Metformin

Metformin is an oral biguanide and one of the historical and contemporary cornerstones of treatment for type 2 diabetes mellitus. Its clinical relevance derives from its action on a central pathophysiological feature of the disease—excessive hepatic glucose production associated with systemic insulin resistance—without intrinsically causing hypoglycemia, together with a profile of cost, clinical experience and manageability matched by few other medications. Although it is no longer the mandatory starting point for every patient, because modern guidelines require initial therapy also to be guided by atherosclerotic cardiovascular disease, heart failure, chronic kidney disease, obesity and individual goals, metformin retains a broad structural role in clinical practice, both as monotherapy in early disease and as the foundation for combination with almost every other glucose-lowering class.

Epidemiologically, metformin is the most widely used oral antidiabetic medication worldwide and the one with the longest prescribing tradition in type 2 diabetes mellitus. Its long-standing central position is not due to cultural inertia, but to a combination of glycemic efficacy, neutral or favorable effects on body weight, low risk of hypoglycemia, possible benefit in diabetes prevention in selected groups and extensive experience across multiple clinical settings. Its modern interpretation nevertheless requires a more nuanced approach than in the past: metformin should be placed within a personalized therapeutic strategy, not used as an automatic response to every new diagnosis, but assessed in light of metabolic phenotype, kidney function, gastrointestinal tolerability, age, comorbidities, nutritional risk, pregnancy and possible vitamin deficiencies.

Metformin became a therapeutic standard because it acts on one of the most consistent abnormalities in type 2 diabetes mellitus: inadequate suppression of hepatic glucose production. In healthy individuals, insulin reduces hepatic gluconeogenesis and glycogenolysis after meals, whereas in type 2 diabetes mellitus this restraint is weakened by the combination of hepatic insulin resistance, inadequate insulin secretion and often relative hyperglucagonemia. The liver therefore continues to release glucose into the circulation when it should not, contributing substantially to fasting glucose in particular. Metformin directly counters this mechanism and is therefore especially effective when the glycemic profile is dominated by the hepatic component of hyperglycemia.

This action also explains its pathophysiological role in the early stages of disease. In many newly diagnosed patients, before severe beta-cell exhaustion develops, much of the metabolic disorder is sustained by excessive hepatic glucose output, reduced insulin efficiency in skeletal muscle and intestinal abnormalities of glucose metabolism. Introducing metformin at this stage targets a central axis of the metabolic defect without imposing artificial secretory stimulation on the beta cell. Unlike secretagogues, metformin does not force the pancreas to release more insulin independently of blood glucose, which explains both its lower hypoglycemic risk and its particular consistency with a pathophysiological strategy.

Metformin’s position in modern guidelines has become more nuanced. In patients without dominant cardiorenal comorbidities, it remains a very common and rational starting option. In those with atherosclerotic cardiovascular disease, heart failure, chronic kidney disease or clinically relevant obesity, contemporary algorithms instead require early consideration of medications with documented specific benefits, such as sodium-glucose cotransporter 2 inhibitors or glucagon-like peptide 1 receptor agonists, without eliminating metformin’s role. In real-world practice, it is often retained as a combination foundation unless contraindicated or not tolerated, because it preserves glycemic efficacy, prescribing simplicity and compatibility with subsequent therapies.

This conceptual evolution is important to avoid two opposite errors. The first is to regard metformin as “outdated,” an interpretation not supported by evidence. The second is to use it mechanically as the first and only medication in every clinical setting, ignoring the precision medicine that now guides diabetes treatment. Its correct place is that of a fundamental but not absolute medication, integrated into a decision hierarchy in which cardiorenal benefit, weight management and organ protection may, in some patients, take priority even over glucose-lowering efficacy alone.

Metformin also retains distinctive historical prognostic value. The pivotal studies that supported its use showed not only improved glycemic control, but also signals of benefit for diabetes-related endpoints in people with overweight, contributing decisively to its establishment as foundational therapy. These findings must now be reinterpreted in the era of cardiorenal medications, but they remain crucial to understanding why metformin continues to be considered a pillar of diabetes treatment rather than merely an inexpensive medication of an older generation.

