Prospective Randomized Double Blind Efficacy of an Analgesic Epidural Paste

April 1999 Volume 90 Number 2

A prospective randomized double–blind controlled trial to evaluate the efficacy of an analgesic epidural paste following lumbar decompressive surgery
 

R. John Hurlbert, M.D., Ph.D., F.R.C.S.(C), Nicholas Theodore, M.D., Janine B. Drabier, R.N., Andrea M. Magwood, R.N. and Volker K.H. Sonntag, M.D.
 
University of Calgary Spine Program, Foothills Hospital and Medical Centre, Calgary, Alberta, Canada; and Department of Neurosurgery, Barrow Neurological Institute, St. Joseph's Hospital and Medical Center, Phoenix, Arizona

OBJECT.   Pain control can often be improved by local (as opposed to systemic) application of analgesic and/or anesthetic medication. The purpose of this study was to evaluate the efficacy of a single–dose epidural analgesic "paste" in the control of postoperative pain in patients who have undergone lumbar decompressive surgery.

METHODS.   Sixty patients undergoing routine elective lumbar decompressive surgery were randomized in a double–blind fashion to one of two groups: those receiving active paste or placebo paste. The paste was applied to the exposed dura at the time of surgery, just prior to wound closure. Patients received follow–up care in the hospital and at home for 3 months postsurgery. Several outcome measures were studied to ascertain differences in pain control and to ensure comparability between groups. Patients who received active paste demonstrated significantly lower pain scores compared with those who received placebo paste for up to 6 weeks postoperatively. General health perception indexed by the Short Form 36 was also significantly better in patients who received active paste for up to 6 weeks. In–hospital and outpatient oral narcotic consumption was significantly lower in the active paste–treated group. Inpatient straight leg raising scores were improved in those patients who received active compared with control paste.

CONCLUSIONS  . Application of an analgesic paste directly to the epidural space during lumbar decompressive surgery significantly improves postoperative pain control, reduces prescribed analgesic drug consumption, and improves overall health perception for up to 6 weeks following surgery. The authors conclude that this postoperative pain control strategy is both effective and safe and may provide a new standard of pain management in patients undergoing lumbar discectomy or laminectomy.

KEY WORDS.   lumbar surgery, epidural analgesia, pain control, randomizedclinical study, patient outcome

 

AFTER lumbar spinal surgery, all patients experience some degree of discomfort associated with tissue dissection and removal at the operative site. Currently, management is largely composed of the systemic administration of narcotic agents, either orally or parenterally, in the acute and convalescing postoperative state. Systemic administration is often accompanied or limited by side effects that include nausea, vomiting, headache, dizziness, mental disturbance, sedation, constipation, respiratory depression, and hypotension. Poor control of postoperative pain and associated side effects resulting from systemic medications can influence patient mobilization, discharge, and in some cases, return to work.

A limited number of studies have been published in which the effect of epidural analgesic agents on pain control following lumbar spinal surgery is examined. Methods that detail the use of single applications of morphine, and morphine and methylprednisolone, directly to the exposed epidural space have been described. 5, 8, 10, Improved pain control has also been reported retrospectively in patients in whom a morphine and methylprednisolone–impregnated absorbable gelatinous sponge was implanted at the time of surgery. 2, More recently, the use of morphine in combination with methylprednisolone, aminocaproic acid, and microfibrillar collagen has been proposed as a successful method of reducing pain for an extended period of time after lumbar surgery. 2,

Of all published methods for improving postoperative pain control after elective lumbar surgery, we found the latter to be the most innovative; its potential for prolonged effect without introducing a foreign epidural mass was intuitively appealing. The purpose of this experiment was to evaluate critically the efficacy of this analgesic paste in patients undergoing lumbar decompressive surgery. Shortfalls of previous studies were taken into account in the design of this trial. Primary outcome measures were chosen to reflect the potential benefits to the patient from the application of the paste. Additional outcome indices were chosen to ensure comparability between groups. We hypothesized that: 1) patients receiving active paste would report less pain compared with patients receiving placebo paste; 2) patients receiving active paste would require fewer prescription analgesic drugs for pain control compared with those receiving placebo; 3) these effects would be prolonged through the convalescent period; and 4) patients from both groups would have similar postoperative motor and sensory scores.

