Opioid analgesics are being increasingly prescribed for chronic noncancer pain, but questions regarding the safety and efficacy of this practice are still being investigated. A common, but often overlooked, potential consequence of long-term therapy is opioid-induced hormonal disturbances, particularly hypogonadism associated with sexual dysfunction in both men and women. Better understandings of these hormonal changes — including prevalence, risk factors, functional significance, and influences of comorbidity — are needed, and two recently reported studies examined this topic.
Association of Opioids for Back Pain & Male Sexual Dysfunction
Men with chronic pain may experience erectile dysfunction for a variety of reasons; however, the prevalence of this problem in back-pain populations and the relative importance of the several risk factors are unknown. Richard Deyo, MD, MPH and colleagues — affiliated with the Oregon Health and Science University, Portland, OR — investigated associations between use of medication for erectile dysfunction or testosterone replacement and the use of opioid therapy, as well as other factors [Deyo et al. 2013].
The researchers conducted a cross-sectional analysis of electronic medical and pharmacy records for males with back pain in the Kaiser Permanente Northwest (KPNW) health maintenance organization during 2004. Prescribed medications for erectile dysfunction or testosterone replacement — dispensed 6 months before and after the index visit — were considered as proxies for sexual dysfunction in males receiving opioid or nonopioid therapy for pain.
During 2004, there were 11,327 males with a diagnosis of back pain (mean age 49 years, 89% white). Subjects receiving medications for erectile dysfunction or testosterone replacement (n = 909, or 8% of the total) were significantly older than those who did not and had greater comorbidity, depression, smoking, and use of sedative-hypnotics.
In logistic regression analyses to control for interacting variables, older age, greater medical comorbidity, depression, and use of sedative-hypnotics were independently associated with the use of medications for erectile dysfunction or testosterone replacement — as might be expected. The long-term use of opioids (>120 days) was significantly associated with greater prescription of medications for erectile dysfunction or testosterone replacement compared with no opioid use (Odds Ratio 1.45; 95% Confidence Interval, 1.12–1.87, p < 0.01). Furthermore, patients prescribed daily opioid doses ≥120 mg of morphine-equivalents had greater use of medication for erectile dysfunction or testosterone replacement than patients without opioid use (OR 1.58; 95% CI 1.03–2.43), even with adjustment for the duration of opioid therapy.
The researchers conclude that age, comorbidity, depression, and use of sedative-hypnotics, appeared to play an independent role in sexual dysfunction. Most especially, however, both long-term use of prescription opioids and high-dose opioid therapy were associated with roughly 50% greater odds of using medications for erectile dysfunction or testosterone replacement after adjusting for the other confounders.
Deyo and colleagues suggest that the potential deleterious effects of opioids on sexual function as reflected in their findings may be an important consideration in deciding on the long-term use of opioids for back pain. However, there were numerous deficiencies in this study that question the significance and validity of the outcomes (see commentary below).
Long-term Opioids Disturb Hormones, But Aid Other Functions
In a presentation at the American Academy of Pain Medicine annual meeting this spring, Forest Tennant, MD, DrPH — of the Veract Intractable Pain Clinic, West Covina, CA — reported on 40 patients (ages 30 to 65 years, 60% female) who had been on daily high-dose opioid therapy (>100 mg morphine equivalents) for 10 or more years [Tennant 2013]. Most patients (65%) had spine-related conditions, while the remainder had head trauma, genetic disorders, autoimmune inflammatory disease, and other disorders; all of them had met the California definition of intractable pain, which is “incurable by any known means.”
Prior to their very long-term treatment with opioids, all patients had multiple nonopioid therapies and still experienced constant, debilitating pain with severe insomnia. During 2012, patients were evaluated using 2 written questionnaires and laboratory testing for serum cortisol, pregnenolone, corticotropin (ACTH), testosterone, erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP). Written questions asked about improvement in 17 physiologic/functional areas, including reading, hearing, concentration, memory, driving, sleep, movement, dressing, and libido. The other questionnaire asked about depression, hopelessness, and quality of life before and during opioid treatment.
All 40 patients reported sustained pain control on a stable opioid dosage, plus improvements in depression, hopelessness, and quality of life. Additionally, all patients reported improvements in at least 1 functional area, including: movement (77.5% of patients), concentration (67.5%), walking (62.5%), sleeping (62.5%), and appetite (50%).
