■ Feature Article
FEBRUARY 2013 | Volume 36 • Number 2 /SUPPLEMENT
abstract
In the United States, more than 600,000 total knee replacements and 285,000 total hip replacements are
performed each year.1,2 At Florida Hos-pital, a 4415 bed hospital system with 22 campuses in central Florida, surgeons
perform more than 2100 primary total hip and knee replacements each year at the 7 campuses in the greater Orlando area. Perioperative pain management proto-cols, which include oral and parenteral analgesics, regional anesthetics, and sur-
gical site injections, designed by physi-cians at the Florida Hospital Winter Park campus have evolved over the past sever-al years. No standardized protocol exist-ed prior to 2006. In 2006, the anesthesia department offered regional anesthesia to all total knee patients. By mid-2008, anesthesiologists at Florida Hospital Winter Park offered every total knee pa-tient a femoral nerve sheath catheter with
Effect of Preoperative Intravenous Methocarbamol and Intravenous Acetaminophen on Opioid Use After Primary Total Hip and Knee ReplacementTHOMAS D. LOOKE, MD, PHD; CAMERON T. KLUTH, MBA
Between 2010 and 2011, a perioperative pain protocol for primary total hip and knee replacement at one Florida medical center replaced preoperative oral an-algesics with intravenous methocarbamol and intravenous acetaminophen. This is a retrospective cohort study of 300 patients, with 150 patients using the new pain protocol and 150 patients using a 2008 pain protocol that did not include these medications. The 2 cohorts were similar in patient gender, age, and body mass index. Opioid consumption was evaluated for a period of 48 hours after incision and was divided into 3 separate time intervals, as well as total 48-hour consumption. Mean opiate use decreased signifi cantly from 2008 to 2011 in all time intervals and total consumption (7.5�3.4 mg to 6.1�3.0 mg; P�.01). Subgroup analysis suggested that changes to the hip protocol were responsible for decreased opioid use in the operating room and the postanesthesia care unit, and changes to the knee protocol were responsible for decreased opioid use on the hospital fl oor and total consumption. The difference between the 2 protocol groups was not due to differences in individual surgeon practice patterns. Physical therapy progress of knee fl exion, average walking distance, and maximum walk-ing distance were signifi cantly improved. Hospital discharge was shorter in the 2011 group (4.0�1.1 days in 2008 group and 3.6�1.0 days in 2011 group). This study shows signifi cant improvement in patient care from 2008 to 2011 that is at least partially due to the change to the use of preoperative intravenous methocar-bamol and intravenous acetaminophen.
The authors are from the Florida Hospital De-partment of Anesthesiology and Florida Hospital, Winter Park Division, Offi ce of Orthopedic Research (TDL); and from the University of Central Florida College of Medicine (CTK), Orlando, Florida.
Dr Looke received investigator-initiated study grant support paid to his institution from Cadence Pharmaceuticals, Inc., and was compensated by SLACK Incorporated for his contribution to this manuscript. Mr Kluth received travel reimburse-ment from the University of Central Florida College of Medicine.
The authors thank Gwen Certain and Cecille Brocato for sharing their extensive knowledge and experience as certifi ed pain nurses and for providing their invaluable assistance with data collection; and Dr Julie W. Pepe with Florida Hospital Offi ce of Re-search Administration for her biostatistics expertise. The authors also thank the leadership team at Flor-ida Hospital Winter Park for providing equipment, offi ce space, and logistical support for this project.
Correspondence should be addressed to: Thom-as D. Looke, MD, PhD, 4609 Jetty St, Orlando, FL 32817 ([email protected]).
