Association of Metformin Use With Risk of Lactic Acidosis Across the Range of Kidney Function: A Community-Based Cohort Study | Chronic Kidney Disease | JAMA Internal Medicine | JAMA Network
[Skip to Navigation]
Table 1.  Baseline Characteristics of 75 413 Patients With Diabetes Mellitus Stratified by Baseline Estimated Glomerular Filtration Rate (eGFR) Category in Geisinger Health Systema
Baseline Characteristics of 75 413 Patients With Diabetes Mellitus Stratified by Baseline Estimated Glomerular Filtration Rate (eGFR) Category in Geisinger Health Systema
Table 2.  Number of Prevalent and Incident Metformin Users in Geisinger Health System, Frequency of Estimated Glomerular Filtration Rate (eGFR) Measurements, and Mean Daily Metformin Dose by Baseline eGFR Category
Number of Prevalent and Incident Metformin Users in Geisinger Health System, Frequency of Estimated Glomerular Filtration Rate (eGFR) Measurements, and Mean Daily Metformin Dose by Baseline eGFR Category
Table 3.  Association of Time-Dependent Metformin Use With Acidosis Hospitalization by Time-Dependent Estimated Glomerular Filtration Rate (eGFR) Category in Geisinger Health System
Association of Time-Dependent Metformin Use With Acidosis Hospitalization by Time-Dependent Estimated Glomerular Filtration Rate (eGFR) Category in Geisinger Health System
Table 4.  Results of Active Comparator and Propensity-Matched Analyses in Geisinger Health System
Results of Active Comparator and Propensity-Matched Analyses in Geisinger Health System
Table 5.  Results of Active Comparator Analyses in the Replication Cohort (MarketScan)
Results of Active Comparator Analyses in the Replication Cohort (MarketScan)
1.
Thomas  MC, Cooper  ME, Zimmet  P.  Changing epidemiology of type 2 diabetes mellitus and associated chronic kidney disease.  Nat Rev Nephrol. 2016;12(2):73-81.PubMedGoogle ScholarCrossref
2.
Inzucchi  SE, Bergenstal  RM, Buse  JB,  et al.  Management of hyperglycemia in type 2 diabetes, 2015: a patient-centered approach: update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.  Diabetes Care. 2015;38(1):140-149.PubMedGoogle ScholarCrossref
3.
Hong  J, Zhang  Y, Lai  S,  et al; SPREAD-DIMCAD Investigators.  Effects of metformin versus glipizide on cardiovascular outcomes in patients with type 2 diabetes and coronary artery disease.  Diabetes Care. 2013;36(5):1304-1311.PubMedGoogle ScholarCrossref
4.
Kooy  A, de Jager  J, Lehert  P,  et al.  Long-term effects of metformin on metabolism and microvascular and macrovascular disease in patients with type 2 diabetes mellitus.  Arch Intern Med. 2009;169(6):616-625.PubMedGoogle ScholarCrossref
5.
Maruthur  NM, Tseng  E, Hutfless  S,  et al.  Diabetes medications as monotherapy or metformin-based combination therapy for type 2 diabetes: a systematic review and meta-analysis.  Ann Intern Med. 2016;164(11):740-751.PubMedGoogle ScholarCrossref
6.
Flory  JH, Hennessy  S.  Metformin use reduction in mild to moderate renal impairment: possible inappropriate curbing of use based on food and drug administration contraindications.  JAMA Intern Med. 2015;175(3):458-459.PubMedGoogle ScholarCrossref
7.
Conlay  LA, Loewenstein  JE.  Phenformin and lactic acidosis.  JAMA. 1976;235(15):1575-1578.PubMedGoogle ScholarCrossref
8.
Kwong  SC, Brubacher  J.  Phenformin and lactic acidosis: a case report and review.  J Emerg Med. 1998;16(6):881-886.PubMedGoogle ScholarCrossref
9.
Altun  E, Kaya  B, Paydaş  S, Sarıakçalı  B, Karayaylalı  I.  Lactic acidosis induced by metformin in a chronic hemodialysis patient with diabetes mellitus type 2.  Hemodial Int. 2014;18(2):529-531.PubMedGoogle ScholarCrossref
10.
Duong  JK, Furlong  TJ, Roberts  DM,  et al.  the role of metformin in metformin-associated lactic acidosis (MALA): case series and formulation of a model of pathogenesis.  Drug Saf. 2013;36(9):733-746.PubMedGoogle ScholarCrossref
11.
Fitzgerald  E, Mathieu  S, Ball  A.  Metformin associated lactic acidosis.  BMJ. 2009;339:b3660.PubMedGoogle ScholarCrossref
12.
Perrone  J, Phillips  C, Gaieski  D.  Occult metformin toxicity in three patients with profound lactic acidosis.  J Emerg Med. 2011;40(3):271-275.PubMedGoogle ScholarCrossref
13.
Silvestre  J, Carvalho  S, Mendes  V,  et al.  Metformin-induced lactic acidosis: a case series.  J Med Case Rep. 2007;1:126.PubMedGoogle ScholarCrossref
14.
Wiholm  BE, Myrhed  M.  Metformin-associated lactic acidosis in Sweden 1977-1991.  Eur J Clin Pharmacol. 1993;44(6):589-591.PubMedGoogle ScholarCrossref
15.
Yeh  HC, Ting  IW, Tsai  CW, Wu  JY, Kuo  CC.  Serum lactate level and mortality in metformin-associated lactic acidosis requiring renal replacement therapy: a systematic review of case reports and case series.  BMC Nephrol. 2017;18(1):229.PubMedGoogle ScholarCrossref
16.
Inzucchi  SE, Lipska  KJ, Mayo  H, Bailey  CJ, McGuire  DK.  Metformin in patients with type 2 diabetes and kidney disease: a systematic review.  JAMA. 2014;312(24):2668-2675.PubMedGoogle ScholarCrossref
17.
Department of Health and Human Services. Metformin Petition Closeout Memorandum from CDER Director. Silver Spring, Maryland: Food and Drug Administration; 2016.
