Everyone uses Metformin, but the mechanism of action still remains a bit shrouded in mystery.

Metformin decreases gluconeogenesis and glycogenolysis   J Clin Endocrinol Metab. 1996 Nov;81(11):4059-67    and to a lesser extent increases glucose uptake by skeletal muscles,  Ann Intern Med. 2002 Jul 2;137(1):25-33   actions which are attributable to the ability of Metformin to activate AMPK.   Metformin activates hepatic and muscle adenosine monophosphate-activated protein kinase (AMPK), an enzyme normally activated by adenosine monophosphate, a cellular signal for increased energy requirements. J Clin Invest. 2001 Oct;108(8):1167-74.    Fatty acid synthesis is blocked by activation of hepatic AMPK resulting in the phosphorylation and inhibition of acetyl-coenzyme A carboxylase, which catalyzes the rate-limiting step of lipogenesis. This in turn, promotes fatty acid oxidation. Activation of hepatic AMPK also decreases expression of sterol-regulatory-element-binding-protein-1 (SREBP-1), a transcription factor implicated in the pathogenesis of insulin resistance, dyslipidaemia and diabetes. A reduction in  SREBP-1 expression results in decreased gene expression of lipogenic enzymes, which leads to decreased triglyceride synthesis and hepatic steatosis.   Chronic chemical activation of AMPK has also been shown to increase GLUT-4 protein, and hexokinase activity which could contribute to Metformin's effects. J Appl Physiol. 1999 Nov;87(5):1990-5.  

 Although metformin clearly has no action on the pancreas directly, improvement of hyperglycaemia might alleviate glucotoxicity with improvement of secretory function accounting for improvements in post prandial glycaemia. Endocr Rev. 1992 Aug;13(3):415-31.   Evidence for an improvement of lipotoxicity by reduction of free fatty acid release from adipose tissue by metformin Diabetes. 1994 Jul;43(7):920-8.  is controversial.   N Engl J Med. 333:550–554 . Metformin reduces circulating triglyceride levels by 10-20%    J Clin Endocrinol Metab. 1992 May;74(5):1020-6.   through reduction in hepatic VLDL synthesis. Diabetes Care. 1990 Jan;13(1):1-8.  Absence of weight gain with metformin remains an alluring property and seems secondary to a reduction in calorie intake. Diabetologia. 1999 Apr;42(4):406-12.;   Obes Res. 1998 Jan;6(1):47-53.

Schematic Overview

 

Does adding sulphonylurea to Metformin increase mortality as shown in the UKPDS?

Patients treated with both Metformin and a sulphonylurea had a statistically significant 96% increase in risk of diabetes-related death compared with patients treated with a sulphonylurea alone.  Lancet. 1998 Sep 12;352(9131):854-65.   This combination Metformin-sulphonylurea group is very likely  aberrant.  A combined analysis of the main UKPDS and the supplementary study found a significant 19% reduction in risk for diabetes-related endpoints with Metformin. Improvement in fibrinolysis through reduction in Plasminogen activator inhibitor  Diabetes Care. 1993 Apr;16(4):621-9    along with reduction in platelet aggregability   Diabete Metab. 1989 Nov-Dec;15(6):420-5.  improves the thromobogenic profile of insulin resistance, and could explain the beneficial effects of metformin on the cardiovascular system.

 

Do obese patients  benefit from  Metformin over and above glycaemic control?

In UKPDS  there was a 16% (p=0.052) reduction in myocardial infarction and an 11% (p=0.52)increase in stroke, though the numbers  are very small.  Overweight patients were randomized to a Metformin group in addition to the two intensive therapy groups or the conventional therapy group; the Metformin group had  39% (p=0.01) reduction in myocardial infarction, and  41% (p=0.13) reduction in stroke.

Metformin has a favourable effect on lipids with reduction of total cholesterol, LDL cholesterol N Engl J Med. 1995 Aug 31;333(9):541-9.  and triglycerides. Diabetes Care. 1996 Jan;19(1):64-6.

Studies seem to suggest that Metformin therapy can decrease cancer mortality. Studies in hamsters  Gastroenterology. 2001 Apr;120(5):1263-70.  as well as diabetic humans BMJ. 2005 Jun 4;330(7503):1304-5;  have borne out this potential action.    Diabetes Care. 2006 Feb;29(2):254-8.

 

What is the mechanism of weight loss with Metformin?

Studies have shown weight gains of 2-5 kilograms with thiazolidinediones  Diabetes Technol Ther. 2004 Dec;6(6):850-63;  Diabetes Obes Metab. 2006 Jan;8(1):110-5. and  up to 8 kilograms with intensive insulin therapy. Diabetes Care. 1994 Jul;17(7):719-21. Despite improved glycaemia, metformin does not produce weight gain unlike sulphonylureas, thiazolidinediones or insulin.  This effect on weight seems secondary to a reduction in calorie intake. Diabetologia. 1999 Apr;42(4):406-12.;   Obes Res. 1998 Jan;6(1):47-53.

 

Should patients with impaired glucose tolerance be started on Metformin?

Yes! Currently metformin is not approved for this indication, but logic suggests that it should be used early on in  insulin resistance. While genetic factors are at work in producing progressive beta cell failure, the contribution of insulin resistance to this inexorable decline could be alleviated with metformin use early on.  The diabetes prevention programme has proved this concept to be true.  Diabetes Care. 1999 Apr;22(4):623-34.   In addition, IGT has been shown to pose a cardiovascular risk in its own right, even before overt diabetes is diagnosed.  Diabetes Care. 2003 Oct;26(10):2910-4    If one would consider metformin in IGT, then there is nothing to stop one from initiating metformin at diagnosis of type 2 diabetes. My own view is that the term diet-controlled diabetic should be abolished, as diet control merely improves glucose values without improving the underlying insulin resistance and hence the pancreatic work load remains at a higher level than ideal. Arguing for diet and lifestyle changes based on results of the DPP is all fair, but everyone is acutely aware that in the real world any degree of adherence to weight maintaining maneuvers is  poorly sustained. 

 

Should type I diabetics who develop central obesity be started on Metformin?

The cardiovascular risk in type I diabetic patients is thought to be partly contributed to by insulin resistance.  Arterioscler Thromb Vasc Biol. 1996 Jun;16(6):720-6.    Insulin sensitivity has been shown to be reduced in type I diabetics compared to BMI matched non-diabetic individuals.  J Clin Endocrinol Metab. 2002 Mar;87(3):1036-40.   No disease specific adverse effects are anticipated in the type I diabetic over and above that in Type 2 diabetics and general rules should still apply.   Treat Endocrinol. 2004;3(6):337-43.     Addition of metformin to insulin in type I diabetic patients seems to avoid some degree of weight gain, with a reduction in insulin requirements, although the latter effect was not sustained in the longer term. Vnitr Lek. 2001 Feb;47(2):81-6.   Increased insulin induced glucose uptake is proposed as a beneficial mechanism in these patients similar to type 2 diabetics.  Eur J Endocrinol. 2003 Oct;149(4):323-9.  I feel that increasing insulin requirements or rising fasting hyperglycaemia should be considered when planning to initiate  metformin in type I diabetics with weight gain.  Recent studies have estimated a prevalence of the metabolic syndrome (as defined by NCEP) of 38% in male and 40% in female type I diabetics in the Finnish population. Diabetes Care. 2005 Aug;28(8):2019-24.

 

At what Creatinine level should Metformin be stopped?

NICE guidelines recommend 130 mmol/L of Creatinine as a cut off for considering Metformin discontinuation. Note word Guidelines. That is all it should be. Definitely a magic number out of some hat.  In the individual patient, depending on whether withdrawal of Metformin would adversely influence short term and long term outcomes, I would tend to, and encourage my diabetes nurses to tend to, persevere with Metformin. A Creatinine above 200 mmol/L seems to be more reasonable to me, as slight fluctuations in the 120-170 range often settle or stay stable for a few more years. Given Metformin's cardiovascular benefits in the obese, and the fact that renal impairment patients are even more at risk for cardiovascular events, I feel guideline revision is warranted. Withdrawing Metformin and fighting a losing battle with rising sugars should give enough time for the osmotic diuresis and dehydration to worsen kidney failure in the short term!.  Where have all those EBM buffs gone when silly guidelines are put out?  More reasonable guidelines including Creatinine clearance based recommendations would be welcome.  Metformin and renal impairment: how much is safe?  Eur J Intern Med. 2002 Oct;13(7):428.

 

How likely is Metformin to produce lactic acidosis?

