We all want to alleviate pain and suffering. Most critically ill patients are treated with opioids for this reason. Unfortunately, opioids have numerous side-effects including delirium, constipation, vomiting, and delayed extubation. Opioid infusions may eventually lead to withdrawal, causing pain, nausea, and depression. This post explores the optimal use of systemic medications to control pain while minimizing complications (1).
The classic World Health Organization analgesia ladder is shown above. Although we don't follow it exactly, it underlies our conceptualization of analgesia. However, it was developed in 1986 and wasn't designed specifically for critically ill patients. Modern management of pain among severely ill patients requires a revised strategy.
Getting started: Drugs that aren’t very useful among critically ill patients
Unhelpful: Weak opioids
Traditionally, weak opioids were utilized as the second rung of the analgesic ladder (e.g. codeine and possibly tramadol). However, there is no apparent benefit of using a weaker opioid, as compared to using a low dose of a more potent opioid. The 2013 SCCM guidelines reported that opioids have equivalent efficacy, provided that they are dosed appropriately.
Tramadol is sometimes given under the guise of being a “non-opioid,” but it does stimulate opioid receptors. It has some unique disadvantages compared to other opioids. Due to differences in patient metabolism, its efficacy varies erratically. Compared to most opioids, it has additional side-effects including seizures and serotonin syndrome (2).
Unhelpful: Nonsteroidal anti-inflammatory drugs (NSAIDs)
The use of NSAIDs in critically ill patients is limited by nephrotoxicity. Kidney injury is extremely common in critical illness, correlating closely with increased mortality. NSAIDs are one of the most common nephrotoxins that contribute to hospital-acquired renal failure (Nash 2002). Other complications from NSAIDs that are problematic include gastric ulceration and coagulopathy.
The role of NSAIDs among critically ill patients remains controversial. It's probably safe to use some agents for limited periods of time, among carefully selected patients. However, my preference is usually to avoid them entirely, for the following reasons:
- It's hard to predict which patients may develop renal failure.
- Renal failure is an insidious process, making it impossible to discontinue the NSAID before a significant amount of injury has occurred.
- Many sources recommend limiting NSAID use to a few days, but pain often lasts longer than that.
Proposed analgesic ladder for critically ill patients
Above is one possible revision to the analgesic ladder. This may not apply to all patients, nor is it intended to be followed rigidly (3). Components of this strategy are explored below.
Acetaminophen
Acetaminophen has traditionally been a front-line agent for analgesia, constituting the first rung of the analgesic ladder. Evidence continues to support that acetaminophen deserves this position, based on moderate efficacy with an outstanding safety profile.
The availability of IV acetaminophen has caused a resurgence of research proving its efficacy. For example, one study of surgical ICU patients showed that IV acetaminophen reduced opioid requirements, improved pain control, accelerated extubation, and reduced nausea/vomiting (Memis 2010). Although oral acetaminophen is less flashy, it probably has similar efficacy.
The main concern with acetaminophen administration is hepatotoxicity, but it may still be used in most patients with liver disease. Patients with cirrhosis or severe alcoholism may receive two grams daily (Imani 2014). In acute liver injury or alcoholic hepatitis, acetaminophen should be avoided. Acetaminophen does suppress fever, but it is generally unclear whether fever suppression is beneficial or harmful. In some situations (e.g. neutropenia), a fever would be diagnostically useful and therefore fever suppression is undesirable.
Pain-dose ketamine infusions
Safety & side-effects
Pain-dose ketamine infusions are very safe, even among profoundly unstable patients. Hemodynamic effects are generally minimal, occasionally with a slight increase in blood pressure. Although ketamine is sometimes feared because it is an “anesthetic” agent, a pain-dose ketamine infusion is safer than opioids. Unlike opioids, ketamine doesn't suppress respiration.
Several studies have reported the successful use of ketamine infusions for critically ill patients as an analgo-sedative agent at dissociative doses (e.g. 1-5 mg/kg/hr, about 10-50 times higher than pain-dose ketamine infusions; Patanwala 2015). Even at these high doses, the safety profile of ketamine was excellent. This is further evidence that pain-dose ketamine infusions are safe.
