Wednesday, January 6, 2010

Reviewing the Evidence for Opioids and Cancer Growth

Eric Widera recently posted about the media's irresponsible headlines implying that opioids cause cancer growth. This is our most read post to date (over 1000 direct visits plus email subscribers).

Why was this post so widely read? Because this issue touches on a critically important therapy and the media were clearly irresponsible in their treatment of this story. The media took a story about an opioid antagonist (methylnaltrexone) potentially inhibiting cancer growth in mice, and converted this into a story about morphine causing cancer growth.

As a follow up to Eric's post I emailed the authors of the study, asking them if they would like to comment or post a response. One of the authors called me to give his opinion. He said they have avoided talking to the media, declined multiple interviews with the press, and would not post a reply on our blog. However, he did state that he believed there is strong reason to suspect that opioids may promote cancer growth, and he wanted me to read the evidence. To that end, he sent me 7 articles to review. I'd like to go over them briefly here.

  1. Singelton, J. et al. Synergistic effects of methylnaltrexone with 5-fluorouracil and bevacizumab on inhibition of vascular endothelial growth factor–induced angiogenesis. Mol Cancer Ther 2008;7(6):1669–79. The authors found that methylnaltrexone reduces the concentration of two chemotherapeutic agents necessary to inhibit cancer cell growth in vitro.
  2. Moss, J. From Bench to Clinic: Our Story with 21st Century Drug Development. Association of University Anesthesiologists Update, Spring 2009. This is a newsletter story discussing the development and testing of methylnaltrexone for treatment of constipation and potential other uses.
  3. Durieux, M. Does Anesthetic Management Affect Cancer Outcome? Anesthesia Patient Safety Foundation newsletter, Winter 2009. A newsletter reviewing the evidence for anesthetic technique influencing cancer surgery outcomes. Regarding opioids, the author notes that one the one hand, opioids have been shown to promote angiogenesis (formation of new blood vessels necessary for tumor growth). On the other hand, opioids have been shown to decrease metastases around the time of surgery in rats, likely due to a decrease in the pain-related stress response (stress can promote cancer spread).
  4. Moss, J, and Israel, R. Effects of Anesthetics on Cancer Recurrence. Letter to the editor. Journal of Clinical Oncology 2009. This letter is a review of prior studies.
  5. Moss, J, and Rosow, C. Development of Peripheral Opioid Antagonists: New Insights Into Opioid Effects. Mayo Clinic Procedings 2008. This is another review.
  6. Singelton, J. et al. Methylnaltrexone inhibits opiate and VEGF-induced angiogenesis:
    Role of receptor transactivation. This was a laboratory study where the authors found that morphine promotes endothelial cell migration.
  7. Mathew, B. et al. The Mu Receptor Regulates Lewis Lung Carcinoma Tumor Growth and Metastasis. This is the abstract that garnered so much press. In lung cancer cell lines and mice, the authors found that methylnaltrexone inhibited lung cancer growth by blocking the opioid receptor.
So 2 published studies, 1 abstract, 2 newsletters, 1 letter to the editor, and 1 review. Note that Eric discussed the two articles from Anesthesia on regional versus general anesthetic technique in the previous post. The author of one of those studies, in a letter to the editor, acknowledged that "Under no circumstances should a small retrospective study be the basis for practice."

So what is the take home message? (and I want to hear yours). I think there is evidence that the opioid receptor shows some potential promise as a target for reducing tumor growth. Is there sufficient evidence to state that opioids promote cancer growth? Absolutely not. Opioids have not been tested as potentiating tumor growth in humans. Pathophysiologically, opioids might promote cancer growth (via angiogenesis) or they might reduce cancer growth (via a reduction in the stress response).

Why is it important to not leap to the conclusion that opioids cause cancer growth? Two recent examples come to mind: hormone replacement therapy for post-menopausal women and vitamin E. Early retrospective studies (in humans, not mice) seemed to demonstrate that hormone replacement therapy led to reduced mortality. Unfortunately, large prospective cohort studies showed the opposite to be true! When researchers went back to re-examine the earlier studies, they found those earlier studies failed to account for important baseline differences in women: those who were healthier were more likely to take hormone replacement therapy, and therefore had better health outcomes that were attributed to the drug. Similarly, there was a great deal of excitement about vitamin E, including bizillions of laboratory studies that showed plausible mechanisms by which vitamin E could reduce heart attacks (antioxidant properties, etc). Again, large studies showed the opposite: people who do not take vitamin E live longer and have fewer heart attacks than people who do take vitamin E. In both the case of hormone replacement therapy and vitamin E, patients were taking the drug and health care providers were recommending treatment based on early evidence. And people died.

Some are already leaping to conclusions.  Here's a quote from a physician on the blog Medpie:
For patients due to undergo surgery for cancer, it is probably worthwhile for them to have a frank discussion with their surgeon and anesthesiologist about whether regional anesthesia is feasable (sic) and safe in their case, and whether opiate analgesia can be minimized. patients should "discuss avoiding general anesthesia and minimizing opioid medications).
As Christian Sinclair (of Pallimed) noticed, if you Google search "morphine cancer" the fourth hit is "Pain drug morphine may accelerate cancer growth | Reuters." Think of all of our patients and caregivers who, seeking information, are clicking on that link.

