Quantitative, crossover, double-blind trial paradigm for patient screening for chronic intraspinal narcotic administration

Robert M. Levy, M.D., Ph.D.

Division of Neurological Surgery, Northwestern University Medical School, Chicago, Illinois

Careful preoperative screening of candidates for indwelling drug administration systems for the relief of intractable pain can help to exclude patients who will not benefit from this technology and predict efficacy in others. Unfortunately, bias on the part of both the treating physician and the patient can inappropriately skew the results of subjective or improperly controlled trials and lead to the implantation of drug administration systems in patients who will not benefit from chronic intrathecal narcotic administration.

The author and his coworkers have designed a quantitative, crossover, double-blind paradigm for screening patients who might otherwise be deemed eligible for chronic intraspinal narcotic administration. This paradigm has been used 31 times in 30 patients; based on the outcome of this testing, 22 patients (73%) underwent implantation of chronic infusion systems. Sixteen (80%) of 20 patients with pain related to cancer underwent pump implantation, whereas only six (60%) of the 10 patients with pain of nonmalignant origin were so treated. Sixteen of the patients (72%) have reported good to excellent relief after pump implantation; this includes 12 (75%) of the 16 patients with pain related to cancer and four (66%) of the six patients with pain of nonmalignant origin.

This screening paradigm thus appears to be both reliable and easily applied and promises to be of assistance in the selection of patients appropriate for this mode of therapy.

Key Words * intraspinal morphine * drug pump * patient selection * chronic pain

Chronic intraspinal narcotic administration using implanted drug pumps for the control of chronic pain has become increasingly widespread. Although initially proposed for the control of intractable pain in the setting of terminal cancer[7,9,14], this mode of therapy has been expanded to include intractable pain of nonmalignant origin.[1,2,10,15,16] Generally accepted patient selection criteria for the implantation of a drug administration system have included: documented sensitivity of the pain to opioid agents; the inability of systemic opioid agents to control the pain either because of an inadequate degree of analgesia or intolerable side effects such as nausea, vomiting, constipation, or central nervous system depression; the lack of allergy or sensitivity to the drug to be infused; life expectancy of 3 months or more; the absence of intercurrent systemic infection; and favorable psychosocial evaluation and a positive response to an intraspinal test dose of narcotic.[8,13,18] Unfortunately, the specific features of the last two inclusion criteria have not been well elucidated. Other investigators[4,11] have attempted to delineate important predictive features of psychosocial evaluation for other neuromodulatory procedures, although there is little unanimity in the field.

Similarly, there has been little written about the nature of the preimplant intraspinal morphine testing paradigm. Testing with a single intraspinal dose of an active agent raises the significant possibility that the strong desire of the physician and other health care personnel to help the patient and the patient's desperation to find some relief from their intractable pain will lead to a significant placebo response to this injection. A placebo response rate of 30% has been reported in some such studies. Attempts to control for patient bias by testing both morphine and saline and blinding the patient to which drug is being infused still do not control for the bias of the health care team. Furthermore, the conclusions arising from preimplantation drug trials are often based on completely subjective criteria. This subjectivity can negatively impact the validity and reliability of screening protocols. We thus propose the use of a quantitative, crossover, double-blind trial for the preimplantation screening of candidates for chronic drug infusion therapy for the control of intractable pain.


Thirty patients with medically intractable chronic pain were included in the study. Twenty of these patients had pain in the setting of cancer, whereas 10 had pain of nonmalignant origin. Aggressive nonsurgical therapy administered by a multidisciplinary pain clinic team had failed in all patients prior to referral for intrathecal drug administration. All patients with pain of nonmalignant origin underwent psychological evaluation including, as a minimum, a structured interview and administration of the Minnesota Multiphasic Personality Inventory [5] and the Beck Depression Inventory.[3] In all cases, patients with severe untreated depression or underlying psychiatric illness were excluded. Patients with pain related to a malignancy were recommended for psychological therapy. Aggressive nonsurgical trials had failed in all patients including pharmacological treatment with narcotic drugs, nonsteroidal antiinflammatory agents, and tricyclic antidepressant medications prior to referral for intraspinal narcotic administration.