18.
Kidney Disease Improving Global Outcomes (KDIGO) Work Group.  KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease.  Kidney Int Suppl. 2013;3(1):1-150.Google ScholarCrossref
19.
National Institute for Health and Care Excellence (NICE). Type 2 diabetes in adults: management. Updated May 2017. https://www.nice.org.uk/guidance/ng28. Accessed October 15, 2017.
20.
American Diabetes Association.  Standards of Medical Care in Diabetes - 2016.  Diabetes Care. 2016;39(1):S1-S112.PubMedGoogle Scholar
21.
Lipska  KJ, Bailey  CJ, Inzucchi  SE.  Use of metformin in the setting of mild-to-moderate renal insufficiency.  Diabetes Care. 2011;34(6):1431-1437.PubMedGoogle ScholarCrossref
22.
Bakris  GL, Molitch  ME.  Should restrictions be relaxed for metformin use in chronic kidney disease? yes, they should be relaxed! what’s the fuss?  Diabetes Care. 2016;39(7):1287-1291.PubMedGoogle ScholarCrossref
23.
Kalantar-Zadeh  K, Kovesdy  CP.  Should restrictions be relaxed for metformin use in chronic kidney disease? no, we should never again compromise safety!  Diabetes Care. 2016;39(7):1281-1286.PubMedGoogle ScholarCrossref
24.
Bodmer  M, Meier  C, Krähenbühl  S, Jick  SS, Meier  CR.  Metformin, sulfonylureas, or other antidiabetes drugs and the risk of lactic acidosis or hypoglycemia: a nested case-control analysis.  Diabetes Care. 2008;31(11):2086-2091.PubMedGoogle ScholarCrossref
25.
Cucchiari  D, Podestà  MA, Merizzoli  E,  et al.  Dose-related effects of metformin on acid-base balance and renal function in patients with diabetes who develop acute renal failure: a cross-sectional study.  Acta Diabetol. 2016;53(4):551-558.PubMedGoogle ScholarCrossref
26.
Ekström  N, Schiöler  L, Svensson  AM,  et al.  Effectiveness and safety of metformin in 51 675 patients with type 2 diabetes and different levels of renal function: a cohort study from the Swedish National Diabetes Register.  BMJ Open. 2012;2(4):e001076.PubMedGoogle ScholarCrossref
27.
Eppenga  WL, Lalmohamed  A, Geerts  AF,  et al.  Risk of lactic acidosis or elevated lactate concentrations in metformin users with renal impairment: a population-based cohort study.  Diabetes Care. 2014;37(8):2218-2224.PubMedGoogle ScholarCrossref
28.
Kuo  CC, Yeh  HC, Chen  B, Tsai  CW, Lin  YS, Huang  CC.  Prevalence of metformin use and the associated risk of metabolic acidosis in US diabetic adults with CKD: a national cross-sectional study.  Medicine (Baltimore). 2015;94(51):e2175.PubMedGoogle ScholarCrossref
29.
Richy  FF, Sabidó-Espin  M, Guedes  S, Corvino  FA, Gottwald-Hostalek  U.  Incidence of lactic acidosis in patients with type 2 diabetes with and without renal impairment treated with metformin: a retrospective cohort study.  Diabetes Care. 2014;37(8):2291-2295.PubMedGoogle ScholarCrossref
30.
Salpeter  SR, Greyber  E, Pasternak  GA, Salpeter  EE.  Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus.  Cochrane Database Syst Rev. 2010;(4):CD002967.PubMedGoogle Scholar
31.
Stang  M, Wysowski  DK, Butler-Jones  D.  Incidence of lactic acidosis in metformin users.  Diabetes Care. 1999;22(6):925-927.PubMedGoogle ScholarCrossref
32.
National Committee for Quality Assurance. HEDIS and Quality Measurement: HEDIS Measures. 2017. http://www.ncqa.org/hedis-quality-measurement/hedis-measures. Accessed October 15, 2017.
33.
Levey  AS, Stevens  LA, Schmid  CH,  et al; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration).  A new equation to estimate glomerular filtration rate.  Ann Intern Med. 2009;150(9):604-612.PubMedGoogle ScholarCrossref
34.
Pongwecharak  J, Tengmeesri  N, Malanusorn  N, Panthong  M, Pawangkapin  N.  Prescribing metformin in type 2 diabetes with a contraindication: prevalence and outcome.  Pharm World Sci. 2009;31(4):481-486.PubMedGoogle ScholarCrossref
35.
Vasisht  KP, Chen  SC, Peng  Y, Bakris  GL.  Limitations of metformin use in patients with kidney disease: are they warranted?  Diabetes Obes Metab. 2010;12(12):1079-1083.PubMedGoogle ScholarCrossref
36.
Rachmani  R, Slavachevski  I, Levi  Z, Zadok  B, Kedar  Y, Ravid  M.  Metformin in patients with type 2 diabetes mellitus: reconsideration of traditional contraindications.  Eur J Intern Med. 2002;13(7):428.PubMedGoogle ScholarCrossref
37.
Marcum  ZA, Forsberg  CW, Moore  KP,  et al.  Mortality associated with metformin versus sulfonylurea initiation: a cohort study of veterans with diabetes and chronic kidney disease.  J Gen Intern Med. 2018;33(2):155-165.PubMedGoogle ScholarCrossref
38.
Holman  RR, Paul  SK, Bethel  MA, Matthews  DR, Neil  HA.  10-year follow-up of intensive glucose control in type 2 diabetes.  N Engl J Med. 2008;359(15):1577-1589.PubMedGoogle ScholarCrossref
39.
Roussel  R, Travert  F, Pasquet  B,  et al; Reduction of Atherothrombosis for Continued Health (REACH) Registry Investigators.  Metformin use and mortality among patients with diabetes and atherothrombosis.  Arch Intern Med. 2010;170(21):1892-1899.PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    July 2018