Metformin’s mechanism of action is far more complex than the traditional summary that it “reduces hepatic glucose production.” That statement is correct but incomplete. The molecule enters cells through organic cation transporters, particularly in the liver, intestine and kidneys, and preferentially distributes to tissues that express these transport systems. Once inside the cell, metformin interferes with mitochondrial energy metabolism, especially respiratory-chain complex I to an extent that depends on biological context and intracellular concentration, reducing adenosine triphosphate production and relatively increasing adenosine monophosphate and adenosine diphosphate.

This change in intracellular energy balance has broad consequences. The increased adenosine monophosphate-to-adenosine triphosphate ratio promotes activation of adenosine monophosphate-activated protein kinase (AMPK), a crucial energy sensor that directs the cell toward energy-conserving programs, reduced lipid synthesis and modulation of glucose metabolism. AMPK activation, however, does not fully account for metformin’s mechanism. An important part of its glucose-lowering effect also appears to depend on AMPK-independent pathways, including inhibition of hepatic gluconeogenesis through redox changes, reduced availability of gluconeogenic substrates and interference with key enzymes of mitochondrial metabolism.

The liver remains the main clinically relevant target organ. Metformin reduces hepatic glucose synthesis from lactate, glycerol and gluconeogenic amino acids, attenuates the hepatic response to progluconeogenic signals and limits glucose production, especially during fasting. This lowers fasting glucose and reduces chronic tissue exposure to hyperglycemia. Secondary effects on hepatic lipid metabolism include a tendency to reduce de novo lipogenesis and, in some patients, improve metabolic dysfunction-associated steatotic liver disease, although metformin should not be considered a specific treatment for steatotic liver disease.

The role of the intestine has also become better understood in recent years. A substantial proportion of metformin’s action occurs in the gastrointestinal tract, where the medication reaches particularly high concentrations. The intestine contributes through changes in local glucose absorption and utilization, as well as effects on enteroendocrine cells, secretion of intestinal peptides such as glucagon-like peptide 1, bile-acid metabolism and the microbiota. This helps explain why modified-release formulations can retain efficacy despite different pharmacokinetic profiles and why some adverse effects are specifically gastrointestinal.

Peripherally, metformin may moderately improve insulin sensitivity, especially in skeletal muscle, but this effect should neither be overstated nor confused with that of true insulin sensitizers such as thiazolidinediones. Its most robust clinical action remains reduction of hepatic glucose output, accompanied by favorable modulation of the gut-liver axis and an overall metabolic effect that does not induce pharmacological hyperinsulinemia. This is a fundamental conceptual distinction: metformin corrects the system rather than merely forcing a secretory response.

This physiology also explains certain adverse effects and precautions. Because lactate is a substrate for gluconeogenesis, it is possible to understand why, when the medication accumulates or renal clearance is severely impaired, the risk of metformin-associated lactic acidosis may theoretically increase. This is not an ordinary toxic effect in a stable patient, but the consequence of a metabolic disturbance that arises primarily when hypoxia, kidney failure, hypoperfusion, sepsis, unstable heart failure or other conditions that favor metformin accumulation and increased lactate coexist. Understanding the mechanism is therefore not merely academic, but directly explains clinical caution in acute situations.

Metformin is absorbed in the intestine with incomplete bioavailability, and the gastrointestinal tract is highly relevant to both efficacy and tolerability. It does not bind significantly to plasma proteins, undergoes no clinically relevant hepatic metabolism and is excreted unchanged by the kidneys through glomerular filtration and tubular secretion. This feature has decisive implications: it reduces the risk of interactions mediated by hepatic metabolism, but makes kidney function the principal determinant of safety, accumulation and the need for therapeutic reassessment.

The absence of significant hepatic metabolism distinguishes metformin from many other agents. The liver is the main target organ, but not the organ that inactivates it. Hepatic impairment therefore affects not so much medication biotransformation as the clinical context in which it is used. In advanced liver disease, the main problem is not reduced metabolism, but the greater likelihood of tissue hypoxia, impaired lactate clearance, malnutrition, alcohol misuse or hemodynamic instability—all factors that increase overall metabolic risk.

