Thursday, March 7, 2013

Can Increasing Opioid Dose Worsen Chronic Pain?

Opioid FearsThere has been considerable debate about whether opioid analgesics are helpful or harmful for the treatment of chronic noncancer pain, and what the optimal dosing should be for pain relief during prolonged therapy. A new study suggests that opioid dose adjustments may not reliably influence changes in pain and, in fact, raising the dose may worsen pain in some patients. However, there were many shortcomings of this research that should be considered before leaping to firm conclusions.

A team of researchers from Harvard Medical School in Boston conducted a study to examine the relationship between opioid analgesic dose and resulting chronic pain relief reflected by changes in pain scores [Chen et al. 2013]. They retrospectively analyzed clinical data in patient charts collected at the Massachusetts General Hospital (MGH) Center for Pain Medicine during a 7-year period.

Subjects were selected primarily from a Research Patient Database Registry (RPDR), consisting of patients who had expressed an interest in participating in clinical research studies. Of 927 charts, 109 were selected for this study. The selected subjects included patients with noncancer chronic pain conditions who were on opioid medications and had information recorded on opioid regimen, pain condition, and clinical pain score (on a 0-10 numeric rating scale, or NRS). The patients were 50 years of age on average (range 14-83 years), 52% were female, and all had attended 10 office visits on average with a mean 704 days of followup (range 29 to 1,866 days).

A primary focus of the study was on outcomes of long-term opioid therapy as determined by comparing differences in opioid dose and pain scores between the recorded initial visit and last charted visit at the MGH Center for Pain Medicine. Primary endpoints included: A) the impact of opioid dose adjustments (increase or decrease) on NRS pain score; B) gender and age differences in response to opioid therapy; and C) the influence of different pain conditions on opioid analgesic efficacy. Results reported in an early online edition of the Journal of Pain indicated the following:

  • Patients were variously prescribed one of 12 single-ingredient opioid analgesics or 21 opioid combination regimens. Across all subjects, overall mean ± standard deviation opioid dose at baseline was ≈170 ± 179 mg/day (in morphine equivalent doses, or MED; range 0-1,920 mg/d). During the respective followup periods, 33% of subjects had a dose decrease, 45% a dose increase, and 22% had no change in dose.

  • At the initial visit the mean NRS pain score was 6.5 ± 2.1 across all subjects. Overall, neither opioid dose increase nor decrease correlated significantly with changes in pain scores, up or down, over a prolonged course of followup (R² = 0.0027; P>0.05). That is, there were variously increases and decreases in NRS score in response to both increasing and decreasing doses, so the average net effect was negligible correlation.

  • The lack of significant correlation between dose and NRS pain score was consistent regardless of the type of chronic pain condition, including neuropathic pain (eg, peripheral neuropathy), nociceptive pain (eg, low back pain), or mixed pain (eg, axial plus radicular back pain).

  • Opioid dose change did not significantly alter NRS pain scores with respect to gender or age differences.

  • The researchers also examined whether there would be differences between patients who were initially receiving low opioid doses (<75 mg/d MED) or high doses (≥75 mg/d MED). In both dose groups, many patients variously had their doses either increased or decreased during the course of therapy, while others remained on stable doses; collectively, however, there were no significant differential influences of high vs low dose on changes in NRS pain scores.

Importantly, the researchers note that their global analyses depicting subgroup averages do not realistically depict the variable changes in dose and subsequent NRS pain score responses of individual patients. That is, for each subgroup assessment, pain scores for some patients improved, some got worse, and some remained stable over time in response to similar dose adjustments (also see figure below).

For example, pain in some patients receiving dose increases improved, whereas pain worsened or stayed the same for others receiving dose increases; and, the opposite occurred for dose decreases, with pain variously improving or worsening. Consequently, the net average effect in each subgroup overall was one of no significant change in pain score and a lack of correlation.

Based on this lack of correlation between opioid dose and NRS pain scores, the researchers conclude that that dose adjustments during opioid therapy for chronic noncancer pain may not necessarily alter long-term clinical pain scores. Therefore, individualized opioid therapy based on each patient’s clinical response and effectiveness should be considered to optimize treatment outcomes.

COMMENTARY: Superficially, this interesting data-mining study may appear to repudiate the pain-relieving benefits of opioid analgesics prescribed long-term at any dose for chronic noncancer pain. And, some critics of opioid therapy may use this misperception as “proof” of that premise.

However, it should be borne in mind that this study by Chen and colleagues also represents low-quality evidence; essentially — as a retrospective, chart review investigation — it is a large series of case reports, or anecdotes, representing a particular population of patients. Further investigations of better quality are likely to alter, or at least refine, the outcomes significantly for a better understanding of who would and would not benefit from opioid therapy for chronic pain, as well as the opioid agents and dosing protocols that might be most effective.

Opioid CorrelationChen et al. provide a number of graphs in their paper that help to illustrate data in the various subgroup analyses. For example, trends in the figure [at right] are typical and, in this case, depict the 21 subjects with mixed pain conditions (eg, axial plus radicular low back pain). The overall average change in pain score was small but favorable (–4.1%); however, there was considerable variation in pain-score increases or decreases in patients receiving opioid increases [the right half of the graph], and similarly with reduced opioid dosing [left side of graph]. Several patients without any change in opioid dose [center portion of the graph] variously reported increased or decreased pain to some extent.