doi: 10.3928/01477447-20130122-54
25
ORTHOPEDICS | Healio.com/Orthopedics
■ Feature Article
48 hours of continuous regional anesthe-sia for postoperative pain management. By 2010, the use of postoperative opioids after total knee surgery decreased signifi -cantly, and patient-controlled analgesia delivery of intravenous hydromorphone was replaced with nursing administration of oral oxycodone plus acetaminophen or intravenous hydromorphone at patient request only. In addition, preoperative treatment with oral analgesics (a sur-geon-ordered combination of 1 or more of celecoxib, oxycodone plus acetamino-phen, or pregabalin) was replaced with intravenous methocarbamol (Robaxin In-jection; Baxter Healthcare Corporation, Deerfi eld, Illinois). The goals with this medication were to reduce oral loading preoperatively and reduce recovery room discharge delays due to muscle spasm pain. At that time, no reports existed on the effectiveness of intravenous metho-carbamol postoperatively for total joint surgery. Two studies of its use after breast augmentation had been performed,3-5 but this medication was ignored by most of the anesthesia pain literature.6
In 2011, intravenous acetaminophen (OFIRMEV; Cadence Pharmaceuticals, Inc., San Diego, California) was added to the protocol. Physicians and nursing staff had anecdotally reported that the combi-nation of intravenous methocarbamol and intravenous acetaminophen improved the early postoperative recovery after total hip and knee replacement. Although in-travenous acetaminophen had been stud-ied for use in major orthopedic surgery,7,8 it had not been studied when added to a perioperative pain protocol that included long acting peripheral regional anesthet-ics and other intravenous adjuncts.
Our retrospective study compares a co-hort of patients from 2011, who received preoperative intravenous methocarbamol and intravenous acetaminophen, with a cohort of patients using the 2008 pain protocol, who did not receive these med-ications. The outcome measures studied were opioid consumption, time to fi rst
opioid rescue, VAS pain scores, physical therapy progress, and hospital discharge times. This was considered a pilot study to determine if a prospective, random-ized, controlled clinical trial would likely show any benefi t from expanding the use of these 2 intravenous medications from 1 preoperative dose to regular dosing for the fi rst 48 hours.
MATERIALS AND METHODSThe Institutional Review Boards of
Florida Hospital and the University of Central Florida approved this retrospec-tive 2 cohort study with HIPAA and in-formed consent waivers for 300 patients. One hundred fi fty patients were included in the study group (using the 2011 proto-col with preoperative intravenous metho-carbamol and intravenous acetamino-phen), and 150 patients in the control group (using the 2008 protocol without these 2 medications). The patient popula-tion was identifi ed from anesthesia bill-ing records, including all patients from Florida Hospital Winter Park who had primary unilateral total knee arthroplasty (CPT 27447) or primary total hip arthro-plasty (CPT 27130) between January 1, 2009, and October 12, 2011. Patients were grouped according to the periopera-tive pain protocol and excluded patients from the 2010 protocol who received in-travenous methocarbamol but not intra-venous acetaminophen.
Within each protocol group, patients were also grouped fi rst by surgeon (from greatest to least surgical volume), next by anesthesiologist (from greatest to least volume), and fi nally by date of surgery (in reverse chronologic order). Since the total knee arthroplasty volume at Florida Hospital Winter Park (700 in 2011) is approximately twice that of total hip arthroplasty volume (375 in 2011), the team decided, a priori, to include 100 knees and 50 hip patients from each protocol group. This made it possible to limit the study to 2 surgeons performing knees (50 using the 2011 study protocol
and 50 using the 2008 control protocol each) and 2 surgeons performing hips (25 using the 2011 study protocol and 25 using the 2008 control protocol each). In total, the study included patients from 3 surgeons, because the second most pro-lifi c knee surgeon was also the second most prolifi c hip surgeon. More empha-sis was placed on limiting the number of surgeons rather than anesthesiologists, because the number of patient-care items left to the discretion of the surgeon is greater than those left to the discretion of the anesthesiologist. In addition, the team wanted to determine whether those items left to surgical discretion, such as intraoperative surgical site injections and perioperative analgesic adjunct selection, infl uenced primary outcome differences.
Patients were evaluated from 2011 (study) and 2008 (control) protocol groups in the order described above and included those patients, between the ages of 18 and 85, who had an American Soci-ety of Anesthesiologists (ASA) physical status of 1 to 3. Patients were excluded if they had chronic pain with patient reported opioid use 1 month preopera-tively; had taken centrally acting muscle relaxants 24-hours preoperatively; re-ceived a regional anesthesia procedure for postoperative pain control (other than a femoral nerve sheath catheter for knees or fascia iliaca block for hips); did not receive both of the study medications; and if there was a notable surgical com-plication as defi ned by the surgeon’s op-erative report. Data was collected for 300 patients. During the data analysis phase, 1 total hip patient in the 2008 protocol group, whose opioid consumption ex-ceeded 5 standard deviations above the mean, was also excluded from analysis, based on a history of chronic pain that had been missed in the initial assessment. Thus, 299 patients with a primary total knee or total hip replacement were in-cluded in the study.
