Safety & Adverse Effect Terms

Anaphylaxis diagnostic criteria (Sampson 2006): acute onset (minutes to hours) with involvement of skin/mucosal tissue (urticaria, flushing, angioedema) PLUS either respiratory compromise (dyspnoea, wheeze, bronchospasm, stridor) OR cardiovascular compromise (hypotension, syncope, end-organ dysfunction). Treatment: epinephrine (adrenaline) IM injection into anterolateral thigh (adult 0.5mg, child 0.01mg/kg — first-line, should be given immediately), position patient supine with legs elevated, establish IV access, administer IV fluids for hypotension, nebulised salbutamol for bronchospasm, and antihistamines/corticosteroids as adjuncts. For peptide drugs administered in healthcare settings, anaphylaxis management equipment must be immediately available. Post-anaphylaxis: the triggering peptide drug is generally contraindicated for future use, though desensitisation protocols exist for essential medications without alternatives.

ADA characterisation includes: binding antibody titre (quantifying ADA level), isotype determination (IgG, IgM, IgE — IgE suggests allergic potential), epitope mapping (identifying which part of the drug the antibody recognises), neutralising capacity (does the ADA block the drug's receptor binding?), and cross-reactivity (does the ADA also bind the endogenous counterpart — critical safety concern). Clinical consequences of ADA: reduced drug exposure (enhanced clearance through immune complex formation), reduced efficacy (neutralising antibodies blocking activity), altered PK (forming drug-antibody complexes that change distribution and elimination), hypersensitivity reactions (IgE-mediated or immune complex-mediated), and rarely, cross-reactive neutralisation of endogenous protein (e.g. ADA against exogenous EPO cross-reacting with endogenous EPO, causing pure red cell aplasia). Most ADA formation is transient and clinically inconsequential.

The FDA can require a boxed warning under 21 CFR 201.57(c)(1) when: a serious adverse reaction can be prevented or its severity reduced through appropriate use, a serious adverse reaction exists that limits the drug's use, or the drug has a specific population risk. Examples beyond GLP-1 RA C-cell warning: corticotropin (immunosuppression risk), cyclosporine systemic (nephrotoxicity, neoplasia, serious infections), and various oncology peptides. Boxed warnings must include: a brief summary of the risk, affected population, recommended monitoring or contraindications, and any REMS requirements. The decision to add, modify, or remove a boxed warning involves FDA internal review, may include advisory committee input, and considers: strength of evidence, severity of risk, ability to mitigate through labelling, and benefit-risk balance.

Standard approach: 2-year rat study + 6-month transgenic mouse study (p53+/-, rasH2, or Tg.AC models). ICH S1A/B guidance: carcinogenicity testing required when drug will be used continuously or intermittently for ≥6 months, or when there are specific concerns (genotoxicity findings, pre-neoplastic lesions in repeat-dose studies, structural similarity to known carcinogens). GLP-1 RA thyroid C-cell findings emerged from 2-year rat carcinogenicity studies — illustrating the value of long-term rodent testing for detecting organ-specific tumour risk. For biotechnology-derived peptides, ICH S6(R1) notes that standard 2-year rodent carcinogenicity studies are generally inappropriate (species differences in receptor expression and signalling may render results non-predictive). Weight-of-evidence assessment considering target biology, chronic toxicology, and human exposure data may be used instead.

Cardiac safety assessment (ICH S7B/E14): non-clinical hERG channel assay (assessing risk of QTc prolongation/torsades de pointes), action potential duration studies, telemetry studies in conscious animals (QTc, heart rate, blood pressure), and thorough QT/QTc study in humans (Phase I study specifically designed to evaluate cardiac repolarisation effects). The thorough QT study (TQT) is typically a randomised, double-blind, placebo- and positive-controlled (moxifloxacin), crossover design in healthy volunteers. For peptide drugs, QTc prolongation risk is generally low because peptides typically do not block ion channels. However, indirect cardiac effects are important: GLP-1 RAs increase heart rate by 2-4 bpm (mechanism debated — possibly related to sympathetic activation or direct sinoatrial node effects), which must be characterised and its clinical significance assessed in cardiovascular outcomes trials.

CRS involves systemic release of pro-inflammatory cytokines (IL-6, IL-1, TNF-α, IFN-γ) from activated immune cells. Grading: Grade 1 (fever only), Grade 2 (fever + hypotension responsive to fluids, hypoxia responsive to low-flow O2), Grade 3 (fever + hypotension requiring vasopressors, hypoxia requiring high-flow O2), Grade 4 (life-threatening, ventilator support needed). CRS is most commonly associated with CAR-T cell therapy, bispecific antibodies, and checkpoint inhibitors rather than peptide drugs. However, any biologic that activates immune cells or complements carries theoretical CRS risk. For peptide drugs with immune-modulating properties (corticotropin, cyclosporine, glatiramer acetate, motixafortide), CRS is a theoretical safety consideration monitored in clinical trials, though clinically significant CRS with approved peptide drugs is rare.

Dose-dependent (Type A) adverse effects are predictable extensions of the drug's pharmacology and typically affect a larger proportion of patients at higher doses. They are the most common type of ADR, accounting for approximately 80% of all adverse drug reactions. For GLP-1 RAs, gastrointestinal effects (nausea, vomiting, diarrhoea) are dose-dependent — incidence increases at each dose escalation step and is highest at the maximum therapeutic dose. This dose-dependency is the rationale for gradual dose titration. Dose-dependent toxicity is characterised during dose escalation studies (Phase I) and dose-ranging studies (Phase II). Management includes dose reduction, temporary dose interruption, or slower titration schedule. In contrast, dose-independent (Type B) reactions (allergic, idiosyncratic) can occur at any dose.

True drug allergy involves adaptive immune system activation against the drug or its metabolites. Evaluation includes: clinical history (timing, symptoms, reproducibility), skin testing (immediate — prick/intradermal testing for IgE; delayed — patch testing for T-cell-mediated), in vitro testing (drug-specific IgE, basophil activation test), and graded challenge/desensitisation (administering incremental doses under medical supervision). For peptide drugs, true allergic reactions are uncommon compared to non-immune adverse effects (pharmacological side effects, injection site reactions). When allergy is suspected, the offending peptide should be discontinued and alternative agents from the same class (which may have different immunogenic epitopes) or different classes should be considered. Desensitisation protocols exist for peptide drugs that are essential and lack alternatives.

Peptide drug interaction profile is generally favourable because: peptides are not metabolised by hepatic CYP450 enzymes (the major source of small molecule drug interactions), they do not induce or inhibit CYP isoforms, and they are not substrates for P-glycoprotein or other drug transporters. However, clinically relevant interactions exist: GLP-1 RA gastric emptying delay can slow absorption of concurrently administered oral drugs — particularly relevant for drugs with narrow therapeutic windows (warfarin, levothyroxine, oral contraceptives) or time-sensitive absorption (antibiotics). For oral semaglutide, the fasting requirement (30 min before food/other medications) manages this interaction. GnRH compounds interact pharmacodynamically with sex hormone-dependent therapies. Somatostatin analogues can alter insulin and glucagon secretion, requiring diabetes medication dose adjustment. Drug interaction studies are a standard component of peptide clinical development programmes.

The GnRH agonist flare occurs because initial receptor binding stimulates gonadotroph signalling before desensitisation/downregulation develops. Duration: testosterone levels peak at approximately day 3-7 and typically return to baseline by day 14-21 before declining to castrate levels by day 28. Clinical significance: in prostate cancer, the testosterone flare can temporarily worsen bone pain, urinary obstruction, and spinal cord compression (from vertebral metastases) — potentially causing clinically dangerous disease flare. Mitigation strategies: pre-treatment or concurrent anti-androgen therapy (bicalutamide, flutamide) to block testosterone action during the flare period, or use of GnRH antagonists (degarelix, relugolix) which produce immediate suppression without flare. For endometriosis and fertility applications, the initial hormonal surge is generally less clinically dangerous.