Hormonal deficiencies — of ACTH, cortisol, testosterone, and/or pregnenolone — were found in 20% of patients, which required hormonal replacement therapy in most cases. Meanwhile, 3 patients (7.5%) had elevations of ACTH, cortisol, and/or pregnenolone, and there were high inflammatory markers (ESR, CRP) present in about 23% of patients.
Tennant noted that most patients (85%) were found to have genetic metabolic defects in opioid metabolism, which influenced higher tolerance and may have accounted for higher dose requirements. Some patients were taking much greater than 100 mg/d morphine-equivalents — as high as 1,000 mg/d.
In conclusion, Tennant observed that the long-term patients who were receiving high-dose opioids in this small study benefitted from pain relief and significant improvements in many functional areas, as well as mentally in terms of less depression and hopelessness and a better quality of life. At the same time, hormonal suppression was a significant but medically manageable complication in these patients, and some patients had elevated hormone levels or inflammatory markers, or both, suggesting the presence of an ongoing painful, inflammatory process that may be characteristic of centralized pain disorders.
Potential effects of the extended use of opioids on the endocrine system have been recognized for quite some time. Of all the endocrine effects, androgen deficiency (hypogonadism) and associated sexual dysfunction particularly in men have been recognized the longest. These effects and their medical management are thoroughly discussed in papers from Pain-Topics [Colameco 2008] and from the Postgraduate Institute for Medicine [Brennan et al. 2012].
The study by Tennant, described above, is noteworthy for its investigation of patients who had been taking high-dose opioid analgesics for an unusually lengthy period of time, as well as for the many clinically meaningful factors that were examined. There appeared to be a favorable balance of sorts in terms of pain relief and functional improvements gained versus risks of hormonal disturbances, some of which were medically manageable via hormonal supplementation.
As Tennant stressed in a news report, “There is not a single person in this group [of patients] who wants to take opioids, and they have tried everything else.” However, the patients “were once bed bound or house bound and totally debilitated, and now can function. Centralized pain is not curable,” he stated, “people die from this type of intractable and centralized pain if they don't get help, and if they're functioning and they're doing well, is there a problem here?"
Of course, Tennant’s study could be faulted for the small number of subjects, only 40 patients, making it more of a large case series with limited external validity or generalizability to larger populations. However, this should not detract from Tennant’s contribution to the field in this study, as well as in other investigations coming from his clinic, that have been pursued without outside funding or support.
In contrast, the study by Deyo and colleagues  — which was funded by NIH/NCRR grants and received institutional support — leaves much to be desired from an evidentiary perspective. For one thing, the present study is a data-mining exercise using a subset of data from 2004, and the larger 9-year-old dataset has been used similarly by this same research team for other articles focusing unfavorably on opioids for chronic noncancer pain [eg, see UPDATE here]. If nothing else, this repeated dredging of the same data increases the probability of Type I errors; that is, statistically finding and reporting false-positive outcomes.
Secondly, the journal in which the Deyo et al.  article appears — Spine — rates the level of evidence as only 4, which is the lowest in its rating system and attests to the poor quality. Here are some further, more specific points of concern:
- Using prescriptions of medications for erectile dysfunction or testosterone replacement as a proxy (ie, surrogate marker) for hypogonadism and sexual dysfunction due to opioid therapy is fraught with potential for error and confounding. Besides the fact that this association cannot be used to infer causation, the researchers themselves concede that, “We cannot equate the use of medications for erectile dysfunction with the occurrence of hypogonadism….”
- Assessments of the various important hormone levels were not available in the database. And, the researchers acknowledge that, even if hormone levels had been measured, hypogonadism and erectile dysfunction may be poorly correlated. Added to that, the singular influence of opioid therapy was muddled by effects of other factors, including age, comorbidity, depression, and use of sedatives.
- The researchers did not examine the influence of pain on sexual dysfunction; either pain scores were not present in the database or not taken into account. And, although all patients had back pain, the original indication for prescribing opioids —possibly involving multiple pain conditions — and the type of opioids used are unknown.
