Highlight
– A large network meta‑analysis of 151 RCTs (58,534 participants) found clinically relevant differences among 30 antidepressants in weight, heart rate and blood pressure.
– Paroxetine, duloxetine, desvenlafaxine and venlafaxine were associated with increases in total cholesterol; duloxetine was associated with increased glucose despite weight loss.
– Duloxetine, desvenlafaxine and levomilnacipran increased liver enzymes (AST, ALT, ALP) though changes were small and not judged clinically significant in trials of median 8 weeks.
– No strong evidence emerged for clinically meaningful QTc prolongation or major changes in electrolytes, urea or creatinine.
Background
Antidepressants are among the most commonly prescribed psychotropic medicines worldwide. Beyond their psychiatric effects, several drug classes—selective serotonin reuptake inhibitors (SSRIs), serotonin‑noradrenaline reuptake inhibitors (SNRIs), tricyclics and newer agents—can influence cardiometabolic and other physiological parameters. These changes matter because weight gain, dyslipidaemia, hyperglycaemia, hepatic enzyme elevations and haemodynamic effects may increase long‑term cardiovascular risk, affect adherence, or complicate treatment in persons with multimorbidity. However, the comparative magnitude and direction of these effects across many antidepressants have been unclear, particularly from randomized evidence pooled across trials.
Study design
This Lancet network meta‑analysis (Pillinger et al., 2025) synthesized single‑ and double‑blind randomized controlled trials comparing antidepressants with placebo in acute monotherapy of any psychiatric disorder. The authors searched major bibliographic databases, ClinicalTrials.gov and the US Food and Drug Administration (FDA) website through April 21, 2025, and included both published RCTs and 17 FDA reports. The final dataset included 151 trials and 58,534 participants exposed to 30 antidepressants or placebo. Median treatment duration was 8 weeks (IQR 6.0–8.5).
Frequentist random‑effects network meta‑analyses assessed treatment‑induced changes in multiple physiological endpoints: bodyweight; total cholesterol; glucose; heart rate; systolic and diastolic blood pressure; corrected QT interval (QTc); serum sodium and potassium; liver enzymes (AST, ALT, ALP); bilirubin; urea; and creatinine. Meta‑regressions explored whether baseline age, sex and weight modified effects. The analysis also examined correlations between change in depressive symptoms and metabolic parameter change.
Key findings
Scope and general considerations
Across a broad antidepressant portfolio and more than 58,000 participants, the authors report robust evidence that antidepressants differ substantially in their short‑term physiological effects. Most RCTs were short (median 8 weeks), so findings primarily reflect acute to subacute changes rather than long‑term outcomes; this is important when extrapolating to chronic treatment.
Weight
Significant between‑drug differences in weight change were observed. The largest contrast reported was an approximate 4 kg difference in mean weight change between agomelatine and maprotiline over trial periods. While some drugs on average reduced body weight, others were neutral or associated with weight gain. The pattern suggests that choice of antidepressant can meaningfully affect short‑term weight trajectory, which may have implications for adherence and cardiometabolic risk when treatment is prolonged.
Heart rate and blood pressure
Antidepressants showed variable haemodynamic effects. The network meta‑analysis found over a 21 beats‑per‑minute difference in heart‑rate change between fluvoxamine and nortriptyline, and more than an 11 mmHg difference in systolic blood pressure between nortriptyline and doxepin. Tricyclic‑class agents and related compounds tended to exert larger haemodynamic effects compared with many SSRIs and newer agents. Such differences could be clinically relevant for patients with baseline cardiovascular disease, orthostatic intolerance, or on concomitant antihypertensive therapy.
Lipids and glucose
Paroxetine, duloxetine, desvenlafaxine and venlafaxine were associated with increases in total cholesterol versus placebo. Notably, duloxetine was also associated with increases in glucose concentration despite concurrent reductions in body weight for some drugs. These dissociations between weight and metabolic markers highlight that mechanisms other than weight change (for example, direct effects on hepatic metabolism or insulin sensitivity) may contribute to metabolic alterations with certain antidepressants.
Liver function tests
The analysis showed strong evidence that duloxetine, desvenlafaxine and levomilnacipran increase AST, ALT and ALP concentrations. Although these increases were statistically robust, the authors judged the magnitudes not to be clinically significant within the short durations of most trials. Still, vigilance is prudent because clinically important antidepressant‑associated hepatic injury has been reported in post‑market surveillance for specific agents and in clinical practice.
QTc interval and electrolytes
Importantly, the study did not find strong evidence that any individual antidepressant produced clinically meaningful QTc prolongation in pooled RCT data. Similarly, concentrations of sodium, potassium, urea and creatinine were not altered to clinically significant degrees in the short term across drugs. These findings provide some reassurance but do not obviate the need for ECG monitoring where other risk factors for QTc prolongation or drug interactions are present.
Effect modifiers
Meta‑regression analyses identified that higher baseline bodyweight was associated with larger antidepressant‑induced increases in systolic blood pressure and in liver enzyme concentrations (ALT, AST). Higher baseline age was associated with larger antidepressant‑induced increases in glucose. There was no observed association between change in depressive symptom severity and metabolic disturbance, suggesting metabolic effects are not merely a secondary consequence of symptomatic improvement.
Clinical significance and magnitude
The authors present multiple comparisons with effect sizes that are clinically relevant (for example, multi‑kg weight differences, >10 mmHg systolic BP differences, tens of beats‑per‑minute differences in heart rate). However, most trial durations were short, and the long‑term trajectory of these changes—and their translation to cardiovascular events or diabetes risk—remain uncertain.
Expert commentary and interpretation
This network meta‑analysis provides the most comprehensive randomized evidence to date comparing physiological effects across a broad range of antidepressants. Strengths include its large sample size, inclusion of FDA trial data, and the use of network methods to estimate many pairwise contrasts that are otherwise unavailable from head‑to‑head RCTs.
From a mechanistic perspective, differing affinities for histaminergic, cholinergic, serotonergic or noradrenergic receptors, and effects on hepatic enzymes, could explain why drugs in the same class nevertheless have heterogeneous metabolic and haemodynamic profiles. For example, agents with prominent noradrenergic effects or anticholinergic activity are biologically plausible candidates for greater increases in heart rate or blood pressure.
Clinically, these results argue for selecting antidepressants with consideration of baseline cardiometabolic risks. For patients with obesity, diabetes, dyslipidaemia or cardiovascular disease, choosing agents with more neutral profiles for weight, lipids and glycaemia may be prudent. Conversely, for underweight patients or those experiencing clinically significant weight loss, agents associated with weight gain may be advantageous.
Limitations and generalizability
Key limitations include the short median trial duration (8 weeks), which constrains conclusions about long‑term cardiometabolic risk. RCT populations often exclude people with significant medical comorbidity, limiting generalizability to multimorbid or frail patients. Heterogeneity in trial methods, measurement timing, and baseline characteristics may influence pooled estimates despite the use of random‑effects models and meta‑regression. Finally, rare but clinically important adverse events (for example, clinically significant liver injury or sudden cardiac events) may not be detected with RCT sample sizes and durations used here; post‑marketing surveillance remains essential.
Clinical implications and practical recommendations
1) Baseline assessment: For all patients starting antidepressant therapy, document weight/BMI, blood pressure, fasting glucose and lipids; assess alcohol use, existing liver disease and cardiovascular history. Consider baseline liver enzymes when prescribing agents with signals for LFT change (for example, duloxetine, desvenlafaxine, levomilnacipran) or when other hepatotoxic risks exist.
2) Individualize drug selection: Use patient‑specific cardiometabolic risk to inform antidepressant choice. Where short‑term haemodynamic stability is critical (advanced cardiac disease, uncontrolled hypertension), prefer agents with smaller observed BP and heart‑rate effects. For patients with baseline dyslipidaemia or at high cardiovascular risk, consider agents associated with smaller lipid perturbations.
3) Monitoring: Repeat weight and blood pressure early in treatment (e.g., 4–8 weeks) and reassess fasting lipids and glucose within 8–12 weeks if drugs with metabolic signals are used or if baseline risk is elevated. Extend monitoring as clinically indicated for long‑term therapy.
4) Interpret changes contextually: Small and statistically significant laboratory changes do not always equate to clinically important harm, especially over short trial durations. Use trends, absolute values and clinical context to guide management decisions (dose adjustment, switching agents, or additional investigations).
Research implications and future directions
Longer‑term randomized trials, pragmatic effectiveness studies and observational cohorts with robust confounding control are needed to quantify how short‑term physiological changes translate into cardiovascular events, diabetes onset or clinically important liver disease. Studies should include older adults, people with multimorbidity, and diverse ethnic groups to improve generalizability. Mechanistic research examining direct drug effects on hepatic metabolism, insulin sensitivity and autonomic regulation will help explain dissociations between weight change and metabolic markers.
Conclusion
This comprehensive network meta‑analysis demonstrates that antidepressants differ markedly in their short‑term physiological effects, particularly for weight, haemodynamics and selected metabolic markers. Clinicians should integrate these differential risk profiles into shared decision‑making with patients, tailoring drug choice and monitoring to individual cardiometabolic risk and treatment goals. Treatment guidelines should reflect these distinctions and provide explicit monitoring pathways, while additional long‑term data are needed to quantify downstream cardiovascular and metabolic consequences.
Funding and trial registration
The underlying study was funded by the National Institute for Health Research, Maudsley Charity, Wellcome Trust and Medical Research Council. The original paper: Pillinger T, Arumuham A, McCutcheon RA, et al. The effects of antidepressants on cardiometabolic and other physiological parameters: a systematic review and network meta‑analysis. Lancet. 2025 Nov 1;406(10515):2063‑2077. doi:10.1016/S0140-6736(25)01293-0.
References
1. Pillinger T, Arumuham A, McCutcheon RA, et al. The effects of antidepressants on cardiometabolic and other physiological parameters: a systematic review and network meta‑analysis. Lancet. 2025;406(10515):2063‑2077. doi:10.1016/S0140-6736(25)01293-0.
AI thumbnail prompt
A clinician at a desk reviewing a patient file with floating graphical overlays showing trends in weight, blood pressure, heart rate, cholesterol and liver enzyme bars; modern clinic room, diverse middle‑aged patient in the background, realistic lighting, professional and contemplative mood, high detail.
