It’s been another busy shift in the ED. You finish discharging one patient as you are pulling up the x-rays of a 24 year old who rolled his ankle while playing basketball. You don’t see a fracture and so you go into the room to tell him the good news that it is only a sprain. As you finish demonstrating some exercises that he can do to rehabilitate his ankle, he says, “Hey doc, what you gonna give me for this?” You tell him that you don’t think he needs narcotics, and given the epidemic of prescription opioid abuse and addiction, you think acetaminophen along with rest and ice should do the trick. You turn to walk out of the room when he responds, “Really, I heard on the radio that that stuff will kill you!” Somewhat shocked, you have a 10 minute conversation and finally agree to give him high dose ibuprofen and discharge him. Still shocked by that encounter, you fire up Dr. Google to try to learn more about this before your next shift.
The Dangers of Acetaminophen
A recent hour-long report on NPR’s This American Life discussed the dangers of acetaminophen. This was also the subject of a recent article on ProPublica. While the goal of making medications as safe as possible is commendable, the underlying message of how dangerous this medication can be may be overstated. They report that approximately 150 patients die each year after an accidental overdose of acetaminophen. While this is 150 patients too many, it also needs to be viewed in the context of how many people use acetaminophen each year, which is reportedly about 50 million people per year. This would mean less than 0.1% of patients exposed to acetaminophen each year die from an accidental overdose.
The report discussed a cohort of 47 patients that presented over 4 years to a single hospital following an overdose of acetaminophen. Of the 47, almost half were reported as unintentional overdoses. Of these 23 patients, three died and two required hepatic transplantation. The unintentional group tended to be sicker and do worse than the intentional group. This data was once again presented to illuminate the dangers of acetaminophen, and how easy it was to overdose as it had such a “narrow therapeutic margin.” However, other conclusions can also be drawn from this data. The unintentional group likely did worse, and was sicker, because of a delay in presentation compared to the intentional group. Also a higher percentage of patients were transferred from outside hospitals in the unintentional group, likely due to the fact that they had already developed acute liver injury and needed to go to a tertiary care center.
How patients were classified as unintentional overdoses may also be flawed. The support for them being unintentional was based on the fact that only 10/23 had a psychiatric evaluation in the hospital, while all the patients classified as an intentional overdose had a psychiatric evaluation. While the difference in psychiatric consultation rates between the two groups may be due to fewer suicide attempts in the unintentional group, there are also other explanations for this. As such, using psychiatric evaluation as a marker for intentional exposures may be flawed.
Much has also made regarding the narrow therapeutic margin of acetaminophen, especially compared to other over-the-counter NSAIDS. While the mechanism of action of acetaminophen is not clear, it appears to indirectly inhibit prostaglandin synthesis, mainly COX-2, in the CNS. Further investigation into their data shows that on average these patients were taking 8.1 grams of acetaminophen/day (range 1.1-18 grams/day). This is double the recommended dose, not simply taking one or two extra tablets a day by accident. While an ingestion of 10 grams a day (approximately 150 mg/kg for a 70 kg individual) is mentioned as the toxic threshold, this is only for a single ingestion. It is well known that the toxic threshold decreases if you continue taking supratherapeutic doses multiple days in a row. More importantly, we know that acetaminophen taken at the maximum recommended dose of 4 grams a day is very safe. Patients were safely given 4 grams of acetaminophen a day for 4-6 weeks in multiple studies (Schnitzer 2005)(Bradley 1991). We even know the recommended dosing is safe in high-risk patients, such as with chronic alcoholics who recently stopped drinking. In addition to their baseline hepatic dysfunction this group was theoretically at an increased risk of an adverse event as they had upregulation of the enzymes responsible for producing the toxic metabolites of acetaminophen. But this group did well while taking 4 grams of acetaminophen a day for 3 consecutive days (Kuffner 2007).
One problem is that the term “narrow therapeutic margin” as used in these news reports is incredibly subjective. Acetaminophen does have a much narrower margin compared to the amount of ibuprofen required in a single overdose to kill someone. However, an argument could also be made that acetaminophen has a much wider margin compared to many other medications such as verapamil, lithium, or colchicine, just to name a few.
In addition to the concern raised by PBS, many patients with chronic liver disease are also warned by their doctors to avoid acetaminophen. While I agree that long term use is probably not good for them, short term use for pain control in the ED and the next couple of days with re-evaluation by their primary care doctor or hepatologist seems reasonable for the majority of these patients. In this population other NSAIDS also have their drawbacks due to inhibiting platelet aggregation in a population with an elevated baseline risk of bleeding.