The main formulations are immediate-release and modified-release. The standard formulation offers widespread availability, dosing flexibility and low cost, but is more often associated with nausea, diarrhea, abdominal pain, bloating and early intolerance. The modified-release formulation improves gastrointestinal tolerability in many patients by reducing peak local exposure in the proximal gastrointestinal tract and improving adherence. In some regulatory systems and recent guidelines, modified release is emphasized from treatment initiation or as the preferred option in patients with gastrointestinal vulnerability.

Plasma half-life does not fully describe the medication’s behavior because metformin tends to accumulate in tissue compartments, particularly the intestinal mucosa and red blood cells. Clinically, what matters most is that clearance depends on kidney function and dose adjustment must follow estimated glomerular filtration rate. Pharmacokinetics therefore explains why dosage cannot be prescribed once and for all, but must be readjusted over time in older patients, people with chronic kidney disease or those who develop intercurrent conditions that reduce renal perfusion.

Another important aspect concerns membrane transporters. Metformin uses systems such as organic cation transporters (OCTs) and renal efflux transporters, so some drug interactions depend not on cytochrome enzymes but on competition for tubular secretion. Agents that reduce renal metformin clearance or alter its transport may increase systemic exposure. This becomes clinically relevant especially in patients with polypharmacy, older adults and people with kidney disease.

Strictly pharmacodynamically, metformin produces a moderate average reduction in glycated hemoglobin, with a greater effect when baseline hyperglycemia is higher. Its action is stronger on fasting glucose than on isolated postprandial excursions, although its intestinal effect and reduction in overall glycemic load contribute to global improvement of the glucose profile. Its neutrality with respect to hypoglycemia stems from the fact that metformin is not an insulin secretagogue, which facilitates its use with many other classes. The medication is not, however, “risk-free” in an absolute sense, because clinical risk depends on kidney function, nutritional status, digestive tolerance and acute intercurrent conditions.

The classic indication for metformin is type 2 diabetes mellitus, particularly in the early stages and in people with insulin resistance, overweight or obesity. This does not mean that it is useful only in that phenotype, but this is the pathophysiological setting in which its biological rationale is strongest: excessive hepatic glucose production, persistent relative or absolute hyperinsulinemia, absence of ketosis and need for a medication that improves glycemic control without increasing body weight. Metformin therefore remains a frequent choice at the onset of type 2 diabetes mellitus, either as monotherapy or from the outset in combination when glucose levels are very high or targets are ambitious.

A second important area of use is prediabetes, or a high-risk state, in selected individuals. Pharmacological prevention does not replace lifestyle intervention, which remains the most powerful and broadly beneficial approach, but metformin has been shown to reduce diabetes incidence in high-risk populations, especially younger people, those with a high body mass index, higher fasting glucose, a history of gestational diabetes or other features of marked insulin resistance. In this setting, metformin should not be viewed as a shortcut, but as additional support when risk is high and nonpharmacological intervention alone is insufficient or difficult to sustain.

Metformin is also used in some nondiabetic settings, but considerable conceptual rigor is required. In polycystic ovary syndrome, it may have a role in subgroups with insulin resistance, metabolic abnormalities or specific reproductive indications, but it should not be described as universal therapy for the condition. In obesity without diabetes, its use is more selective and generally less effective than therapies now available for weight loss. In metabolic dysfunction-associated steatotic liver disease, it may improve the metabolic context but is not a specific antifibrotic treatment. Numerous research avenues exist in oncology, geriatrics, neurology and other fields, but these should not be turned into established indications without adequate clinical evidence.

In type 1 diabetes mellitus, metformin is not indicated as replacement therapy and does not correct the essential defect of insulin absence. It has been studied as an adjunct in specific subgroups with insulin resistance, excess weight or high insulin requirements, but benefits are generally limited and do not justify universal routine use. The distinction is important because it prevents prescribing drift based more on pathophysiological analogies than on robust clinical advantages.

Metformin occupies a more delicate position in gestational diabetes and preexisting diabetes during pregnancy. In some settings it may be used, sometimes alone and more often with insulin, but selection depends on the clinical picture, the guidelines followed, glycemic status, gestational age, patient preferences and considerations regarding placental transfer. This use therefore requires obstetric-diabetological contextualization and cannot be treated as an automatic extension of routine prescribing in type 2 diabetes mellitus.

In patients with chronic kidney disease, metformin is no longer regarded as indiscriminately contraindicated as it was in the past. The current approach is based on estimated glomerular filtration rate and clinical stability. This change has expanded the number of patients who can benefit, but has also made closer surveillance indispensable. The ideal phenotype is therefore defined not only by type 2 diabetes mellitus, but by the balance among metabolic benefit, renal safety, tolerability and the overall combination strategy.

Metformin lowers glycated hemoglobin by an average of approximately one percentage point, with variability depending on baseline levels, adherence, the dose achieved, kidney function and the combination-therapy setting. The reduction is often greater when the patient starts with substantial hyperglycemia and a profile dominated by increased hepatic glucose production. Its effect on fasting glucose tends to be more evident than its effect on isolated postprandial excursions, although improvement in the overall metabolic profile is also reflected in mean postprandial values.

With regard to weight, metformin is generally neutral or associated with modest weight loss, especially compared with medications that increase insulinemia or promote fat accumulation. This is clinically valuable in type 2 diabetes mellitus, in which body weight is not merely an anthropometric measurement but an integral part of the pathophysiology. A medication that controls blood glucose without promoting weight gain interrupts the vicious cycle in which more insulin and more weight worsen insulin resistance, subsequently requiring increasing treatment doses.

The risk of hypoglycemia is very low when metformin is used alone because it does not directly stimulate insulin secretion independently of blood glucose. This makes the medication particularly useful in selected older adults, workers exposed to occupational hazards, patients highly vulnerable to hypoglycemia and, more generally, every setting in which day-to-day safety is as important as lowering glycated hemoglobin. The profile naturally changes when metformin is combined with insulin or secretagogues, because the risk of hypoglycemia is then driven mainly by the other medications.

Historical studies also attributed favorable prognostic value to metformin for diabetes-related outcomes in newly diagnosed patients with overweight, helping establish its image as a medication with benefits beyond glucose lowering alone. In diabetes prevention, the Diabetes Prevention Program showed a reduction in disease incidence compared with placebo, although smaller than that achieved with intensive lifestyle intervention. This defined a precise field of use: metformin may prevent or delay diabetes in selected groups, but it does not replace behavioral modification as the primary strategy.

Metformin’s limitations must nevertheless be defined precisely. It is not the most potent medication for weight loss, does not provide the most specific cardiorenal benefits in high-risk populations and is insufficient by itself when beta-cell function is severely impaired or hyperglycemia is severe at presentation. In patients with very high glycated hemoglobin, catabolic symptoms, ketosis or marked metabolic decompensation, starting metformin alone is often inadequate. Similarly, in advanced cardiovascular or kidney disease, other medications may need to take priority for prognostic reasons even if metformin can be retained as metabolic support.

An inappropriate interpretation of possible “pleiotropic” effects reported in experimental or observational literature should also be avoided. The medication has interesting data regarding inflammation, lipid metabolism, the gut-liver axis, the microbiota and other biological pathways, but the clinical relevance of these effects varies greatly by context. In a rigorous therapeutic review, the correct conclusion is that metformin is highly valuable because it is effective, safe in many situations and metabolically coherent, but it remains one component of a broader therapeutic system rather than a universal solution to every metabolic problem.

Metformin should be started progressively. The general principle is to begin with low doses and increase gradually to improve gastrointestinal tolerability. A typical strategy is to start with 500 mg once daily during or immediately after the main meal and, after several days or one to two weeks, consider increasing to 500 mg twice daily or to a higher evening dose of the modified-release formulation. The speed of titration depends on age, digestive symptoms, nutritional status, patient frailty and glycemic goals.

