Metformin is widely used for treatment of type 2 diabetes mellitus. Side-effects are few and mainly from the gastrointestinal tract. Since metformin is cleared from the blood exclusively via the kidneys reduced renal function is a relative contraindication. We have earlier demonstrated that metformin safely can be used to a lower GFR level of 30 ml/min/1.73. Below that level the risk of lactacidosis, a severe complication, increases.

In the present study we plan to analyse serum levels of metformin repeatedly in patients with moderate renal failure (CKD = GFR of 30-60 ml/min/1.73). Blood samples will be taken as trough values in the morning, week 0, 2, 4, and 8 and at four weeks a blood sample will be taken two hours after intake of the morning dose of metformin. Renal function will be estimated with creatinine and cystatin C at each occasion. The intraindividual variation of metformin will be calculated.

The study rests on a new method for measuring metformin. The technique uses Liquid Chromatography Tandem Mass Spectometry (LCMSMS). Proteins are removed from serum by adding acetonitrile to the sample. After centrifugation a diluted portion of the supernatant is injected into the LCMSMS-system. The total runtime for a sample is 6 minutes.

The study will show if variation in serum levels of metformin measured in the same patient is high or low and thus give us better understanding whether a change i serum level is due to biological variation or to increased retention caused by progressive renal failure.

Inclusion Criteria:

• Patients above the age of 18 years with type 2 diabetes mellitus.
• Reduced renal function measured as estimated GFR (below 60 ml/min/1.73).
• Treatment with metformin on going.

Exclusion Criteria:

• Severely reduced GFR (< 30 ml(min/1.73)

Patients with type 2 DM followed at the out-patient department at the hospital will be asked to join the study