The safety of NSAIDS such as ibuprofen has also been questioned by the media. On one hand, it is hard to kill yourself with an acute ingestion of ibuprofen – you’re generally talking about swallowing hundreds of pills to develop life-threatening toxicity such as seizures. On the other hand, chronic use of NSAIDS does result in significant morbidity such as ulcers, GI bleeding, and renal failure. A recent article in The Lancet investigated this very issue (Coxib 2013). The meta-analyses included 280 trials of NSAIDS versus placebo and 474 trials of one NSAID compared to another NSAID and characterized the vascular (MI, stroke, vascular death) and gastrointestinal (bleed, perforation, obstruction) side effects. Few patients had prior history of atherosclerosis (9%) or peptic ulcer disease (7%), but patients were older with a mean age of 61 years. NSAIDS included in the analysis were any coxib (selective COX-2 inhibitor), diclofenac 150 mg/day, ibuprofen 2,400 mg/day, and naproxen 1000 mg/day. All trials were required to be at least 1 month in duration to be included but could be longer. Selective COX-2 inhibitors were thought to be beneficial for the management of chronic pain as they would not cause the GI side effects associated with traditional NSAIDS such as ibuprofen. Unfortunately, they were associated with a higher risk of atherothrombotic vascular and cardiac events leading to the withdrawal of Vioxx® (rofecoxib) from the market. The assumption by many, including some involved in the discussion regarding the inherent dangers of acetaminophen, is that chronic use of NSAIDS is safe.
Not surprisingly, in the meta-analysis there were approximately a third more major vascular events in the group allocated to a coxib versus a placebo, mainly from an increase in the rate of major coronary events. Diclofenac was also associated with an increase in major vascular events similar to those patients on the average coxib regimen. But the patients taking the other NSAIDS did just fine, right? Well actually ibuprofen was associated with an increased number of major coronary events (adjusted rate ratio 2.22, 95% CI 1.10-4.48, p=0.0253) but a non-significant increase in the number of major vascular events (1.44, 95% CI 0.89-2.33, P=0.14). High dose naproxen on the other hand fared much better and was not associated with a significant risk of major vascular events (0.93, 95% CI 0.69-1.27) or coronary events (0.84, 95% CI 0.52-1.35). Both the coxibs and diclofenac were associated with a significantly increased risk of vascular death. This was not seen in the group receiving naproxen, while the group receiving ibuprofen had a non-significant increase in the rate of vascular death. In more concise terms, this would equate to 3 additional major vascular events and 1 death per 1000 patients per year in the groups taking either a coxib or diclofenac compared to a placebo. For higher risk individuals receiving a coxib or diclofenac, the risk of a vascular event or death was further increased, and patients taking high dose ibuprofen may have also had a similar risk, too. The risk of hospitalization due to heart failure was roughly doubled by all NSAID regimens. All regimens were also associated with an increased risk of gastrointestinal complications, mostly bleeding, of which 2% were fatal. Coxibs had the lowest risk of GI complications compared to the other NSAIDs. This data should be interpreted with caution as only patients chronically using NSAIDS were included, unlike the many patients seen in the ED only requiring acute pain control. For these patients, short term use of almost any NSAID is likely safe. If they are older or have risk factors for or have cardiovascular disease, naproxen may be a better choice. It also has the advantage of twice a day dosing. If the patient is presenting for management of chronic pain, naproxen is a better choice compared to ibuprofen, diclofenac, or one of the coxibs. However, if the patient is also at risk for a GI bleed, it is probably best to just avoid NSAIDS altogether. Development of renal insufficiency was not included in this meta-analysis. The patient’s renal function should be considered before prescribing any NSAID, especially if they are going to receive multiple doses.
Does this mean that every patient with a stubbed toe needs their creatinine checked before you order ibuprofen? No. But if they are older, and could have renal insufficiency, it’s prudent to either check their creatinine or review their records for recent laboratory results. Or better yet, just use acetaminophen.
In their concluding remarks, one of the investigators interviewed on This American Life quotes a patient as saying, “If I had known it would make me this sick, I never would have taken it.” Is this the message that they should be sending? Any medication is associated with some risk, and anything done or used inappropriately can be problematic. Swimming is an inherently safe activity enjoyed by millions of people a year. However it is dangerous to swim at night, by yourself, when you’re intoxicated. This doesn’t mean people should stop swimming, it just means that they need to be educated on how to swim safely. The same applies to over the counter analgesics. When used responsibly and in the appropriate patients, they are safe and helpful medications. Is more education needed to prevent the morbidity associated with them? Absolutely. Is more education needed to make sure they are being used in the correct patient population? Probably. But the message of all this should not be that either ibuprofen or acetaminophen is dangerous, but that when used appropriately, they are very safe and useful medications. In fact, just thinking about all this has made my head hurt. Now what should I take…
1. Schnitzer TJ, Weaver AL, Polis AB, et al. Efficacy of rofecoxib, celecoxib, and acetaminophen in patients with osteoarthritis of the knee. A combined analysis of the VACT studies. J Rheumatol 2005;32:1093-105.
2. Bradley JD, Brandt KD, Katz BP, et al. Comparison of an anti-inflammatory dose of ibuprofen, an analgesic dose of ibuprofen, and acetaminophen in the treatment of patients with osteoarthritis of the knee. N Engl J Med 1991;325:87-91.
3. Kuffner EK, Green JL, Bogdan GM, et al. The effect of acetaminophen (four grams a day for three consecutive days) on hepatic tests in alcoholic patients-a multicenter randomized study. BMC Medicine 2007;5:13
4. Coxib and traditional NSAID Trialists’ (CNT) Collaboration. Vascular and upper gastrointestinal effects of non-steroidal anti-inflammatory drugs: meta-analyses of individual participant data from randomized trials. The Lancet 2013;382(9894):769-779.