Clinical Material and Methods

Patient Population and Sample Size

Over a period of 20 consecutive months, patients undergoing routine elective lumbar surgery for discectomy or spinal stenosis were recruited (Table 1). The study protocol received ethical and scientific approval after formal internal review. All patients provided informed consent prior to entering the study. Based on sample–size calculations obtained from pilot data we predicted that 50 patients per group were necessary to detect a significant difference in narcotic analgesic administration and McGill Pain Questionnaire (MPQ) scores. However, the study was terminated when analyses performed off–site established significance of the primary outcome measures at 30 patients for each of the two study groups.

 
Surgical Protocol

After enrollment into the study, patients underwent a detailed neurological examination and were given three questionnaires (MPQ, Aberdeen Back Pain Index [ABPI], and Short Form 36 [SF–36] General Health Survey). Participants were stratified into two groups based on the magnitude of the planned surgery (discectomy or laminectomy).

On the day of surgery patients were randomized into a placebo or control group. All patients received a standard inhalational anesthetic with supplemental oxygen and nitrous oxide. All surgeries were performed by the recruiting surgeon (V.K.H.S.) at a single institution. After the decompressive procedure was performed, the paste was mixed on an adjacent sterile field by a physician not involved in patient care or follow up (N.T.) and applied to the exposed dura prior to wound closure by the surgeon. The ingredients in the active paste consisted of microfibrillar collagen (Avitene; Davol Inc., Cranston RI), methylprednisolone (Depomedrol; Upjohn Company, Kalamazoo, MI), morphine (Duramorph; Elkins–Sinn Inc., Cherry Hill, NJ), and aminocaproic acid (Amicar; Immunex Corp., Seattle, WA) and were combined using the technique previously described by Needham. 7 The placebo mixture lacked both mophine and methylprednisolone, which were replaced volume for volume with normal saline.

Postoperatively, all patients recovered on a regular surgical ward and their vital signs were charted every 2 hours throughout the hospital stay. Oxygen saturation was monitored continuously by pulse oximetry until 24 hours postoperatively. Morphine was administered intravenously for pain control in 2– to 4–mg doses every 1 to 2 hours as requested in the recovery room and on the surgical ward for the first 24 hours postsurgery. Thereafter, all patients were treated with an oral oxycodone/acetaminophen preparation, one or two tablets every 4 to 6 hours on request until discharge. Hospital discharge was undertaken when patients could ambulate, had adequate pain control, and transportation was available. The patient, surgeon, support staff, and on–site study nurse were blinded to treatment groups and remained so for the duration of the follow–up period.

 
Follow–Up Period

A follow–up period of 1 year from the time of surgery was planned for all patients. In this report we detail the early (12–week) results. All patients were available for follow–up examination at 3 months. Long–term outcomes will be reported separately.

 
Outcome Measures

Physician–derived outcomes included a detailed motor and sensory examination that was quantified according to the American Spinal Injury Association scale, modified to include spinal segments from the L–1 level and below. These data were obtained preoperatively and on Day 1 postoperatively to ensure that one group did not exhibit poorer neurological status than the other, either before, or as a result of surgery. Bilateral passive straight leg raising tests were also recorded at these times.

Patient outcomes were derived from three questionnaires, administered in person or by telephone: the MPQ, the ABPI, and the SF–36. Three components of the MPQ were used quantitatively to assess patients' pain at the time when the questionnaire was administered. The rank value pain rating index (PRI[R]), number of words chosen (NWC), and present pain intensity (PPI) were given preoperatively and then postoperatively on Day 1 and on Weeks 1, 3, 6, and 12. 6 The ABPI was used pre– and postoperatively to estimate the influence of low–back and radicular symptoms on daily activities in our patient population. 9, The SF–36 provided an index of overall general health perception. 4, 11, Both instruments were administered preoperatively and postoperatively at Weeks 3, 6, and 12.

