Serum Opioid Monitoring; where’s the evidence?

Share with others

There is no shortage of guidelines for opioid use in chronic noncancer pain.1-4  All of them advocate in some form that clinicians use universal precautions that include urine drug testing (UDT), controlled substance agreements (CSA), and in most cases incorporate validated risk stratification tools prior to and during opioid therapy.  All attempt a definition at what constitutes “high dose” morphine equivalents and what that means in terms of risk, but few agree on a standard opioid conversion (good topic for another blog).

Clinicians sometimes avoid UDTs and CSAs because it is uncomfortable broaching the topic with long term patients and/or they don’t want to jeopardize the patient-provider relationship or alienate their patients.  The issue reaches a whole different level when I approach the topic with oncologists, palliative care, and hospice providers.  The sad truth is that cancer should not be a free pass to substance diversion; that’s not to say that it is a huge problem, but clearly it has never been quantified and without a doubt, just because a patient or family member develops cancer, it does not cancel out a previous history of substance abuse.  There is after all a subset of substance abusers that do develop cancer, just like any other disease (also a topic for another blog…stay tuned).

In my experience, honest patients generally have no problem participating in various testing or signed agreements when it is carefully explained that all patients are treated equally and that there is an unfortunate public health risk and liability to providers.  Certainly when given the choice of the “pharmacy crawl” or acquiescing to appropriate tests, most good patients would happily select the latter.


No guidelines require or suggest that serum opioid testing is useful or should be a standard of care.  In this blog, I intend to explain why this could be useful and I hope to dispel certain myths on the topic, will wrap up with a couple of cases, and encourage input from my esteemed colleagues and patient followers.

Before making my case however, Pharmacy Pain and Palliative Care PGY2 Resident Dr. Tim Atkinson helped scour the literature to prove that there is limited evidence.  Why you might ask?  (Keep your foot on the clutch for now)

PubMed was searched using blood levels + opioid monitoring yielding 145 results.  All studies mentioning opioids were reviewed for inclusion or excluded if they focused on anesthesia, behavorial monitoring, or dose increases without measuring blood levels.  Most studies included describe various testing technologies (LC-MS/MS vs GC-MS), post-mortem and antemortem concentration correlation, and methadone serum monitoring.  The only articles mentioning serum monitoring were regarding methadone treatment programs focusing on compliance and attempting to correlate doses with levels and predict the effect of various drug interactions.

A linear dose-response can be predicted according to Gergov and colleagues.5 It appears that when they attempted to use a quadratic regression it only correlated well with the highest concentrations perhaps even toxic.  They ended up diluted those higher samples and correlating them with linear reference standards further reinforcing utilization.  In addition, among other challenges they list blood thickening as a problem with using post-mortem samples which obviously won’t be a problem for our patients.

There are at least four reasons to monitor serum opioid levels:

  1. Some patients will have lower or higher than expected serum levels due to polymorphic variations.
  2. Some patients will have lower or higher than expected levels due to significant drug interactions.
  3. Elevated or diminished serum levels could help guide the clinician to select a drug dose for opioid rotation and prevent overdose or underdose (withdrawal).  Consider a referred patient on fentanyl 100mcg/hour patch requiring conversion to OxyContin.  If the patient is either not using the fentanyl, or not absorbing the fentanyl, a conversion to OxyContin could result in death.
  4. Some patients may be taking part or none of their medications…this is a public safety risk in the case of diversion, a significant cost to society, and just plain problematic on many levels.  UDTs do not quantify or predict ingested dose.

The first two above are very important therapeutic issues that could help clinicians to better understand why a patient isn’t responding to therapy or why a patient has toxic side effects that might otherwise be misinterpreted as overuse or misuse by the patient.  This could be a natural progression to pharmacogenomic testing or at least switching to an opioid with a differing metabolic pathway.

Who will pay for this and who is inconvenienced? 

These are questions I am often asked.  First, patients must understand that it is at least as inconvenient for the prescriber to order and interpret the tests as it is for the patient to have blood drawn.  We don’t want to do this anymore than the patient wants it done. Third party payment for serum monitoring is really a “no brainer” and requires education to the payers.  Many years ago I approached certain laboratory folks and administrators about serum testing, but nobody wanted to pay for the tests.  Once I “showed them the money” it became quite clear that serum testing was worth every penny.  Not doing it may be penny wise and pound foolish.