Psychomimetic effects are occasionally seen at the higher end of the dose range. These may include somnolence, agitation, euphoria, or hallucinations. Such symptoms aren’t nearly as dramatic as a classic ketamine re-emergence reaction. If the symptoms are bothersome, they may be managed by holding the infusion temporarily and then restarting at a lower rate. The rate of side-effects with an infusion is lower than with a bolus of 0.3 mg/kg ketamine (Motov 2017). Furthermore, a patient who has an aversive reaction to a ketamine bolus should be able to tolerate a low-dose infusion (no, they don't have a “ketamine allergy”).
Psychomimetic effects are a greater concern among intubated patients who may be unable to communicate if they are having disturbing hallucinations. This risk may be reduced by co-administration of propofol and/or antipsychotics, which such patients will often be receiving already (4). As with any infusion in critically ill patients, clinical effects should be monitored carefully.
Benefits of ketamine infusion
Ketamine infusions allow for a lower opioid dose to be used, avoiding opioid toxicity. This may be particularly beneficial among patients at higher risk for opioid side-effects. The classic example are patients with COPD or obesity hypoventilation syndrome, who are at high risk for respiratory suppression. Another example would be patients with ileus or pancreatitis, who are at high risk for bowel hypomotility.
Ketamine attenuates the development of opioid tolerance and opioid-induced hyperalgesia (Barr 2013; Angst 2003). This is particularly useful among patients on opioid infusions, where tolerance and hyperalgesia may become problematic:
Ketamine has been shown to be rapidly effective for depression, even at low doses (Rasmussen 2013, Zarate 2006). Depression is a substantial problem among critically ill patients, but we generally overlook it because we don't have a good treatment for it. It is possible that ketamine exposure could improve depression, thereby accelerating physical therapy and rehabilitation.
Evidence
Dozens of RCTs support the use of ketamine for post-operative pain. These studies show that ketamine reduces the required dose of opioids by ~40%, reduces nausea/vomiting, and has an excellent safety profile (Jouguelet-Lacoste 2015). There is increasing experience regarding the use of low-dose ketamine infusions among critically ill patients, for example:
- Guillou 2003 performed a randomized, double-blind trial of ketamine infusion among 101 patients in a surgical ICU. Ketamine caused a 25% reduction in morphine consumption (p<0.05), without any difference in adverse events.
- Moitra 2016 described the use of ketamine infusions in four patients with prolonged ICU admission at Columbia University Medical Center. Perhaps most notable was one case where ketamine (0.18 mg/kg/hr) was successfully used to treat depression.
- Buchheit 2017 published a retrospective report of 40 intubated patients on opioid infusions who were treated with pain-dose ketamine infusions within the surgical ICU at Massachusetts General Hospital. Ketamine was found to be effective in reducing opioid consumption, without any reported side-effects.
Ketamine at low doses was able to serve as a backbone for weaning both opioids and sedatives while maintaining adequate comfort – Buchheit 2017
Where should ketamine fit in the algorithm?
Most evidence describes using a pain-dose ketamine infusion in combination with opioids (Steps 3-4 on the ladder above). In practice, this is the most common application of ketamine infusions.
However, ketamine provides analgesia on its own right and doesn't necessarily need to be combined with opioids (Motov 2015). For example, acetaminophen plus ketamine infusion could be used to provide a moderate amount of analgesia in a patient with hypercapnia and very weak respiratory drive (Step 2b on the ladder above).
PRN opioid
Among critically ill patients, the most commonly used PRN opioids are intravenous morphine or hydromorphone. Hydromorphone may be superior because it causes less histamine release. Although fentanyl is excellent for peri-procedural analgesia, it’s not great for longer-duration pain control. Fentanyl has a short half-life, so when used PRN it must be given very frequently.
Patient-controlled analgesia (PCA) pumps are occasionally useful for patients who are conscious enough to operate them. When used for patients who aren't on chronic opioids, they should be programmed without any basal rate. Using a basal rate doesn’t improve pain control, but it does increase the risk of respiratory suppression (Palmer 2010).
Among intubated patients, it may be reasonable to use a relatively generous dose of PRN opioid (e.g. 1 mg IV hydromorphone). Using an adequate PRN dose increases the likelihood that a PRN strategy will succeed, thereby avoiding an opioid infusion. Among non-intubated patients, a lower PRN dose may be preferable to avoid respiratory suppression.