The most common error at initiation is to aim immediately for the full dose. This increases the risk of nausea, diarrhea and early discontinuation. Metformin rewards gradual titration: a slower increase often produces a better result than rapid escalation. Patients should be told that some initial gastrointestinal discomfort may occur and that the medication should be taken with food. This simple measure significantly reduces early intolerance and improves adherence.

The maintenance dose varies according to efficacy and tolerability. In many patients, substantial glycemic benefit is achieved at intermediate doses, whereas higher doses may disproportionately increase adverse effects relative to metabolic gain. It is therefore incorrect to identify the “maximum dose” automatically with the “ideal dose.” The optimal dose is the one that provides the best balance among glycemic control, kidney function, digestive tolerability and regimen simplicity.

With the immediate-release formulation, dividing the dose into two or three daily administrations may reduce digestive symptoms and improve distribution of the effect. With modified release, once-daily evening or otherwise simplified administration often improves adherence, especially in patients receiving multiple medications. When the standard formulation is not tolerated, switching to modified release is one of the most useful strategies before declaring metformin definitively intolerable.

Dose adjustment must always take estimated glomerular filtration rate into account. With moderately reduced kidney function, metformin can often be continued at a lower dose with closer monitoring. When eGFR falls below recommended safety thresholds, the medication should not be started or should be discontinued if already in use. Isolated serum creatinine is insufficient because, in older adults or people with low muscle mass, it may underestimate actual impairment. The correct reference is estimated glomerular filtration rate and, even more importantly, its trend over time.

Correct practical use also requires the so-called sick-day rules, meaning rules for temporary discontinuation during acute illness. Patients should know that metformin may need to be stopped temporarily in the presence of vomiting, significant diarrhea, high fever with dehydration, reduced fluid intake, hypoxia, sepsis, shock, sudden worsening of kidney function or hospitalization for severe acute illness. This education is as essential as the prescription itself because much of the clinical risk arises not from routine use during stability, but from failure to adapt treatment during intercurrent events.

Metformin’s most frequent adverse effects are gastrointestinal. Nausea, diarrhea, abdominal pain or cramping, bloating, early satiety and sometimes a metallic taste are common, especially at treatment initiation or after overly rapid dose increases. These symptoms do not indicate systemic toxicity, but local and functional intolerance largely related to the medication’s intestinal action. In most cases, they can be reduced by taking metformin with meals, titrating slowly, temporarily reducing the dose or switching to a modified-release formulation.

One of the most important long-term safety issues is vitamin B12 deficiency. Chronic metformin use is associated with reduced cobalamin absorption, probably through interference at the ileal level and complex mechanisms involving calcium and intestinal physiology. Deficiency may remain subclinical for years or present with macrocytic anemia, peripheral neuropathy, worsening of preexisting neurological symptoms or cognitive changes in vulnerable individuals. This issue is particularly important in diabetes because diabetic neuropathy may mask or overlap with neuropathy caused by vitamin B12 deficiency, leading to missed diagnosis. Periodic vitamin B12 monitoring is clinically appropriate in patients receiving long-term treatment, particularly older adults, vegetarians, malnourished patients, people using proton-pump inhibitors or those with neurological signs.

The most feared event is metformin-associated lactic acidosis, but it must be described precisely to avoid both inappropriate alarm and dangerous trivialization. It is rare and generally promoted not by the mere presence of the medication, but by its accumulation in a patient with predisposing conditions such as substantial kidney failure, hypoperfusion, shock, sepsis, tissue hypoxia, unstable heart failure, severe respiratory failure, major alcohol misuse or advanced liver disease in a frail setting. Metformin does not automatically “cause” lactic acidosis in a stable patient; rather, it makes it more likely when lactate metabolism and renal clearance are profoundly impaired.

Contraindications and precautions should therefore be interpreted as tools for metabolic prevention. Severe kidney failure remains the main absolute contraindication. Acute conditions with a risk of rapid renal deterioration or hypoxia also require temporary discontinuation. Excess alcohol is relevant because it alters hepatic lactate metabolism and may amplify risk, especially with fasting or malnutrition. Advanced liver disease warrants caution not so much because of medication metabolism, but because of the patient’s overall metabolic fragility.

Management around iodinated contrast procedures is more selective than in the past. Metformin does not need to be stopped indiscriminately in every patient, but discontinuation should be considered when there is a risk of contrast-associated acute kidney injury or when kidney function is reduced according to current recommendations. The central point is simple: the problem is not a direct chemical interaction between metformin and contrast medium, but possible acute worsening of kidney function that could promote medication accumulation. The decision should therefore be tailored to estimated glomerular filtration rate, procedure type and clinical stability.

Risk management ultimately requires a nonbinary view. Metformin is neither a “harmless” medication that can be ignored after prescribing nor a dangerous medication to be used with excessive fear. It is very safe when prescribed to the right patient, at the right dose, with reassessment of kidney function, education regarding temporary discontinuation, vitamin B12 monitoring and attention to acute clinical settings. The quality of prescribing therefore depends not only on the initial choice, but on the ability to follow its trajectory over time.

Metformin interactions are driven not by hepatic metabolism, but by transport and kidney function. Medications that reduce renal perfusion, alter glomerular function or compete for tubular secretion may increase systemic metformin exposure and, in susceptible patients, increase the risk of accumulation. The interaction should therefore be interpreted not as an abstract laboratory phenomenon, but as a concrete change in clinical risk for the individual patient.

Iodinated contrast media are the best-known setting. As already explained, the issue is possible acute deterioration of kidney function after the procedure. In patients with normal kidney function and low risk of acute kidney injury, systematic discontinuation is not always necessary under more recent guidance. In patients with reduced filtration, instability or high-risk procedures, temporary interruption and reassessment of kidney function before restarting are prudent. This strategy avoids both unnecessary discontinuation and imprudent exposure.

Alcohol warrants separate consideration. Modest occasional consumption does not automatically constitute an absolute contraindication, but excessive acute or chronic intake increases the risk of altered lactate metabolism and adverse events, especially when combined with fasting, malnutrition or liver disease. Counseling should be concrete: it is not enough to say “avoid alcohol”; patients should understand that risk increases with binge drinking, periods of low caloric intake and dehydration.

Combinations with other antidiabetic medications also require careful interpretation. Metformin integrates very well with sodium-glucose cotransporter 2 inhibitors, glucagon-like peptide 1 receptor agonists and dipeptidyl peptidase-4 inhibitors because their mechanisms are complementary and hypoglycemic risk remains low unless the regimen is more complex. With insulin or sulfonylureas, however, the risk of hypoglycemia is driven mainly by those agents and may require adjustment of the entire regimen.

Medications that promote dehydration or hemodynamic deterioration may indirectly alter metformin safety. Intensive diuretic therapy, nephrotoxic medications, hypotension or treatments associated with reduced fluid intake warrant greater attention, especially in frail older adults. Similarly, intercurrent illnesses such as gastroenteritis, severe infections, major surgery and acute heart failure may transform a stable patient into one who is temporarily unsuitable for treatment.

Interactions should therefore not be reduced to a package-insert list. In clinical practice, they include every condition that increases the risk of accumulation, worsens digestive tolerance, alters fluid balance or makes lactate metabolism more vulnerable. Truly safe prescribing depends on anticipating these scenarios and teaching the patient when the medication should be continued, reduced or stopped.

Metformin often remains useful in older adults, but requires assessment beyond chronological age alone. The crucial point is to distinguish a robust older person from a frail, sarcopenic, malnourished or dehydration-prone patient with frequent acute exacerbations. In many functionally preserved older adults, metformin retains an excellent risk-benefit profile. In frail older adults, the main concerns become reduced renal reserve, risk of acute kidney injury, greater frequency of anorexia or gastrointestinal symptoms, and overlap between diabetic neuropathy and vitamin B12 deficiency. In these cases, the medication should not automatically be excluded, but should often be used at a more cautious dose with closer monitoring.

