Dose–response relationship Study Guide

📖 Core Concepts Dose‑response relationship – how the magnitude of a biological response changes with the amount of a stimulus (usually a chemical). Dose‑response curve – graph of dose (often log‑scaled) on the x‑axis vs. response on the y‑axis; most are sigmoidal (S‑shaped). Quantal vs. graded – quantal records all‑or‑none outcomes (e.g., death, seizure); graded records continuous outputs (e.g., enzyme activity). EC₅₀ / ED₅₀ / IC₅₀ – dose producing 50 % of the maximal effect (potency metric). Hill coefficient (n) – describes curve steepness; larger n → sharper transition. Emax – maximal achievable response; “ceiling effect” occurs when response plateaus despite higher doses. Threshold dose – lowest dose that yields a measurable response above control. Non‑monotonic curves – U‑shaped or inverted‑U patterns (common with endocrine disruptors, hormesis). --- 📌 Must Remember Sigmoidal shape → log‑dose on x‑axis, steep middle, plateaus at low & high doses. EC₅₀ = inflection point of the sigmoid; lower EC₅₀ = higher potency. Hill equation: \[ E = \frac{E{\text{max}}[D]^n}{EC{50}^n + [D]^n} \] Emax model (baseline effect allowed): \[ E = E{\text{max}}\frac{[D]}{EC{50} + [D]} \] Quantal data → analyzed with probit or logit regression; graded data → non‑linear regression preferred. Hormesis = low‑dose stimulation + high‑dose inhibition. Linear no‑threshold (LNT) model is not universally valid; low‑dose non‑monotonicity can occur. --- 🔄 Key Processes Generate dose‑response data Choose assay (organ‑bath, ligand‑binding, clinical trial). Apply a series of doses spanning several orders of magnitude. Plot the curve Plot dose (log scale) vs. response. Identify baseline, threshold, EC₅₀, and plateau. Fit a model Select appropriate model (Hill, Emax, multiphasic). Perform non‑linear regression to estimate \(E{\text{max}}, EC{50}, n\). Interpret parameters Potency → EC₅₀ (or ED₅₀/IC₅₀). Efficacy → \(E{\text{max}}\). Steepness → Hill coefficient n. Apply to risk/therapeutic decisions Compare EC₅₀ to exposure levels → safety margin. Use quantal analysis for lethal dose (LD₅₀) or effective dose (ED₅₀). --- 🔍 Key Comparisons Quantal vs. Graded Quantal: discrete outcomes, % of subjects responding → probit/logit analysis. Graded: continuous response magnitude → non‑linear regression (Hill/Emax). Hill Equation vs. Emax Model Hill: includes Hill coefficient n; assumes zero baseline effect. Emax: baseline effect may be non‑zero; simpler (no n term). Monotonic vs. Non‑Monotonic Curves Monotonic: response always increases (or decreases) with dose. Non‑Monotonic: U‑shaped or inverted‑U; low doses can stimulate, higher doses inhibit. Linear No‑Threshold vs. Threshold Models LNT: assumes any dose carries some risk, linear extrapolation from high doses. Threshold: posits a dose below which no observable effect occurs. --- ⚠️ Common Misunderstandings “EC₅₀ = safe dose.” – EC₅₀ is a potency measure, not a safety limit; adverse effects may appear at lower or higher doses. “All sigmoidal curves are monotonic.” – Some sigmoids are non‑monotonic (U‑shaped) in endocrine disruption contexts. “Higher Hill coefficient always means stronger drug.” – It only indicates a steeper transition; potency is still governed by EC₅₀. “Linear regression works for dose‑response data.” – For most biochemical data, non‑linear regression fits the underlying biology better than forced linearization. --- 🧠 Mental Models / Intuition “Slider” analogy: Imagine a volume slider (dose) controlling a speaker (response). Low doses → barely audible (baseline). Mid‑range → rapid climb (steep part). High doses → max volume (plateau). The slider’s position where sound is half‑max is the EC₅₀. “Hill hill” – Think of n as the number of “steps” the system must take to flip on; more steps → steeper hill. --- 🚩 Exceptions & Edge Cases Low‑dose endocrine disruptors → may produce a U‑shaped response (hormesis) contrary to classic monotonic expectations. Route of exposure (inhalation vs. oral) can shift EC₅₀ and curve shape. Time‑dependent effects – prolonged exposure can lower the apparent threshold dose. Multiphasic responses – when more than one mechanism dominates, curves may have multiple sigmoidal segments. --- 📍 When to Use Which | Situation | Preferred Model / Analysis | |-----------|-----------------------------| | Discrete outcome (death, seizure) | Quantal analysis → probit or logit regression | | Continuous response with clear baseline | Emax model (baseline ≠ 0) | | Steep, cooperative binding | Hill equation with n > 1 | | Evidence of multiple phases | Multiphasic or composite models | | Need to estimate antagonist affinity | Schild analysis (for competitive antagonism) | | Assessing low‑dose hormesis | Fit biphasic (U‑shaped) model; avoid linear extrapolation | --- 👀 Patterns to Recognize Sigmoid with inflection at 50 % response → likely Hill‑type behavior. Flat response at low doses, sudden rise, then plateau → classic dose‑response; locate EC₅₀ at midpoint. U‑shaped curve → suspect hormesis or endocrine disruption; check for low‑dose stimulation. Straight‑line on a log‑dose vs. probit plot → data are well‑described by a probit model (quantal). --- 🗂️ Exam Traps Choosing EC₅₀ vs. ED₅₀ – EC₅₀ refers to concentration; ED₅₀ refers to dose administered to a population. Assuming Hill coefficient = 1 always – Many receptors exhibit cooperativity (n ≠ 1). Misreading “threshold” – A graphical threshold (visible rise) is not the same as a formal safety threshold. Confusing “ceiling effect” with “maximal efficacy” – Ceiling is the observed plateau; maximal efficacy may be lower due to physiological limits. Applying LNT model to hormetic data – Leads to over‑estimation of risk at low exposures. ---