- Although the researchers suggest that long-term use and high-dose opioids were separately “associated with roughly 50% greater odds of using medications for erectile dysfunction or testosterone replacement,” they fail to emphasize that they are referring to Odds Ratios, not Risk Ratios, and the reported ORs of 1.45 and 1.58 are actually only relatively small effect sizes (comparable to Cohen’s d of 0.29 and 0.36, respectively), despite the fact that they are statistically significant. Additionally, the lower limits of the OR Confidence Intervals are approaching the point of being statistically nonsignificant, which further emphasizes that this is weak evidence.
The Number-Needed-to-Harm (NNH) associated with long-term opioid use, calculated from study data [using PTCalcs here], would be about 37. That is, for every 37 patients with chronic back pain treated long-term with opioids, rather than without opioids, 1 additional patient might be prescribed medications for sexual dysfunction. This is a statistically small effect size and the clinical magnitude of this NNH is questionable; especially, for a physical condition that is readily managed medically.
- Of some importance, only 8% of the men in the study population were receiving medications for erectile dysfunction or testosterone replacement, presumably due to hypogonadism and associated sexual dysfunction. According to other sources, the prevalence of symptomatic hypogonadism in the general male population is about 5% to 6% [Brennan et al. 2012]. The 3% to 2% absolute risk difference would amount to an NNH of 33 to 50; that is, of every 33 to 50 men with chronic back pain in the KPNW patient population, 1 additional patient might have sexual dysfunction than typically expected. This NNH is a small effect size that does not seem clinically astonishing, and the addition of opioid therapy has relatively little impact.
- The researchers concede that sexual dysfunction (and related medication use) may have preceded the prescription of opioid therapy in a proportion of cases, particularly when considering patient age and other pre-existing cofactors. Furthermore, the database did not include all diagnoses, such as diabetes, which could have been confounding causes of sexual dysfunction.
In sum, hormonal disturbances are a well-known potential adverse effect of long-term opioid therapy, and patients should be advised of this when soliciting their informed consent to treatment. Along with that, a fairly-balanced discussion would include mention of the many possible benefits of such therapy, as noted in the study by Tennant.
Also, in this regard, potential risks of alternative analgesic therapy should be considered and discussed with patients. For example, in a large prospective study of men who were regular NSAID users [see UPDATE here], the frequent use of NSAIDs compared with no such use increased the odds of self-reported erectile dysfunction by about 70% (Odds Ratio = 1.72; 95% CI: 1.67-1.77; NNH=9). This is a considerably greater negative effect than is associated with opioids in the Deyo et al. study.
While the study by Tennant is woefully inadequate in terms of its size and rigor, the data-mining approach used by Deyo and colleagues may be useful for generating hypotheses for further study, but can otherwise be misleading. This type of research, which is becoming overly popular in the literature, often perpetrates a cum hoc, ergo propter hoc (ie, with this, therefore because of this) fallacy. That is, assumptions are made on the basis of potentially spurious relationships — a form of “guilt by association” — leading to deceptively erroneous conclusions [see discussion of Fallacies of Evidence in Pain Research here].
The sort of sophistry that data-mining research may engender should be considered with great caution. While it may help to support the agendas of those opposing opioid analgesics for chronic noncancer pain, the arguments are based for the most part on weak, unreliable, and/or invalid evidence. The pain management field, and patients with pain, deserve much better.
> Brennan MJ, Guay AT, Morgentaler A. Opioid-Induced Androgen Deficiency: Approaches to Diagnosis and Management. Postgrad Inst Med. 2012; available online [here].
> Colameco S. Opioid-Induced Sexual Dysfunction: Causes, Diagnosis, & Treatment. Pain Treatment Topics. 2008 [available here].
> Deyo RA, Smith DHM, Johnson ES, et al. Prescription Opioids for Back Pain and Use of Medications for Erectile Dysfunction. Spine. 2013(May 15);38(11):909-915 [abstract here].
> Tennant F. Improvement in Physiologic but Not Hormonal Functions After 10 Years of Opioid Treatment. Presentation at the 2013 meeting of the American Association for Pain Medicine; April 11, 2013; Fort Lauderdale, FL. Poster 135 [abstract here].
Don’t Miss Out. Stay Up-to-Date on Pain-Topics UPDATES!
Register [here] to receive a once-weekly e-Notification of new postings.