Alcohol and Flagyl = disulfiram rxn? Where're the data, dood?!

I think probably the first "real" counseling point any pharmacy student learns is "Don't drink alcohol with Flagyl!" If it's not the first thing, it's easily the second or third. In fact, I've seen this hand-written on prescription labels for added emphasis, even though the auxiliary labels that print out already say it. You don't often see "Take with food" hand-written, even though it would probably provide more real-world benefit to the patient than the standard "Don't drink alcohol" mantra.

"Heresy!" you shout. Well, hear me out…

You see, there's almost no data to support the assertion that alcohol and metronidazole combine to create a disulfiram-like reaction. It's crazy, I know. How could this age-old advice be wrong? The reason this is drilled into pharmacy and med students' heads is because the conventional wisdom is old. It got here because "everyone knows" that ethanol + metronidazole = A Bad Time. Even though there's no meaningful evidence to support this conclusion.

Regular readers know my distaste (hah!) for metronidazole. In fact, I missed out on my best friend's 21st birthday drunkfestcelebration because of it. As it turns out, I missed out for naught. Alas.

Exhibit A is a meta-analysis of published anecdotes, "Do Ethanol and Metronidazole Interact to Produce a Disulfiram-Like Reaction" published in The Annals of Pharmacotherapy. Exhibit B is a double-blind, placebo-controlled study out of Finland, also published in TAOP entitled "Lack of Disulfiram-Like Reaction with Metronidazole and Ethanol" which is a bit more science-y and a little less meta-analysis-y.

This is a long entry, so here's a ToC.

1. Bits and bobs from Exhibit A
2. Bits and bobs from Exhibit B
3. Final thoughts

Bits and bobs from Exhibit A (Back to top)

The first account of a metronidazole-ethanol interaction was noted in 1964. At that time, folks wondered if Flagyl could be used to curb alcohol abuse. A sort of proto-Antabuse, if you will. 8 of 17 studies found it to be marginally effective. But only marginally, and only 2 of the 8 positive studies were double-blind, and these 2 studies were statistically significant only when dropouts had been excluded.

My commentary after each quote.

revealed six case reports involving a total of eight patients. Tunguy-Desmerais reported on a two-year-old child taking acetaminophen and amoxicillin for pharyngitis. After a febrile seizure, ulcerative gingivostomatitis was diagnosed, phenobarbital–vitamin B6 syrup was added, and oral metronidazole was started. The next evening, the child was flushed but not febrile and, because both the analgesic and phenobarbital–vitamin B6 syrup contained ethanol, a metronidazole–ethanol reaction was considered likely.

I'm not an expert in pediatric liver function, but I do know that it takes longer for children to metabolize EtOH than it does for an adult. And the flushing is certainly a side effect of plain old alcohol consumption. Going right for the metronidazole-ethanol "reaction" seems a little too convenient, and more than a little irresponsible.

 

Another report involved three patients: a woman prescribed rectal metronidazole following hysterectomy, who became nauseous, pale, and dyspneic 36 and 60 hours postoperatively and was discovered to have taken a large amount of whiskey just prior to each episode; an 18-year old patient being treated with amoxicillin and metronidazole for pelvic inflammatory disease who experienced nausea, flushing, and headaches each evening after drinking ethanol; and a man who, after a 1g dose of metronidazole and a shared bottle of wine at his evening meal, vomited violently.

1. Taking whiskey in a post-operative state is probably not advisable regardless of the circumstance. Hysterectomy, while relatively common, isn't a walk in the park.
2. Metronidazole can nausea and even vomiting all by itself. It doesn't need any help from alcohol.
3. See #2. 1g is a lot of metronidazole all at once, especially if you're unused its GI effects. Would the man have vomited even if he'd not had a bottle of wine? We can't be sure, but you can bet that it wouldn't have been mentioned in the literature if it hadn't. It would have been chalked up to a bad reaction to the medication.

 

Plosker reported a reaction following intravaginal use of metronidazole. This case involved a female pharmacist who, after two or three cocktails (each contained ~1 oz of (vodka), inserted a single vaginal suppository of metronidazole 500 mg and went to sleep. She awakened an hour later with a burning sensation in her stomach, nausea, and a severe headache accompanied by a cold sweat, which she believed was a metronidazole–ethanol reaction.

I could go for the easy ad hominem attack here, but it doesn't serve any real purpose — and in any case, I am on my way towards showing that EtOH and metronidazole is not necessarily the End of the World. That said, the burning sensation and nausea could have been from the medication itself. Flagyl can cause this type of reaction, even when it's not administered orally.

 

A potentially serious reaction involved a 16-year-old male who, nine days after resection of an hepatic echinococcal cyst, developed a staphylococcal infection. He was treated initially with intravenous vancomycin, followed by a combination of intravenous nafcillin, clindamycin, and gentamicin. This regimen was then changed to intravenous trimethoprim/sulfamethoxazole (TMP/SMX) and intravenous metronidazole, in addition to his chronic therapy with albendazole, docusate sodium, and ferrous sulfate. After 12 hours and for the following 60 hours until the TMP/SMX was switched to oral therapy, he vomited and experienced flushing, which was attributed to a metronidazole interaction with the alcohol in intravenous TMP/SMX.

Attributing these side effects to the EtOH in the TMP/SMX is easy. It's also irresponsible, because a mechanism of action had been proposed by this point in time, but not substantiated. (Exhibit B will cover this more.) This reaction is not completely out of the question — metronidazole can indeed increase the levels of intracolonic alcohol dehydrogenase — but it's still irresponsible to put down, for certain, that this was the cause of the boy's distress. I wonder if the heavy antibiotics he was on could have contributed to this. It seems possible that by mowing down his normal intestinal flora, he's in a position to experience these effects anyway.