The testing paradigm consisted of two phases: first, establishing a dose of narcotic that was sufficient to decrease the patient's pain significantly and second, a crossover, double-blind trial of this dose of narcotic versus saline was initiated.

Phase I: Dose Escalation Trial

Patients were admitted to a constant observation unit and a lumbar intrathecal catheter was implanted in the patient at the bedside using aseptic technique. Intravenous access was established and intravenous prophylactic antibiotic agents (vancomycin) were administered throughout the period of externalized intrathecal cannulation. Ampules of naloxone (Narcan) were made readily available to treat narcotic overdosage. Pulse oximetry was used to monitor selected patients. A baseline visual analog pain scale (VAPS) score determination was made and followed by intrathecal administration of 0.5 mg morphine sulfate diluted in preservative-free saline. The catheter was then flushed with 0.5 ml saline. Thirty minutes later, another VAPS score determination was made; if there was less than a three-point drop on the VAPS, the 0.5 mg dose was repeated. Thirty minutes later, the VAPS was repeated; at this point a total of 1 mg morphine had been administered. If there was less than a three-point drop on the VAPS, an additional 1.0 mg morphine was administered. The VAPS was again administered 30 minutes later; if the VAPS score had decreased by less than three points, a final 1 mg morphine bolus was given. The VAPS was adminstered hourly thereafter for 12 hours (Table 1).

If at any point during the dose escalation trial, the VAPS score decreased by three or more points, the trial was completed and the total dose was used in the second phase of testing. If a total dose of 3 mg morphine given intrathecally did not significantly impact the VAPS score, it was our experience that there was insufficient effect to warrant pump implantation. The trial was aborted and the catheter was removed.

Phase II: Crossover, Double-Blind Trial of Intrathecal Saline Versus Morphine Sulfate

Phase I testing provided a dose of morphine sulfate, which on intrathecal injection, produced a three point or more decrease in the VAPS score. During Phase II, this dose of morphine sulfate was tested against saline in a crossover, double-blind paradigm.

Prior to the 1st day of testing, an order was written to the pharmacy to supply drugs labeled "Drug A" or "Drug B" to the testing physician. The labels and delivery sequences were assigned by reference to a random assignment table. Both agents were drawn up in a 0.6 ml volume. In this paradigm, the testing physician, the nursing staff, and the patient were not aware of the nature of the drug being administered.

At the start of the 1st day of testing, Drug A was administered as a bolus injection and the catheter was flushed with 0.5 ml saline. Patients were encouraged to engage in a normal level of activity so that they might experience their pain much as they would outside the hospital. Visual analog pain scale scores were obtained twice at 30-minute intervals and hourly thereafter. Six hours after injection, the previous two VAPS scores were compared to the baseline VAPS score. If the VAPS score had returned to within one point of the baseline level, a second injection (either Drug A or Drug B) was given. If not, the testing physician waited until two hourly VAPS had returned to within one point of baseline before administering the next injection (Table 2).

On the 2nd day of testing, the order of drug administration was reversed so that diurnal variations in drug sensitivity were eliminated as possible confounding factors. Thus, if on Day 1, Drug A was given first, on the second day of testing Drug B was given as the first blinded injection. This paradigm allowed for a maximum of four daily injections; because of our hesitation in testing late in the night and the fact that active injections often provide analgesia lasting several hours, patients tended to receive either two or three injections per day.

A record of the VAPS scores was kept on a flow sheet in the patient's medical record. After a minimum of four injections (if it appeared that there was an unequivocal impact of one test drug and not the other) or a maximum of eight injections, the drug code was broken and the mean change in VAPS scores was calculated for each drug. If there was less than a two-point drop in the mean VAPS score following saline administration and if the drop in the VAPS score with morphine was more than two points greater than that following saline administration, the patient was considered a candidate for pump implantation. The catheter was removed and surgery was performed the following day.