By way of example, a prescription for #60 OxyContin 60mg costs approximately $747.00 at CVS in upstate NY which is a one month supply assuming a dose of OxyContin 60mg every 12 hours.  A one year supply for a single patient costs $8964.00 per year.  An oxycodone serum test using liquid chromatography–tandem mass spectrometry costs $134.00 in upstate NY (price will be lower on contract).  The yearly cost for two tests would be $268.00.

If 50 patients received the dose above, OxyContin over one year = $448,200
Two tests per year for each patient (100 tests total) = $13,400

If 4% (2 out of 50 patients) were not taking their OxyContin, the savings in one year for discontinuing therapy = $17,928.  For just two non-compliant patients, the net gain to third party payers is $4528 per year for testing all 50 patients twice yearly.  Believe me, 4% is probably a fraction of the actual percentage, and the net profit could help to offset serum testing for less expensive extended release products such as morphine and also methadone (which of course is not extended release).


  1. Last year, my colleagues and I reported a case of low serum morphine 6  in a patient that was 75% less than expected by calculation.  The patient had a history of infectious endocarditis and was being treated with rifampin.  As most of you know, since morphine does not engage in cytochrome P450 metabolism, we wouldn’t expect an interaction.  But, since rifampin is a potent p-glycoprotein inducer, morphine absorption was severely impacted by stimulating the gastric efflux pump and preventing absorption. 6 (If requested, you may qualify for a copy)  This concept was hit home with a subsequent publication in an even more frightening interaction with methadone and telaprevir.7  Imagine if this patient had been tested in a clinic with a previous history of heroin abuse and his serum levels were 75% less than predicted…he would have been falsely accused of selling his morphine.  That’s why any test, UDTs or serums, require a keen knowledge of how to interpret unexpected results. And how about this?  What happens when the rifampin therapy is complete and the serrum morphne levels jump by 75%?
  2. In a second case, a patient was receiving oxycodone as follows: OxyContin 80mg PO q12h; OxyContin 20mg PO q8h; oxycodone IR 10mg PO q4h.  Pt uses #240 per month. The UDT for opiates is negative. Quantitative urine confirmation for oxycodone yielded a value of 438ng/mL.  A follow-up serum was reported at 19ng/mL, a level consistent with a daily oxycodone dose of perhaps 30mg per day (but prescribed daily dose was 280mg per day).  “Dose proportionality and/or bioavailability has been well established for 10mg, 20mg, 40mg, 80mg, and 160mg tablets strengths for both peak plasma levels (Cmax) and extent of absorption (area under the curve).  At a dose of 10mg sustained release oxycodone PO q12h, the mean maximum serum concentrations (+/- SD) for sustained release oxycodone has been reported as 15.1 +/- 4.7 ng/mL”.  We can expect 90% confidence levels with these numbers. (Per OxyContin package insert, Reder RF, Oshlack B, Miotto JB, Benziger DD, Kaiko RF)

This patient is taking no more than 30mg of immediate release oxycodone per day.  Street value of the medication balance is over $6000.00 per month, tax free of course!

I think it’s about time something is done to validate opioid serum analysis clinically, educate the opioid prescribing world, and do some prospective studies.  Data is in fact available, as seen HERE which has been available on for many years.  You can be sure that no new opioid or new opioid formulation has been approved in the last several years or has/will come to market without having significant serum pharmacokinetic data. So, yes, we do have the data, but clinicians need to know how and why it is clinically useful and the pitfalls of utilizing such tests.  Nobody squawks about serum carbamazepine, valproate, phenytoin, or LFTs for statins.

In patients that are poor methadone metabolizers, someone with a higher than expected methadone level is at higher risk of widened QTc, the side effect of which is death.  This may also be problematic in certain patients receiving low dose methadone that have elevated serum levels because of a drug interaction.  A good example is the antiretroviral nevirapine; I have seen widened QTc’s in some of these patients, and low and behold, when doing a serum, it is often 4 to 10-fold higher than expected.  And we wonder why 30% of all opioid deaths involve methadone?  Maybe a significant portion of that is not just polymorphism and poor conversion ratios, but also significant and sometimes unexpected drug interactions as noted above with telaprevir.6

And finally, when a coroner reports cause of death to be “opioid overdose” do to serum levels, it is unconscionable to say such a thing without factoring in for “redistribution”8,time and location of sampling, and the amount of opioid a patient was able to tolerate for the last several days, weeks, or years preceding death.