In patients with chronic kidney disease, metformin should be managed according to estimated glomerular filtration rate. Modern recommendations permit use at levels of renal impairment that would previously have led to discontinuation, but only within precise limits and with dose reduction when necessary. Severe reduction in filtration remains the contraindication threshold. Clinical stability is as important as the numerical value: a patient with a moderately reduced but stable eGFR is not equivalent to one with rapidly deteriorating kidney function or a high risk of acute kidney injury.

In patients with liver disease, the decision should be individualized. Simple hepatic steatosis or stable chronic liver disease without severe failure does not necessarily constitute an absolute contraindication. The problem arises when liver disease is accompanied by reduced lactate clearance, hypoxia, alcohol misuse, malnutrition or systemic instability. In other words, the metabolic and hemodynamic context accompanying liver disease matters more than “a diseased liver” itself.

During pregnancy, metformin’s role requires close integration between the diabetes and obstetric teams. The medication crosses the placenta and cannot simply be continued from prepregnancy therapy without reassessment. In some situations it is used in preexisting type 2 diabetes, gestational diabetes or specific syndromes, sometimes with insulin, but the choice depends on local guidance, patient profile and obstetric and metabolic goals. The correct formulation is therefore not simply “it is used” or “it is not used,” but “it is used in selected settings on specialist indication.”

In the perioperative period, metformin should be reassessed according to the type of procedure, expected fasting, hemodynamic risk and kidney function. Minor procedures in stable patients do not have the same significance as major surgery with a risk of hypotension, sepsis, hemorrhage or acute kidney injury. Here too, modern management is individualized rather than based on rigid, identical discontinuation rules for everyone, while maintaining caution in more unstable settings.

Finally, in people with prediabetes, polycystic ovary syndrome, obesity or other metabolic conditions, metformin should always be linked to a clear objective. It should not become the default medication for every mild metabolic disorder. Prescribing quality depends on selecting the setting in which the medication’s mechanism genuinely translates into a clinically relevant and sustainable advantage.

Metformin is one of the most easily combined medications in diabetology. The reason lies in its physiology: it acts mainly on the liver and intestine, does not independently cause hypoglycemia and does not increase body weight. This makes it a natural partner for medications acting on the kidneys, incretin pathways, beta cells or exogenous insulin. In clinical practice, it is often retained even when therapy is intensified, provided it is tolerated and compatible with kidney function.

In combination with sodium-glucose cotransporter 2 inhibitors, metformin creates a highly rational pair: the former reduces renal glucose reabsorption and provides selective cardiorenal benefits, while metformin reduces hepatic glucose output and contributes to overall metabolic control. Combination with glucagon-like peptide 1 receptor agonists is equally coherent because it adds glucose lowering, weight benefit and complementary actions on the gut-pancreas-liver axis. In these settings, metformin often serves as a therapeutic platform onto which medications with additional prognostic objectives are added.

Combination with dipeptidyl peptidase-4 inhibitors has a different, more specifically glycemic rationale. The main advantages are good tolerability, simplicity and a low risk of hypoglycemia, although effects on body weight and cardiorenal outcomes are more limited than with other classes. With sulfonylureas or meglitinides, metformin may improve glycemic control at the cost of greater hypoglycemic risk and, with sulfonylureas, a possible tendency toward weight gain. Although still used, these combinations have therefore lost prominence in contemporary algorithms.

With insulin, metformin retains an important role, especially in type 2 diabetes mellitus. Reducing hepatic glucose output and improving insulin efficiency often limits insulin requirements, attenuates weight gain and improves fasting glucose control. Its continuation in an insulin regimen should nevertheless be reassessed if significant digestive intolerance, progressive eGFR decline or increasing clinical frailty develops.

Metformin’s position in current algorithms is therefore dynamic. It is no longer correct to describe it invariably as the “mandatory first medication,” but neither is it correct to relegate it to a marginal secondary option. In patients without specific comorbidities, it often remains an excellent starting point. In those with priority cardiorenal or weight indications, it may be combined with or sometimes preceded by other classes, but is often continued. The correct decision depends on the hierarchy of objectives: glycemia, organ protection, weight, hypoglycemic risk, cost, tolerability and therapeutic complexity.

