significant differences between groups with respect to sex, age, height, weight, and American Society of Anesthesiologists physical status (table 1), nor were there any differences in surgical process times (table 2).

Main outcome measures were eligibility to bypass phase 1 PACU to phase 2 PACU and eligibility and timing for same-day discharge. More patients who received ISB (76%) were able to bypass phase 1 PACU than those who received GA (16%) (table 3). Four patients (all in the GA group) were unable to be discharged because of refractory pain and were admitted to the hospital, whereas no patients in the ISB group were (P = 0.05). Among patients who were discharged, time to home readiness and time to discharge were more than 2.5 h sooner for patients who had received ISB versus  those who had received GA (table 3). Number-needed-to-treat analysis for the specific aims is shown in table 4.

For the secondary aims, moderate/severe pain (VAS > 3) was not reported by any ISB patients, whereas 80% of all GA patients requested treatment with analgesics (P < 0.001; table 3). PONV and sore throat were significantly less frequent in the ISB group (P < 0.05), whereas times to ambulation and oral intake were significantly less in the ISB group (P < 0.005; table 3)

For patients reached by phone at 24, 48, and 72 h, there was no significant difference between groups in pain scores and pill counts (table 5). The ISB and GA groups did not differ with regard to difficulties with sleep or appetite, self-care, or ambulation within the first 72 h after surgery. However, these findings are statistically underpowered.

Two weeks after surgery, 1 GA patient and 3 ISB patients reported backache (P = not significant); similar results were reported for headache. There were no reports of prolonged numbness, radiating pain in the distribution of the brachial plexus, or motor weakness. However, these findings are statistically underpowered. Global patient satisfaction with anesthesia care was higher in the ISB than in the GA group, and significantly more patients in the ISB group reported that they would choose the same anesthetic technique again (P = 0.014 for each; table 6).

The change from inpatient to ambulatory surgical care represents a significant advance.16Rapid recovery, adequate analgesia, prevention of PONV, and timely discharge are essential to a successful ambulatory anesthesia practice.17–19Our data suggest that from both the hospital’s and the patient’s perspective, there are advantages to using nerve block anesthesia versus  GA for outpatient rotator cuff surgery. Nerve block anesthesia was associated with a greater number of patients being able to bypass phase 1 PACU, fewer unplanned hospital admissions, and faster time to discharge.

Fast tracking/PACU bypass (being able to bypass the more costly and labor-intensive phase 1 PACU to go directly to phase 2 PACU) is frequently used as a benchmark test for success in ambulatory surgery (as is same-day discharge).20,21In our study, we ensured that all patients were free from pain, nausea, and vomiting before being transferred from phase 1 PACU to phase 2 PACU recovery, regardless of the Aldrete score.22This is important to prevent shifting of workload from phase 1 to phase 2 nurses.

In other studies, bypass of acute phase 1 PACU recovery has been associated with a $400 cost savings per patient,20and successful same-day discharge decreases hospital costs by an estimated $400–1,000.20,23Our findings of more successful PACU bypass and same-day discharge after ISB could be due, in part, to the fact that patients who had received ISB were more alert at the time of arrival to the PACU and had significantly less pain than did patients in the GA group.

None of the patients in the ISB group required treatment for pain before discharge home, whereas 80% of patients in the GA group required pain management despite wound infiltration and intraarticular instillation of local anesthetic by the surgeon. This finding is similar to another study in which continuous interscalene analgesia was superior to continuous wound infiltration with surgically placed incisional/intracapsular catheters.24 

Some perceived disadvantages of ISB versus  GA include the additional time required to perform the block, the possibility of block failure, and the potential that patients undergoing blocks ultimately may have more pain when the blocks wear off. None of these disadvantages were apparent in our study, although our sample size was underpowered to determine these specific outcomes. It is possible that OR times could have been even shorter for the ISB group if blocks had been placed in the preoperative area while the OR was being prepared.25Anesthesia-controlled time for emergence (time from end of the application of surgical dressing until OR exit) has been reported to be shorter after regional anesthesia versus  GA.25,26Applying these findings would help to offset any additional time needed to place the block in the OR before the procedure begins.