We have used this screening protocol 31 times in 30 patients. Twenty of these patients have experienced pain in the setting of cancer and 10 have experienced pain of nonmalignant origin. Based on the outcome of the crossover, double-blind testing, 22 (73%) drug pumps were implanted in these 30 patients. Sixteen (80%) of 20 patients with pain related to cancer underwent pump implantation, whereas only six (60%) of the 10 patients with pain of nonmalignant origin were treated long term. Sixteen (73%) of the patients have reported good to excellent relief after pump implantation; this includes 12 (75%) of the 16 patients with pain related to cancer and four (66%) of the six patients with pain of nonmalignant origin.


Chronic intraspinal narcotic administration using implanted drug pumps is a potentially effective adjunct in the treatment of intractable pain; however, the success of this treatment modality relies to a great degree on the care with which patients are selected. In light of the costs associated with chronic intrathecal drug administration, it becomes even more important that only those patients in whom there is a significant probability of effectiveness undergo pump implantation. Most investigators have suggested that after meeting the generally accepted patient selection criteria, the analgesic efficacy of the drug should also be demonstrated by trial intraspinal injection or infusion. The two most frequently encountered problems with such trials are the lack of quantitative data on which to base the decision to proceed with chronic infusion therapy and the lack of control for bias on the part of either the treating health care provider or the patient. To address these problems, we have proposed the use of this quantitative, crossover, double-blind trial for the preimplantation screening of candidates for chronic drug infusion therapy for the control of intractable pain.

This approach is not, however, without its limitations. The screening protocol requires a minimum of 3, and more often 4 days of hospitalization prior to pump implantation. Although this represents a significant expense, the overall costs of the implanted hardware, surgical fees, hospitalization, and postoperative care are such that if a lack of efficacy is demonstrated in only one in five potential candidates, there is still an economic benefit favoring the preimplantation testing paradigm. Our experience indicates that more than 20% of patients can be eliminated from surgical consideration; in patients with pain related to cancer, 20% of potential candidates did not receive an implanted pump as a result of trial screening, in patients with pain of nonmalignant origin, 33% of patients did not respond adequately and thus did not receive an implanted pump. Overall, 27% of patients avoided potentially ineffective surgery as a result of this screening procedure. Thus, despite the expense of this testing paradigm, it appears to be cost effective.

There are several issues that arise related to the reliability of the VAPS and its quantitation. The VAPS is a simple and efficient tool that consists of a 10-cm line anchored at one end by the label "no pain" and at the other end by the label "worst pain possible." Scoring is accomplished by having the patient mark the line to indicate pain intensity and the line is then measured to the mark on a 0- to 10-point scale. This tool is relatively straightforward and over 90% of patients are able to comprehend and complete it.[17] Although some have questioned the validity of the VAPS,[6] much research has suggested that it is reliable and valid as both a sensitive measure of pain and, more significantly, as a measure of change in pain.[12] There are clearly more comprehensive pain assessment tools than this unidimensional measure, and all candidates for interventional therapy should undergo such rigorous pain assessment. The VAPS is ideally suited for such a drug screening trial because of its reliability, simplicity, and quantitative nature.

A further issue arising from the unidimensional nature of the VAPS is the lack of linearity of effect across the extremes of the scale. Thus, a three-point drop from the top of the scale represents 30% pain relief, whereas a parallel drop toward the other extreme (such as a drop from four to one) represents 75% relief. Thus, the degree of impact necessary to label a trial successful in patients with extremely high reports of pain on the VAPS is much less than that required in a patient with a relatively low report of pain intensity. The argument that percentage decrease in VAPS scores rather than the raw scores should be used as a final measure is a reasonable one. We choose not to perform the trial in patients when their reports of pain intensity are low. We delay testing until VAPS reports of pain intensity approach their maximum; if a patient consistently reports a low VAPS score, then their pain may well be adequately treated without the addition of interventional techniques.