    Association of Metformin Use With Risk of Lactic Acidosis Across the Range of Kidney Function: A Community-Based Cohort Study

    Author Affiliations
    • 1Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland
    • 2Division of Nephrology, Monash Medical Centre, Clayton, Australia
    • 3Department of Medicine, Johns Hopkins University, Baltimore, Maryland
    • 4Division of Nephrology, Tufts Medical Center, Boston, Massachusetts
    • 5Kidney Health Research Institute, Geisinger Health System, Danville, Pennsylvania
    JAMA Intern Med. 2018;178(7):903-910. doi:10.1001/jamainternmed.2018.0292
    Key Points

    Question  Does metformin use increase the risk of acidosis in patients with chronic kidney disease?

    Findings  In 2 large retrospective cohorts of patients with diabetes mellitus (Geisinger Health System, n = 75 413 and MarketScan, n = 82 017), metformin use was not significantly associated with incident acidosis at estimated glomerular filtration rate (eGFR) greater than 30 mL/min/1.73 m2.

    Meaning  Metformin therapy may be safe in patients with type 2 diabetes and eGFR 30 to 60 mL/min/1.73 m2.

    Abstract

    Importance  Approximately 1 million patients in the United States with type 2 diabetes mellitus and mild-to-moderate kidney disease do not receive guideline-directed therapy with metformin. This may reflect uncertainty regarding the risk of acidosis in patients with chronic kidney disease.

    Objective  To quantify the association between metformin use and hospitalization with acidosis across the range of estimated glomerular filtration rate (eGFR), accounting for change in eGFR stage over time.

    Design, Setting, and Participants  Community-based cohort of 75 413 patients with diabetes in Geisinger Health System, with time-dependent assessment of eGFR stage from January 2004 until January 2017. Results were replicated in 67 578 new metformin users and 14 439 new sulfonylurea users from 2010 to 2015, sourced from 350 private US health systems.

    Exposures  Metformin use.

    Main Outcomes and Measures  Hospitalization with acidosis (International Classification of Diseases, Ninth Revision, Clinical Modification code of 276.2).

    Results  In the primary cohort (n = 75 413), mean (SD) patient age was 60.4 (15.5) years, and 51% (n = 38 480) of the participants were female. There were 2335 hospitalizations with acidosis over a median follow-up of 5.7 years (interquartile range, 2.5-9.9 years). Compared with alternative diabetes management, time-dependent metformin use was not associated with incident acidosis overall (adjusted hazard ratio [HR], 0.98; 95% CI, 0.89-1.08) or in patients with eGFR 45 to 59 mL/min/1.73 m2 (adjusted HR, 1.16; 95% CI, 0.95-1.41) and eGFR 30 to 44 mL/min/1.73 m2 (adjusted HR, 1.09; 95% CI, 0.83-1.44). On the other hand, metformin use was associated with an increased risk of acidosis at eGFR less than 30 mL/min/1.73 m2 (adjusted HR, 2.07; 95% CI, 1.33-3.22). Results were consistent when new metformin users were compared with new sulfonylurea users (adjusted HR for eGFR 30-44 mL/min/1.73 m2, 0.77; 95% CI, 0.29-2.05), in a propensity-matched cohort (adjusted HR for eGFR 30-44 mL/min/1.73 m2, 0.71; 95% CI, 0.45-1.12), when baseline insulin users were excluded (adjusted HR for eGFR 30-44 mL/min/1.73 m2, 1.16; 95% CI, 0.87-1.57), and in the replication cohort (adjusted HR for eGFR 30-44 mL/min/1.73 m2, 0.86; 95% CI, 0.37-2.01).

    Conclusions and Relevance  In 2 real-world clinical settings, metformin use was associated with acidosis only at eGFR less than 30 mL/min/1.73 m2. Our results support cautious use of metformin in patients with type 2 diabetes and eGFR of at least 30 mL/min/1.73 m2.