I would be especially grateful for any input from our pain clinician colleagues here and invite comments from patients as well.


  1. Manchikanti L, Abdi S, Atluri S et al.  American Society of Interventional Pain Physicians (ASIPP) guidelines for responsible opioid prescribing in chronic non-cancer pain: Part 2-guidance. Pain Physician. 2012; 15(3 suppl):S67-S116.
  2. Chou R, Fanciullo G, Fine P et al.  Clinical guidelines for the use of chronic opioid therapy in chronic noncancer pain. J Pain. 2009; 10(2):113-130.
  3. American Society of Anesthesiologists (ASA) and American Society of Regional Anesthesia and Pain Medicine(ASRAPM). Practice Guidelines for Chronic Pain Management: An Updated Report by the ASA Task Force on Chronic Pain Management and ASRAPM. Washington, DC: ASA & ASRAPM, 2010.
  4. Model Policy on the Use of Opioid Analgesics in the Treatment of Chronic Pain. 2013. Federation of State Medical Boards.
  5. Gergov M, Nokua P, Vuori E, Ojanpera I.  Simultaneous screening  and quantification of 25 opioid drugs in post-mortem blood and urine by liquid chromatography-tandem mass spectrometry. Forsen Sci Intl. 2009; 186:36-43.
  6. Fudin J, Fontenelle DV, Payne A. Rifampin Reduces Oral Morphine Absorption; A Case of Transdermal Buprenorphine Selection Based on Morphine Pharmacokinetics. Journal of Pain & Palliative Care Pharmacotherapy. 2012;26:362–367.
  7. Fudin J, Fontenelle DV, Fudin HR, Carlyn C, Ashley CC, Hinden DA. Potential P-glycoprotein Pharmacokinetic Interaction of Telaprevir with Morphine or Methadone. Journal of Pain and Palliative Care Pharmacotherapy. August 2013, Vol. 27, No. 3, Pages 261-267.
  8. Cook DS, Braithwaite RA, Hale KA. Estimating antemortem drug concentrations from postmortem blood samples: the influence of postmortem redistribution. J Clin Pathol 2000;53:282–285.

19 thoughts on “Serum Opioid Monitoring; where’s the evidence?

  1. How reliable are vitreous levels of tapentadol on a post mortem exam.?
    My pubmed search produced no results on monitoring serum opioid levels. This is very disappointing given the difficulty of managing chronic pain in complex patients.

  2. We have a patient that is nearly identical to the Case Perspective #2. Serum tests reveal 23ng/mL concentrations with 5) were ‘failed’ indicating no oxycodone ingestion, but positive for cocaine. I am absolutely convinced that this patient is diverting his Rx, but when I contacted DEA, they referred me to local police, who are not interested, as his official residence is in a neighboring town. As a staff physician, our managing physician continues to Rx this patient’s medications. It’s frustrating. The information and effort by Dr Fudin is very much appreciated.

    1. Thank you “staff physician”. If you’re involved with this case, I’d write a comprehensive note and recommend that the patient be enrolled in counseling for cocaine abuse. The oxycodone can be abruptly stopped, and if the patient uses a small amount with any regularity, you could provide clonidine or Lucmyra. Your State Bureau of Narcotics may have interest in pursuing the case.

  3. Is a 3.86 mg/L cardiac blood serum of hydrocodone an.amount that would lead one to think the decedent had intentionally or accidentally overdosed? Can someone direct me to where I might find evidence of varying blood serum levels of hydrocodone if I need to make my own educated guess? Please, even if it is simply a guess. I would appreciate an answer.

    1. Tracy, That is a very very high level and could be cause of death but there are other factors. Cardiac blood is not as accurate as arterial blood because cardiac blood levels of drugs can become more concentrated postmortem. Also, hydrocodone as cause of death depends in the patients that are opioid tolerant must be considered. If the decedent was on high doses of opioids prior to death or there are other factors involved, it isn’t as simple as hydrocodone levels alone as a determinant of death.

