The diarrhoea on week one was brutal. Is the cheap 850 mg chip just not for me?
Cheap 850 mg, then the eGFR gate
Price is not the hard part. The hard part is not surprising the gut or missing the kidney number.
The US label offers two adult opens for immediate-release metformin: 500 mg twice daily or 850 mg once daily, both with meals. This site's SERP lock is the 850 mg chip, so that is the open I will keep on the blotter. Food cuts peak level and AUC - for a single 850 mg tablet, Cmax falls about 40 percent and AUC about 25 percent versus fasting - and you still take it with food, because the gut is the reason people quit. The clinical relevance of those PK dips is listed as unknown. The clinical relevance of diarrhoea on an empty stomach is not.
Before the first swallow, pull an eGFR. Below 30 the tablet is contraindicated. Between 30 and 45 you should not start. Above 45 the cheap climb can begin, with at least yearly kidney checks, more often in older people or anyone whose creatinine is already sliding. The labeled card for strengths and holds is the Glucophage 850 mg monograph.
The 850 mg ladder is every two weeks, not every two days
| Step | 850 mg-centred example | With |
|---|---|---|
| Open | 850 mg once daily | Evening meal |
| Week 3 | 850 mg twice daily | Breakfast and dinner |
| If needed | 850 mg three times daily | Toward the 2550 mg IR ceiling |
| XR rescue | Up to 2000 mg once daily | Evening meal, if the gut flagged |
| Hard stop | Do not start or continue | eGFR below 30 |
Label climb: raise by 500 mg weekly or by 850 mg every two weeks, steered by glucose and by what the stomach will tolerate, up to 2550 mg a day in divided doses. Doses above 2000 mg are often easier as three meals, not two. A typical settled zone is 1500 to 2000 mg. Stretching from there to the 2550 mg ceiling returns less glucose drop and more gut cost. There is no prize for arriving angry.
Extended-release is a different ladder. Usual ceiling 2000 mg a day, often once daily with the evening meal. It does not deliver more drug. It delivers a slower release that many guts will accept after the immediate-release chip failed. If someone truly needs the last 550 mg of the IR maximum, the ordinary tablet is the one that allows it. Most people never get that far. The evidence for why the climb is worth finishing is in the UKPDS event journal.
Diarrhoea is the common flag. XR is the usual rescue.
If the first fortnight is miserable, look at pace and formulation before you look at a different class.
Diarrhoea, nausea, cramping, bloating, and a metallic taste track the dose. They are worst when the open is too high or the climb is too fast. For most people they fade across a few weeks. A smaller group gets a lasting loose-stool tax at higher totals. Taking every swallow with food is not soft advice. It is the lever. Splitting a large total across three meals is the next lever.
Switching to extended-release once daily with dinner clears or sharply cuts the gut tax for a lot of people who were ready to quit. Try that before you abandon a cheap foundation that UKPDS actually moved events on. If even a reduced XR dose is intolerable, that is a fair exit. A formulation switch rescues more patients than a dramatic speech about adherence. The plant-to-pill file that explains why this dull tablet was worth saving is the goat's rue trace.
A metallic taste is annoying and not dangerous. Persistent watery stool after a slow climb is a formulation problem until proven otherwise. Blood in the stool, fever, or night pain is not 'just metformin' - that is a different flag. Do not let a cheap 850 mg chip become the explanation for every gut complaint. And do not let a fixable gut complaint become the reason a person never gets the event benefit the tablet actually earned.
EGFR is the gate. Everything else is commentary.
| eGFR (mL/min/1.73 m2) | Start? | If already on it |
|---|---|---|
| 60 or above | Yes | Yearly kidney check |
| 45 to 59 | Yes, with eyes open | At least yearly; sooner if sliding |
| 30 to 44 | Do not start | Reassess; often halve and watch |
| Below 30 | No | Stop |
Metformin is not metabolised. About 90 percent of absorbed drug leaves in the urine in the first 24 hours. Plasma half-life is about 6.2 hours. Whole-blood half-life is about 17.6 hours because red cells hold some of it. Renal clearance runs about 3.5 times creatinine clearance, which means tubular secretion does the heavy work. Fail the kidney and the tablet accumulates. Accumulation is the road to the rare lactic-acidosis crisis.
Label gates, not folklore. eGFR below 30: contraindicated, stop. eGFR 30 to 45: do not initiate; if a person already on the drug slips here, weigh benefit against risk, and many clinics halve the dose and watch more often. eGFR 45 and above: ordinary use with at least yearly checks. Obtain eGFR before the first chip. Recheck when the person ages, gets sick, or starts a drug that nicks the kidney. This is where careless prescribing does real harm.
Iodinated contrast is a hold, not a vibe
Intravascular iodinated dye can dip kidney function. If that dip happens while metformin is on board, the tablet can accumulate. The label is specific. Stop at the time of, or before, the scan if eGFR is 30 to 60, if there is a history of liver disease, heavy alcohol use, or heart failure, or if the dye will be intra-arterial. Recheck eGFR 48 hours later. Restart only if the number is stable.
A person with a normal eGFR and a routine venous scan often does not need a hold. 'Always stop for every CT' is a blunt rule some radiology desks still use. 'Never mention it' is the opposite miss. Raise the scan before it happens. The 48-hour restart is a kidney check, not a punishment.
The same hold logic applies when a person is properly sick - vomiting, unable to keep water down, septic, or in heart-failure decompensation. That is not a contrast rule. It is the same accumulation problem from the other side: kidneys or perfusion drop, the 850 mg chip stays in, lactate can rise. Sick-day stops are part of the cheap climb, not an optional footnote. Restart when the person is eating, drinking, and the eGFR has been seen again.
Years on the tablet can quietly lower B12
The long-term miss is not lactate. It is a vitamin you can measure.
Metformin interferes with B12 absorption. Across years, a meaningful fraction of people drift low. Low B12 can look like anaemia or like the same foot tingling diabetes already causes, which is why it gets blamed on 'just neuropathy' and left alone. Periodic B12 checks in long-term users - especially anyone with anaemia, numbness, or a vegan diet - are cheap. Replacement, sometimes by injection, is straightforward once you look.
This is not a reason to refuse a cheap 850 mg foundation. It is a reason to put B12 on the same list as the yearly eGFR. The history of why a gut-taxing, kidney-cleared tablet became that foundation is in the 1957 name file.
Bioavailability of a fasting 500 mg tablet is about 50 to 60 percent, and absorption does not scale linearly as the dose climbs toward 2550 mg - more milligrams, a smaller fraction absorbed. That is another reason the last slice of the IR ceiling buys less glucose drop than the first 1500 mg. Steady state arrives in a day or two. The gut tax arrives faster if you ignore food. Treat those PK facts as a pace argument, not as a reason to crush tablets or chase a peak.
Lactic acidosis is rare, serious, and mostly a hold problem
The black box is about metformin-associated lactic acidosis: rising lactate, falling pH, anion-gap trouble, high death rate when it happens. With metformin it is genuinely rare - far rarer than it was with phenformin, the cousin pulled in 1977. The situations that raise the risk are specific: failing kidneys, severe acute illness, dehydration, heavy alcohol, liver failure, or a sudden drop in oxygen or blood pressure. That is why the tablet is held when a person is properly sick, held around certain contrast, and stopped below eGFR 30.
For a stable person with a decent eGFR taking the 850 mg climb as labeled, the risk is very low. The rules exist to keep a rare event rare, not to decorate a carton. Anyone who turns very unwell - vomiting, rapid breathing, muscle pain - should stop the tablet and be assessed. Label text lives on DailyMed. Plain overviews sit at MedlinePlus. Neither is a substitute for the person who has actually examined you.
Reader mail
Reader questions on this article
Answered by Dr. Nathan Whitfield, MD · Internal medicine & clinical pharmacology
Readers wrote after the climb. These are the gut and kidney questions worth a full answer.
Often it is pace, not destiny. The label allows an 850 mg once-daily open with food, but if your gut is already loud I would rather see 500 mg with dinner for a week than a heroic 850 mg on an empty stomach. Climb by 500 mg weekly or 850 mg every two weeks, not faster. If the ordinary tablet still loosens you after a settled month, ask for extended-release once daily with the evening meal. That switch rescues a lot of people who were ready to quit. Symptoms also fade for many in the first fortnight. Quitting on day three is how a cheap foundation gets abandoned for the wrong reason.
They held my metformin before a CT with dye. Did I do something wrong?
No. The dye can briefly stress the kidneys, and metformin is cleared by the kidneys. If eGFR dips while the tablet is still on board, levels can rise, and that is the lactic-acidosis road. The label says stop at or before the scan if eGFR is 30 to 60, if you have liver disease, heavy alcohol use, or heart failure, or if the dye is intra-arterial. Restart about 48 hours later after a repeat eGFR looks stable. If your kidneys are fully normal and it was a routine venous scan, some desks still hold it out of habit. A hold is a precaution, not a verdict that the 850 mg chip failed you.
I have been on metformin for years and my feet buzz. Could the tablet be the cause?
It is worth a B12, because the tablet can lower that vitamin over years, and low B12 causes exactly this kind of tingling, sometimes with anaemia. Diabetes itself also damages nerves, so I would not pin it on metformin alone. The test is cheap. Replacement is straightforward. I check B12 in long-term users when numbness, anaemia, or a restricted diet shows up, and I do not wait for a crisis. If B12 is normal, we look at glucose history, other deficiencies, and whether the neuropathy pattern fits diabetes. Do not stop a working 850 mg foundation on a hunch before the blood is drawn.
My eGFR has been sliding. When does the cheap climb have to stop?
The hard floor is 30. Below that, stop. Between 30 and 45 we would not start a new person, and if you are already on it we reassess - often a lower total, closer labs, and a hard look at whether a different class should take more of the load. Above 45 the 850 mg climb is generally fine with at least yearly eGFR, sooner if you are older or the creatinine is moving. There is no single magic number for every body, but 30 is the contraindication and 45 is where extra care begins. Bring the actual eGFR to the visit, not a memory of 'a bit low last year'.
How scared should I be of the lactic-acidosis warning on the carton?
Respect it. Do not let it freeze you. The crisis is serious when it happens and rare when the holds are kept. Risk climbs with failing kidneys, a bad acute illness, dehydration, heavy drinking, liver failure, or a crash in oxygen or blood pressure. For a stable person with a reasonable eGFR on a labeled 850 mg climb, the risk is very low. Hold the tablet when you are properly sick. Hold it around the contrast situations above. Stop it below eGFR 30. If you ever turn very unwell with vomiting, fast breathing, or muscle pain, stop it and get checked. The box is a hold list, not a reason to refuse a cheap foundation.
Is extended-release 'better' than the ordinary 850 mg tablet?
Better for many guts. Not stronger on glucose. XR releases slower, which is why it is the rescue when immediate-release diarrhoea will not settle, and why it can be once daily with dinner. Its usual ceiling is 2000 mg, lower than the 2550 mg IR maximum. If someone truly needs that last slice of IR dose, the ordinary tablet is the one that allows 850 mg three times a day. Most people never need that. I pick XR when the gut flagged, not because a cart called it 'advanced'. Same molecule. Different release. Different ceiling.
Can I drink on metformin, or is that another hidden hold?
Light, occasional drinking with food is generally low-risk if your kidneys and liver are healthy. Heavy use and binges are a real problem. Alcohol interferes with lactate clearance, can stress the liver, and can drop glucose, so a binge on metformin stacks the exact risks the black box is about. The sensible line is moderation, not a theatrical ban on a single glass. If alcohol is a large part of your week, that is a clinic conversation, not a comment-thread guess. I would rather you tell the truth about intake than hide it and keep the 850 mg chip.
Why take it with food if food lowers the blood level of an 850 mg tablet?
Because the gut tax is the thing that actually stops the drug, and the PK dip has unknown clinical meaning on the label. A single 850 mg tablet with food shows a lower peak and a smaller AUC than the same tablet fasting. We still dose with meals. The glucose effect builds over days to weeks, not on the shape of one peak. An empty-stomach 850 mg that you vomit or abandon is a zero. A fed 850 mg you can climb is a foundation. Food is a tolerability tool here, not a bioavailability hack.
I saw 850 mg twice daily in a prevention study. Can I just start there?
That was DPP, a prevention trial in people who did not yet have diabetes, and even there the 850 mg twice-daily dose sat on a protocol, not on a dare. The labeled adult open for treatment is 500 mg twice daily or 850 mg once daily with meals, then a climb. Jumping to 1700 mg on day one is how you buy a gut surprise. If a clinic is using metformin off-label to delay diabetes, I still want the same slow ladder and the same eGFR gate. The trial dose is a target some people reach. It is not an opening move. The event context for that 850 mg BID is in the UKPDS and DPP journal.
General education from a clinician, not personal medical advice. Bring your own history to your own prescriber.
Last Updated