    Introduction

    More than 380 million people worldwide are affected by type 2 diabetes mellitus (DM) and approximately 20% have an estimated glomerular filtration rate (eGFR) less than 60 mL/min/1.73 m2.1 Metformin hydrochloride is recommended as the first-line medication for type 2 DM because of its low cost, favorable adverse effect profile, and a possible beneficial effect on cardiovascular risk.2-5 However, it is frequently avoided in patients with chronic kidney disease (CKD) because of the concern of drug accumulation and lactic acidosis,6 which arose following the withdrawal of phenformin from the US market in 1978,7,8 and case reports of metformin-associated lactic acidosis in patients with CKD.9-15

    Regulatory and professional society guidelines suggest that metformin may be an option in patients with mild to moderate CKD.16 The US Food and Drug Administration (FDA) recently changed the metformin label from contraindicated at serum creatinine level greater than 1.5 mg/dL in men and greater than 1.4 mg/dL in women (to convert to micromoles per liter, multiply by 88.4) to contraindicated at eGFR less than 30 mL/min/1.73 m2, and not recommended to initiate metformin at eGFR less than 45 mL/min/1.73 m2.17 Other guidelines cautiously support the use of metformin at eGFR 30 to 60 mL/min/1.73 m2, recommending that metformin be reviewed18,19 at eGFR 30 to 45 mL/min/1.73 m2 and that dose adjustment be considered.20,21

    Despite recent changes to metformin labeling, data addressing the safety of metformin in patients with eGFR less than 60 mL/min/1.73 m2 are inconclusive.22,23 Large retrospective studies have reported conflicting results regarding the association between metformin therapy and acidosis at eGFR less than 60 mL/min/1.73 m2, and have not addressed individual variation in eGFR over time, or possible confounding by concomitant insulin use.24-31 Although a systematic review of 347 trials and cohort studies found no evidence that metformin therapy increased the risk of acidosis compared with alternative antihyperglycemic therapies, 191 of the studies specifically excluded patients with CKD, and few of the remaining reported subgroup analyses for the participants with lower eGFR.24,30

    We aimed to investigate the relationship between metformin therapy and acidosis across the full spectrum of eGFR in a large, integrated electronic medical record cohort, accounting for time-dependent eGFR stage, and for potential confounding from multiple variables, including concomitant insulin use. We sought to replicate findings in a separate nationwide cohort derived from 350 private health systems. In both cohorts, we compared acidosis risk during metformin use with the risk during alternative management of DM, hypothesizing that acidosis would be no more common among metformin users within categories of eGFR.

    Methods
    Study Population and Design

    We studied a community-based cohort of patients with a diagnosis of DM and a postdiagnosis serum creatinine measurement between January 1, 2004, and January 20, 2017, receiving primary care in Geisinger Health System. Deidentified individual patient data from inpatient and outpatient encounters, including problem lists, prescriptions, diagnostic codes, and laboratory measurements, were used. We classified DM using the Health Plan Employer Data and Information Set (HEDIS) criteria, defined as 1 inpatient or 2 outpatient diagnostic codes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]) of 250.x, 357.2, 362.0, and 366.41, or a prescription for a medication to treat DM other than metformin monotherapy.32 The baseline date for the primary analysis was the first serum creatinine measurement on or after DM diagnosis and January 1, 2004. We excluded patients with missing serum creatinine level (n = 2232), end-stage renal disease status at baseline (n = 399), or initial eGFR less than 15 mL/min/1.73 m2 (n = 711), for a total study population of 75 413. We censored patient time at death or at the end of study follow-up (February 2, 2017).

    Outcome, Exposure, and Covariate Definitions

    The primary outcome was a hospitalized acidosis event, defined as an inpatient ICD-9-CM code of 276.2.29,31 This diagnostic code broadly encapsulates acidosis but specifically excludes diabetic ketoacidosis. We also evaluated the occurrence of ICD-9-CM 276.2 as a primary code.

    Metformin use and daily dose were ascertained from electronic prescription records. Start time was defined as the date of electronic prescription, and stop time was the end date of the prescription, or the date that a clinician discontinued the medication, whichever came earlier. A gap between prescriptions of less than 60 days was not considered a medication discontinuation to allow for the possibility of stockpiling of medications.

    The eGFR was estimated from serum creatinine level using the CKD-Epidemiology equation and staged according to the Kidney Disease Improving Global Outcomes guidelines (<30, 30-44, 45-59, 60-89, and ≥90 mL/min/1.73 m2).33 In analyses using time-dependent eGFR stage, all outpatient serum creatinine levels were assessed. Each time a participant changed eGFR stage, the variable was updated and subsequent follow-up time was categorized accordingly.

    Other covariates included age, sex, race, smoking status, body mass index (BMI, calculated as weight in kilograms divided by height in meters squared), serum bicarbonate, hemoglobin A1c (HbA1c), comorbid cardiovascular disease, heart failure, hypertension, and medication use. We categorized patients as ever or never cigarette smokers, and determined comorbid cardiovascular disease (coronary artery disease, stroke, or peripheral arterial disease), heart failure, and hypertension by the presence of relevant diagnostic codes at any time prior to cohort inclusion (eTable 1 in the Supplement). We abstracted serum bicarbonate and HbA1c from outpatient laboratory data within 1 year prior to index date. We recorded baseline and time-dependent use of statins, renin-angiotensin system inhibitors, diuretics, nonsteroidal anti-inflammatory drugs, insulin, sulfonylureas (eg, glimepiride, glyburide, glipizide), and other hypoglycemic medications in a manner similar to metformin use. We determined deaths by linkage to the National Death Index.

    Statistical Analysis

    We compared baseline characteristics of patients according to baseline eGFR stage. Differences across categories were determined using linear regression for continuous variables and logistic regression for binary variables on the median eGFR value within each eGFR category.

    We analyzed risk of acidosis in metformin use, compared with no metformin use, using Cox proportional hazards regression. We used a model in which metformin use and eGFR category were captured as time-dependent variables. In other words, values were allowed to change over time, with metformin use initiated or discontinued with prescription start and stop time, and eGFR category updated when outpatient eGFR changed. Models were run unadjusted, demographic-adjusted, and fully adjusted for baseline covariates (demographic characteristics, eGFR, serum bicarbonate level, smoking status, BMI, cardiovascular disease, heart failure, hypertension, and medication use). A linear spline was used for bicarbonate level, with a knot at the mean (27.5 mEq/L [1:1 conversion to millimoles per liter]), to address the nonlinear association between bicarbonate and acidosis. We tested for modification of the association between metformin and acidosis by eGFR stage by including a product term with eGFR stage modeled as an ordinal variable. In secondary analysis, we also captured the use of other medications as time-dependent variables.