Opioid infusion
Theoretical exploration of opioid infusions
Let's take a moment and consider different opioid strategies from a purely theoretical standpoint. Patients typically have some pain at baseline, upon which is superimposed intermittent bouts of pain. Our goal is to treat the patient's pain with the minimal exposure to opioid possible.
The simplest strategy is to use an opioid infusion with no PRN doses (#1 above). This will effectively treat the patient's pain, but it exposes them to much greater opioid dose than is actually needed.
The second strategy is to only use PRN boluses of an opioid with a half-life of ~2 hours (e.g. hydromorphone or morphine). This works reasonably well. Bolus doses are used to treat bouts of pain. Because the opioid takes a while to wear off, these bolus doses also end up covering the patient's baseline pain. Patients are exposed to a bit more opioid than they need, but not much.
The third strategy is a fentanyl infusion plus PRN fentanyl boluses. This is theoretically the most attractive strategy (#3 above), which may be fit most precisely to the patient's level of pain. Fentanyl has a short half-life, so boluses may be used to increase the opioid level briefly to treat a bout of pain.
Unfortunately, in practice a fentanyl infusion strategy usually ends up looking more like #4 above. It is difficult to adjust the infusion precisely to match the basal level of pain. Generally, the infusion ends up being left at an unnecessarily high rate (e.g. the infusion is up-titrated during a bout of pain and continued at an excessive rate). Accumulation of fentanyl in fat tissue will also tend to cause drug levels to rise over time. Eventually, this will result in exposing the patient to unnecessarily high drug levels.
In summary, a fentanyl infusion could theoretically be the best strategy, but it must be meticulously titrated with very diligent weaning of the basal rate (otherwise it will overdose the patient). In contrast, a PRN-only strategy using morphine or hydromorphone is a good strategy, which is harder to screw up (it will naturally auto-wean itself, making large overdose impossible).
Evidence regarding opioid infusions?
There are numerous studies evaluating the benefit of pain-dose ketamine infusions under various circumstances. In contrast, there doesn't seem to be any data showing a benefit to adding an opioid infusion. The closest I could find are the following:
- As mentioned above, for patients being treated with PCAs, using a basal opioid infusion has been shown not to improve pain control compared to PRN doses only (Palmer 2010). This seems counterintuitive, but may be explained by figure #2 above: the PRN boluses used to cover bouts of pain last long enough to cover both breakthrough pain and baseline pain.
- Strom 2010 documented the ability to safely manage intubated ICU patients with PRN morphine boluses only, proving that the routine use of opioid infusions isn't needed.
- Wanzuita 2012 showed that replacing fentanyl infusions with methadone accelerated extubation, implying that fentanyl infusions prolong intubation.
Risks of opioid infusions and how to mitigate them
Opioid infusions are very powerful, but with great power comes a great number of side-effects:
- Continuous opioid infusions lead to the development of tolerance, dependence, and eventually withdrawal (when the infusion is reduced).
- High-dose infusions may cause opioid-induced hyperalgesia, a phenomenon wherein opioids paradoxically exacerbate pain (Lyons 2015).
- Ideally, opioid infusions should be down-titrated on a regular basis to ensure that an unnecessarily high dose of medication isn’t being used. However, in practice this often doesn’t happen as frequently or aggressively as it should. Failure to down-titrate fentanyl infusions may cause accumulation in fat tissues, leading to over-sedation that delays extubation.
In fairness, fentanyl infusions generally work nicely for a few days (before tolerance, dependence, and fat accumulation become problematic). So fentanyl infusions are fine for initial patient stabilization. They may be especially useful for patients with ARDS or asthma, wherein suppressing respiratory drive facilitates ventilator synchrony. However, once the dust has settled, if the patient is going to stay on the ventilator for more than a couple days then a more balanced analgesic strategy may be better.
When an infusion is necessary, its safety may be maximized using the following strategies:
- Opioid infusions should be used to manage baseline pain, not intermittent pain (e.g. pain due to turning and suctioning). Thus, a fair amount of opioid should still be required in the form of boluses. If the patient is receiving very few boluses, this indicates that the opioid infusion is being used to treat all pain, which results in the use of unnecessarily high opioid doses (#1 in the figure above).