Let's set the record straight.

11 comments:

Eric Widera said...

Great review Alex. I just did a google search on "morphine" and "cancer". Just the first page of results revealed these jaw dropping headlines:

- Morphine causes cancer to spread faster
- Pain drug morphine may accelerate cancer growth | Reuters
-Common Pain Medication Accelerates Growth of Cancer Tumors‎

I can't imaging the fear a cancer patient would have after doing a similar google search. Unfortunately, as you so nicely put forth in your post, this fear is not based on any clinical evidence.

Dan Matlock said...

What if morphine does promote cancer growth?

This will be nearly impossible to confirm or refute in clinical studies. An RCT of cancer patients with a morphine arm and a placebo arm would almost surely be plagued by crossover and the confounder of stress effects from untreated pain on cancer growth. (not to mention the fact that an RCT with an arm of untreated pain in cancer shouldn't pass an IRB).

Likewise, large observational studies will have selection bias as patients with growing tumors will almost certainly be more likely to receive narcotics.

What do we say to patients then? I think there are two ways we could handle this problem. And I'm honestly not sure which is better yet.

1) We could, as Eric and Alex have suggested, tell our patients that there is no evidence in clinical studies. I think this is clearly an accurate statement and I thank Eric and Alex for their thorough review.

2) Alternatively, we could acknowledge that there is some weak evidence that narcotics may cause cancer growth and then have a shared decision with the patient regarding the risks and benefits. I, for one, would still take narcotics if I had painful metastatic CA even if it did promote cancer growth. Unfortunately, we really have no alternatives for pain control.

I think both approaches are valid but I'm leaning towards the latter as I think it would foster more trust among patients.

Eric Widera said...

Our recommendations should change depending on what we are calling weak evidence. If we are to trust the clinical data and consider this to be enough to warrant having an open discussion with our patients regarding the risk, then we should really discuss the benefits of regional vs general anesthesia, not the harms just of opioids with cancer recurrence. This line of evidence isn't strong enough for me, but if others think it is than that would be reasonable. Again, the benefits seen in these trials can be due to a multitude of reasons, including, lower opioid usage, lack of volatile anesthetic use, or confounders not adjusted for.

I am more concerned, though, if we rely just on the important but preliminary bench research on this subject, especially since the most recent data on methylnaltrexone has not even been published in a peer review journal yet. I view my role as a physician as a filter for patients when discussing promising yet prelimiary lines of evidence. This wouldn't even come close for passing through my filter.

brimcmike said...

What if opioids promote cell survival and angiogenesis among some types of malignant cells?

I think there is some convincing cell biology data that supports this hypothesis in this unreplicated paper: Gupta K, Kshirsagar S, Chang L, Schwartz R, Law PY, Yee D, Hebbel RP. 2002. Morphine stimulates angiogenesis by activating proangiogenic and survival-promoting signaling and promotes breast tumor growth. Cancer Res. 62(15):4491-8.

There is also a lot of theoretical cell signaling information that supports the notion that stimulation of Gi-subtype GPCRs, in general, and opioid receptors, in particular, can promote cytoskeletal remodeling, promote cell survival, change gene regulation/protein expression, including promoting cell migration and mitogenic proliferation, along with angiogenesis. Jordan B, Devi LA. 1998. Molecular mechanisms of opioid receptor signal transduction. British Journal of Anaesthesia. 81:12-19. See, too, canonical cell signaling pathways, for Gi-subtype GPCR and VEGFR, posted all over the internet.

Consider too, the well-documented phenomena of GPCR-RTK cross-talk, cross-regulation and synergy, and there are added theoretical underpinnings for this possible side effect. Gavi S, Shumay E, Wang H Malbon CC. 2006. G-protein-coupled receptors and tyrosine kinases: crossroads in cell signaling and regulation. Trends in Endocrinology and Metabolism. 17(2):48-54.

Add the concepts of derrangement of physiological processes by interposition of disregulated intracellular proteins in transformed cells, e.g., overexpression of Gαs subunits competitively inhibiting association with Gβγ subunits by physiologic Gαi subunits, or expression of mutated oncoproteins/tumor suppressors, paraneoplastic syndromes, etc., and then all bets are off. Dorsam RT, Gutkind JS. 2007. G-protein-coupled receptors and cancer. Nature Reviews Cancer. 7:79-94.

I do regret and resent the sensationalization of science and medicine at the expense of the public interest of patients' interests. That being said, I'm with Dan, I think we need to be honest with patients and earn their trust, and acknowledge that there is a theoretical basis for an even more insidious double-effect of opioids. Tempered with Alex's well-reasoned and well-documented argument against changing standards of treatment based on plausible hypotheses.

What if opioids promote cell survival and angiogenesis among some types of malignant cells? On a deeper level, if true, what does that mean about pathogenesis, virulence, natural history, evolution, co-evolution, not to mention some uncanny teleology. I can see the function of inducing diarrhea, sneezing and coughing for infectious organisms, but cancer producing malignant pain to coerce the provision of opioids? Talk about selfish genes.

Christian Sinclair, MD said...

I would take the approach Dan offered as the first one:

1) We could, as Eric and Alex have suggested, tell our patients that there is no evidence in clinical studies.

But that discussion would not simply dismiss the patient/family concern as I am afraid some hospice and palliative care staff might do if they have not seen these two blog posts here. I would think the average HPM staff would say, 'That's not true, and go into the classic reasonings why opioids are helpful for cancer pain without acknowledging the headlines the patient/family so clearly read. That approach would be sure to lose trust in the staff.

But with an informed HPM staff, you could offer a simple rebuttal based on a simplified version of what Eric and Alex has shared. I may work on that script and post it to Pallimed with a big fat link to this great example of medical journalism via blogging.

Do you think we will see this type of discussion pop up in medical journals? Probably not. Blogging is a great medium for these types of discoveries and discussions.

CH said...

It's shocking to me that physicians are out there on the internet spreading this false information to cancer patients to "avoid general anesthesia and opioid medications." As an anesthesiologist, I have fortunately not yet met any patients who have asked about this, and I have taken care of many patients with cancer. Sure, many could benefit from regional anesthesia, but I'm not sure it's because it decreases their risk of metastases in the future.

I actually think it would not be too difficult to do a randomized controlled trial, pitting general anesthesia with and without regional anesthesia for pain relief, and following the patients for development of mets. There is no "standard" or "best" way to provide analgesia for many of these postop patients - whether with local anesthetics or with opioids. So we may have our answer to this debate sooner rather than later.

Chrissy Kistler said...

What a great discussion! I read the MedPie article as well and was really disturbed by their lack of insight into "confounding by indication".

What I wonder is, even if there is a biologically plausible pathway for opioids accelerating cancer, how much of a true mortality impact does that have. There have been plenty of instances when we though there were biologically plausible reasons to think something might work, like acetylcholinesterase inhibitors in ALZ, only to find that they don't pan out in the real world, their effects are modest at best.

The actual size of the effect may be a drop in the ocean compared to other tumor promotors going on in the body. If we take out the teaspoon, will it stop the bathtub from overflowing?

I worry people may be reaching for something to hold on to, if I just suffer through it, I can get better; or conversely something to blame, if I die it's b/c of the morphine the docs gave me... This black and white thinking is dangerous all round. I think it is much too early to come down on either side of this.

As an aside, I have been interested in the amount of cancer the health care system causes via overutilization of radiologic technologies (principally CT). However, the data suggests that it's still only about 2% of all cancers... you'd do much better to stop smoking, get colonoscopies, etc.

I googled "CT scan cancer" and the first 5 hits were all about CTs increasing cancer risk. The one that stuck out was http://www.cnn.com/2010/HEALTH/expert.q.a/01/04/ct.scan.radiation.exposure.shu/
where a woman asks about the risk to her 3 yr old who had 6-7 scans due to a head injury at age 11 months. The physician's response seemed well reasoned:
The bottom line: Ultimately, the risks of radiation exposure must be weighed against the risks of not getting medical tests that are needed to help care for you or your family. I encourage all patients to ask questions like these before any medical procedures as you weigh all the risks and benefits involved.

You could craft something for opioids if/when the time comes, but I'm not sure the time is now.

WhyMommy (Susan) said...

As a cancer patient, I have very much appreciated my doctors' frank assessment of recent studies. It lets me know that she's up-to-date and helps me understand the science behind the sensationalism.

Thank you for your writing here; it is setting my fears that perhaps my pain medicine use caused my cancer to grow faster (raised by the headlines) to rest.

Anonymous said...

What a great resource!

Marian Grant said...

I think this is another bump on the road to true evidence-based practice. There is still the impulse to jump on preliminary data, without checking whether it's "good", valid, generalizable, etc. We in the profession need to keep asking those questions so that others, including the media, will see that just because something's "published" doesn't mean it's valid data. Keep up the good work!

Rogue Medic said...

If, instead of a theoretical risk of a treatment, this were a theoretical benefit of an unused treatment, would we be viewing the publicity over this new cure for cancer any differently from any other wonder cure/scare story?

Would anyone feel the need to provide a full disclosure of weak evidence that an unused treatment might provide some benefit, if it were used?

Is there any evidence that the theoretical harm is greater than the known benefits of opioid medications?

Surrogate endpoints are misleading. Is this anything more than a surrogate endpoint?

Is there a good reason to believe this is providing useful information, rather than just increasing the noise in the data?

When the argument is mostly supported by letters to the editor and newsletters, why should we pay attention? If there is ever good research, we should pay attention to it. Where is the good research?

Maybe this will give new life to the mixed agonist/antagonist opioid advocates. From what I have seen of mixed agonist/antagonist opioid use, they are not effective at pain management. The research on mixed agonist/antagonist opioids also does not promote optimism.