Causal relationship between an event and a drug

The most important problem in evaluating Adverse Drug Reactions (ADRs) is to determine the causal relationship between the drug and undesirable clinical events. To analyse an event, one has to consider temporality, medical or pharmacological plausibility, confounders, dechallenge and rechallenge. Several evaluation methods and algorithms have been published and used. Here is an indicative list followed by some discussion

The Karch and Lasagna’s Algorithm: Prepared in 1977, presents the user with a number of questions with offer only two options as answers (yes/no, true/false) and are arranged in three tables. The combination of results not only establishes causality but identifies atypical prescription compliance, prescription error, drug-drug interaction, terminal illness and appropriate use of the drug. Its reliability and validity are not as yet well established.

The Kramer’s Scale: Published in 1970s, it consists of 56 questions. It employs exhaustive flowcharts, is very intricate and not very practical for routine use.

The Naranjo’s Algorithm: It was developed in 1981 is a simpler method for the non-specialists’ use. It has 10 questions with “Yes or No” answers and has positive as well as negative scores. It is simple, brief, most extensively used, but its dependability and validity has not been established in children.

The Begaud’s Algorithm: This is a French scale, with a three-stage flowchart. It involves clinical criteria and the chronology of events, is simple to apply and does not require specific knowledge of the pharmacology of the involved drug. It has no specific advantage over the other methods.

The WHO Causality assessment method: Proposed by the Upsala Monitoring centre tries to simplify the whole issue, but gets lost in verbiage. Terms like reasonable, less than reasonable etc are difficult to interpret.

All these methods can be used., but each has some deficiencies. Hence many institutions and companies therefore, end up using the Global Introspection (Clinical Judgement) method. It is sensitive and is the most common approach. However, since it is based on clinical judgement there could be inter and intra rater differences. Let us review global introspection more closely. One has to look at the following aspects.

The Drug

• Is the event mentioned in the reference safety information of the drug? If yes, it means the event is a known possibility.
• Is it predictable based on its pharmacology? Is it an extension of its pharmacological effect? Has there been an overdose or a medication error?

The Event

• Some events like aplastic anaemia, torsade’s de pointes, fixed-drug eruption, neuroleptic malignant syndrome, are often drug related. Over the past decades they have been known to occur with many drugs or classes
• Some events are specifically related to certain drugs or class of drugs like statins and rhabdomyolysis, SSRIs and the serotoninergic syndrome, anti-HIVs and lipodystrophy.
• Some laboratory features like liver tests conforming to Hy’s law in case of hepatitis clearly indicate the drug as a cause.
• Has the reaction occurred locally as on skin or eye in case of topical applications, or in the oesophagus during transit?

The Patient

• Did the event appear earlier? In which context? Was that due to exposure to the same drug or class? This could be considered as a prechallenge

The Confounders

• Are there any potential suspects amongst the concomitant medications being administered to the patient? Is there a potential for drug-drug interaction?
• Can the event be explained by the current disease itself or any comorbidities?
• Could the patient’s age, sex, weight, history of allergies, renal and hepatic condition predispose to the event?

The Timeline:

• In relation to drug administration, when did the event occur? After a single dose, sometime after a few doses or after dose escalation. The time gap should be plausible based on the pharmacokinetics and pharmacodynamics of the drug.
• How long did the event last before any action like stopping the drug or reducing the dose was taken?
• What happened after stopping the drug or reducing the dose? An improvement would be considered positive dechallenge and would favour causality
• What happened if and when the drug was administered again? A repeat would indicate positive rechallenge and also favour causality. This factor often cannot be considered because there may not be a reason to rechallenge

Can causality assessed by investigator and company medical officer be different?

Investigator is aware of the patient’s clinical condition, the conmeds, comorbidities and is involved in the day-to-day management. He is provided an investigator’s brochure (IB) by the company that lists out all the known adverse events of the product under investigation till date. He considers the case temporality, comorbidities, conmeds and all clinical factors/ confounders before making a decision.

But the company has a much larger database of the product and has systematic documentation of all known adverse events. These have been analysed and some have been marked off as expected events, some as events under study and some rare ones have already been flagged for future study. Based on the additional information the company medical officer might opine differently.

This information about new events occurring after publication of the IB, as and when they arose, had been sent to the investigators too, but they are not expected to remember all these cases until they are included in the next version of the Investigator’s brochure.

So, although full weightage is given to the investigator’s opinion, in some cases the company relatedness decision could be different from that of the investigator. A detailed rationale has to be given in such cases. While reporting, both decisions are conveyed to the regulators.

References

1. Marante KB, Challenges of Adverse Drug Reaction Evaluation, J Pharmacovigil 2018, 6:3
2. The use of the WHO-UMC system for standardised case causality assessment. The Upsala Monitoring Centre
3. https://www.c3isolutions.com/blog/guide-to-causality-assessment/ authored by Bart Cobert 2019, downloaded in June 2020

Dr. Vishwas Sovani (June 2020)