- A relatively low infusion rate should probably be used for most medical ICU patients. Providers often don't realize the total cumulative dose of opioid that they are exposing the patient to. For example, a continuous infusion of fentanyl 100 mcg/hr is equivalent to giving the patient ~480 mg of oral oxycodone daily. Most opioid-naïve patients in a medical ICU probably shouldn't need this much opioid.
- Care should be taken to titrate the opioid infusion for control of pain, not control of agitation or delirium (which are better managed with a sedative or antipsychotic). Too often, high-dose opioid infusions are used as a nonspecific “calm down” drug, abusing their weak sedative and euphoric properties.
- Any infused drug will require 4-5 half-lives to reach steady state concentrations. Therefore, increasing the infusion rate will have little immediate impact on pain control. For persistently inadequate pain control, it's better to give a large bolus with a small increase in the infusion rate, rather than vice versa.
- There should be a protocoled approach to regularly down-titrate the infusion.
- Co-infusion with ketamine may limit tolerance and opioid-induced hyperalgesia.
- Uncontrolled pain and opioid side-effects are both major problems for critically ill patients.
- Multimodal analgesia that utilizes non-opioids to minimize opioid toxicity is broadly accepted in anesthesiology and postoperative pain management. However, there has been poor penetration of this concept among critically ill patients in general (Payen 2013).
- An analgesic strategy utilizing different classes of medication (acetaminophen, ketamine, and opioids) may allow for effective pain control with fewer side-effects.
- There is little evidence supporting the use of prolonged opioid infusions among the critically ill. Opioid infusions should be avoided when possible, or used with caution.
- An analgesic ladder for critically ill patients is proposed:
Related
- Fentanyl infusions for sedation: the opioid pendulum swings astray? (PulmCrit)
- Opioid-free ED with Sergey Motov (EMCrit)
- Ketamine Brain Continuum (Reuben Strayer)
- Editorials about tramadol from Matthew DeLaney & David Juurlink
Notes
- Nonpharmacologic therapy (e.g. repositioning, lidocaine patches, music therapy) and epidural/regional anesthesia are excellent approaches to analgesia.
- The risk of seizure or serotonin syndrome is low, but this may become a greater problem among critically ill patients who are often on a number of medications. Overall tramadol seems to add a lot of drug-drug interactions and potential toxicity without being very effective.
- Please note that this ladder is designed to deal with ongoing pain, not procedural pain. For a pre-planned procedure, different strategies will be helpful (e.g. ketamine boluses).
- Ketamine emergence reactions respond nicely to low doses of benzodiazepines. Propofol has the same mechanism of action as benzodiazepines (GABA-receptor stimulation), so it would be expected to prevent/treat psychiatric reactions from ketamine. Some anesthesiology literature supports the concept that the combination of ketamine plus propofol causes a low rate of psychiatric reactions (Ishahara 1999).
- PulmCrit Blogitorial – Use of ECGs for management of (sub)massive PE - March 24, 2024
- PulmCrit Wee: Propofol induced eyelid opening apraxia – the struggle is real - March 20, 2024
- PulmCrit wee: Why I like central lines for GI bleed resuscitation - March 13, 2024
Hi Josh
Great post. On my Unit we’re using remifentanil infusions to provide the baseline analgesia. This is used as an infusion only and never bolused. This is typically combined with an infusion of 2% propofol.
Do you view this as a suboptimal analgesia strategy?
It seems pretty good and the only difficulty I’ve encountered is patients becoming very jittery when it’s discontinued abruptly- they appear to be withdrawing.
What are your thoughts on this strategy?
I don’t have much experience with remifentanyl. Based on my reading it has similar pharmacodynamics but superior pharmacokinetics compared to fentanyl. So remifentanyl will cause tolerance, withdrawal, and hyperalgesia if continued long enough and at high enough doses. However, remifentanyl does have a major advantage compared to fentanyl in that it doesn’t accumulate in fat tissue. This probably depends a lot on the doses you are using and duration. For short-stay patients remifentanyl is probably fine. For patients who are going to be in the ICU for longer, you may run into dependence/withdrawal. Overall I think the general concepts discussed… Read more »
This is absolutely awesome! I was just about to start sketching out a new pain management regime for the ICU, and the first thing I ran across was this write-up. I don’t think I have much to add, to be honest. As an anaesthetist working in the ICU, I am a bit biased towards nerve-blocks and epidurals, but I know that is not available everywhere (and I know you mentioned it as well). The one drug that I would add is probably Clonidine as an adjuvant to the opioids. I’ve had some great experience with that in the past. Anyway,… Read more »
Anders, Many of us in the US have switched to dexmedetomidine as our sedative of choice to gain the same opioid-sparing effects as the clonidine
Thanks Scott,
Agree about the dex, and I use that too. But clonidine can be given orally as well, which is sometimes more practical, especially when preparing to transfer out of the ICU.
fantastic. what kind of doses do you use IV and Oral. Any problems with BP
Sorry for delayed response. The subscription seems out of order…
The starting dose depends a bit on how much opiods (and for how long) the patient has been on, but i find that 50-100 mcg q6 hours rarely causes any issues. I feel quite comfortable increasing it up to 150 mcg q4hr (900mcg/day).
Of course, it is safer to start switching over while the patient is still monitored in the ICU. In fact, that is what I always try to do.
Hey Josh — great post! I implemented a sub-anesthetic (0.1 – 0.3 mg/kg/hr) ketamine infusion protocol for critical care areas a little over a year ago. My practice has been to initiate a fentanyl/propofol strategy for 24-48 hours after intubation — you are exactly correct that the fentanyl component is often titrated up very quickly and usually more for “agitation” reasons, not judicious application of behavor-based pain scales. Around day 2-3, my practice has been to facilitate a transition of stable patients to a KetaDex strategy (ketamine/precedex infusions) with PRN opioid boluses. This limits opioid and anesthetic exposure, and maintains… Read more »
KetaDex sounds extremely cool. Have you had any problems with tolerance/withdrawal from dexmedetomidine? I’ve encountered some patients who withdrew from dexmedetomidine after >4-5 days exposure. Combined with the cost of dexmedetomidine, this has pushed me away from using dexmedetomidine as a first-line agent for sedation (my reading of the literature is that RCTs generally show similar outcomes comparing propofol vs. dexmedetomidine). I do like dexmedetomidine a lot as a short-term agent to facilitate extubation because as you mentioned it can be continued throughout the extubation period. I would love to see some more evidence regarding the use of dexmedetomidine for… Read more »
Your point is well taken that RCTs (mainly ProDex) haven’t shown big differences between propofol and Precedex in outcomes. However, being self-aware (thanks in part to your post) of the flaw of averages, I interpret this as the following: some patients will do better with Precedex and some patients with do better with propofol, so personalizing sedation based on individual characteristics, and modifying based on response, is probably the best approach. I tend to start with fentanyl-propofol, but will modify quickly based on RASS/CPOT, patient risk-factors and hemodynamics, and of course, where they are in their ventilator course. My practice… Read more »
Great stuff, thanks for the article. Very nice strategy.
There is a wee bit of literature that ketamine might prevent tolerance/dependence to dexmedetomidine (similar to the activity of ketamine in preventing tolerance to opioid): http://anesthesiology.pubs.asahq.org/article.aspx?articleid=1944516
If this is true, it would be game-changing because it would imply that you could use a ket-dex strategy upfront without problems with regards to tolerance/withdrawal from dexmedetomidine.
Thanks for the reference! Cool stuff.
Hi! Great post. I wonder, is it really so dangerous to use NSAIDs? I mean, bearing in mind that for example in post surgical patients the inflamatory component is very important. On the other hand, what are your thoughts on the potential for nefrotoxicity with paracetamol (giving its similarities with NSAIDs). Finally, what are your thoughts on analgesia-based sedation with initial remifentanil drip? Thanks! Fede (Argentina)
The risk from NSAIDs isn’t entirely clear, probably varying a lot depending on the patient. The sicker the patient is, the more likely they are to develop renal failure. In less ill patients you could probably get away with NSAIDs, but at the same time they’re not massively effective either. So ultimately I don’t think the risk is worth the benefit.
Acetaminophen doesn’t appear to cause nephrotoxicity at therapeutic doses. In large overdose it can absolutely cause renal failure. I don’t know the exact physiology behind these differences.