It seems paradoxical that the subject with the greatest increase in opioid dose (≈450%, far right) experienced an increase in pain. At the next highest dose-increase level (≈300%), 2 of 3 patients had decreases in pain. It would be important to better understand what might have affected such outcomes, but the study does not provide individual patient data for assessment and it cannot be assumed that opioid dose alone was the primary factor.

Overall, in this subgroup with mixed pain conditions, there was no clinically significant relationship between changes in opioid dose and pain score (R²=0.0005; P>0.05). Yet, some patients did well with dose increases or decreases, while others did not. Similar patterns of variation were displayed for all other subgroup comparisons; however, there were too few patients in each subgroup to eliminate the random play of chance in the overall results, and there could be significant false negative findings (type II error) due to such a lack of statistical power.

In addition to small group sizes and the subsequent lack of statistical power there were other important shortcomings in this study, some of which the researchers themselves conceded:

  • Due to the retrospective study design, clinical conditions over time were not controlled or taken into account, including: disease progression, comorbidities, adjunctive medications prescribed, and interventional pain procedures. These clinical factors were not documented in the charts reviewed in this study, but this does not necessarily mean that they did not occur.

  • Similarly, there were other potentially confounding patient-centered factors not accounted for in the study data, including: current opioid abuse/addiction or aberrant drug-use behavior, nonadherence to the prescribed opioid regimen, variations in opioid metabolism and drug interactions, unreported use of OTC analgesics, or adverse events associated with therapy. Any of these might have affected clinical response to opioid therapy.

  • Subjects in this study represent only a small group of patients with chronic pain treated with opioids at a teaching hospital and the results cannot be generalized to larger populations of patients. Furthermore, almost all subjects had specifically expressed a prior interest in participating in research studies, which suggests a strong self-selection bias.

  • The study was not intended to examine how dose adjustment decisions were made by clinicians during a course of opioid therapy, and such decisions can be complex. Additionally, dose adjustments may not have been made only in accordance with patient pain response; other variables — such as functionality or quality of life — that can be important in gauging therapeutic success were not recorded in the data, but may have affected dosing decisions more than NRS pain scores in many cases.

  • Basing the analyses on only the initial and last recorded office visit in each chart has its limitations, and subjects variously had from 3 to 58 visits with up to 10 dose adjustments during this period. Considerable variations in pain relief may have occurred during the lengthy courses of treatment, with the final NRS pain score used for analysis representing only a transition point for many patients.

  • The researchers postulate that some of the paradoxical findings — eg, pain escalation in response to opioid dose increase, and vice versa — might be attributable to opioid tolerance or opioid-induced hyperalgesia. However, they also acknowledge that their study was not designed to detect or differentiate these clinical effects, so it would be prudent not to speculate on the importance (or existence) of these factors based on this study.

  • Finally, there were 33 different opioid regimens — individual medications and combinations — recorded in the charts, and the study authors do not report if opioid rotation also was used in some patients during the course of therapy. Plus, the treating practitioners were not following any standardized prescribing protocol, so it is possible that the type of opioid or combination, as well as dosage, were not optimized on an individual patient basis to maximize pain relief. With so few subjects and so many different opioid therapies, analyses of efficacy for any regimen could not be performed.

If anything — and assuming that there is some validity in this study’s observations — Chen and colleagues reinforce the notion that long-term opioid therapy is likely not an optimal solution for all patients with all types of chronic pain conditions. Prior systematic reviews and data meta-analyses of clinical trials [eg, Kalso et al. 2004; Noble et al. 2010] have similarly demonstrated that significant proportions of patients respond favorably to ongoing opioid therapy for chronic noncancer pain and continue this treatment indefinitely; at the same time, equally important percentages experience intolerable adverse effects and/or disappointing therapeutic outcomes and elect to discontinue the treatment.

The challenge for the future is in designing and conducting high-quality clinical trials that will help to better determine which patients, with what sort of pain conditions, can most benefit long-term from what type and dose of opioid therapy. At present, eschewing the potential importance of opioid therapy for individual patients with chronic pain solely on the basis of current clinical evidence, or lack thereof, might be considered tantamount to medical malfeasance.

> Chen L, Vo T, Seefeld L, et al. Lack of Correlation Between Opioid Dose Adjustment and Pain Score Change in a Group of Chronic Pain Patients. J Pain. 2013(Feb 26); online ahead of print [
abstract here].
> Kalso E, Edwards JE, Moore RA, Mc-Quay HJ. Opioids in chronic non-cancer pain: Systematic review of efficacy and safety. Pain. 2004; 112:372-380.
> Noble M, Treadwell JR, Tregear SJ, et al. Long-term opioid management for chronic noncancer pain. Cochrane Database of Systematic Reviews. 2010;1(CD006605).

eNotifications Don’t Miss Out. Stay Up-to-Date on Pain-Topics UPDATES!
Register [here] to receive a once-weekly e-Notification of new postings.