Deprescribing also has a role. In very old or frail patients with poor intake, weight loss, frequent acute illnesses or worsening kidney function, routine continuation of metformin out of habit alone may become irrational. As with every chronic medication, true expertise consists not only in knowing how to start it, but also when to reduce, replace or discontinue it.

Metformin monitoring does not end with measurement of glycated hemoglobin. Glycemic response remains central, but should be integrated with kidney function, digestive tolerability, weight trend, nutritional status and signs of possible vitamin B12 deficiency. Well-conducted follow-up distinguishes patients who obtain full benefit from the medication from those who continue taking it under progressively less favorable conditions.

Kidney function should be reassessed periodically in everyone, more frequently in older adults, patients with kidney disease, those with polypharmacy and people predisposed to acute kidney injury. Estimated glomerular filtration rate is the key parameter both for initiation and for maintaining or changing the dose. Any significant decline requires reassessment of the risk-benefit balance, especially when associated with dehydration episodes or recent hospitalizations.

Glycated hemoglobin estimates overall treatment efficacy, but does not tell the whole story. In patients with digestive symptoms, weight loss or irregular food intake, it is necessary to assess whether apparently better glycemic control actually results from reduced caloric intake. Similarly, in patients reporting weakness, paresthesias, worsening vibration sensation or anemia, monitoring should include vitamin B12 and, when necessary, further hematological and neurological investigations.

Vitamin B12 testing should not be treated as a bureaucratic formality. It is particularly important during prolonged treatment, in patients with diabetic neuropathy, older adults, vegetarians, people with unexplained anemia and those taking medications that reduce gastric and ileal absorption. Early identification permits targeted correction without necessarily stopping metformin, provided the rest of the clinical profile remains favorable.

Monitoring should also include continuing education. At every visit, it is useful to verify that the patient understands temporary-discontinuation rules for vomiting, significant diarrhea, fever with dehydration, contrast studies in high-risk settings or acute deterioration in general condition. A substantial part of medication safety depends not on sophisticated tests, but on the quality of the instructions received and remembered by the patient.

Long-term success with metformin is therefore measured in three domains: metabolic efficacy, sustainable tolerability and absence of hazardous accumulation. Only integrated monitoring of all three makes it possible to keep the medication central to therapy without turning a chronic benefit into an avoidable source of complications.

Metformin continues to occupy an exceptionally important position in the treatment of type 2 diabetes mellitus because it combines historical experience, genuine efficacy, low cost and pathophysiological coherence. In an era when diabetology has medications that are more selective for cardiovascular, renal and weight protection, its role has not disappeared but become more refined. Metformin’s contemporary value lies not in insisting that it must always be the first medication in a rigid sense, but in remaining an extremely useful therapeutic foundation when the patient’s profile permits.

Its deepest clinical significance may be this: metformin represents the transition from merely lowering a numerical glucose value to therapy that corrects a central metabolic abnormality with a low risk of hypoglycemia and without promoting weight gain. This explains why, despite the arrival of new classes, it remains a benchmark for almost every pharmacological strategy in type 2 diabetes mellitus.

Future perspectives concern less the discovery of a “miraculous new use” for metformin than a more precise understanding of the patients who derive the greatest benefit, the role of the intestine and microbiota, genetic determinants of response and optimal integration with incretin-based and cardiorenal therapies. The direction of research is not to replace an effective medication at all costs, but to position it more precisely within personalized treatment pathways.

In an individual patient, the prognosis of metformin treatment depends on the quality of initial selection and follow-up. In an appropriate person with adequate glomerular filtration, good digestive tolerance and correct monitoring, the medication can remain a stable, safe and useful component of therapy for years. In the wrong patient, especially one who is frail, prone to dehydration, malnourished or has severe kidney disease, even a historically reliable medication can become unsuitable. The final lesson is therefore methodological: metformin should be neither mythologized nor undervalued, but prescribed with clinical precision.