In my ICU , people donot not agree to use injection ketamin in view of the fear of causing adrenal insufficiency in patient s who are very sick. is it really evidence-based. I tried to discuss with them the articles given by doctor cliff reid in his blog. Pls share your thoughts. And how abt your experiences on airway secretions with ketamine
Adrenal insufficiency may be a problem with etomidate, but not ketamine. I haven’t found secretions to be an issue when intubating with ketamine, perhaps this takes a while to occur and doesn’t become a problem during the time-frame involved in rapid sequence intubation.
Good review.
But I believe that the revolution in pain management in the critical patient is undeniably tied with loco-regional anesthetic techniques, wheter neuraxial or peripheric nerves. Local anesthetics have a supreme safety profile, minor systemic absorption and low complications when eco-guided. And studies in anesthesia show that virtually any nerve can be blocked. Playing with concentration will provide analgesia up to surgical anesthesia. Single-shot techniques for lower duration, insertion of epidural, subarachnoid or perineural cathethers for lasting infusions. The possibilities are endless. And it allows for sparing of opioids and their side-effects n great scale.
Agree that regional blocks are great. However this isn’t applicable to diffuse pain which is sometimes encountered in ICU patients (e.g. pain experienced with turning). Ultimately no single modality is probably sufficient to control pain entirely on its own.
Hey Josh,
great post as usual.
Quick question: as with subanaesthetic dosing are we talking ketamine 0.1-0.3 mg/kg/hr or esketamine? Any rationale for using on or the other?
Thanks much. I think the ketamine I’m using is racemic ketamine (not s-ketamine). I haven’t looked into the difference in detail, I’m pretty sure the only form of ketamine availible at Genius General is the racemic variety.
I’d love to hear your views on geriatric analgesia and how that demographic might alter this ladder. I’m imaging the classic multiple boney fractured blunt trauma patient perhaps just a ground level fall with and without complicating co-mobidities and poly-pharmacy. I’ve been looking for evidence that ketamine might be less risky in this kind of demographic as opposed to regional anaesthesia as the traditional opioid strategy. Delerium seems to the ubiquitous no matter what we do???
I think the general strategy would be relatively similar, since most patients in an ICU are elderly. Thus, the baseline population in a medical ICU could probably be described as “geriatric” or close to it. Probably would use a combination of regional analgesia plus the analgesic ladder described above. You’re right, delirium is a major problem even with perfect management.
Dr Farkas, Im a critical care fellow that will be doing my grand rounds on alternative analgesia in the ICU in the next 2 months. My interest was sparked due to this post. Since then, has anything significantly changed, like according lets say to the newest SCCM pain guidelines and what not? Or any other source you recommend for me to look into??
Hi Carlos, I don’t think anything major has changed. I don’t think the SCCM guidelines change anything (they take a rather narrow view of the evidence they’re willing to consider). The DEXACET trial came out, which supports early use of acetaminophen (again, no big change; https://emcrit.org/pulmcrit/dexacet/). I would take a look at the KetaDex post, that’s probably the biggest update since this post (https://emcrit.org/pulmcrit/ketadex/). There is more and more evidence about lidocaine infusions and I think that will be the next major move in this field. Also possibly adding a bit more gabapentin. Good luck with your GR!
Thank you so much Dr Farkas…
I have a couple of points of disagreement – maybe we can get to a common point 1) fentanyl does not have a short half life; its half life is actually longer than that of morphine. Its clinical effectiveness lasts only 45-60 minutes however, while its potential adverse effects can last much longer. That is why infusions of more than 24 hours lead to difficulty getting patients off ventilators – fentanyl accumulates due to its long half life. 2) tolerance develops to opioids from breaks in mu receptor blockade, primarily those in the DRG (both pre and post synaptically). A… Read more »
Hi Jim and Josh
Our unit has run into some supply problems with remifentanil (the opioid is choice for infusion). We’ve been discussing alternatives to fentanyl, including alfentanil.
How would you compare the pharmacokinetics of alfentanil when compared to fentanyl in terms of it’s accumulation and half life effects?
I personally don’t have much experience of using alfentanil, outside of the operating theatre.
Best wishes
Dean