Gastroparesis is diagnosed by delayed gastric emptying (>10% retention at 4 hours on scintigraphic gastric emptying study) in the absence of mechanical obstruction. GLP-1 RAs deliberately slow gastric emptying through vagal and direct smooth muscle mechanisms — this is a therapeutic effect (reducing postprandial glucose spikes and prolonging satiety). The clinical question is whether pharmacological delayed gastric emptying can progress to pathological gastroparesis in susceptible individuals. FDA investigation: FAERS analysis identified gastroparesis signals; the FDA requested additional data from manufacturers. Tachyphylaxis to the gastric emptying effect may occur with chronic GLP-1 RA use (the initial slowing attenuates over weeks). Risk factors for GLP-1 RA-associated gastroparesis may include: pre-existing gastric motility disorders, diabetes-related autonomic neuropathy, and concurrent use of other motility-reducing drugs.

Standard genotoxicity battery (ICH S2(R1)): Test 1 — bacterial reverse mutation assay (Ames test, detecting point mutations in Salmonella typhimurium and E. coli); Test 2 — in vitro mammalian cell assay (chromosomal aberration in CHO/CHL cells OR mouse lymphoma tk assay, detecting clastogenicity and aneuploidy); Test 3 — in vivo micronucleus test (detecting clastogenicity/aneuploidy in rodent bone marrow or peripheral blood). For biotechnology-derived peptides (recombinant proteins, long peptides), ICH S6(R1) states that genotoxicity testing is generally not needed because: peptides do not interact with DNA directly, they are degraded to amino acids, and standard bacterial assays are not relevant to large molecule mechanisms. However, for synthetic peptides with non-natural components (linkers, chemical modifications, conjugates), standard genotoxicity testing may be warranted.

Drug-induced liver injury (DILI) is classified as: hepatocellular (elevated ALT >3× ULN, indicating hepatocyte damage), cholestatic (elevated ALP >2× ULN, indicating bile duct/flow impairment), or mixed pattern. Hy's Law (FDA) identifies cases at highest risk of severe outcomes: ALT >3× ULN + bilirubin >2× ULN without cholestasis or other cause — this combination predicts approximately 10-50% risk of fatal liver failure. Liver function monitoring (ALT, AST, ALP, bilirubin) is standard in clinical trials and post-marketing for most drugs. For peptide drugs, hepatotoxicity is less common than with small molecules because peptides are not metabolised by CYP450 enzymes (a major source of reactive metabolite-mediated liver injury). However, some peptide drugs affect liver-related pathways: somatostatin analogues can cause gallstones (inhibiting gallbladder contraction), and pasireotide can cause hyperglycaemia (hepatic glucose metabolism effects).

Gell-Coombs classification: Type I (immediate/IgE-mediated — mast cell degranulation → histamine, leukotrienes → urticaria, angioedema, anaphylaxis; onset minutes to hours), Type II (antibody-mediated cytotoxicity — IgG/IgM against cell-surface antigens → complement activation, ADCC; onset hours to days), Type III (immune complex — drug-antibody complexes deposit in tissues → complement activation, inflammation; onset days to weeks; serum sickness-like reactions), Type IV (delayed/T-cell-mediated — sensitised T cells release cytokines → inflammation; onset 48-72 hours; contact dermatitis). For peptide drugs, Type I reactions (including anaphylaxis) are the most clinically significant acute risk. All injectable peptide drug labels include warnings about hypersensitivity. Skin testing and desensitisation protocols exist for some peptide drugs.

Idiosyncratic reactions are by definition unpredictable from standard pharmacological or toxicological testing. Proposed mechanisms include: genetic polymorphisms affecting drug metabolism or immune response (pharmacogenomic susceptibility), hapten formation (drug or metabolite covalently binds to endogenous proteins, creating neo-antigens recognised by the immune system), danger hypothesis (drug causes cellular stress that provides co-stimulatory signals for immune activation), and mitochondrial dysfunction (drug impairs mitochondrial function in genetically susceptible individuals). For peptide drugs, idiosyncratic reactions are less common than for small molecules (which have more complex hepatic metabolism producing reactive intermediates). However, immune-mediated idiosyncratic reactions to biologic peptides can occur through anti-drug antibody formation with immune complex deposition or through direct T-cell-mediated responses.