Hospital–derived outcomes were obtained by nursing staff (also blinded to treatment group) who were involved in the care of patients during their hospital stay. Collected data included vital signs (blood pressure, heart rate, and respiratory rate) and oxygen saturation. All parameters were charted every 2 hours throughout the patients' stay. Quantitative records of the analgesic drugs administered in the recovery room and on the ward were obtained from patient medication records for the duration of hospitalization by the on–site study nurse. After discharge, patients reported the analgesic medication that they took at home to the on–site study nurse at Weeks 1, 3, 6 and 12 based on prescription renewal and remaining pills. Each patient was surveyed at these intervals for postoperative complications such as urinary retention, wound difficulties, pruritus, and symptom recurrence.

 
Data Analysis

All data were forwarded to the second institution for ongoing analyses. Questionnaires were scored initially by the blinded on–site study nurse (J.D.) and then rechecked by the off–site nurse (A.M.). Discrepancies were resolved by the principal investigator according to the published guidelines for each questionnaire. Descriptive statistics were compiled for all demographic, primary, and secondary outcome measures. Data were checked for normality by using the Kolmogorov–Smirnov test with Lilliefors' correction. 1, Parametric tests of significance were performed by repeated–measures analysis of variance (ANOVA) or paired t–test. Nonparametric data (resistant to transformation) were analyzed by the Mann–Whitney rank sum test. Examination of categorical data was performed by chi square test or Fisher's exact test. Probability levels for significance were defined as a value less than or equal to 0.05. Descriptive values presented in this paper are expressed as mean standard error of the mean unless otherwise specified.

Results

Sixty patients were successfully randomized to the treatment arms (30 placebo, 30 paste) and stratified for operative procedure (15 laminectomy and 15 discectomy procedures in each treatment arm). All patients remained in the study throughout their follow up. All patients were available for 3–month follow–up assessment.

 
Comparability Between Groups

Both groups of patients were similar with respect to demographic characteristics. There were 11 women and 19 men in the placebo group and 12 women and 18 men in the active paste–treated group. The average age of patients in the placebo group was 50 3 years, whereas the average age in the paste group was 53 3 years (t –0.752, df 58; p = 0.46). The average patient weight was 79 4 kg and 86 5 kg in the placebo and paste groups, respectively (t –1.181, df 40; p = 0.25). There were no statistical differences between the two groups with regard to duration of symptoms, medical insurance carrier, employment status, marital situation, smoking history, or incidence of comorbidity (p 0.05, Fisher's exact test). Results of preoperative physical examinations indicated similar motor and sensory scores for both groups (motor scores were 49 0.2 for the placebo group and 48 0.6 for the paste group; sensory scores were 78 0.7 for the placebo group, and 78 0.8 for the paste group). General health perception as measured by the SF–36 questionnaire was similar between groups (322 17 for the placebo group, and 313 20 for the paste group; t 0.355, df 58; p = 0.72). Patients in both groups experienced similar amounts of pain preoperatively as indexed by all three facets of the MPQ and the ABPI (p 0.05, ANOVA).

 
Treament Complications

There were no intraoperative complications recognized at the time of surgery. Eight patients in the placebo group required intermittent catheterization in the immediate postoperative period to relieve temporary urinary retention. Six patients were similarly afflicted in the active paste–treated group. All patients regained normal bladder function within 24 hours of receiving the anesthetic. One patient in the active paste–treated group who exhibited serous drainage from the wound postoperatively was treated with an extra skin suture. Two additional patients from this group suffered superficial wound infections that were managed successfully with antibiotic agents and local treatment measures at home. There were no reported instances of itching or rash in either group. Within the 3–month follow–up period one patient from each group suffered recurrence of symptoms that required additional surgery. Results from both of these patients were included in our analyses.

 
Hospital Course

All patients were closely monitored for alterations in oxygen saturation, respiratory rate, and blood pressure during their hospital stay. Minimum oxygen saturation was comparable between placebo–treated (95 1%) and active paste–treated (93 3%) groups (p = 0.74, Mann–Whitney rank sum). The minimum respiratory rate observed was also similar for placebo– (14 1) and paste–treated (14 1) patients (p = 0.63, Mann–Whitney rank sum). There was no significant difference in mean maximum systolic blood pressure (142 4 mm Hg, placebo group; 142 3 mm Hg, paste group) or diastolic blood pressure (72 2 mm Hg, placebo group; 74 2 mm Hg, paste group) between patients. Similarly minimum systolic blood pressure between placebo– and paste–treated groups was not notably different (109 2 mm Hg, and 111 3 mm Hg, respectively). However, the placebo–treated patients tended to have a lower minimum diastolic blood pressure (60 2 mm Hg) compared with paste–treated patients (68 2 mm Hg; t –2.437, df 58; p = 0.018). We observed no clinical consequences attributable to this difference.

Length of hospital stay was similar in the two groups (1.8 0.2 days for placebo– and 1.7 0.2 days for paste–treated groups; 55 cumulative days for all patients in the placebo group and 52 cumulative days for all patients in the paste group).

 
Primary Outcome Measures

Patients' self–perceived discomfort, indexed by the MPQ, was significantly lower throughout the postoperative period for patients who received active paste compared with those who received the placebo paste. The PRI(R) and NWC indices were markedly less on postoperative Day 1 in the actively treated compared with control–treated patients. These differences lessened but nonetheless persisted through Weeks 3 and 6, becoming comparatively similar at Week 12 (Fig. 1). Statistical significance was achieved through the 6–week time point for differences between groups for both the PRI(R) (F = 5.903, df 1239; p = 0.022) and the NWC (F = 4.882, df 1239; p = 0.035). The PPI (a more volatile index) indicated lower postoperative pain levels in paste–treated patients and at Weeks 1 and 3 compared with control patients, becoming similar at Week 6. The differences in PPI were significant through 3 weeks of follow up (F = 4.528, df 3239; p = 0.005). Surgery was clearly effective in reducing pain levels in patients from both groups (p 0.001).

Those patients who received active paste required fewer narcotic analgesic drugs while hospitalized and while at home throughout the 3–month follow–up period. The mean inpatient cumulative morphine dose was 19 5 mg for placebo– and 14 4 mg for paste–treated patients. This difference did not reach statistical significance (p = 0.16, Mann–Whitney rank sum). More striking was the difference in oral analgesic consumption (Fig. 2). Placebo–treated patients consumed more than three times as many percocet tablets during their hospital stay compared with paste–treated patients (10 2–unit doses and 3 1–unit doses, respectively; p = 0.007, Mann–Whitney rank sum).

This trend continued into the postoperative convalescence period at home. Patients who reported taking controlled analgesic medication in the paste–treated group consumed half the number of pills compared with patients in the placebo–treated group (37 9 and 75 18–unit doses, respectively; p = 0.044, Mann–Whitney rank sum). Results of the post hoc analysis demonstrated that prescription analgesic consumption in placebo–treated patients was higher than for those receiving the active paste through Weeks 3 and 6, dropping to comparable levels by Week 12 (Fig. 3).

 
Secondary Outcome Measures

On postoperative neurological examination similar motor (49 0.2 and 50 0.1) and sensory scores (79 0.3 and 78 0.8) were revealed between the placebo– and paste–treated groups, respectively. Interestingly, the mean left and right straight leg raising results improved from 80 3° to 85 2° and 75 4° to 85 2°, respectively, in the paste–treated group. However, in the placebo–treated group mean left and right straight leg raising scores decreased by 5 4° on postoperative day 1. Fig. 4). These differences were statistically significant (left straight leg raising F = 8.139, df 1114; p = 0.006; right straight leg raising F = 5.906, df 1114; p = 0.019).

General health perception was assessed by evaluating results of the SF–36 questionnaire. Somewhat unexpectedly, we observed an improvement in overall health perception postoperatively in patients receiving active paste compared with those receiving placebo paste. A statistically significant interaction was found between group and time over the 3–month duration of our follow–up period (F = 3.565, df 3239; p = 0.015). General health perception was better in active paste–treated patients 3 and 6 weeks postoperatively compared with placebo–treated patients (Fig. 5). At 12 weeks it was determined that both groups shared a similar general health profile. The results of post hoc subsection analyses demonstrated trends toward improved scores in physical functioning, role functioning, bodily pain, general health, social functioning, vitality, emotional health, and mental health in paste–treated patients compared with placebo–treated patients. These trends were most prominent in the general health, bodily pain, and vitality categories. There was no statistically significant difference in ABPI scores between both groups (F = 1.074, df 3239; p = 0.361). The SF–36 (F = 51.675, df 3239; p 0.001) and ABPI (F = 52.581 df 2239; p 0.001) significantly improved in both groups as a result of surgery.

 

Discussion

Based on a combination of objective and subjective outcome measures, these results indicate that improved pain control is possible in the immediate postoperative period and for up to 6 weeks posttreatment when relatively small amounts of morphine, methylprednisolone, and aminocaproic acid mixed with microfibrillar collagen are applied to the epidural space. Those patients who received actively treated paste took fewer narcotic analgesic drugs, had better straight leg raise scores, reported quantitatively less pain, and had better general health perception compared with placebo–treated patients who received standard postoperative pain control therapy.

The credibility of these observations is substantiated by internal consistencies both within and between outcome measures, as well as by validity between actual observed and theoretically expected improvements conferred through use of the active paste. Internal consistency is evident within the patterns of analgesic consumption, MPQ scores, and SF–36 results. Each of these outcome measures demonstrated improved pain control in the active paste–treated compared with placebo–treated groups, commencing in the immediate postoperative period and persisting to Week 6. In each instance, both groups achieved similar outcomes at 12 weeks. Validity (expected compared with observed) can be appreciated on two levels. The first involves overall patient performance; as indexed by MPQ, ABPI, and SF–36 scores, surgery conferred improvement that was independent of treatment group. This improvement was graduated, persisting for 6 to 12 weeks in the follow–up period. It was not erratic in nature. Such improvement is usually observed in a normal postoperative course after lumbar decompressive surgery. The presence of this improvement lends credibility to the differences between groups that were demonstrated using the same measurement tools. The second level of validity pertains to patient performance by group. It seems reasonable to expect that if improved pain control were to be observed, it should be reflected through reduced medication use, lower pain scores, and improved straight leg raising scores. It would also seem reasonable that neurological status would be comparable regardless of treatment group.

The mean minimum diastolic blood pressure observed in placebo–treated patients was significantly lower than that in the active paste–treated group. A possible explanation may rest with secondary hypotension induced by a threefold increase in oral narcotic usage in this group. Although relative diastolic hypertension might be considered in the pastetreated group, the mean diastolic pressure of 68 mm Hg actually observed in this group does not support such reasoning. Alternatively, this difference might be spurious, arising from uncorrected multiple comparisons.

It is interesting to speculate on the failure of the ABPI to detect a difference between treatment groups, despite differences in MPQ and SF–36 scores. Questions from the ABPI are weighted toward both back pain and radicular pain as reflected in activities of daily living. However, after lumbar decompressive surgery, most patients experience near–complete resolution of their radicular pain. Clearly, the sensitivity of the instrument becomes diminished when questions no longer apply to the study groups. Although it is recommended that an "item–specific" (disease–specific) questionnaire be administered in tandem with the SF–36, similar problems might be expected to arise with the use of other questionnaires, such as the Oswestry and Roland– Morris functional disability scales. 3,

The responsiveness of the SF–36 to differences observed between our groups was somewhat surprising. The SF–36 has become a valuable tool in detecting self–perceived general health differences among various treated and nontreated groups of patients. However, we were reluctant to predict sensitivity in detecting differences between our two groups because both were to receive a "greater" benefit from undergoing decompressive surgery. To the best of our knowledge, the results of this report are the first to demonstrate sensitivity of the SF–36 to different treatments of back pain. Trends toward improvement in paste– compared with placebo–treated patients were seen in all eight subcategories of the SF–36, most notably in measures of general health, bodily pain, and vitality. Our results imply that pain plays a very substantial role in determining patients' responses to general health–oriented issues.

