Diazepam at high doses has long been thought to be an antidote to (hydroxy)chloroquine poisoning and is routinely administered in emergency care for this purpose. The PPeH expressly warns about this:
Looking at the evidence base, though, reveals that it is scant. It consists of case reports of people surviving a chloroquine overdose after also ingesting a large amount of diazepam, and a few studies that compared people who had taken chloroquine + diazepam as opposed to chloroquine only (case-control studies).
It seems that most of these studies were published in the 1980s - 1990s. Surprisingly, despite the fact that chloroquine was widely used during the first year of the Covid-19 pandemic, only one unrelated case report of chloroquine overdose has come out since (Störmann et al, 2021).
Below is a brief summary of the evidence base, taken from a commentary on a 2020 review of the management of chloroquine poisoning (Megarbane & Schicchi, 2021) and another one from a commentary on a 2004 case report (Yanturali, 2004; both commentaries were peer-reviewed):
————
A note about the control group in the study by Riou et al. mentioned above:
Normally, to investigate the efficacy of a drug, one needs to run a randomized controlled trial (RCT). The active treatment group receives the drug, while the control group receives either a placebo or another drug that is considered the standard of care at the time the trial is conducted (since it would be unethical not to give it to patients in need). The crucial point is that randomizing patients to receiving either the drug or the placebo / other drug ensures that both the active and control groups are well matched in terms of age, sex, weight, ethnicity, disease severity, comorbidities and other unmeasured factors (a.k.a. confounders, see below) that could influence the results of the trial. In other words, randomization allows for a reliable comparison between both groups.
In contrast, participants are not assigned to one treatment or another in a case-control study; the investigators only observe how much and how often they use a treatment (if at all). This lack of randomization makes case-control studies prone to systematic bias(es), as illustrated below.
Imagine that a case-control study investigates the effect of celery juice on heart size. The investigators record other measures like age, sex and weight, and eventually find that those who drink celery juice have a larger heart. Generally, celery juice is associated with wellness so fit people are more likely to drink it. As a result, even if the participants are well matched on age, sex and weight, increased fitness is much more likely to explain a larger heart than drinking celery juice (the heart is a muscle, so it grows by working out). Here, fitness level is a confounding variable because it is an unmeasured factor that influences the results of the study. It would then be wrong to conclude that the observed association between celery juice consumption and heart size is caused by the former.
In general, case-control studies cannot infer causation, and this is even more true when they contain an inadequate control group — unfortunately, this often happens. Of course, on ethical grounds, one cannot run a RCT of chloroquine poisoning so we are forced to rely on case-control studies. Failing that, we have to rely on experimental studies on animals.
————
In 2020, a well-conducted experimental study on rats that tested the efficacy of diazepam, clonazepam and 4'-chlorodiazepam on reducing the cardiac toxicity of chloroquine poisoning was published in the British Journal of Pharmacology (Hughes, 2020). The results were clear; none of the tested benzodiazepines succeeded at this task (Hughes, 2020):
Although Hughes found that diazepam in combination with adrenaline, but not alone, had a small positive effect on cardiac contractility (Fig. 2, Hughes, 2020), it was at the expense of lower potassium blood level and increased cardiac arrythmias, both of which are dangerous. There also are two limitations to this finding:
1) Its statistical significance was borderline (p < 0.05 but >= 0.01). This means that if the study were to be repeated on a larger sample of rats (here, there were only 8 per group: control / diazepam / diazepam + adrenaline), this finding may or may not replicate.
2) The author attributes the effect of diazepam on cardiac contractility to PDE4 inhibition but the expression of PDE4 is lower in humans than in rats, so whether this effect is relevant in humans is unclear (Hughes, 2020):
Based on this finding as well as the negative results of a RCT of diazepam on moderate chloroquine intoxication, Megarbane and Schicchi conclude that "the belief that diazepam may improve CQ/HoCQ-induced-vasodilation or dysrhythmic effects is illusive". Instead, they call for intubation, ventilation and epinephrine (Megarbane and Schicchi, 2021):
Finally, Yanturali refers to the same RCT mentioned above and points out that unlike the case he was commenting on, a patient of his who overdosed on 22 grams of hydroxychloroquine (compared to the PPeH's recommended 12g) survived without diazepam:
In contrast with the PPeH, the authors of the Guide to a Humane Self-Chosen Death (wozz Foundation, led by the anesthesiologist Pieter Admiraal, MD) were skeptical of the relevance of diazepam in chloroquine overdose. Instead, they recommended combining it with a short-acting benzodiazepine such as midazolam, but only because the symptoms of chloroquine toxicity should kick in before diazepam puts one to sleep (as it has an onset of about an hour):
To conclude, there is no robust evidence that diazepam has a protective effect against chloroquine overdose. Rather, the current evidence base strongly suggests the opposite. For CTBing with chloroquine, it or another long-acting benzodiazepine should absolutely be taken to avoid any suffering, in combination with a shorter-acting one that kicks in before the symptoms of cardiac toxicity from chloroquine do.
