COVID19 update, May 14, 2020: drug cocktails greater than the sum of their parts

(1) The term “drug cocktail” is best known from AIDS, where the introduction of “cocktails” of (usually three from at least two different classes) antiretrovirals helped turn HIV from a death sentence into a long-term manageable disease.
Now (h/t: Mrs. Arbel) a team from Hong Kong has achieved excellent results for COVID19 using a different cocktail, reports the Jerusalem Post. The full medical article in The Lancet can be read here:

The cocktail in question has three components:

  • The HIV drug Kaletra, itself a mixture of two protease inhibitors, Lopinavir and Ritonavir.
  • The hepatitis drug Ribavirin, a nucleoside analog that can mimic both the letters A and G of the genetic code, and thus messes with copying of the viral RNA (cf. my earlier posts on Remdesivir)
  • The immunomodulator Interferon beta-1b, better known to multiple sclerosis patients as REBIF.

The control group was given just Kaletra. Otherwise, both groups received standard supportive care, including antibiotics for secondary bacterial infections.

What’s with the protease inhibitor? Many of these viruses (including SARS-NCoV-2 have their envelope etc. Proteins encoded as a single long “protein sausage” on their RNA. After protein synthesis in the ribosome (the cell organelle that assembles proteins from amino acids according to the ‘program tape’ on the RNA), a protease then splits the ‘sausage’ into individual ‘links’.

So we have two drugs that tamper with the ability of the virus to make its envelope, plus one that inserts junk ‘letters’ in the copied RNA. Even if each partially successful, they will slow down viral reproduction. So what is the role of the beta-interferon? To tell the body’s immune system: “don’t go berserk, take it easy!” and prevent cytokine storm.

If treatment was started less than 7 days after onset of symptoms, the “triple cocktail” group showed better clinical and virological outcomes than the control group across all meaningful measured variables. For the subgroups of patients where treatment was delayed longer, there was no statistically significant difference in outcomes between the cocktail and control group. So early intervention is worth a lot.

Median time to negative RT-PCR test was 7 days for the “cocktail” group, compared to 12 days for the control group.

It’s not a magic bullet drug: every doctor dreams of what Frederick Banting experiences when he first administered insulin to boys in diabetic coma, where the first boys were waking up before he’d finished injecting the last. But that kind of spectacular success is the rare exception in drug research.

What we can safely say we have here, I believe, is a ‘cocktail’ that is greater than the sum of the parts. And a nice thing about cocktails of existing drugs: each component already has undergone clinical trials and obtained FDA (or foreign equivalent) approval individually.

What about side-effects? Reading Table 4 in the paper, the difference with the control group for nausea, diarrhea,… is not statistically significant. No patients in the ‘cocktail’ group suffered severe adverse events, vs. one in the ‘Kaletra only’ control group.

(2) My friend “masgramondou” weighs in on the source code of the “Ferguson Model”: All models are wrong, and some are useless. Or worse than useless, in this case. He points to another model that might at least be somewhat more transparent than the others:, developed by the group of Prof. Richard Neher at U. of Basel, Switzerland.