  4. I routinely test my patients and am left feeling like I’m missing something when levels come back consistently absent despite adequate dosing. Example: OxyContin 80mg q 8 hrs plus oxycodone 30mg q 6 hrs PRN and a <5 level repeatedly – I cannot help but feel that the patient is not taking any of the prescribed doses. Same with Free Morphine – 100 mg q 8 hours and consistently an undetectable serum level after months of therapy. This makes me feel crazy but I do several samples and counsel along the way before I quit writing narcotics – did I mention I take care of sickle cell patients?

    1. I can tell you with no uncertainty that these patient are not taking any medication. There prescriptions should be stopped abruptly with no taper – you have a huge liability with that information in the record if you continue to prescribe!

    2. Are you saying the serum test showed ZERO drug in their system?? And yet you continued to prescribe or taper? What does their urine screen say? Have you read any medical literature that says zero serum levels are common – even when med is taken? Did you base your decision on the patients taking other drugs and/or other factors discussed in this blog post?

      I applaud your willingness to taper patients BUT doing so under the circumstances you’ve stated is more than likely doing more harm than good and not just to the patient, medical community, and general public but to yourself and your other patients as well.

      May I suggest in the future that when a serum test returns negative that you run the test again, run a urine screen, prescribe just a weeks worth and then call the patient in for a pill count three days in. I’m TOTALLY against pill counts but in this case even I’d make exception.

      Again, I do applaud your compassion and willingness to give the benefit of the doubt. The chronic pain community NEEDS MORE practitioners like yourself.

      However, please also protect yourself from unscrupulous patient dealers. I believe they’re far and few and in between BUT they, sadly, do exist rather by cruel intent to their fellow beings or out of self preservation (ie: needing food, fuel, electricity, other meds, car repair & so forth). Not that self preservation is an excuse or viable reason to divert but as an old proverb says: “Society prepares the crime and the criminal commits it”. If our government wants to stop the the rate of self preservation through moral disengagement then it should offer better opportunities for the disenfranchised, marginalized, depressed and otherwise hopeless population. Better access to income assistance programs, lower the cut off threshold to qualify for assistance programs, FREE secondary education (at least undergraduate & student loan forgiveness), automatic qualification for fuel assistance, food, and transportation when a person must choose to go on disability. Help the working class poor by implementing the aforementioned.

      Anyway, sorry for digressing – going a bit off topic but I really think we NEED to look at diversion as a symptom of poverty and our country’s seemingly inability to see that it sets its people up for failure.

      I BELIEVE people are innately GOOD but when their very survival is at stake … what wouldn’t some people do not to go hungry or homeless?

      I’m probably just too much of an idealist but maybe, just maybe when a patient’s drug test (of any type), returns negative a practitioner could assess the patient’s financial means and then refer the patient to financial assistance programs without any severe or harsh punishments. But, of course create a type of probation where the patient receives one week of medication at a time, random monthly urine/blood screens, medication counts, and psychological counsel. I believe such measures are draconian for the historically compliant patient but hey if someone is breaking the law and betraying their provider they get what they get. I just don’t believe we ought to punish people for being poor and for being placed into a situation where they’re basically forced to go against their own moral code and that of the collective societal moral code.

      We CAN do better! We CAN keep people safe. We just need more compassion, less greed and a willingness to not only think outside of the box but to act outside of the box.

      Thank you for your time.
      My reply is not entirely a response solely for the OP. Just got a little carried away. Bored on the sofa thanks to forced titration. Least today I can concentrate on writing for a little while.

  5. Your posts are always educational and much appreciated. I have done serum drug levels for patients on “high” dose opioids for several years. The reason was to potentially protect myself from accusation of malpractice and overprescribing if a coroner stated the cause of death was an “overdose” of the medication I was prescribing. If I had chart documentation of a serum drug level that a lab may have labeled as “lethal”, yet the patient was obviously alive and functioning well on the opioid, it would be more difficult to blame me. With current legislative changes, it seems even more people are in the business of pointing fingers at prescribing physicians, including coroners.

    1. Debora,

      I’m so sorry for all the stress and nonsense you must deal with as a practitioner willing to help people with chronic intractable pain. But, THANK YOU so MUCH for going the extra mile rather than throwing in the towel! I applaud you!