    Sensitivity Analyses

    First, we performed an active comparator study, in which new (first-time) metformin users were compared with new sulfonylurea users without accounting for time-dependent eGFR. Second, we constructed propensity-matched cohorts for prevalent metformin users within each category of eGFR, censoring at metformin discontinuation for cases, metformin initiation for controls, or eGFR category change. The propensity score was created using logistic regression of metformin status on the aforementioned covariates. Cases (metformin users) and controls (non–metformin users) were matched using nearest-neighbor matching on a 1:1 basis within each category of eGFR using a caliper of one-fifth the standard error (0.044). Third, we tested associations after excluding the 12 971 baseline insulin users. Fourth, we included baseline HbA1c in the fully adjusted model. Fifth, we replicated the analysis in persons with a new diagnosis of DM after January 1, 2004. Finally, to address the possibility of incomplete capture of metformin discontinuation and thus overestimation of metformin exposure during which no events occurred, we censored all metformin users without acidosis events 90 days earlier.

    Replication in Commercial Claims Database

    We replicated active comparator results in the MarketScan database, an individual-level inpatient and outpatient claims data source from 350 private health systems, linked to outpatient laboratory test results for some patients. We included adults with DM defined by the HEDIS criteria and at least 1 creatinine level assessed on or after diagnosis. Medication use was sourced from pharmacy dispensing claims. Covariates, laboratory values, and acidosis outcomes were determined from diagnostic codes in a parallel manner to Geisinger Health System. Time at risk began at prescription for either metformin or sulfonylurea after January 1, 2010, and ended at hospitalization with acidosis, use of the opposite medication class, cessation of medication use, death, or end of study follow-up on December 31, 2014, whichever came first. Cessation of medication use was recorded as the date of last prescription, plus days of supply. Baseline eGFR was defined as the closest serum eGFR measurement within 1 year prior to first medication use. We compared risk of acidosis in metformin users with sulfonylurea users using Cox proportional hazards regression, adjusted for baseline age, sex, eGFR, cardiovascular disease, heart failure, hypertension, and use of insulin, renin-angiotensin system inhibitors, diuretics, and nonsteroidal anti-inflammatory drugs. All calculations were performed using statistical software (Stata, version 13.1/14.2; StataCorp LP).

    Patient Involvement

    Patients were not involved in the design or implementation of the study. The study was reviewed and deemed exempt by the Geisinger Health System and Johns Hopkins University Institutional Review Boards.

    Results
    Study Population

    There were 75 413 individuals included in the Geisinger study population. Mean (SD) age at baseline was 60.4 (15.5) years, 51% were female, and mean (SD) BMI was 34.1 (8.1) (Table 1). At cohort enrollment, 14 662 patients had an eGFR less than 60 mL/min/1.73 m2 and 1765 had an eGFR less than 30 mL/min/1.73 m2. Median follow-up duration was 5.7 years (interquartile range [IQR], 2.5-9.9 years). The median number of creatinine measurements per year was 2.1 (IQR, 1.1-3.4), which increased with lower eGFR (Table 2).

    Time-Dependent Metformin Use and Acidosis Events

    Forty-five percent of patients were taking metformin at enrollment in the study (n = 34 095), and 13 781 of the remaining patients were subsequently prescribed metformin during follow-up (Table 2). The median duration of metformin use was 2.8 years (IQR, 0.9-6.2 years).

    Overall, there were 2335 hospitalizations with acidosis over 470 114 person-years of follow-up: 737 events occurred over 188 578 person-years of metformin use and 1598 events occurred over 281 536 person-years of no metformin use. Of these events, only 29 had an acidosis diagnostic code in the primary position.

    Metformin Use and Risk of Acidosis by Time-Dependent eGFR Category

    The adjusted hazard ratio (HR) of acidosis during metformin use compared with nonuse was 0.98 (95% CI, 0.89-1.08) (Table 3). There was a higher acidosis risk associated with metformin at lower GFR (P = .01 for interaction). However, the risk associated with metformin use was not statistically significant at eGFR greater than 90 mL/min/1.73 m2 (adjusted HR, 0.88; 95% CI, 0.73-1.05), eGFR 60 to 89 mL/min/1.73 m2 (adjusted HR, 0.87; 95% CI, 0.75-1.02), eGFR 45 to 59 mL/min/1.73 m2 (adjusted HR, 1.16; 95% CI, 0.95-1.41), and eGFR 30 to 44 mL/min/1.73 m2 (adjusted HR, 1.09; 95% CI, 0.83-1.44). There was an increased risk of acidosis associated with metformin use at eGFR less than 30 mL/min/1.73 m2 (adjusted HR, 2.07; 95% CI, 1.33-3.22). Results were consistent when adjusted for other time-dependent medication use, including diuretics, renin-angiotensin system inhibitors, statins, nonsteroidal anti-inflammatory drugs, insulin, and other antidiabetic medications: eGFR 30 to 44 mL/min/1.73 m2 had an adjusted HR of 1.13 (95% CI, 0.85-1.49), and eGFR less than 30 mL/min/1.73 m2 had an adjusted HR of 2.21 (95% CI, 1.42-3.44).

