A combination is graded by its riskiest ingredient: each part of the score is the lowest of the ingredients' scores. This grade answers one narrow question: how strong is the public evidence and safety record behind this drug, and how clean is its money trail and court record? It is not a verdict on whether the drug is right for you.
Each ingredient has its own full page with the FDA label, side-effect reports, recalls, evidence, and payments. Open them for the detail behind this grade.
Score breakdown (54/100)
Safety ×35%
F
Evidence ×25%
F
Regulatory ×15%
B
Money & lawsuits ×15%
A
Options ×10%
A
Safety: 17
Graded by its riskiest ingredient, codeine (17)
Carries an FDA boxed warning, the strongest warning the FDA issues (-30)
FDA issued 2 Drug Safety Communication(s) naming this drug since 2010, most recently 2017 (-16)
Subject to a REMS safety program: FDA requires special controls to use it (-10)
Federally controlled substance: dependence and misuse risk (-10)
Death appears in 2.6x the share of reports seen across all of FAERS (-12)
Off-label use is among the five most-reported items: widely used beyond its approved evidence (-5)
Evidence: 48
Graded by its riskiest ingredient, promethazine (48)
91 randomized trials and 5 systematic reviews/meta-analyses indexed in PubMed
3 independent Cochrane review(s) exist
Only 41% of completed registered trials posted results (-8)
Regulatory history: 77
Graded by its riskiest ingredient, codeine (77)
No recalls on record since 2016
REMS program in force (-10)
6 label changes approved since 2021: the safety story is still moving (-8)
Label lists recent major changes to warnings or indications (-5)
Money & lawsuits: 100
Graded by its riskiest ingredient, codeine (100)
Pharma paid doctors $0 tied to this drug last year: negligible
Not named in any pending federal MDL
Options: 100
Graded by its riskiest ingredient, codeine (100)
Generic versions approved (1 applications)
6 same-class alternatives exist
Money and lawsuits
In 2024, drug companies reported $0 in payments to US physicians and hospitals tied to this combination's brands across 0 transactions.
No pending federal MDL names this combination or its brands.
The ingredients' own payments and lawsuits are on their pages and already count toward this grade.
Who takes it and what it costs
848Medicare Part D patients (2024)
1,137prescriptions filled
$8average cost per fill
$0Mtotal Medicare spend
Questions to ask your prescriber
This is a fixed combination of codeine and promethazine. Why both together, instead of only the one I need?
In 2017 the FDA issued a safety communication about this drug (FDA restricts use of prescription codeine pain and cough medicines and tramadol pain medicines in children; re). Does that risk apply to me?
This drug carries a boxed warning. Which specific risk does it cover, and does it apply to me?
This is a controlled substance. What is the plan for how long I take it and how I stop?
This drug has a REMS safety program. What monitoring does that require from us?
The most-reported problems in FDA's adverse event data are drug hypersensitivity, pain. How would we recognize and handle those?
Off-label use is among the most-reported items for this drug. Is my use on-label, and if not, what evidence supports it?
Why this one over dextromethorphan, ethylmorphine, hydrocodone? What would make you pick one of those instead?
What result are we expecting, by when, and what is the plan if it does not happen?
If 100 people like me took this drug, how many would actually get better because of it, and how many would be harmed by it? (Doctors call these the number needed to treat and the number needed to harm.)
An independent reviewer (Prescrire) lists this as a drug to avoid because the harms outweigh the benefits versus alternatives. What is your view on that?
The most-reported problems in FDA's adverse event data are intentional overdose, nausea, intentional self-injury, fatigue. How would we recognize and handle those?
Why this one over bamipine, chloropyramine, chlorphenoxamine? What would make you pick one of those instead?