For the study time periods, the sur-gical incision was defi ned as time zero.
26
FEBRUARY 2013 | Volume 36 • Number 2/ SUPPLEMENT
EFFECT OF METHOCARBAMOL AND ACETAMINOPHEN ON OPIOID USE AFTER THA AND TKA | LOOKE & KLUTH
All outcome data for the fi rst 48 hours after surgical incision was placed into 1 of 10 distinct time periods: intraoperative OR (incision time to time of admission to the postanesthesia care unit), postan-esthesia care unit PACU (recovery room admission time to discharge time), and 8 6-hour hospital FLOOR intervals with interval 1 starting at PACU discharge and ending 6 hours after incision time. Any opioids given prior to surgical in-cision (preoperative or for induction of anesthesia) and included in surgical site injections were excluded. All data were
obtained from electronic medical records including scanned anesthesia and nurs-ing notes, dictated operative reports, and electronically entered nursing, pharma-cy, and physical therapy notes.
The primary study endpoint was opi-oid consumption, which was document-ed in the electronic medical record, and all opioid medication was converted to hydromorphone equivalent milligrams (Table 1). The secondary study endpoints were: VAS pain scores (recorded by nurs-ing staff at regular intervals and prior to administration of PRN opioids); rescue
time (in minutes from incision to fi rst request for opioid medication); physical therapy progress (range of motion after total knee in degrees of fl exion and ex-tension, and ambulation in feet after both total hip and total knee) recorded by a physical therapist once or twice daily; PACU discharge time (in minutes from time of arrival to the time that PACU discharge criteria was met); and hospital discharge time (in days from time of inci-sion to time offi cially recorded as end of hospital admission).
Statistical analysis used a chi-square test for each categorical variable (gender and ASA class) and considered noncat-egorical variables to be continuous. The team then evaluated noncategorical vari-ables using either parametric or nonpara-metric tests depending on whether the data were found to be normally or non-normally distributed. Parametric tests included an independent 2-sample t test for the primary group (protocol) compar-isons and analysis of variance (ANOVA) for the multiple subgroup comparisons. The team used Levene’s test of equality of error variances for the ANOVA com-parisons to determine if the subgroup populations had comparable statistics. A signifi cant Levene’s test result means that subgroup populations have different characteristics and thus ANOVA signifi -cance values may not be valid. The team then applied the Mann-Whitney U test to evaluate nonparametric data.
RESULTSThe patients in the 2 protocol cohorts
showed similar demographic data for gender, age, and body mass index (BMI), but there was a signifi cantly greater num-ber of ASA 3 (and fewer ASA 1) patients in the 2008 protocol group (Table 2).
Four time intervals contained enough data to be evaluated for the primary out-come, opiate consumption: OR, PACU, FLOOR, and TOTAL (Table 3). The FLOOR period combined the 8 6-hour
Table 1
Opiate Equivalency Table Used to Convert Opiates Encountered During the Study to Hydromorphone Equivalents
Opiate Generic Name Route Equivalent Dose
Fentanyl Intravenous 100 mcg
Hydrocodone Oral 30 mg
Hydromorphone Intravenous 1.5 mg
Hydromorphone Oral 7.5 mg
Meperidine Intravenous 75 mg
Morphine Intravenous 10 mg
Morphine Oral 30 mg
Oxycodone Oral 20 mg
Propoxyphene Oral 150 mg
Table 2
Demographics of Study Patients
Group
Sex, No.(%) Age, y, Mean (SD)
BMI, Mean (SD)
ASA Class No. (%)
Female Male 1 2 3
All 174 (58) 125 (42) 65.5 (9.8) 30.8(5.9) 23 (8) 190 (63)
86 (29)
2008 84 (56) 65 (44) 66.1 (9.4) 30.8(5.4) 5 (3) 95 (64)
49 (33)
2011 90 (60) 60 (40) 64.9 (10.3) 30.9(6.3) 18 (12)
95 (63)
37 (25)
P value .559 .276 .959 .011a
Abbreviations: ASA class, American Society of Anesthesiology physical status; BMI, body mass index. aMeets criteria for signifi cance at P<.05.
