Open at 500 mg, then climb on a clock
Rushing the dose is what makes people quit. The label already wrote a slower path. A 500 mg open is cheap because it is small, not because it is optional.
| Step | Labeled or ward move | Why it exists |
|---|---|---|
| IR open | 500 mg twice daily or 850 mg once daily, with meals | US Glucophage start |
| IR climb | +500 mg weekly or +850 mg every 2 weeks | Glucose and gut decide |
| IR ceiling | 2550 mg/day, divided | Above 2000 mg often tid with meals |
| XR open | 500 mg once daily with the evening meal | Same molecule, slower release |
| XR ceiling | 2000 mg once daily | Different cap than IR |
| Useful floor | About 1500 mg/day | Below this, many people see little glucose move |
Gut effects - nausea, cramping, diarrhoea, metallic taste - are dose-related and worst in the first weeks. The Glucophage label starts adults at 500 mg twice daily with meals, or 850 mg once daily with meals. Increase by 500 mg weekly, or by 850 mg every two weeks, on glucose and tolerability. Immediate-release ceiling is 2550 mg per day in divided doses. Doses above 2000 mg often sit better as three meals. Clinically useful responses are uncommon below about 1500 mg per day, which is why a 500 mg chip is a start, not the destination. People who quit at 500 mg because 'it did nothing' often never reached the useful floor.
A common ward climb that respects that label: 500 mg with dinner for a week if a twice-daily open feels like too much on a sensitive gut, then 500 mg twice daily, then toward 2000 mg split with meals. If a step flares, drop to the last comfortable dose and hold before trying again. That is titration, not failure. Extended-release opens at 500 mg once daily and tops out at 2000 mg once daily. Guesswork is skipping food, jumping to 2000 mg on day one, or treating a pharmacy XR swap as a new molecule. The molecule behind the climb is in the Glucophage reference.
Food is the hold, not a comfort tip
Take metformin with meals. That is not optional courtesy. Food blunts the gastrointestinal effects that drive most abandonment. An empty stomach is a reliable route to feeling sick. Immediate-release is two or three times daily; anchoring each dose to breakfast, lunch, or dinner both settles the gut and aids memory. If tablet count is the problem, that is exactly when once-daily extended-release earns its place. A person who 'cannot tolerate metformin' after three empty-stomach mornings has not failed the drug. They have failed the hold.
People skip food because they want the drug to 'work faster.' Metformin's glucose effect is not a same-hour rescue. It is a days-to-weeks quieting of hepatic glucose output. Starving the stomach to chase a faster A1c is a strike. Eat, then swallow. If the first weeks are still rough after that, the next lever is a slower climb or an XR switch, not a silent quit. Tell the prescriber before you throw the bottle away. Most of those bottles can be saved.
Immediate-release or extended-release is a gut decision
Immediate-release is the original: two or three times daily with meals, cheap, what most people start on. Extended-release drifts out slowly - once daily for many, gentler stomach for many. Same active drug. Different release speed. Practical rule on this desk: start IR unless the prescriber already knows the gut will rebel; switch to XR if GI effects dominate or once-daily dosing would keep the person on the drug. Formulation is a gut decision, not a potency upgrade.
One patient quirk that generates panicked mail: some XR shells appear whole in stool. The medicine is already gone. It is an empty matrix, not a failed pill. XR maximum is about 2000 mg once daily. Do not assume a pharmacy switch changed your total milligrams without reading the new label. Ask if the target dose moved. A silent swap that drops you from 2000 mg IR to 500 mg XR is a lost year of control, not a convenience win.
EGFR writes the start and the stop
The claim that metformin poisons kidneys is the one I strike first. Open the clearance line.
| eGFR (mL/min/1.73 m2) | Labeled action | Desk note |
|---|---|---|
| 60 and above | Standard dosing | Yearly eGFR at minimum |
| 45 to 59 | Continue; watch closer | Often recheck every 3 to 6 months |
| 30 to 44 | Do not start new; reassess if already on | Dose reduction often considered |
| Below 30 | Contraindicated - stop | Accumulation risk is the reason |
Metformin leaves through the kidneys largely unchanged. When filtration drops, the drug can accumulate. Accumulation is what raises the rare lactic-acidosis risk. That is why the label uses estimated glomerular filtration rate, not a folklore creatinine cutoff. Obtain eGFR before the first tablet. Do not use the drug when eGFR is below 30 mL/min/1.73 m2. Initiation is not recommended between 30 and 45. If a person already on therapy falls below 45, reassess benefit and risk; if they fall below 30, stop. Above 45, standard dosing is the usual file, with closer watch in the 45-59 band. The 2016 FDA safety update is what moved US practice from a hard creatinine line to this eGFR ladder. Older notes that say 'never if creatinine is above X' are stale.
Check at least yearly in everyone, more often if older or if other kidney risks sit on the chart. Older rules leaned on serum creatinine alone. eGFR is the better guide and the current labels use it. This is not paperwork theatre. It is the check that keeps a cheap 500 mg start as safe as the reputation claims. A falling eGFR is not proof the tablet attacked the kidney. It is a reason to change the dose or stop so the tablet cannot pile up. The FDA spelling of the thresholds is at the FDA.
Contrast dye, and the boxed lactic-acidosis line
Iodinated contrast for CT and angiography can transiently worsen kidney function. Metformin depends on kidneys to leave. The combination is where trouble can stack. Standard practice: hold metformin around contrast when kidney function is reduced; restart a couple of days later once eGFR is checked and stable. Normal kidneys get a lighter caution; holding remains a safe default when function is borderline. Tell the scan desk you take metformin when you book. Do not guess the hold from a forum. A radiology sheet that never asks about the tablet is an incomplete sheet.
The boxed warning is lactic acidosis - dangerous lactic acid build-up. Genuinely rare with metformin, far rarer than withdrawn phenformin, serious when it happens. Strategy: remove the precipitating situations - severe kidney impairment, dehydration, acute illness, contrast on shaky kidneys, heavy alcohol, significant liver disease. Warning signs that need urgent care: unusual muscle pain, trouble breathing, severe fatigue, feeling very cold. The point is not that the 500 mg chip is dangerous. A few specific circumstances raise risk enough to act. The cousin that taught the class this fear is in the plant-to-pill file. Fear without those circumstances is leftover 1977 noise.
B12 is the quiet long-haul check
Long-term metformin interferes with vitamin B12 absorption. Over years that can become a real deficiency. Low B12 causes anaemia and, more insidiously, numbness and tingling in the feet that mimics diabetic nerve damage. The two look alike. That is why the check slips. Periodic B12 on long use, especially with new neuropathy or anaemia, is the file. Supplement if low - usually without stopping the tablet. Stopping a first-line diabetes drug to 'fix tingling' without a B12 number is guesswork in the other direction.
People write new tingling off as 'just diabetes' for months. Ask for the B12 number. It is a blood test, not a riddle. Event context for why we bother keeping people on the drug at all sits in the UKPDS event review. The origin story does not change the lab order. A cheap 500 mg start that never gets a yearly eGFR or a later B12 is only half a file.