    In both metformin users and nonusers, lower eGFR itself was associated with a higher incidence of acidosis. Incidence increased from 4 events per 1000 person-years at eGFR 60 to 89 mL/min/1.73 m2, to 7 events per 1000 person-years at eGFR 45 to 59 mL/min/1.73 m2, to 10 events per 1000 person-years at eGFR 30 to 44 mL/min/1.73 m2, to 24 events per 1000 person-years at eGFR less than 30 mL/min/1.73 m2.

    Active Comparator and Propensity-Score Matched Analyses

    In the active comparator cohort analysis in Geisinger Health System (n = 12 690), the crude incidence of acidosis among metformin users was 4.1 events per 1000 person-years and for sulfonylurea users was 6.9 events per 1000 person-years. Sulfonylurea use was more common at lower eGFR. Compared with sulfonylurea use, metformin use had similar associations with acidosis overall (adjusted HR, 0.91; 95% CI, 0.70-1.18), and in eGFR 45 to 59 mL/min/1.73 m2 (adjusted HR, 1.03; 95% CI, 0.60-1.77) and eGFR 30 to 44 mL/min/1.73 m2 (adjusted HR, 0.77; 95% CI, 0.29-2.05) (Table 4). For eGFR less than 30 mL/min/1.73 m2, there was higher risk for acidosis associated with metformin use, although there were only 3 events among metformin users, limiting power.

    In propensity score–matched cohorts within strata of eGFR, metformin use was associated with similar risk of acidosis compared with use of alternative hypoglycemic medications at eGFR 45 to 59 mL/min/1.73 m2 (adjusted HR, 0.82; 95% CI, 0.55-1.23) and eGFR 30 to 44 mL/min/1.73 m2 (adjusted HR, 0.71; 95% CI, 0.45-1.12), and there was an increased risk of acidosis at eGFR less than 30 mL/min/1.73 m2 that was not statistically significant (adjusted HR, 1.46; 95% CI, 0.86-2.48).

    Additional Sensitivity Analyses

    Excluding baseline insulin users resulted in a slightly lower prevalence of low eGFR (eTable 2 in the Supplement), but no increase in risk for metformin use at eGFR of at least 30 mL/min/1.73 m2 (Table 3). Adjusting for baseline HbA1c yielded similar results, as did analyses in 49 839 patients with incident DM. Early censoring of metformin use showed some increased risk in patients with eGFR 45 to 60 mL/min/1.73 m2 but not eGFR 30 to 44 mL/min/1.73 m2. However, higher acidosis risk in patients with eGFR less than 30 mL/min/1.73 m2 was seen in each of these analyses.

    Replication Analysis

    There were 67 578 new metformin users and 14 439 new sulfonylurea users in MarketScan. Metformin users were slightly younger and more often female compared with sulfonylurea users (eTable 3 in the Supplement). Median follow-up of metformin users was 12.0 months (IQR, 5.5-22.6 months) and of sulfonylurea users was 11.5 months (IQR, 4.6-22.4 months).

    There were 238 acidosis events among metformin users, and 94 among sulfonylurea users. The incidence of acidosis was 2.7 events per 1000 person-years in metformin users, and 5.0 events per 1000 person-years in sulfonylurea users (Table 5). Lower eGFR was a risk factor for acidosis in both metformin and sulfonylurea users. The risk of acidosis associated with metformin use was slightly lower overall (adjusted HR, 0.75; 95% CI, 0.58-0.97), and not increased in patients with eGFR 45 to 59 mL/min/1.73 m2 (adjusted HR, 0.83; 95% CI, 0.42-1.62) and eGFR 30 to 44 mL/min/1.73 m2 (adjusted HR, 0.86; 95% CI, 0.37-2.01). There was a higher risk in patients with eGFR less than 30 mL/min/1.73 m2, but this was not statistically significant (adjusted HR, 1.83; 95% CI, 0.57-5.88; P = .49 for interaction).

    Discussion

    In 2 large retrospective cohorts of patients with DM, metformin use in those with eGFR of at least 30 mL/min/1.73 m2 was not associated with incident hospitalization with acidosis, even after accounting for change in eGFR stage over time and for potential confounding variables, including demographic characteristics, cardiovascular risk factors, HbA1c, and concomitant medications, including insulin. This was true in adjusted, active comparator, and propensity score–matched analyses, and in a replication cohort of 82 017 patients from a nationwide database of health care claims. These findings support the recent expansion of the eGFR thresholds for metformin use by the FDA, and recommendations from other regulatory bodies, which suggest that metformin can be used when eGFR is 45 to 59 mL/min/1.73 m2 and cautiously when eGFR is 30 to 44 mL/min/1.73 m2.18-20

    Our finding of no association between metformin use and acidosis at eGFR of 30 mL/min/1.73 m2 or greater adds to existing literature by rigorously accounting for changes in eGFR over time and addressing confounding through propensity score matching and multiple sensitivity analyses. Our findings support results from smaller cohorts that assessed lactic acidosis through medical record review,16,24 and using a wide range of diagnostic codes.26,34,35 Although randomized clinical trials are the benchmark in assessing causality, they are generally underpowered to evaluate rare events. The 1 trial that randomized 393 participants with moderate CKD (serum creatinine level, 1.47-2.49 mg/dL) to cease or continue metformin therapy observed no events of lactic acidosis.36 On the other hand, our observation of increased risk among patients with eGFR less than 30 mL/min/1.73 m2 strengthens the evidence against metformin use in this group, which previously consisted of case reports and limited evidence from cohort studies.