Continued on page 29
27
ORTHOPEDICS | Healio.com/Orthopedics
■ Feature Article
Table 3
Opiate Use by Time Period for Protocol and Subgroupa
Group ORb PACUc FLOORd TOTALe
Protocol group
2008 .53 (.99) mg .56 (.80) mg 6.4 (2.7) mg 7.5 (3.4) mg
2011 .16 (.49) mg .30 (.60) mg 5.6 (2.7) mg 6.1 (3.0) mg
NP P valuef .000 .002 .003 .000
Total Hip Arthroplasty Only
2008 1.24 (1.20) mg 1.10 (.92) mg 6.6 (2.7) mg 8.9 (3.8) mg
2011 .19 (.54) mg .54 (.79) mg 6.9 (3.3) mg 7.6 (3.6) mg
NP P valuef .000 .001 .978 .074
Total Knee Arthroplasty Only
2008 .18 (.65) mg .29 (.58) mg 6.3 (2.7) mg 6.8 (3.0) mg
2011 .15 (46) mg .17 (.43) mg 5.0 (2.0) mg 5.3 (2.2) mg
NP P valuef .572 .179 .000 .000
Subgroup Analysis by Surgery Type, Surgeon, and Protocol Group
THA, surgeon 1 2008 2011
6.8 (3.2) mg8.0 (4.0) mg
10.0 (4.3) mg8.9 (4.6) mg
THA, surgeon 2 2008 2011
6.4 (2.3) mg5.8 (1.7) mg
7.9 (2.9) mg6.3 (1.6) mg
ANOVA P valuesg
Surgeon vs surgeon 2008 protocol vs 2011 protocol Surgeon infl uence on protocol
.024
.604.143h
.002
.067.717h
TKA, surgeon 3 2008 2011
5.3 (2.3) mg4.5 (1.8) mg
5.7 (2.5) mg4.8 (2.1) mg
TKA, surgeon 2 2008 2011
7.3 (2.8) mg5.5 (2.2) mg
7.8 (3.1) mg5.8 (2.3) mg
ANOVA P valuesg
Surgeon vs surgeon 2008 protocol vs 2011 protocol Surgeon infl uence on protocol
.000
.000.146h
.000
.000.134h
Abbreviations: ANOVA, analysis of variance; NP, nonparametric; OR, operating room; PACU, postanesthesia care unit; THA, total hip arthroplasty; TKA, total knee arthroplasty. aAnalysis given as mean (standard deviation) in hydromorphone equivalent mg.bOR time period is defi ned as incision time to PACU admission time. cPACU time period is defi ned as PACU admission time to PACU discharge time.dFLOOR time period is defi ned as PACU discharge to 48 hours after incision. eTOTAL time period is defi ned as incision time to 48 hours after incision time. fNP P value refers to nonparametric test of signifi cance (Mann-Whitney U test) with .05 considered signifi cant, subgroup analysis by surgeon only completed for FLOOR and TOTAL time periods. gANOVA P values for surgeon subgroup analysis show that for THA there was a statistically signifi cant difference between the 2 surgeons but not the 2 protocol groups and that differences between the surgeons did not signifi cantly infl uence analysis of the protocol groups; for TKA, statistically signifi cant differences were found between the surgeons and the protocol groups but the surgeon differences did not signifi cantly impact the protocol groups. hLevene’s test of equality of error variances was positive when comparing hip surgeons. Therefore, comparing the hip surgeon populations by ANOVA may not be valid due to signifi cant differences in the population characteristics in the 2 protocol groups.
28
FEBRUARY 2013 | Volume 36 • Number 2 /SUPPLEMENT
EFFECT OF METHOCARBAMOL AND ACETAMINOPHEN ON OPIOID USE AFTER THA AND TKA | LOOKE & KLUTH
FLOOR intervals (as described above) into 1, a necessary adjustment due to spo-radic FLOOR data. The TOTAL period combined all data for the 48-hour period starting with incision time.
Our patient population demonstrated a non-normally distributed opioid use that was heavily skewed to the left with the median much greater than the mean in all
time periods. Even with patients in both protocol groups requiring minimal opi-oids in the OR and PACU (Figure 1, for the OR), nonparametric analysis showed signifi cant reduction of opioid use in the 2011 protocol group in all time periods (Figure 2).
The results were also analyzed focus-ing on surgery type (hip or knee) and surgeon (Table 3). For surgery type, data
proved adequate for all time periods for both hip and knee surgeries and showed the same skew to the left as the full pro-tocol groups, necessitating a nonparamet-ric approach for this analysis as well. The 2011 protocol group demonstrated signifi -cantly reduced opioid use in the OR and PACU time periods after hip surgery but not knee surgery. There was signifi cantly reduced opioid use on the FLOOR and
Figure 1: Demonstration of skewed opiate data. The vast majority of patients receive no opiate medication in the operating room. The y-axis is opioid use in hydromorphone equivalent mg. The x-axis, labeled Frequency, shows number of patients receiving the amount of hydromorphone in the interval shown with the bottom interval (highest frequency) being the number of patients receiving hydromorphone 0 to 0.2 mg.
1 Figure 2: Opiate usage by time interval and protocol group. OR is the time period in the operating room from time of incision to time of PACU admission; PACU is time in postanesthesia care unit from time of admission to time of discharge from PACU; FLOOR is the time period from PACU discharge to 48 hours after incision; and TO-TAL is the combined period from incision time to 48 hours after incision. The height of the bars represents mean with error line (standard deviation) rising above mean demonstrating that data is not normally distributed (skewed left). Mann-Whitney U test P values shown above each time interval.
2
Figure 3: FLOOR opiate use by surgeon. Comparison of 2008 and 2011 protocol groups opiate use on the FLOOR (after discharge from PACU) broken down by surgeon. The height of the bars (error lines) represent mean (standard deviation) opiate use in hydromorphone equivalent mg. Table 4 gives ANOVA P values. For THA, Surgeon 1 and 2 populations have different opiate needs but the combined protocol groups are similar and the differences between the surgeons did not signifi cantly affect protocol group changes. For TKA, Surgeon 3 and 2 popula-tions have different opiate needs, combined 2008 and 2011 protocol groups have different needs, but the difference between the surgeons did not impact the dif-ference between the protocol groups.
3Figure 4: Comparison of 2008 and 2011 protocol groups total opiate use broken down by surgeon. The height of the bars (error lines) represent mean (standard deviation) opiate use in hydromorphone equivalent mg. Table 4 gives ANOVA P values. For THA, Surgeon 1 and 2 populations have different opiate needs but the combined protocol groups are similar and the differences between the surgeons did not signifi cantly affect protocol group changes. For TKA, Surgeon 3 and 2 popu-lations have different opiate needs, combined 2008 and 2011 protocol groups have different needs, but the difference between the surgeons did not impact the differ-ence between the protocol groups.
4
Opiate Use in OR by Protocol Group
2008 Protocol (Blue) 2011 Protocol (Green)
Continued from page 27
29
ORTHOPEDICS | Healio.com/Orthopedics
■ Feature Article
TOTAL in the knee surgery group but not the hip surgery group. Examining the in-dividual surgeons (identifi ed only as sur-geons 1, 2, and 3), FLOOR and TOTAL opiate use were normally distributed data and ANOVA was used to determine if changes in opioid use (FLOOR and TO-TAL) from the 2008 protocol to the 2011 protocol could be accounted for by those items left to surgeon discretion (Table 3). All subgroups showed reduced opioid use from the 2008 protocol to the 2011 pro-tocol, except for increased FLOOR opi-oid use for surgeon 1, hip (Figures 3, 4). ANOVA analysis showed that differences in opioid use from the 2 protocols cannot be accounted for (statistically) by differ-ences among the surgeons. This assess-ment is tempered by a positive Levene’s test of equality of error variances for the hip surgeon population groups. A positive Levene’s test means that the 2 hip surgeon population groups have such different sta-tistical characteristics that ANOVA may not be valid.
Examining secondary outcomes (Table 4), the average pain intensity increased signifi cantly from the 2008 protocol to the 2011 protocol. Of the 10 time intervals of the study, only time interval 5 (from 24 to 30 hours post-incision) had enough data points for separate evaluation. This showed no signifi cant difference between pain in-tensity scores in this time interval. In ad-dition, no signifi cant differences in rescue time were found. All physical therapy met-rics, except for knee extension, improved in the 2011 protocol group. PACU dis-charge times were unchanged. There was approximately a half-day reduction in hos-pital length of stay between the patients in the 2008 and 2011 protocols.
DISCUSSIONThis retrospective cohort study showed
that patient care for primary hip and knee replacement at Florida Hospital Winter Park improved from the 2008 protocol to the 2011 protocol. Opioid consumption decreased signifi cantly, physical therapy
progress accelerated, and length of hos-pital stay decreased. The study was de-signed to isolate, as much as possible, the infl uence of adding a single, preoperative dose of intravenous methocarbamol and intravenous acetaminophen on these out-comes. However, due to the retrospective nature of the study, it was not possible to eliminate all confounding variables, in-cluding differences in patient populations, planned changes in the protocols other than the study medications, and changes in surgeon practices.
The 2008 protocol group had sig-nifi cantly more ASA 3 patients than the 2011 group (33% vs 25%, P=.011). This factor could possibly account for an in-creased length of stay in the 2008 group. This 8% difference, although statistically
signifi cant, would not impact a power analysis or design a future randomized, controlled clinical trial.
In addition to adding a single, pre-operative dose of intravenous methocar-bamol and intravenous acetaminophen, other changes to the perioperative pain protocols occurred from the 2008 group to the 2011 group, including type of an-esthesia, spinal morphine, and fascia iliaca blocks. In the 2008 protocol group, 80% of the patients received general anesthesia (Table 5). In the 2011 proto-col, 80% of the patients received spinal anesthesia. The expected result of spi-nal anesthesia is less or no pain in the early postoperative period and a shift in opioid use from intraoperative load-ing to post-PACU (FLOOR) dosing. The
Table 4
Secondary Outcome Variablesa
2008 Protocol Group
2011 Protocol Group P
Pain; all periodsb 4.9 (1.0) VAS 5.5 (1.2) VAS <.001c
Pain; period 5b,d 5.4 (1.5) VAS 5.4 (1.8) VAS .846
Rescue timee 403 (332) min 440 (386) min .464 (NP)
PT-extensionf 3.8° (15.8°) 3.5° (17.9°) .582 (NP)
PT-fl exionf 66° (21°) 76° (17°) .001 (NP)c
PT-walk, averageg,h
52 (55) ft 71 (83) ft .020 (NP)c
PT-walk, maximumg,i
65 (76) ft 87 (95) ft .027 (NP)c
PACU dischargej 93 (41) min 88 (35) min .446 (NP)
Hospital dischargek
4.0 (1.1) d 3.6 (1.0) d <.001 (NP)c
Abbreviations: NP, nonparametric analysis (required due to nonstandard distribution of data); PACU, postanesthesia care unit; PRN, as needed; PT, physical therapy; VAS, visual analog scale.aAll values given as mean (standard deviation). bPain measured on standard VAS scale of 0-10 with only ordinal values reported.cSignifi cance standard of <.05 met. dPeriod 5 was the FLOOR time period 24 to 30 hours after incision. eTime from surgical incision until fi rst dose of any PRN opiate. fValues for knee extension and fl exion as reported in physical therapy note. gDistance reported in physical therapy note. hCombination of all distances reported for every patient in the fi rst 48 hours after incision.iMaximum distance walked in the fi rst 48 hours after incision.jTime from PACU admission until PACU discharge criteria were met. kTime in days from incision until patient is offi cially discharged from hospital.