Our results of nerve block success differ from those reported in a retrospective study by Weber and Jain.27In that study, the authors reported that 13% of ISB failed and that 92% of patients receiving ISB required additional opioid analgesics. In addition, potentially severe complications were reported, including seizures, cardiovascular collapse, respiratory distress, and neurologic injuries.27The difference between our study and that of Weber and Jain could be explained by the fact that the latter used retrospective methodology, whereas our study was blinded and randomized. Second, nerve block performance in our study was limited to a team of university-based anesthesiology trainees medically directed by anesthesiologists with substantial experience with ISB and nerve block anesthesia in general. It has been reported that training in nerve block anesthesia is inadequate in many residency programs.28Nonetheless, ISB nerve block is successfully used as a sole anesthetic, with a low risk of complications, in institutions where the staff is experienced.4 

Postoperative pain is a common reason for unexpected hospital admission or delay in discharge.29It is possible that a multimodal approach to postoperative pain management (including the addition of antiinflammatory drugs) could have resulted in better analgesia for patients in the GA group.30,31Unfortunately, multimodal analgesia using newer nonopioid analgesics (specifically, cyclooxygenase-2 inhibitors) is not a universal practice.

Several studies have suggested that use of nerve block anesthesia or local anesthetics may produce a preemptive effect in reducing sensitization of nerve endings after surgical incision, potentially reducing postoperative pain.32,33A single-injection ISB using a long-acting local anesthetic such as ropivacaine (as in the current study) has been described as providing 12–14 h of analgesia after shoulder surgery.12More recently, ISB with a continuous infusion of local anesthetic during the postoperative period has resulted in excellent analgesia with minimal opioid requirement as long as the infusion is maintained.5,10,34–38 

Postoperative nausea and vomiting remains a common problem after anesthesia; these symptoms commonly result in discharge delays after ambulatory surgery.17,29,39,40In the current study, patients receiving ISB had a significantly lower incidence of PONV, despite the fact that they did not receive prophylactic dolasetron, an antiemetic, as did patients receiving GA. The odds ratio of experiencing PONV after GA with volatile agents (e.g. , desflurane, when compared with propofol sedation/anesthesia) has been reported to be 2.7–10.6.41,42It is probable that the use of volatile anesthetics as the primary maintenance technique, when superimposed on the significant postoperative pain and opioid requirements after shoulder surgery,20may have predisposed patients to PONV. Using antiemetics with different sites of action may have reduced the risk of PONV in the GA group.42,43However, in our study, all patients receiving GA who were admitted had a primary reason of refractory pain, not nausea or vomiting.

It may be argued that the advantages of nerve block anesthesia in this study would be less pronounced if another GA technique had been used. However, the GA protocol in this study is commonly accepted as a conventional model for fast-track GA in patients undergoing outpatient shoulder surgery. A systematic analysis of the literature comparing postoperative recovery after propofol-, isoflurane-, desflurane-, and sevoflurane-based anesthesia in adults demonstrated that early recovery was faster in the desflurane and sevoflurane groups. However, PONV was less frequent with propofol.44In addition, our data may not be reproducible in institutions without extensive expertise in performing peripheral nerve blocks. The training and practice of peripheral nerve blocks varies significantly from institution to institution, and in-depth training is a prerequisite for the success and safety of peripheral nerve blocks.28,45 

We assessed our patients’ perception of their recovery using questions about their sleep, appetite, self-care, ambulation, interest in activities, and anxiety after surgery. The functions assessed by our questioning addressed most aspects covered in two validated surveys.46,47Unfortunately, these validated survey instruments were not available at the time of our study design. Nonetheless, there were no differences between the two anesthetic regimens in these daily functions, although our sample size was underpowered to definitively show no differences between treatment groups.

In summary, under the conditions of our clinical practice, interscalene block with long-acting local anesthetic in outpatients undergoing rotator cuff surgery provided efficient and reliable surgical conditions. Compared with GA with wound infiltration, nerve block anesthesia with a long-acting local anesthetic also resulted in increased eligibility for PACU bypass and same-day discharge, faster same-day recovery, fewer adverse events on the day of surgery, better analgesia immediately after surgery, and greater patient acceptance.