    Our results partially contradict a large cohort study investigating the association between metformin use and acidosis at low eGFR.27 The UK General Practice Database, with 258 539 patients, found that current metformin users with eGFR less than 60 mL/min/1.73 m2 had a higher risk of lactic acidosis, defined by diagnostic code or plasma lactate concentration greater than 5 mmol/L (45 mg/dL), compared with never metformin users (adjusted HR, 6.37; 95% CI, 1.48-27.5). However, this study was limited by sparse eGFR data and did not account for changes in eGFR over time, leaving the possibility of confounding by GFR, which we show is itself a strong risk factor for acidosis. In the UK study, more than 70% of the never metformin users had either eGFR greater than 60 mL/min/1.73 m2 or an unknown eGFR status.

    From a public health perspective, the potential benefits of using metformin for patients with DM and CKD are vast, given the increasing number of people affected with both diseases worldwide.1,6 Metformin is the first-line therapy for type 2 DM owing to its cost-effectiveness and favorable adverse effect profile. Metformin may also have pleiotropic health benefits beyond its effect on glycemic control.16 In an observational study, patients with type 2 DM and eGFR 30 to 44 mL/min/1.73 m2 had less mortality with the use of metformin than sulfonylureas.37 Patients using metformin may have less weight gain, lower risk of myocardial infarction, and even lower long-term mortality than those receiving other hypoglycemic therapies.36,38,39

    Limitations

    The present study, while benefiting from a large sample size, has certain limitations. First, as with all observational studies, residual confounding is possible. Patients receiving metformin could have different acidosis risk than those who did not receive metformin, for reasons other than their metformin use. The previous contraindication to metformin use in patients with elevated serum creatinine (>1.5 mg/dL in men and >1.4 mg/dL in women) may have introduced a channeling bias, where patients with reduced eGFR might have been more likely to receive metformin therapy if they were healthier. However, we observed consistent results in propensity-matched analyses, with multivariable adjustment, and when excluding insulin users and directly comparing with alternative therapies. Second, the diagnostic code that we used to measure acidosis (ICD-9-CM 276.2) is not specific for lactic acidosis and may dilute the power of the study. On the other hand, the diagnostic code maintains consistency with earlier epidemiological studies,26,29 and allows for the possibility that metformin may contribute to acidosis from other metabolic factors such as malnutrition, liver disease, sepsis, and drug toxicity, which are all clinically relevant for patients. Furthermore, we showed increased risk associated with metformin use in patients with eGFR less than 30 mL/min/1.73 m2, which is consistent with current FDA recommendations. Third, if the effect of metformin use on lactic acidosis were mediated by decline in eGFR, the use of time-dependent eGFR in our analyses could bias results to the null. However, we found similar results when only adjusting for baseline eGFR in active comparator analyses. Fourth, we could not differentiate whether a change in eGFR stage occurred due to CKD progression or an acute kidney injury. In theory, metformin accumulation may be enhanced in prerenal acute kidney injury compared with CKD, where tubular function and active secretion may be disproportionately impaired due to concurrent volume depletion.23 Fifth, we could not discern the specialty of the prescribing clinician, and medication use was derived from electronic medical records or claims and not verified by patient report. Finally, although we have no reason to suspect that results would be different by race, most of the Geisinger Health System population was white, which may limit generalizability.

    Conclusions

    Metformin use was not associated with incident acidosis in patients with eGFR 30 to 60 mL/min/1.73 m2 in 2 large and diverse cohorts, but there was increased risk at eGFR less than 30 mL/min/1.73 m2. Our results support cautious use of metformin in patients with type 2 DM and eGFR of at least 30 mL/min/1.73 m2.

    Back to top
    Article Information

    Accepted for Publication: March 6, 2018.

    Corresponding Author: Morgan E. Grams, MD, PhD, Department of Medicine, Johns Hopkins University, 2024 E Monument St, Baltimore, MD 21205 (mgrams2@jhmi.edu).

    Published Online: June 4, 2018. doi:10.1001/jamainternmed.2018.0292

    Author Contributions: Drs Chang and Grams had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Chang and Grams served as co–senior authors, each with equal contribution to the manuscript.

    Study concept and design: Lazarus, Wu, Secora, Grams.

    Acquisition, analysis, or interpretation of data: Lazarus, Wu, Shin, Sang, Alexander, Inker, Coresh, Chang, Grams.

    Drafting of the manuscript: Lazarus, Wu.

    Critical revision of the manuscript for important intellectual content: Wu, Shin, Sang, Alexander, Secora, Inker, Coresh, Chang, Grams.

    Statistical analysis: Lazarus, Wu, Shin, Sang, Coresh, Grams.

    Obtained funding: Coresh.

    Administrative, technical, or material support: Secora, Coresh, Grams.

    Study supervision: Grams.

    Conflict of Interest Disclosures: Dr Alexander is Chair of the FDA’s Peripheral and Central Nervous System Advisory Committee; serves on the Advisory Board of MesaRx Innovations; holds equity in Monument Analytics, a health care consultancy whose clients include the life sciences industry as well as plaintiffs in opioid litigation; and serves as a member of OptumRx’s P&T Committee. This arrangement has been reviewed and approved by Johns Hopkins University in accordance with its conflict of interest policies. No other disclosures are reported.

    Funding/Support: Ms Sang and Dr Coresh receive funding from the National Institutes of Health via grant R01DK100446. Dr Grams receives funding from the National Institutes of Health via grants R01DK100446 and K08DK092287.