Looking at the evidence base, though, reveals that it is scant. It consists of case reports of people surviving a chloroquine overdose after also ingesting a large amount of diazepam, and a few studies that compared people who had taken chloroquine + diazepam as opposed to chloroquine only (case-control studies).
It seems that most of these studies were published in the 1980s - 1990s. Surprisingly, despite the fact that chloroquine was widely used during the first year of the Covid-19 pandemic, only one unrelated case report of chloroquine overdose has come out since (Störmann et al, 2021).
Below is a brief summary of the evidence base, taken from a commentary on a 2020 review of the management of chloroquine poisoning (Megarbane & Schicchi, 2021) and another one from a commentary on a 2004 case report (Yanturali, 2004; both commentaries were peer-reviewed):
————
A note about the control group in the study by Riou et al. mentioned above:
Normally, to investigate the efficacy of a drug, one needs to run a randomized controlled trial (RCT). The active treatment group receives the drug, while the control group receives either a placebo or another drug that is considered the standard of care at the time the trial is conducted (since it would be unethical not to give it to patients in need). The crucial point is that randomizing patients to receiving either the drug or the placebo / other drug ensures that both the active and control groups are well matched in terms of age, sex, weight, ethnicity, disease severity, comorbidities and other unmeasured factors (a.k.a. confounders, see below) that could influence the results of the trial. In other words, randomization allows for a reliable comparison between both groups.
In contrast, participants are not assigned to one treatment or another in a case-control study; the investigators only observe how much and how often they use a treatment (if at all). This lack of randomization makes case-control studies prone to systematic bias(es), as illustrated below.
Imagine that a case-control study investigates the effect of celery juice on heart size. The investigators record other measures like age, sex and weight, and eventually find that those who drink celery juice have a larger heart. Generally, celery juice is associated with wellness so fit people are more likely to drink it. As a result, even if the participants are well matched on age, sex and weight, increased fitness is much more likely to explain a larger heart than drinking celery juice (the heart is a muscle, so it grows by working out). Here, fitness level is a confounding variable because it is an unmeasured factor that influences the results of the study. It would then be wrong to conclude that the observed association between celery juice consumption and heart size is caused by the former.
In general, case-control studies cannot infer causation, and this is even more true when they contain an inadequate control group — unfortunately, this often happens. Of course, on ethical grounds, one cannot run a RCT of chloroquine poisoning so we are forced to rely on case-control studies. Failing that, we have to rely on experimental studies on animals.
————
In 2020, a well-conducted experimental study on rats that tested the efficacy of diazepam, clonazepam and 4'-chlorodiazepam on reducing the cardiac toxicity of chloroquine poisoning was published in the British Journal of Pharmacology (Hughes, 2020). The results were clear; none of the tested benzodiazepines succeeded at this task (Hughes, 2020):
Although Hughes found that diazepam in combination with adrenaline, but not alone, had a small positive effect on cardiac contractility (Fig. 2, Hughes, 2020), it was at the expense of lower potassium blood level and increased cardiac arrythmias, both of which are dangerous. There also are two limitations to this finding:
1) Its statistical significance was borderline (p < 0.05 but >= 0.01). This means that if the study were to be repeated on a larger sample of rats (here, there were only 8 per group: control / diazepam / diazepam + adrenaline), this finding may or may not replicate.
2) The author attributes the effect of diazepam on cardiac contractility to PDE4 inhibition but the expression of PDE4 is lower in humans than in rats, so whether this effect is relevant in humans is unclear (Hughes, 2020):
Based on this finding as well as the negative results of a RCT of diazepam on moderate chloroquine intoxication, Megarbane and Schicchi conclude that "the belief that diazepam may improve CQ/HoCQ-induced-vasodilation or dysrhythmic effects is illusive". Instead, they call for intubation, ventilation and epinephrine (Megarbane and Schicchi, 2021):
Finally, Yanturali refers to the same RCT mentioned above and points out that unlike the case he was commenting on, a patient of his who overdosed on 22 grams of hydroxychloroquine (compared to the PPeH's recommended 12g) survived without diazepam:
In contrast with the PPeH, the authors of the Guide to a Humane Self-Chosen Death (wozz Foundation, led by the anesthesiologist Pieter Admiraal, MD) were skeptical of the relevance of diazepam in chloroquine overdose. Instead, they recommended combining it with a short-acting benzodiazepine such as midazolam, but only because the symptoms of chloroquine toxicity should kick in before diazepam puts one to sleep (as it has an onset of about an hour):
To conclude, there is no robust evidence that diazepam has a protective effect against chloroquine overdose. Rather, the current evidence base strongly suggests the opposite. For CTBing with chloroquine, it or another long-acting benzodiazepine should absolutely be taken to avoid any suffering, in combination with a shorter-acting one that kicks in before the symptoms of cardiac toxicity from chloroquine do.
Last edited:
