How do we determine the Level of Evidence and Level of Severity for interactions?
The Level of Evidence of a specific interaction is determined based on the research that’s available. For example, an interaction between a supplement and a medication is theoretical if it’s only supported by in vitro studies, and it’s probable if it has been observed in multiple clinical trials. The Level of Severity is assigned based on various specific criteria, including but not limited to the type of medication that’s involved, the clinical consequences of the interaction (e.g., if it’s life threatening or if it doesn’t require any intervention), and the degree of the interaction (e.g., strong, weak).
The following tables detail the criteria that our team of pharmacists uses to determine the Level of Evidence and the Level of Severity for each interaction.
Level of Evidence in detail
| Level of Evidence (LoE) | If the interaction is supported by one or more of the following criteria: |
|---|
| Probable | - Multiple clinical trials
- Multiple case reports with causality testing (e.g., de-challenge, rechallenge tests). To establish the causality and to determine the sufficiency of case report evidence, we refer to the drug interaction probability scale (DIPS). A DIP score between 5 and 10 indicates a probable causal relationship.
- FDA/government agency warnings
- Pharmacodynamic interactions where the supplement was given to people also taking medications to manage a condition, such as diabetes or hypertension.
Example: If an additional reduction in blood glucose or blood pressure was observed in multiple clinical trials or multiple case reports with a DIP score between 5 and 10 when the supplement was given to people who were also taking medications to lower blood glucose or blood pressure (even if this did not lead to side effects, or if this was seen to be a beneficial outcome), this can be considered as evidence of a probable interaction.
|
| Possible | - Case report(s) with a DIP score of 2 or higher, because a DIP score less than 2 indicates a low probability of an interaction.
- A single clinical trial where an effect that was not present at baseline appeared after administering a supplement along with a medication or another supplement.
- Pharmacodynamic interactions where the effect of the intervention alone is large or there is a clear understanding of the mechanism of the effect/interaction.
- Pharmacodynamic interactions where the supplement was given to people who were also taking medications to manage a condition such as diabetes or hypertension.
Example: If an additional reduction in blood glucose or blood pressure was observed in a single clinical trial or case report(s) with a DIP score of 2 or higher, when the supplement was given to people who were also taking medications to lower blood glucose or blood pressure (even if this did not lead to side effects, or if this was seen to be a beneficial outcome), this can be considered as evidence of a possible interaction.
|
| Theoretical | - Predicted pharmacodynamic interactions where there isn't evidence from clinical trials.
Example: A supplement has blood-pressure-lowering effects on its own and theoretically might interact with blood-pressure-lowering agents, but studies administering this supplement along with blood-pressure-lowering agents have not been conducted, or in studies of this supplement, the participants were not taking any medications to lower blood pressure. - Animal studies
- In vitro studies:
- In vitro studies using human liver microsomes and cultured hepatocytes that assess possible interactions of a supplement with cytochrome P450 (CYP) enzymes.
- In vitro studies that assess interactions of a supplement with membrane transporters: P-glycoprotein (P-gp), breast cancer resistance protein (BCRP), organic cation transporters (OCTs), organic anion transporters (OATs), and organic anion transporting polypeptides (OATPs).
|
Level of Severity in detail
| Level of Severity (LoS) | If the interaction is supported by one or more of the following criteria: |
|---|
| Severe | - The interaction may be life-threatening or cause permanent damage. Example: clinical evidence of serious adverse events such as serious bleeding, liver damage, or kidney damage.
- The interaction requires discontinuation/change of therapy. Example: clinical evidence of strong CYP inhibition/induction (and therefore changes in the efficacy of drugs that can lead to treatment failure or medication toxicity).
- Pharmacodynamic interactions where the supplement was given to people who were also taking medications to manage a condition, and the interaction was clinically significant and caused harm or could have caused harm. Example: fenugreek was given to people with diabetes who were taking blood-glucose-lowering medications, and an additional reduction in blood glucose was found. This interaction led to hypoglycemia; therefore, the LoS is "severe".
- No clinical outcomes have been measured, the area under the curve (AUC) and the confidence interval (CI) of the interacting drug fall outside of the no-effect boundary, and the consequences of increased or decreased drug effects are deemed "severe" (e.g., serious therapeutic failures and/or toxicity). When no clinical outcomes have been measured (e.g., no measured effect on blood pressure, heart rate, risk of bleeding), the LoS of an interaction may be estimated based on changes in the geometric mean ratio of the AUC before and after coadministration of the supplement with a medication or another supplement.