Immunogenicity risk factors include: non-human sequences (lower homology to endogenous peptides increases risk), route of administration (SC > IV > topical for immunogenicity), impurities and aggregates (misfolded peptides and HCP contaminants act as adjuvants), dose and frequency (higher doses and more frequent administration increase risk), patient factors (genetic predisposition, immune status). Assessment uses a tiered approach: screening assay (sensitive but less specific — detects binding antibodies) → confirmatory assay (verifies true positives) → neutralising antibody assay (determines functional impact) → titre determination. Regulatory guidance (FDA 2014, EMA 2017) requires immunogenicity assessment throughout clinical development. For approved peptide drugs, ADA incidence ranges from <1% to >50% depending on the compound.

Infusion reactions are classified by timing: immediate (during or within 1 hour — likely IgE-mediated or non-immune mast cell activation), delayed (1-24 hours — often cytokine-mediated or complement activation). Management: for mild reactions (flushing, chills, mild urticaria) — slow the infusion rate, administer antihistamines; for moderate reactions (significant urticaria, dyspnoea, hypotension responsive to fluids) — stop infusion, treat symptoms, consider restarting at slower rate with premedication; for severe reactions (anaphylaxis features) — stop infusion, treat as anaphylaxis. Premedication protocols (acetaminophen, antihistamines, corticosteroids) are used for drugs with known infusion reaction risk. Carfilzomib has notable infusion reaction risk (~15-20%) requiring premedication with dexamethasone and infusion rate management.

ISR mechanisms include: physical trauma (needle insertion causing tissue disruption), foreign body response (immune reaction to deposited material), pharmacological effect (local tissue response to the drug or excipients), and volume-related (distension of tissue by injected fluid — depot injections with larger volumes typically cause more ISRs). ISR management: site rotation (preventing cumulative damage), allowing refrigerated products to warm to room temperature before injection (cold products cause more pain), proper injection technique (correct depth, angle, speed), and application of cold packs post-injection for swelling. For GLP-1 RA pens, ISR rates are typically 1-5% in clinical trials. Depot injections (octreotide LAR, leuprolide depot) have higher ISR rates due to larger needle gauge, injection volume, and microsphere formulation.

Factors affecting breast milk drug transfer: molecular weight (peptides >1000 Da generally transfer poorly), protein binding (highly albumin-bound peptides transfer less), lipophilicity (hydrophilic peptides transfer less), half-life, and maternal plasma levels. Relative infant dose (RID) = (infant dose via milk / maternal dose) × 100; RID <10% is generally considered compatible with breastfeeding. For peptide drugs, even if small amounts transfer into milk, degradation in the infant's GI tract would be expected to inactivate most peptides (the same GI proteases that prevent oral bioavailability in adults). However, clinical data on breastfeeding safety of peptide drugs are limited because nursing women are typically excluded from clinical trials. Individual risk-benefit assessment considers: medical necessity of the drug, infant age and health, availability of alternatives, and the specific peptide's physicochemical properties.

MTD determination uses formalised dose escalation designs. In the 3+3 design: MTD = highest dose where ≤1/6 patients experience DLT. In model-based designs (CRM, mTPI, BOIN): MTD = dose closest to the target DLT rate (typically 20-33%) based on accumulated dose-toxicity data modelled using Bayesian or likelihood methods. For peptide drugs, the MTD established in Phase I may not be the therapeutic dose — many drugs are most effective at doses below MTD. The optimal biological dose (OBD) concept is increasingly used alongside MTD, particularly for targeted therapies where maximum efficacy may plateau before maximum toxicity. GLP-1 RA dose titration schedules essentially implement a personalised MTD approach — patients escalate to the highest tolerated dose within the approved range.

MTC arises from thyroid C-cells and accounts for approximately 3-5% of all thyroid cancers. It can be sporadic (~75%) or hereditary (~25%, associated with MEN2A, MEN2B, or familial MTC — all caused by RET proto-oncogene mutations). MTC produces calcitonin (used as tumour marker) and sometimes CEA. Five-year survival varies by stage: localised ~97%, regional ~90%, distant ~40%. The contraindication of GLP-1 RAs in patients with personal/family history of MTC or MEN2 is a precautionary measure based on rodent C-cell tumour findings. Genetic testing for RET mutations is recommended for patients diagnosed with MTC. Pre-treatment screening for MTC before initiating GLP-1 RA therapy is not recommended in general guidelines — the boxed warning relies on patient/family history rather than universal screening.

GLP-1 RA nausea mechanisms: central (GLP-1R activation in the area postrema and nucleus tractus solitarius — brainstem emetic centres), peripheral (delayed gastric emptying causing gastric distension and vagal afferent stimulation), and conditioned (anticipatory nausea in patients who experienced initial episodes). Incidence varies by compound and dose: semaglutide 2.4mg ~44% (STEP trials), tirzepatide 15mg ~31% (SURMOUNT), liraglutide 3.0mg ~39% (SCALE). Temporal pattern: peaks during first 4-8 weeks and dose escalation periods, then diminishes (tachyphylaxis of the emetic response). Management strategies: gradual dose titration (the primary approach), smaller meal portions, avoiding high-fat foods, eating slowly, staying hydrated, and anti-emetic medications (ondansetron, prochlorperazine) for persistent symptoms. Nausea is the most common reason for GLP-1 RA discontinuation in clinical practice.

Nephrotoxicity is assessed by monitoring serum creatinine, blood urea nitrogen (BUN), and estimated glomerular filtration rate (eGFR). Acute kidney injury (AKI) is classified by KDIGO criteria based on creatinine elevation or urine output reduction. Antimicrobial peptides with significant nephrotoxic potential: colistin/polymyxin B (dose-dependent proximal tubular necrosis — incidence 20-60% depending on dose and duration, requiring careful dose adjustment and monitoring), vancomycin (acute interstitial nephritis and/or acute tubular necrosis — trough monitoring targets 15-20 μg/mL for serious infections to balance efficacy vs nephrotoxicity). For GLP-1 RAs, initial reports of acute kidney injury prompted FDA warnings, though the mechanism appears related to dehydration from GI side effects (nausea, vomiting, diarrhoea) rather than direct nephrotoxicity. Adequate hydration counselling is important during GLP-1 RA initiation.

CNS safety pharmacology (ICH S7A): functional observational battery (FOB) in rodents assessing behaviour, motor activity, sensory function, and reflexes, plus assessment of body temperature. For peptide drugs designed to act on neurological targets, additional studies may include: electroencephalography (EEG), learning/memory assessments, and specific neurobehavioural testing. Peripheral neurotoxicity: some peptide drugs can affect peripheral nerves — proteasome inhibitors (bortezomib) cause dose-limiting peripheral neuropathy through disruption of intracellular protein homeostasis in neurons. Bortezomib-induced peripheral neuropathy occurs in approximately 30-40% of patients and is the most common reason for dose reduction. Carfilzomib has lower neurotoxicity rates (~15%) due to its different proteasome binding profile. For peptides targeting the BBB or CNS, thorough neurotoxicity assessment is essential.

Neutralising antibody assays use either cell-based functional assays (measuring whether ADA-containing serum blocks the drug's ability to activate its receptor in a cell system) or competitive ligand-binding assays (measuring whether ADA blocks drug-receptor binding). Cell-based assays are more physiologically relevant but technically demanding. The clinical threshold for clinically significant neutralisation varies by drug — low-titre transient NAbs may have no clinical impact, while high-titre persistent NAbs can render treatment ineffective. For biosimilar development, comparative immunogenicity (ADA/NAb incidence and clinical impact in the biosimilar vs reference product) is a key regulatory requirement. Risk mitigation strategies include: humanising peptide sequences, removing T-cell epitopes (deimmunisation), optimising formulation to minimise aggregation, and selecting administration routes with lower immunogenicity risk.

NOAEL is determined from the highest dose group in preclinical toxicology studies that shows no statistically or biologically significant increase in adverse findings compared to control. The NOAEL is used to calculate the human equivalent dose (HED) using allometric scaling (body surface area adjustment: HED = animal NOAEL × (animal Km / human Km), where Km = body weight/body surface area). The maximum recommended starting dose (MRSD) for first-in-human trials is typically 1/10th of the HED from the most sensitive species (applying a 10-fold safety factor). For peptides with species-specific pharmacology (e.g. human-specific receptor binding), the NOAEL from pharmacologically relevant species is prioritised. Additional safety factors may be applied for: irreversible toxicity, steep dose-response curves, or novel mechanisms.

Acute pancreatitis: diagnosed by 2 of 3 criteria — characteristic abdominal pain (severe epigastric, radiating to back), serum amylase or lipase >3× upper limit of normal, and imaging findings (CT/MRI). Severity ranges from mild (interstitial oedematous, self-limiting) to severe (necrotising, organ failure, mortality 20-40%). GLP-1 RA-pancreatitis relationship: early post-marketing reports raised concern. Large CVOTs provided mixed evidence — most showed no statistically significant increase in adjudicated pancreatitis, though numerical imbalances were observed in some trials. Meta-analyses suggest a small potential increase in risk (OR ~1.1-1.5). Mechanistic plausibility: GLP-1R activation stimulates pancreatic exocrine secretion and may promote ductal cell proliferation. Current position: pancreatitis is listed as a warning/precaution (not contraindication) for all GLP-1 RAs. Patients should report persistent severe abdominal pain. GLP-1 RA should be discontinued if pancreatitis is confirmed.

Legacy FDA categories (phased out 2015, replaced by narrative labelling under the Pregnancy and Lactation Labelling Rule/PLLR): Category A (adequate human studies show no risk), B (animal studies no risk + no adequate human studies), C (animal studies show risk + no adequate human studies OR no studies in either), D (evidence of human risk but benefits may justify use), X (evidence of human/animal risk, risks clearly outweigh benefits — CONTRAINDICATED). Under PLLR (effective June 2015 for new drugs), labels include: Pregnancy subsection (risk summary, clinical considerations, data), Lactation subsection (risk summary, clinical considerations, data), and Females/Males of Reproductive Potential subsection (contraception requirements, fertility effects). Most peptide drugs affecting hormonal systems are contraindicated in pregnancy. GLP-1 RAs recommend discontinuing at least 2 months before planned pregnancy (based on semaglutide's ~5-week half-life requiring ~5 half-lives for washout).

Rebound occurs because the body adapts to the drug's presence through compensatory mechanisms (receptor upregulation, counter-regulatory pathway activation, altered set points). Upon discontinuation, these compensatory changes persist temporarily while the drug effect disappears, causing symptoms that exceed pre-treatment levels. For peptide drugs: GLP-1 RA discontinuation is followed by rapid appetite return and weight regain (the hypothalamic appetite circuits re-emerge from drug suppression), GnRH agonist discontinuation restores gonadotropin pulsatility and sex steroid production (gradual recovery over weeks-months, exploited in fertility protocols where post-flare recovery produces ovulation), and somatostatin analogue discontinuation may cause temporary GH/hormone rebound. Gradual dose tapering can attenuate rebound effects for some drugs, though this is not standard practice for all peptide drug classes.

ICH S5(R3) reproductive toxicity studies: Segment I — fertility and early embryonic development (assessing effects on male/female fertility, mating behaviour, conception, and pre-implantation/early post-implantation development); Segment II — embryo-foetal development (EFD, assessing teratogenicity — structural abnormalities in offspring when dosed during organogenesis); Segment III — peri/postnatal development (PPND, assessing effects on late pregnancy, parturition, lactation, and offspring development). For GnRH agonists/antagonists, reproductive toxicity is expected based on the mechanism (suppressing sex hormones disrupts fertility and embryo-foetal development) — they carry pregnancy Category X classification. GLP-1 RAs: animal EFD studies showed embryo-foetal effects at high doses; pregnancy is listed as a contraindication or precaution. Women of childbearing potential are advised to use contraception during treatment and for a washout period after discontinuation.

Teratogenicity assessment uses animal models: typically rats (representing rodents) and rabbits (representing non-rodents, particularly sensitive to some teratogens). Dosing during the organogenesis window (days 6-17 in rats, days 6-18 in rabbits) identifies structural malformations. Endpoints include: external malformations (limb defects, neural tube defects), visceral abnormalities (cardiac, urogenital), skeletal abnormalities (vertebral, rib, sternal), and developmental variations (delayed ossification). For peptide drugs, teratogenic risk may relate to: direct embryotoxicity, disruption of growth factor signalling essential for development, or hormonal effects that alter the uterine environment. The thalidomide tragedy established the requirement for thorough reproductive toxicity testing before drugs are used in women of childbearing potential. Modern developmental toxicity testing follows ICH S5(R3) guidelines.

TI = TD50/ED50 (ratio of dose causing toxicity in 50% of population to dose producing desired effect in 50%). A TI >10 is generally considered wide; TI <2 is narrow. Drugs with narrow TI include: vancomycin (TI ~2-4, requiring therapeutic drug monitoring), warfarin (TI ~2), lithium (TI ~2-3), and digoxin (TI ~2). Most peptide drugs have wider TI due to receptor specificity, but TI varies by indication — semaglutide's TI for glycaemic control may differ from its TI for weight management (where higher doses are used). The therapeutic index concept is also applied to the separation between desired effects and specific adverse effects: for GLP-1 RAs, the 'therapeutic index' for weight loss vs nausea is relatively narrow during dose titration.

Preclinical data: GLP-1 RAs caused dose-dependent, treatment-duration-dependent thyroid C-cell hyperplasia, adenomas, and carcinomas in rats and mice at clinically relevant exposures. Mechanism: sustained GLP-1R activation on rodent C-cells stimulates calcitonin release and C-cell proliferation. Species difference: human thyroid C-cells express far fewer GLP-1 receptors than rodent C-cells, and human C-cells show minimal calcitonin response to GLP-1R agonism. Epidemiological data: large database studies and post-marketing surveillance have not demonstrated a clear increase in MTC incidence in GLP-1 RA-treated patients, though follow-up duration remains limited for detecting rare, slow-growing tumours. Calcitonin monitoring (routine serum calcitonin measurement during GLP-1 RA therapy) is NOT recommended by most guidelines due to low specificity and high false positive rates, but unexplained calcitonin elevation should prompt thyroid investigation.

Preclinical toxicology programme for peptide drugs (ICH M3/S6 guidelines): single-dose toxicity (acute effects, determine maximum non-lethal dose), repeat-dose toxicity (2-4 weeks for Phase I, 13 weeks for Phase II, 26 weeks for Phase III/marketing — in two species, one non-rodent), genotoxicity battery (Ames bacterial mutation, in vitro chromosomal aberration or mouse lymphoma, in vivo micronucleus — note: for biotechnology-derived peptides/proteins, genotoxicity may be waived per ICH S6(R1) if no DNA-reactive components), reproductive toxicity (fertility and early embryonic development, embryo-foetal development, peri/postnatal development), and carcinogenicity (2-year rodent studies — required for chronically administered drugs; may use 6-month transgenic mouse + 2-year rat, or justified waiver). For peptides ≥40 amino acids produced by recombinant technology, ICH S6(R1) provides specific guidance acknowledging that standard small molecule toxicology approaches may not be appropriate.

Peptide drug withdrawal is generally not associated with the classical physical withdrawal syndromes seen with opioids, benzodiazepines, or alcohol (because peptide drugs don't typically cause physical dependence). However, disease-related withdrawal effects occur: GLP-1 RA discontinuation leads to weight regain in 50-70% of patients within 1-2 years (STEP 4 extension data showed approximately 2/3 of weight lost was regained after stopping semaglutide); GH replacement discontinuation causes return of GH deficiency symptoms (increased visceral fat, decreased lean mass, fatigue, reduced quality of life); and GnRH agonist discontinuation restores the hormonal milieu that was therapeutically suppressed. These withdrawal effects reflect reversal of pharmacological effects rather than true pharmacological dependence.