It’s my first day again locally psychological well being clinic working as a psychiatric resident physician. Brompton within the nook, Kelston Roundhill out the window. Jim’s telling me that the threatening voices have began coming again. After some dialogue, I’m telling Jim I feel he ought to restart an antipsychotic. He explains that aripiprazole made him stressed and solely olanzapine labored effectively final time, however he gained plenty of weight. What ought to we do?
The 2020 Maudsley Superior Prescribing in Psychosis Pointers inform me prevention of weight achieve is best than remedy, and that weight loss program, train, switching antipsychotics or including aripiprazole, metformin, topiramate, orlistat or Glucagon-Like Peptide-1 (GLP-1) receptor agonists might help (Morrison et al., 2020). A 2022 Cochrane Evaluate tells me that there’s low certainty proof from 4 randomised managed trials (RCTs) that metformin could also be efficient in stopping weight achieve (Agarwal et al., 2022). A Psychological Elf weblog from final month highlights a 2025 Lancet Fee report, which recommends including metformin if beginning clozapine or olanzapine (Lawson, 2025).
Effectively, these assets are suggesting metformin, however I can’t keep in mind ever seeing a affected person beginning metformin concurrently an antipsychotic. What’s going on? Am I a foul physician? Including to my confusion, the 2025 Maudsley Prescribing Pointers in Psychiatry say that medicines to deal with weight achieve ought to solely be thought-about if life-style interventions or antipsychotic switching haven’t labored or weight problems is a direct bodily threat (Taylor et al., 2025). It might be nice to see what clinicians are doing across the UK. Fortunately, some wonderful psychiatric epidemiologists have already had that very same thought (Trajano et al., 2025).
Regardless of sturdy guideline assist for utilizing metformin to stop weight achieve attributable to antipsychotic remedy, its use in apply stays unusual, elevating questions on medical habits and real-world prescribing.
Strategies
Trajano and colleagues used two main care databases from the Medical Observe Analysis Datalink to establish 26,537 sufferers who began aripiprazole, olanzapine, quetiapine or risperidone between 2005 and 2019, and had beforehand had schizophrenia, bipolar dysfunction or different non-organic psychoses famous of their information.
In Half 1, they measured whether or not these folks have been ever prescribed metformin and when. Then, for people who weren’t initially prescribed metformin, they measured the cumulative incidence of being prescribed metformin 1 and a pair of years down the road. Additionally they measured the proportion of antipsychotic customers prescribed metformin annually.
In Half 2, they in contrast the traits of these prescribed metformin inside 2 years of beginning the antipsychotic in comparison with these not ever prescribed any metformin.
In Half 3, they in contrast the adjustments in weight between sufferers who began metformin when beginning the antipsychotic and people who by no means took metformin. Additionally they stratified by substance in addition to utilizing multivariate linear regression fashions to estimate the impact of beginning metformin. Mannequin 1 was adjusted for baseline weight alone and mannequin 2 included baseline weight plus antipsychotic medicine, age at index date, ethnicity, social deprivation, prior prognosis of diabetes and prior prognosis of poly-cystic ovarian syndrome (PCOS).
Outcomes
Half 1
Of the 26,537 sufferers who began an antipsychotic, 2,873 had began metformin earlier than beginning the antipsychotic. Among the many remaining 23,664, there was a cumulative incidence of 1.9% beginning metformin inside a 12 months and three.3% beginning it inside 2 years. This equated to 696 folks beginning metformin inside 2 years. 1,083 began metformin greater than 2 years after beginning an antipsychotic, and 21,885 by no means took metformin.
The proportion of individuals prescribed metformin elevated from 1.31% in 2005 to five.84% in 2017.
Half 2
In comparison with the 21,885 that by no means took metformin, the 696 that took metformin inside 2 years of beginning the antipsychotic have been extra prone to have diabetes, have hypertension, be overweight, and take lipid-regulating medicines. There have been additionally extra delicate variations in recorded ethnicity and native relative deprivation and the proportion with schizophrenia, dyslipidaemia, PCOS (polycystic ovarian syndrome), insulin-prescriptions, in addition to imply HbA1c and glucose ranges. Distributions of intercourse, ages, geographical areas and different well being situations and medicines appeared related between the 2 teams. It appeared like 51% of these beginning metformin had documented diabetes or PCOS as possible indications for its use.
Half 3
The 21,885 sufferers who by no means began metformin have been 76.1kg on common when beginning the antipsychotic and 78.7kg two years later. So, their weight elevated 4.2% on common. 212 sufferers began metformin between 1 month earlier than and three months after beginning the antipsychotic. On common they have been 90.4kg in the beginning and 89.3kg at 2 years, thus shedding 0.7% on common. There have been no clear variations between completely different antipsychotics within the stratified analyses.