30
FEBRUARY 2013 | Volume 36 • Number 2 /SUPPLEMENT
EFFECT OF METHOCARBAMOL AND ACETAMINOPHEN ON OPIOID USE AFTER THA AND TKA | LOOKE & KLUTH
desired result is less time in the PACU adjusting for inadequate or excessive opiate loading in the OR and less opiate use overall. The increased use of spinal anesthesia from 2008 to 2011 certainly contributed to the signifi cant reduction in OR and PACU opioid use in the total hip subgroup (Table 3) and probably to the slight (insignifi cant) increase in opioid use on the fl oor. Overall, the TOTAL opi-ate use after hip surgery was numerically,
but not statistically less (P=.074). In the 2008 total hip protocol, 33% of patients (all from surgeon 2) also received spinal morphine, but this increased to 80% of patients in the 2011 hip protocol (from both surgeons). The expected benefi t is 6 to 12 hours of postoperative pain relief. None of the knee patients received spi-nal morphine. Ultrasound-guided fascia iliaca blocks were also added to all pa-tients in the 2011 protocol for total hip
surgery, which may provide 12 hours of postoperative pain relief. We are not able to determine if the increased use of spi-nal morphine or fascia iliaca blocks to the 2011 THA protocol contributed to the fi ndings. It is possible that a statistically signifi cant increase in FLOOR opioid use was averted by these techniques.
Other changes in the use of patient-controlled analgesia, preoperative oral analgesics, and surgical site injections may have infl uenced the results. Ten months after initiating the 2008 protocol, most total knee patients who received spi-nal anesthesia and a femoral nerve cath-eter controlled their postoperative pain with only PRN oral opioids. The patient-controlled analgesia for total knee sur-gery postoperatively was phased out over the next year. In the 2008 protocol group, 33% of patients (who had their surgery in 2009 and early 2010) received a patient-controlled analgesia, and with the 2011 protocol, no patients (except chronic pain patients) received a patient-controlled analgesia. This was also phased out for total hip patients, with 80% receiving patient-controlled analgesias in the 2008 protocol group and only 24% receiving patient-controlled analgesias in the 2011 protocol group (Table 6). There was a signifi cant reduction in late opiate use (FLOOR and TOTAL) and a concomitant increase in average pain scores (Table 4). This may be related to eliminating the patient-controlled analgesia. The pres-ence of a patient-controlled analgesia is an independent (second only to pain it-self) driver of opioid consumption and its elimination will decrease opioid use without signifi cantly changing patient satisfaction. Thus, we were surprised to see a 12% increase in pain scores and are uncertain if the increase in average pain scores from 4.9 to 5.5 is accompanied with reduced patient satisfaction.
In the 2008 protocol, both knee sur-geons (surgeon 2 and surgeon 3), routinely ordered preoperative and postoperative oral analgesics, including celecoxib and a
Table 5
Changes in Intraoperative Anesthesia Preference for Total Hip Surgery
20082008 2011
No. GA Spinala Bothb No. GA Spinala Bothb
Surgeon 1, No(%)
24 24 (100)
0 0 25 13(52)
17(68)
5(20)
Surgeon 2, No(%)
25 15 (60)
16 (64)
6 (24)
25 5 (20)
23(92)
3(12)
Total, No(%)
49 39 (80)
16 (33)
6 (12)
50 18(36)
40(80)
8(16)
Abbreviation: GA, general anesthesia. aSingle injection of spinal anesthesia. bPatient received both spinal and GA. Reason for receiving both was unspecifi ed but could be failed or slow onset spinal or other patient factors, such as patient preference, morbid obesity, or sleep apnea).
Table 6
Use of Patient-controlled Analgesia Postoperatively by Surgery Type, Surgeon, and Protocol Group
2008 2011
Procedure PCAa No PCA PCA No PCA
THA
Surgeon 1 (n=49) 14 10 12 13
Surgeon 2 (n=50) 25 0 0 25
Total, No.(%) 39 (80%) 10 (20%) 12 (24%) 38 (76%)
TKA
Surgeon 3 (n=100) 12 38 0 50
Surgeon 2 (n=100) 21 29 0 50
Total, No.(%) 33 (33%) 67 (67%) 0 100
All (n=299), No.(%) 72 (48%) 77 (52%) 12 (8%) 138 (92%)
Abbreviations: N, number of patients; PCA, patient-controlled analgesia; THA, total hip arthroplasty; TKA, total knee arthroplasty. aNumbers given in PCA columns represent the number of patients with documented initiation of a PCA in nursing records; totals given as number (%).