The no-effect boundary represents the interval within which a change in AUC for a specific drug is unlikely to be clinically relevant (i.e., the change in exposure is unlikely to affect clinical outcomes, require additional monitoring, or lead to a dosing adjustment). A default range of 80% to 125% can be used as the no-effect boundary for most medications, or 90% to 111.11% for drugs with narrow therapeutic index (drugs, like warfarin, with a narrow window between the minimum effective concentration and the minimum toxic concentration). This means that when the AUC and its 90% CI after coadministration fall outside the 80%–125% boundary, caution is warranted, because an AUC lower than 80% suggests that there may be a clinically significant reduction in drug exposure, that could lead to a loss of efficacy. An AUC higher than 125% suggests that there may be a clinically significant increase in drug exposure that could lead to a higher risk of toxicity. - The interaction is based on data from animal studies that used doses or routes of administration that roughly reflect human use (e.g., a dose that’s equivalent to a human dose by body weight; the same route of administration) and the outcome of the interaction is deemed “severe” based on clinical judgment.
|
| Moderate | - The interaction may worsen a condition, is bothersome, or limits activities of daily living but does not require additional care or extended hospitalization. Example: theoretically, zinc supplements can bind with bisphosphonates, which reduces their absorption and thus their effectiveness, which may then result in inadequately treated osteoporosis.
- The interaction requires some change in therapy or monitoring.
- Pharmacodynamic interactions where the supplement was given to people who were also taking medications to manage a condition, and the interaction was clinically significant but did not cause harm. Example: fenugreek was given to people with diabetes who were taking blood-glucose-lowering medications, and an additional reduction in blood glucose was found. This interaction led to changes in their medication dose but did not cause hypoglycemia.
- No clinical outcomes have been measured, the area under the curve (AUC) and the confidence interval (CI) of the interacting drug fall outside of the no-effect boundary, and the consequences of increased or decreased drug effects are deemed "moderate". Please see above, under LoS "severe, for a more detailed explanation of the no-effect boundary.
- The interaction is based on data from animal studies that used doses or routes of administration that roughly reflects human use (e.g., a dose that’s equivalent to a human dose by body weight; the same route of administration) and the outcome of the interaction is deemed “moderate” based on clinical judgment.
|
| Minor | - The interaction causes few or no clear adverse effects and may cause no symptoms (e.g., changes in serum markers). Example: coadministration of hesperidin and celiprolol led to an increase in the AUC of celiprolol, but no significant changes in blood pressure or heart rate were observed.
- The interaction does not require a change in therapy or additional monitoring (e.g., a change in a clinical outcome that is not clinically relevant).
- Pharmacodynamic interactions where the supplement was given to people who were also taking medications to manage a condition, but the interaction was not clinically significant. Example: fenugreek was given to people with diabetes who were taking blood-glucose-lowering medications, and an additional reduction in blood glucose was found. If the reduction was not clinically significant (e.g., did not lead to changes in their medication dose), the LoS is minor.
- No clinical outcomes have been measured, the area under the curve (AUC) and the confidence interval (CI) of the interacting drug fall within the no-effect boundary (80%–125% or 90%–111.11%). Please see above, under LoS “severe”, for a more detailed explanation of the no-effect boundary.
- The interaction is based on data from animal studies that used doses or routes of administration that roughly reflects human use (e.g., a dose that’s equivalent to a human dose by body weight; the same route of administration) and the outcome of the interaction is deemed “minor” based on clinical judgment.
|
| Unknown | - Severity cannot be predicted at this time.
- Predicted pharmacodynamic interactions. Example: A supplement has blood-pressure-lowering effects on its own and theoretically might interact with blood-pressure-lowering agents, but studies administering supplement along with blood-pressure-lowering agents have not been conducted, or in studies of this supplement, the participants were not taking any medications to lower blood pressure.
- The interaction is based on data from in vitro research or from animal studies that used doses or routes of administration that do not apply to human use (e.g., the dose by body weight is 50 times greater than a human-equivalent dose; an orally administered supplement was injected).
If an animal study did use a human-equivalent dose and route of administration, the LoS can be upgraded accordingly.
|