  6. Thanks for covering this important topic and for highlighting the fact that urine measurements cannot be used to judge compliance to a prescribed regimen. In addition to extensive literature precedent, federal case law now exists on this subject. Hence, oral fluids or serum are the only choices left standing.

    There’s a problem, however. Most mass-spectrometry labs have been built for the current gold rush, with accuracy being an irritant rather than a goal. Urine concentrations are so much higher than corresponding serum and oral fluids concentrations. With a huge quality gap for urine specimens, few of these labs have any hope of achieving meaningful accuracy for other specimen types. Hence, we are stuck with a industrial focus on urine as the easiest specimen – the specimen type does not control the insurer payment.

    1. Jim,
      Thank you for your insight.

      What does “the specimen type does not control the insurer payment.” mean? I’m just curious.

      Thanks for your time.

  7. Physicians are often concerned about patients’ feelings about drug testing worrying that patients may not feel trusted or respected by a doctor who asks them to submit a UDT.
    One limitation of urine drug testing is that it provides no information on how much drug has been taken. However this issue really doesn’t come into play it’s more a matter of determining whether they’re in the patient’s system or not. But providers are also concerned that their patients may be taking only some of their prescribed pain medication, they may be taking only one pill of their prescribed medication each morning in order to pass their urine drug test and then selling the rest. Because there is no linear relationship between what’s detected in the urine and how much drug was actually ingested, as long as the drug shows up in their urine, that patient is going to appear compliant. To find out how much drug was taken one must go back to pharmacokinetics and look at steady state blood levels.

  8. Dr. Fudin,

    Great post! I believe that monitoring serum opioid levels is a smart and effective method to enhance patient outcomes, especially in patients prescribed opioids for chronic pain. It allows one in healthcare profession to assess quantitative data not just to provide therapeutic levels for the patient, but to actually determine if the patient is adherent to the instructions as means to provide safer and smarter prescribing of opioids. I have experienced the importance of monitoring a patient’s serum opioid levels first hand through completing your 6 week pain management rotation at Stratton VAMC. I have also learned the importance of using tools such as the SOAPP and COMM to assess the likelihood of abuse in a patient who may have predisposing risk factors for medication harm (another potential for a great blog post!). All in all, I believe providers should be notified of the importance of utilizing serum opioid levels. Hopefully in the near future, physicians and pharmacists could take advantage of this technology as rationale behind prescribing opioids safely and effectively for their patients.

  9. I did use serum levels in some patients when I was in practice. They can also be useful for use when the patient needs emergency room care where many if not most chronic pain patients are treated as addicts or undeserving drug seeker/dealers first and people with medical problems second or never. They can also be useful as guidelines for post operative pain management. A patient adequately controlled and functioning well on a given dose with a specific serum level cannot be expected to do well post-operatively if they are given a dose of opioids that produces a lower serum level.

    Second, the issue of variable absorption is given little attention but may be a significant factor in some patients. I had one patient who literally required a double dose to achieve any measurable serum level of hydrocodone, with the patient being witnessed taking the pills and using pre-pill and post pill peak and trough levels. The standard dose of hydrocodone produced no measurable serum level. He did well using an intermittent “high” single dose, while never exceeding daily maximal acetaminophen dose. He was able to work full time. A subsequent blog on factors affecting absorption of opioids would be appreciated.

    1. Doc for the people:
      Your request for “A subsequent blog on factors affecting absorption of opioids would be appreciated” is a great idea. While I did not cover food or specific pH type drug-drug interactions (rare for opioids) or first pass prodrug issues (only affects codeine), I did cover p-glycoprotein absorption issues in great detail in the article below. If you need a copy, let me know.

      Fudin J, Fontenelle DV, Fudin HR, Carlyn C, Ashley CC, Hinden DA. Potential P-glycoprotein Pharmacokinetic Interaction of Telaprevir with Morphine or Methadone. Journal of Pain and Palliative Care Pharmacotherapy. August 2013, Vol. 27, No. 3, Pages 261-267.

  10. This article is great. I will keep it as a reference.

    For my practice (Chronic Pain), we only use serum levels when a patient is unable to offer a urine sample. I find the serum levels to be useful, but in a different way than the urine levels. Regardless, understanding the pharmacokinetics is necessary to clinically interpret the serum results in a way that is fair to the patient.

    Thanks for all the hard work in this article.

    – James Patrick Murphy, MD, MMM

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.