Fatal Drug Interaction or Incidental Co-Occurrence?

Here's a question that was asked to me by a frequent correspondent in an e-mail:

One quick comment on your post and related conversations: You often emphasize the issue of drug interactions, especially w/ benzodiazepines.
Question: how do we know these are interactions as opposed to just co-occurrence?

And this was my answer, slightly edited:

 Excellent question! I think the answer is that obviously we don’t know. In fact I argue in several of my posts that the proliferation of substances on death records casts doubt on the idea that they are truly establishing the causes of death, versus just jotting down everything they find in a toxicology screening. Was cannabis use really a contributing cause of death? Probably not. If a record shows “cocaine” and “benzodiazepines”, I’m thinking that the medical examiner is just listing what s/he sees, because I don’t think there’s a known interaction between these substances. The things on the death certificate are supposed to be “contributing causes of death”, but I think we both know those records are fallible.
 On the other hand, I believe the interaction between opioids and benzodiazepines and alcohol (any combination thereof) is pretty well understood. Taken in combination, these substances suppress respiration more than if taken alone. Also, I think it’s pretty hard to “overdose” on any one of these substances. I’m sure some opioid users occasionally take way too much and accidentally kill themselves, but it’s probably more likely that they take a somewhat normal dose (given their tolerance) with some other substance and it has an unpredictable effect on their respiration. From the 2016 data, 90% of benzodiazepine poisonings involve some kind of opioid. Of course I have no way of knowing if benzodiazepines were just incidental or actually contributed to what proportion of these, but since the pharmacology of this interaction is fairly well understood I’d say “probably most of them.”
 For the deaths involving heroin and other synthetic opioids, I’m more sympathetic to the idea that the other substances listed on the death record are incidental. These strong opioids can be easy to overdose on, for sure. Still, it looks suspicious that such a large proportion of them involve multiple substances. 

I asked my wife, who is a pediatrician, and she said the medical examiner would probably just list everything found in a toxicology screening. But she also said there’s no standard (something I think you and I have both figured out), so you would get different records from different examiners looking at the same information.

I thought this clarification might be of interest to some of my readers. This correspondent is working on a paper on the opioid epidemic. He shares my skepticism about the official narrative. I'll definitely write a post when his paper comes out. When I got his question I was worried I was missing something, but he was basically pleased with my answer. (Apologies for the vagueness; I'm trying to avoid name-dropping for the sake of name-dropping here.)