The mechanism by which the epidural paste confers its prolonged analgesic effects has not been determined. It is our belief that the microfibrillar collagen acts as a reservoir that allows release of small amounts of morphine and methylprednisolone over time, as the collagen is absorbed. Aminocaproic acid, which inhibits fibrinolysis and, therefore, breakdown of the local blood clot, may help prolong this release. Application of morphine and methylprednisolone–impregnated gelatin sponge to the lumbar epidural space has been reported to provide "extended" analgesia in patients who have undergone discectomy in an uncontrolled retrospective series, 3, lending further support to the concept of a reservoir system. Conversely, a single application of morphine alone to the epidural space during lumbar decompressive surgery has been shown to provide postoperative pain control that lasts, on average, 6 hours compared with 2 hours in patients receiving saline application. 2, These observations also serve to provide evidence against a hypothesis of central neural plasticity and antinociceptive effects due to preemptive analgesia.

Following paste application, it is believed to be important to irrigate the subcutaneous tissues thoroughly after watertight closure of the lumbodorsal fascia (prior to skin closure). This eliminates the potential for the compound to act as a hyperosmolar agent, producing a sometimes painful but sterile and self–limiting subcutaneous fluid collection (CW Needham, personal communication, 1998). Long–term (1–year) follow–up review of these patients is in progress to elucidate any potential harmful side effects. Although we are aware of no anecdotal reports of excessive postoperative scarring, further research is near completion to evaluate more quantitatively this potential. In addition, a second study is also underway to address the relative contribution of each constituent within the paste.

 

Conclusions

In summary, the results of this study demonstrate that application of an analgesic paste to the exposed dura after lumbar decompressive surgery not only reduces the need for in– and outpatient analgesic administration (by a factor of 2–3 times) but also provides significantly better pain control compared with the current standard of medical practice. In addition patients' perception of general health status is also improved. The beneficial effects of the analgesic paste are present within 24 hours of application and persist for approximately 6 weeks postsurgery. Long–term pain control of this magnitude with a single application has not previously been reported. In our hands, the analgesic paste has proven safe; we encountered no complications or complaints attributable to use of this compound. These observations suggest this analgesic paste may have the potential to become a new standard of care for pain control in patients undergoing lumbar decompressive surgery. Potential applications in other surgical procedures should be considered.

 
 
Acknowledgments:

The authors would like to express their sincere gratitude to Dr. Charles Needham for providing the insight and opportunity to perform this research. Partial funding of our Phoenix–based research nurse was made possible through the kind generosity of Davol Pharmaceuticals.

 

References

 
 
1. Fox E, Shotton K, Ulrich C: SigmaStat Statistical Software User's Manual. San Rafael, California: Jandel Corp, 1995.
 
2. Gibbons KJ, Barth AP, Ahuja A: Lumbar discectomy: use of an epidural morphine sponge for postoperative pain control. Neurosurgery 36:1131–1135, 1995.
 
3. Leclaire R, Blier F, Fortin L: A cross–sectional study comparing the Oswestry and Roland–Morris functional disability scales in two populations of patients with low back pain of different levels of severity. Spine 22:68–71, 1997.
 
4. McHorney CA, Ware JE, Raczek AE: The MOS 36–item short–form health survey (SF–36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 31:247–263, 1993.
 
5. McNeill TW, Andersson GJ, Schell B: Epidural administration of methylprednisolone and morphine for pain after a spinal operation: a randomized prospective, comparative study. J Bone Joint Surg (Am) 77:1814–1818, 1995.
 
6. Melzack R: The McGill pain questionnaire: major properties and scoring methods. Pain 1:277–299, 1975.
 
7. Needham CW: Painless lumbar surgery: morphine nerve paste. Conn Med 60:141–143, 1996.
 
8. Rechtine GR, Reinert CM, Bohlman HH: The use of epidural morphine to decrease postoperative pain in patients undergoing lumbar laminectomy. J Bone Joint Surg (Am) 66:113–116, 1984.
 
9. Ruta DA, Garratt AM, Wardlaw D: Developing a valid and reliable measure of health outcome for patients with low back pain. Spine 19:1887–1896, 1994.
 
10. Waikakul W, Chumniprasas K: Direct epidural morphine injection during lumbar discectomy for postoperative analgesia. J Med Assoc Thailand 75:428–433, 1992.
 
11. Ware JE Jr, Sherbourne CD: The MOS 36–item short–form health survey (SF–36) I. Conceptual framework and item selection. Med Care 30:473–483, 1992.
 
 
Manuscript received October 1, 1998.
Accepted in final form January 7, 1999.