The authors then estimated weight at 2 years utilizing a linear regression mannequin that included solely metformin use and baseline weight as unbiased variables. Females that took metformin have been predicted to be 2.04 kg lighter (95percentCI: -0.47 to 4.55) than females that didn’t, and males 3.02 kg lighter (95percentCI: 0.20 to five.84). When utilizing a mannequin that additionally included antipsychotic medicine, age at index date, ethnicity, social deprivation, prior prognosis of diabetes and prior prognosis of PCOS, the values have been 1.48 kg (95percentCI: -1.07 to 4.03) amongst females and 1.84 kg (95percentCI: -0.98 to 4.67) amongst males.
Amongst over 26,000 sufferers beginning antipsychotics, metformin was hardly ever prescribed, sometimes reserved for these with diabetes or polycystic ovarian syndrome, and its use was related to modest weight reduction in comparison with those that by no means obtained it.
Conclusions
The authors concluded that:
- There may be low utilisation of metformin, regardless of tips supporting its use,
- Additional research are required to reliably estimate the impact of metformin exterior of trial settings and who would possibly profit most, and
- Clinicians want clear steering on implementing present steering round cardiometabolic care when utilizing second era antipsychotics.
Metformin stays underused regardless of guideline assist, and clearer implementation methods and additional real-world analysis are wanted to information its function in cardiometabolic care alongside second era antipsychotics.
Strengths and limitations
It’s thrilling to see the researchers harness giant population-representative datasets to evaluate clinically vital questions:
- Are we prescribing metformin?
- Who’re we prescribing it to?
- Are we seeing the anticipated results on weight achieve?
Strengths of the research embody the big research inhabitants and detailed comparability between these beginning metformin and never. Being primarily based on routinely collected knowledge in main care, it needs to be at low threat of the wholesome volunteer bias and Hawthorne impact (i.e., behaviour change in response to being noticed) and the sharing of analytical code is invaluable for understanding and constructing upon this work.
The information supply and broad inclusion standards imply that research inhabitants also needs to be consultant of the broader inhabitants beginning antipsychotics. Nevertheless, with the typical age of sufferers coming into this research being over 50 years previous, the outcomes might not be generalisable to youthful adults of the ages at which extreme psychological sickness (SMI) is probably to be first recognized. The exclusion of over 120,000 sufferers due being registered on the GP for lower than 6 months or not having a document of SMI earlier than the antipsychotic, probably contributed to underrepresentation of younger adults. Whereas these exclusion standards have been vital for research validity and relevance, the exclusion of 36,800 sufferers (as a result of there have been no lipids or HbA1c recorded in prior 2 years) was probably pointless and a contributor to choice bias.
One other limitation is that at every time level after antipsychotic initiation solely 33 to 41% of members had weight recorded (Richards-Belle et al., 2025). The authors stuffed within the lacking knowledge primarily based on the idea that the chance of a variable being lacking will be calculated from noticed knowledge. Nevertheless, this might not be legitimate as individuals who achieve weight are maybe extra prone to get their weight measured. Sensitivity analyses may very well be used to estimate the believable affect of this (Madley-Dowd et al., 2025), however it will be further difficult as a result of the connection between weight achieve and whether or not weight is measured might be affected by whether or not somebody is taking metformin.
Ideally, for a fuller image of healthcare practices this research would additionally embody data on secondary care prescribing or different weight administration interventions. Further limitations regarding half 3 of the research are the inherent problem in causal analyses of weight adjustments (Tennant et al., 2023) and the inevitable presence of unmeasured confounding. Lastly, it’s a disgrace that there was a scarcity of energy to match outcomes between antipsychotics: we might count on the usefulness of metformin to be larger for these beginning olanzapine than aripiprazole.
This huge research makes use of consultant knowledge to discover metformin prescribing and weight outcomes, however exclusions, lacking information, and confounding restrict its relevance to youthful sufferers and broader prescribing selections.
Implications for apply
Half 1 of this research needs to be a wake-up name that we’re most likely not doing sufficient to stop or handle antipsychotic induced weight achieve.
Half 2 of this research signifies that even among the many few folks co-prescribed metformin, the explanation was largely diabetes or PCOS, once more reinforcing the message that we don’t appear to be specializing in the prevention of weight achieve sufficient.
Half 3 of the research serves as a contribution to triangulating the proof across the impact of metformin in stopping antipsychotic-induced weight achieve. We had proof from RCTs that metformin helps stop weight achieve, however these research could not have been very generalisable and should have been affected by volunteer bias or the Hawthorne impact. That this research of main care information attracts related conclusions, regardless of having completely different sources of bias, strengthens our confidence within the conclusions (Lawlor et al., 2016).
What different types of observational proof could reinforce the conclusion that metformin helps stop weight achieve? We might carry out cross-context comparisons utilizing different nations with completely different healthcare methods and confounding buildings. Alternatively, we might evaluate the outcomes to these for a destructive management publicity that we don’t count on to trigger giant weight reduction, equivalent to statins, or a destructive management end result that we don’t count on to be affected by metformin, for instance blood strain in non-diabetic sufferers. Finally, there stays sufficient uncertainty in regards to the execs, cons and affected person experiences round beginning metformin with antipsychotics that inclusive RCTs with substantial follow-up instances seem warranted.
So, ought to Jim begin metformin? If he’s going to start out olanzapine, most likely sure. However first it will likely be vital to debate the effectiveness of life-style interventions, the potential unintended effects of metformin, together with lactic acidosis and vitamin B12 deficiency, and be clear in regards to the ongoing monitoring of this medicine use, whether or not in main or secondary care. Will I like to recommend metformin for everybody beginning second-generation antipsychotics primarily based on the outcomes of this research? No, we’ll make shared selections primarily based on particular person threat elements and preferences, with the assistance of the most recent tips (e.g. Carolan et al., 2025). Will the rules look the identical in 5 years’ time? GLP-1 agonists will certainly characteristic extra prominently, gained’t they?
Clinicians ought to take into account metformin for sufferers beginning olanzapine or clozapine, however stay aware of particular person dangers, the worth of life-style interventions, and the evolving function of newer therapies like GLP-1 agonists.
Assertion of pursuits
I’ve by no means met the authors, but additionally use CPRD for psychiatric epidemiology analysis, and have taken a lot inspiration from their earlier work.
Hyperlinks
Main paper
Farache Trajano L, Hayes JF, Launders N, Davies NM, Osborn DPJ, Richards-Belle A. (2025) Co-prescription of metformin and antipsychotics in extreme psychological sickness: a UK main care cohort research. BMJ Psychological Well being. 2025;28:e301505. https://doi.org/10.1136/bmjment-2024-301505
Different references
Agarwal, S. M., Stogios, N., Ahsan, Z. A., Lockwood, J. T., Duncan, M. J., Takeuchi, H., Cohn, T., Taylor, V. H., Remington, G., Faulkner, G. E. J., & Hahn, M. (2022). Pharmacological interventions for prevention of weight achieve in folks with schizophrenia. The Cochrane Database of Systematic Opinions, 10(10), CD013337. https://doi.org/10.1002/14651858.CD013337.pub2
Carolan, A., Hynes-Ryan, C., Agarwal, S. M., Bourke, R., Cullen, W., Gaughran, F., Hahn, M. Okay., Krivoy, A., Lally, J., Leucht, S., Lyne, J., McCutcheon, R. A., Norton, M. J., O’Connor, Okay., Perry, B. I., Pillinger, T., Shiers, D., Siskind, D., Thompson, A., … O’Donoghue, B. (2025). Metformin for the Prevention of Antipsychotic-Induced Weight Achieve: Guideline Improvement and Consensus Validation. Schizophrenia Bulletin, 51(5), 1193–1205. https://doi.org/10.1093/schbul/sbae205
Lawlor, D. A., Tilling, Okay., & Davey Smith, G. (2016). Triangulation in aetiological epidemiology. Worldwide Journal of Epidemiology, 45(6), 1866–1886. https://doi.org/10.1093/ije/dyw314
Lawson, Okay. (2025, August 13). Bodily well being unintended effects of psychotropic medicine. Nationwide Elf Service. https://www.nationalelfservice.web/remedy/antipsychotics/physical-health-side-effects-of-psychotropic-medication-holistic-prevention-and-management/
Madley-Dowd, P., Hughes, R. A., Mathur, M. B., Heron, J., & Tilling, Okay. (2025). Utilizing directed acyclic graphs to find out whether or not a number of imputation or subsample a number of imputation estimates of an exposure-outcome affiliation are unbiased (No. arXiv:2503.24035). arXiv. https://doi.org/10.48550/arXiv.2503.24035
Morrison, P., Taylor, D. M., & McGuire, P. (2020). The Maudsley tips on superior prescribing in psychosis (pp. xii, 93). Wiley Blackwell.
Richards-Belle, A., Launders, N., Hardoon, S., Richards, A., Man, Okay. Okay. C., Davies, N. M., Bramon, E., Hayes, J. F., & Osborn, D. P. J. (2025). Comparative cardiometabolic security and effectiveness of aripiprazole in folks with extreme psychological sickness: A goal trial emulation. PLOS Drugs, 22(1), e1004520. https://doi.org/10.1371/journal.pmed.1004520
Taylor, D. M., Barnes, T. R. E., & Younger, A. H. (2025). The Maudsley Prescribing Pointers in Psychiatry. John Wiley & Sons.
Tennant, P. W. G., Tomova, G. D., Murray, E. J., Arnold, Okay. F., Fox, M. P., & Gilthorpe, M. S. (2023). Lord’s ‘paradox’ defined: The 50-year warning on using ‘change scores’ in observational knowledge (No. arXiv:2302.01822). arXiv. https://doi.org/10.48550/arXiv.2302.01822
Trajano, L. F., Hayes, J. F., Launders, N., Davies, N. M., Osborn, D. P. J., & Richards-Belle, A. (2025). Co-prescription of metformin and antipsychotics in extreme psychological sickness: A UK main care cohort research. BMJ Psychological Well being, 28(1). https://doi.org/10.1136/bmjment-2024-301505