31
ORTHOPEDICS | Healio.com/Orthopedics
■ Feature Article
combination oxycodone-acetaminophen tablet. In the 2011 protocol, both surgeons eliminated these medications and added oral methocarbamol (500 mg 3 times daily) to the postoperative routine. This change may also have contributed to the lower opi-ate use on the fl oor for total knee surgery postoperatively. Another variable evaluated was intraoperative surgical site injections, as both total knee surgeons routinely in-jected either the “subcutaneous tissues” or the “posterior capsule.” Surgeon 3 used the same formula in both 2008 and 2011: bu-pivacaine 0.25% 50 mL, preservative-free morphine 5 mg, and ketorolac 30 mg. Sur-geon 2 injected bupivacaine 0.25% 50 mL and preservative-free morphine 5 mg in 2008 and bupivacaine 0.25% 50 mL alone in 2011. Subgroup analysis focusing on the type of surgery subdivided by surgeon (Table 3; Figures 3,4) revealed that the opi-oid use group means were signifi cantly dif-ferent (surgeon 1 compared with surgeon 2 for total hip and surgeon 3 compared with surgeon 2 for total knee). The ANOVA test to determine if the surgeon differences were responsible for overall group changes from 2008 protocol to 2011 protocol was not signifi cant. Thus, the minor differences in surgeon practice probably infl uenced the opioid use, but did not affect the out-come changes and were not confounding variables.
The study found no difference in PACU discharge times but a signifi cant decrease in hospital length of stay. This may be re-lated to the difference in ASA grades in the 2 groups, or a possible change in round-ing patterns by the hospitalists. However, there was no evidence other than improved pain control to explain the decreased length of stay. There was no change in physical therapy practice between the 2 cohorts and improved pain control seems to be respon-sible for improvements in knee fl exion and walking distances.
CONCLUSIONThis retrospective 2-cohort pilot study
showed that the addition of preoperative intravenous methocarbamol and intrave-nous acetaminophen led to signifi cant im-provements in the care of patients under-going total knee and hip surgery at Florida Hospital Winter Park. The expanded use of these medications postoperatively for 48 hours should be studied in a prospec-tive, randomized, controlled clinical trial of total hip and knee patients. Based on the data of this study, in order to reduce opioid use by at least 10%, reduce pain scores by at least 10%, and avoid con-founding variables, the proposed 2-armed randomized controlled trial would need 90 patients in each arm. This study is pres-ently in the planning stage.
REFERENCES1. Total knee replacement. American Academy
of Orthopaedic Surgeons Web site. http://orthoinfo.aaos.org/topic.cfm?topic=a00389. Updated December 2011. Accessed October 20, 2012.
2. Total hip replacement. American Academy of Orthopaedic Surgeons Web site. http://orthoinfo.aaos.org/topic.cfm?topic=a00377. Updated December 2011. Accessed October 20, 2012.
3. Hidalgo DA, Pusic AL. The role of methocar-bamol and intercostal nerve blocks for pain management in breast augmentation. Aesthet Surg J. 2005; 25(6):571-575.
4. Schneider MS. Methocarbamol: adjunct ther-apy for pain management in breast augmen-tation. Aesthet Surg J. 2002; 22(4):380-381.
5. Schneider MS. Pain reduction in breast aug-mentation using methocarbamol. Aesthetic Plast Surg. 1997; 21(1):23-24.
6. White PF. The changing role of non-opioid analgesic techniques in the management of postoperative pain. Anesth Analg. 2005; 101(5 Suppl):S5-22.
7. Sinatra RS, Jahr JS, Reynolds LW, Viscusi ER, Groudine SB, Payen-Champenois C. Effi cacy and safety of single and repeated administration of 1 gram intravenous acet-aminophen injection (paracetamol) for pain management after major orthopedic surgery. Anesthesiology. 2005; 102(4):822-831.
8. Sinatra RS, Jahr JS, Reynolds L, et al. Intra-venous acetaminophen for pain after major orthopedic surgery: an expanded analysis. Pain Pract. 2012; 12(5):357-365.
32