 

Another potentially serious reaction reported by Harries et al. also involved intravenous metronidazole, this time combined with cefotaxime and papaveretum in a patient who had been drinking heavily and stabbed in the chest and abdomen. A chest drain was inserted and 500 mL of blood was drained; peritoneal lavage produced clear fluid only and the patient was admitted for observation. Four hours after an initial improvement, he became short of breath and nauseous; he vomited, had a headache, and was profoundly acidotic. This metabolic disturbance was attributed to a metronidazole–ethanol interaction.

Drunken guy stabbed in the chest and abdomen. Chest drain removes half a liter of fluid, and four hours later the patient complains of nausea, vomits, has a headache, and is acidotic. And then it's attributed to a metronidazole-ethanol interaction.

I could think of a few things that're more likely to have caused this. Can you? He's been stabbed and he's probably developing a hangover spring readily to mind.

 

This last one is a real doozy:

Toxicity due to an ethanol–metronidazole interaction appeared on the death certificate of a 31-year-old woman. Cina et al. described the case of an alcohol abuser who had been in frail health for four years following a serious car accident involving severe chest, abdominal, and closed head injuries. After being assaulted by a man, the patient had collapsed and died. Medications found at the scene included propoxyphene, acetaminophen, naproxen, metaxalone, carisoprodol, amitriptyline, hydroxyzine, vitamins, and cough syrup. No metronidazole or empty metronidazole container was found and she had not recently been prescribed metronidazole by her doctor. Because high concentrations of ethanol and acetaldehyde were found, assays were performed for disulfiram and metronidazole. There was no discussion of the methodology that reportedly found metronidazole in her serum, despite the fact that this was not one of the drugs found at the scene. In addition, there was no discussion about whether any of the other drugs in her possession, for which she was apparently not tested, may have cross-reacted in the metronidazole assay. It was concluded that she had probably ingested metronidazole without the knowledge of her physician and had discarded the bottle before she died.

I think the idiocy here speaks for itself. "It was concluded that she had probably ingested metronidazole without the knowledge of her physician and had discarded the bottle" my ass. Flagyl ain't a drug people are likely to abuse.

Indeed the common thread throughout these cases has been the assumption that metronidazole and ethanol are the culprits without any real testing to verify whether or not this is the case. This study was published in 2000, and while there was a proposed mechanism of action for this EtOH-metronidazole reaction, there was no real evidence yet to support the proposed mechanism. There are also possible — and I would be so bold as to suggest more probable — causes for each of these adverse events, and I have done my best to explain them.

Now on to Exhibit B which talks about the science of the proposed alcohol-metronidazole interaction.

Bits and Bobs from Exhibit B (Back to top)

Disulfiram works its magic by inhibiting the hepatic low aldehyde dehydrogenase (ALDH) which increases blood acetaldehyde concentrations after alcohol is consumed. This is exactly like "Asian flush" — a common, genetic condition wherein those affected are unable to effectively metabolize alcohol completely leading to flushing, nausea, and a quickened pulse.

It was theorized that metronidazole may have a similar effect on ALDH. However, studies have shown that this is not true in rats. So, what about people?

Well, it doesn't do it in humans, either. In fact, it had the opposite effect. Metronidazole caused a reduction in acetaldehyde production, opposite to the effect of disulfiram(!). Indeed, in Exhibit B, the graphs show (probably clinically insignificant) that the participants in the double-blind, placebo-controlled, alcohol-metronidazole study that the blood alcohol levels for the participants taking metronidazole were slightly lower than their placebo-controlled counterparts between the 40 and 80 minute marks:

None of the participants noted any dyspnea, flushing, vertigo, or headache during the test. Interestingly, the heart rates for the metronidazole group tended to be about 10bpm lower than the control group throughout the test. I don't know that this is clinically significant, but it is interesting:

Final thoughts (Back to top)

So we've got some age-old advice that doesn't stand up when tested properly, and we have a proposed mechanism of action that doesn't hold up to closer scrutiny, either. Where does that leave us?

It leaves us with a couple things… Some people experience GI distress while on Flagyl. Sometimes they vomit. Sometimes they drink alcohol and vomit. Sometimes they drink alcohol and don't vomit. That tells us that:

1. Flagyl is hard on the stomach. It's a difficult medication to tolerate for a good percentage of folks, regardless of its effect on blood acetaldehyde levels.
2. Flagyl is not a disulfiram-like drug, and should not be referred to as such.
3. Flagyl will not absolutely cause the vomiting associated with Antabuse when consumed in conjunction with alcohol. Pharmacists should stop counseling that it will.
4. Adverse events are too often attributed to metronidazole because it is convenient, and "everyone knows" that alcohol and Flagyl are a recipe for disaster.
5. Flagyl remains an unpleasant drug to take; its side effects are real and often severe, and should not be downplayed.
6. Patients should be educated about these side effects, and how to minimize them. Avoiding alcohol is one way to do this. Taking it with food is probably more effective. Doing both is obviously better than doing just one of the two.

It is my suspicion that early researchers thought there was a link between metronidazole and alcoholism due to the medication's side effects. Taste perversion. Smell perversion. These things will cause folks to avoid certain foods for the duration of the drug therapy. Alcohol is one of those things. It's coincidental, and its usefulness in the real world is questionable.

Copyright © 2013 OnThePharm