Very unlikely. Diabet Med. 1999 Apr;16(4):273-81      ; Can Med Assoc J. 1983 Jan 1;128(1):24-6.  Lactic acidosis occurs in 0.024-0.15 cases/1000 patient years in Metformin treated patients. (Swedish study). Compared to its predecessor phenformin, metformin is 20 times less likely to produce lactic acidosis. Metformin does not increase peripheral lactate production. Neither does it decrease lactate oxidation.    J Clin Endocrinol Metab. 1996 Nov;81(11):4059-67. Studies have shown that physicians (probably having recognised this low tendency for lactic acidosis) rarely stick to guidelines rigorously Arch Intern Med. 2002 Feb 25;162(4):434-7   and often patients with contraindications are on Metformin. Diabet Med. 2001 Jun;18(6):483-8.

So how often does this much hyped MALA (Metformin associated Lactic acidosis) occur in the studies using the drug?-

Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus.
 Cochrane Database Syst Rev. 2002;(2):CD002967.

 Risk of Fatal and Nonfatal Lactic Acidosis With Metformin Use in Type 2 Diabetes Mellitus Systematic Review and Meta-analysis
 Arch Intern Med. 2003;163:2594-2602.

 

Should Metformin be stopped prior to radiographic investigations involving contrast?

The Royal College of Radiologists published advice in November, 1996 (Advice to Members and Fellows with regard to Metformin-induced lactic acidosis and X-ray contrast medium agents, RCR Publication) supporting the manufacturers' advice that Metformin should not be used in the 48 h before or after intravenous (i.v.) contrast medium.

Non evidence based practice!!. This recommendation is based on case reports of lactic acidosis in patients with renal failure on Metformin.  Definitely there is no evidence that Metformin needs to be discontinued in diabetics with normal renal function. Eur Radiol. 1999;9(4):738-40.   If anything, the recommendations should be to stop Metformin for 48 hours after the contrast use; NOT BEFORE!.  If no renal failure occurs, there is no reason to stop Metformin. Clin Radiol. 1998 May;53(5):342-4.    Monitoring renal failure post contrast use should really suffice. I can only equate this ludicrous recommendation for stopping Metformin  by the royal college,  to the guidelines that recommend orlistat safe only for 2 years! At the same time, nobody has a problem using Rosuvastatin without any evidence of improved hard end points or  its safety!

A revised (?improved) version of the original RCR recommendations and an altered statement is made available here - ROYAL COLLEGE OF RADIOLOGISTS' GUIDELINES WITH REGARD TO METFORMIN-INDUCED LACTIC ACIDOSIS AND X-RAY CONTRAST MEDIUM AGENTS.  A more informative and reasonable recommendations from the Royal College of Australian Radiologists is available. 

 

How likely is Metformin to produce Megaloblastic Anemia?

Megaloblastic anemia due to Metformin has been found to occur in up to 9% of 600 type 2 diabetic patients assessed,  Aust Fam Physician. 2003 May;32(5):383-4  all of who promptly responded to B12 replacement.  Since the Metformin-induced B12 deficiency is mediated through a calcium-dependent ileal membrane antagonism, Diabetes Care. 2000 Sep;23(9):1227-31.    supplementation of calcium seems to be an effective preventative measure.

 

Is metformin therapy safe in pregnancy and breast feeding?

The use of Metformin in pregnant patients with polycystic ovarian disease seems to be accepted among the common clinician based on small studies although views remain varied  Curr Opin Obstet Gynecol. 2004 Jun;16(3):245-50  in the absence of randomised control trials around which the whole world revolves these days. While miscarriage and gestational diabetes seem to be more common in polycystic ovary syndrome, evidence that these can be improved with metformin is awaited from trials in the future. Good old personal experiences supported by peer group experiences should still find a strong place in decision making to use these drugs while awaiting further evidence through trials. (personal view)  Remembering that "Absence of Evidence is not Evidence of Absence", is crucial to avoid denying the benefits of drugs that are "more likely" to be beneficial than harmful based on current evidence, while waiting years for delayed evidence to come through at the mercy of the  whims and competitions of pharmaceutical organizations. Metformin use seems safe during lactation as well.   MJA 2004; 181 (3): 174-175  Realistic sane practical advice  from Australia is available at MJA 2004; 180 (9): 462-464

Read the Cochrane analysis of Metformin monotherapy for type 2 diabetes mellitus