Thus, we have designed a quantitative, crossover, double-blind paradigm for the preimplantation screening of patients with intractable pain who are otherwise candidates for chronic infusion of intraspinal narcotic agents. This paradigm eliminates the inherent bias on the part of the health care providers and the patient and provides quantitative data on which to base decision making. It further helps to suggest beginning intrathecal drug dose requirements. Finally, such a paradigm appears to be cost-effective and its general application should increase the efficacy of this potentially valuable mode of therapy.


1. Auld AW, Maki-Jokela A, Murdoch DM: Intraspinal narcotic analgesia in the treatment of chronic pain. Spine 10:778­781, 1985

2. Auld AW, Murdoch DM, O'Laughlin KA: Intraspinal narcotic analgesia: Pain management in the failed laminectomy syndrome. Spine 12:953­954

3. Beck AT, Steer RA: Beck Depression Inventory Manual. Chicago: Harcourt Brace Jovanovich, 1987

4. Burchiel KJ, Anderson VC, Wilson BJ, et al: Prognostic factors of spinal cord stimulation for chronic back and leg pain. Neurosurgery 36:1101­1111, 1995

5. Butcher JN, Dahlstrom WG, Graham JR, et al: MMPI­2 Manual for Administration and Scoring. Minneapolis: University of Minnesota Press, 1989

6. Carlsson AM: Assessment of chronic pain. I. Aspects of the reliability and validity of the visual analogue scale. Pain 16:87, 1983

7. Coombs DW, Saunders RL, Gaylor MS, et al: Relief of continuous chronic pain by intraspinal narcotics infusion via an implanted reservoir. JAMA 250:2336­2339, 1983

8. Hassenbusch SJ, Pillay PK, Magdinec M, et al: Constant infusion of morphine for intractable cancer pain using an implanted pump. J Neurosurg 73:405­409, 1990

9. Krames ES, Gershow J, Glassberg A, et al: Continuous infusion of spinally administered narcotics for the relief of pain due to malignant disorders. Cancer 56:696­702, 1985

10. Maron J, Loeser JD: Spinal opioid infusions in the treatment of chronic pain of nonmalignant origin. Clin J Pain 12:174­179, 1996

11. North RB, Kidd DH, Zahurak M, et al: Spinal cord stimulation for chronic, intractable pain: experience over two decades. Neurosurgery 32:384­395, 1993

12. Ohnhaus EE, Adler R: Methodological problems in the measurement of pain: a comparison between the verbal rating scale and the visual analogue scale. Pain 1:379­384, 1975

13. Onofrio BM, Yaksh TL: Long-term pain relief produced by intrathecal morphine infusion in 53 patients. J Neurosurg 72:200­209, 1990

14. Onofrio BM, Yaksh TL, Arnold PG: Continuous low-dose intrathecal morphine administration in the treatment of chronic pain of malignant origin. Mayo Clin Proc 56:516­520, 1981

15. Paice JA, Penn RD, Schott S: Intraspinal morphine for chronic pain: a retrospective, multicenter study. J Pain Symptom Man 11:71­80, 1996

16. Penn RD, Paice JA: Chronic intrathecal morphine for intractable pain. J Neurosurg 67:182­186, 1987

17. Revill SI, Robinson JO, Rosen M, et al: The reliability of a linear analogue for evaluating pain. Anaesthesia 31:1191­1198, 1976

18. Shetter AG, Handley MN, Wilkinson E: Administration of intraspinal morphine sulfate for the treatment of intractable cancer pain. Neurosurgery 18:740­747, 1986

Manuscript received November 26, 1996.

Accepted in final form December 31, 1996.

Address reprint requests to: Robert M. Levy, M.D., Ph.D., Division of Neurological Surgery, Northwestern University Medical School, 233 East Erie Street, Suite 614, Chicago, Illinois, 60611.

Click here to return to Neurosurgical Focus Table of Contents.