    Role of the Funder/Sponsor: The National Institutes of Health had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Thomas  MC, Cooper  ME, Zimmet  P.  Changing epidemiology of type 2 diabetes mellitus and associated chronic kidney disease.  Nat Rev Nephrol. 2016;12(2):73-81.PubMedGoogle ScholarCrossref
    2.
    Inzucchi  SE, Bergenstal  RM, Buse  JB,  et al.  Management of hyperglycemia in type 2 diabetes, 2015: a patient-centered approach: update to a position statement of the American Diabetes Association and the European Association for the Study of Diabetes.  Diabetes Care. 2015;38(1):140-149.PubMedGoogle ScholarCrossref
    3.
    Hong  J, Zhang  Y, Lai  S,  et al; SPREAD-DIMCAD Investigators.  Effects of metformin versus glipizide on cardiovascular outcomes in patients with type 2 diabetes and coronary artery disease.  Diabetes Care. 2013;36(5):1304-1311.PubMedGoogle ScholarCrossref
    4.
    Kooy  A, de Jager  J, Lehert  P,  et al.  Long-term effects of metformin on metabolism and microvascular and macrovascular disease in patients with type 2 diabetes mellitus.  Arch Intern Med. 2009;169(6):616-625.PubMedGoogle ScholarCrossref
    5.
    Maruthur  NM, Tseng  E, Hutfless  S,  et al.  Diabetes medications as monotherapy or metformin-based combination therapy for type 2 diabetes: a systematic review and meta-analysis.  Ann Intern Med. 2016;164(11):740-751.PubMedGoogle ScholarCrossref
    6.
    Flory  JH, Hennessy  S.  Metformin use reduction in mild to moderate renal impairment: possible inappropriate curbing of use based on food and drug administration contraindications.  JAMA Intern Med. 2015;175(3):458-459.PubMedGoogle ScholarCrossref
    7.
    Conlay  LA, Loewenstein  JE.  Phenformin and lactic acidosis.  JAMA. 1976;235(15):1575-1578.PubMedGoogle ScholarCrossref
    8.
    Kwong  SC, Brubacher  J.  Phenformin and lactic acidosis: a case report and review.  J Emerg Med. 1998;16(6):881-886.PubMedGoogle ScholarCrossref
    9.
    Altun  E, Kaya  B, Paydaş  S, Sarıakçalı  B, Karayaylalı  I.  Lactic acidosis induced by metformin in a chronic hemodialysis patient with diabetes mellitus type 2.  Hemodial Int. 2014;18(2):529-531.PubMedGoogle ScholarCrossref
    10.
    Duong  JK, Furlong  TJ, Roberts  DM,  et al.  the role of metformin in metformin-associated lactic acidosis (MALA): case series and formulation of a model of pathogenesis.  Drug Saf. 2013;36(9):733-746.PubMedGoogle ScholarCrossref
    11.
    Fitzgerald  E, Mathieu  S, Ball  A.  Metformin associated lactic acidosis.  BMJ. 2009;339:b3660.PubMedGoogle ScholarCrossref
    12.
    Perrone  J, Phillips  C, Gaieski  D.  Occult metformin toxicity in three patients with profound lactic acidosis.  J Emerg Med. 2011;40(3):271-275.PubMedGoogle ScholarCrossref
    13.
    Silvestre  J, Carvalho  S, Mendes  V,  et al.  Metformin-induced lactic acidosis: a case series.  J Med Case Rep. 2007;1:126.PubMedGoogle ScholarCrossref
    14.
    Wiholm  BE, Myrhed  M.  Metformin-associated lactic acidosis in Sweden 1977-1991.  Eur J Clin Pharmacol. 1993;44(6):589-591.PubMedGoogle ScholarCrossref
    15.
    Yeh  HC, Ting  IW, Tsai  CW, Wu  JY, Kuo  CC.  Serum lactate level and mortality in metformin-associated lactic acidosis requiring renal replacement therapy: a systematic review of case reports and case series.  BMC Nephrol. 2017;18(1):229.PubMedGoogle ScholarCrossref
    16.
    Inzucchi  SE, Lipska  KJ, Mayo  H, Bailey  CJ, McGuire  DK.  Metformin in patients with type 2 diabetes and kidney disease: a systematic review.  JAMA. 2014;312(24):2668-2675.PubMedGoogle ScholarCrossref
    17.
    Department of Health and Human Services. Metformin Petition Closeout Memorandum from CDER Director. Silver Spring, Maryland: Food and Drug Administration; 2016.
    18.
    Kidney Disease Improving Global Outcomes (KDIGO) Work Group.  KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease.  Kidney Int Suppl. 2013;3(1):1-150.Google ScholarCrossref
    19.
    National Institute for Health and Care Excellence (NICE). Type 2 diabetes in adults: management. Updated May 2017. https://www.nice.org.uk/guidance/ng28. Accessed October 15, 2017.
    20.
    American Diabetes Association.  Standards of Medical Care in Diabetes - 2016.  Diabetes Care. 2016;39(1):S1-S112.PubMedGoogle Scholar
    21.
    Lipska  KJ, Bailey  CJ, Inzucchi  SE.  Use of metformin in the setting of mild-to-moderate renal insufficiency.  Diabetes Care. 2011;34(6):1431-1437.PubMedGoogle ScholarCrossref
    22.
    Bakris  GL, Molitch  ME.  Should restrictions be relaxed for metformin use in chronic kidney disease? yes, they should be relaxed! what’s the fuss?  Diabetes Care. 2016;39(7):1287-1291.PubMedGoogle ScholarCrossref
    23.
    Kalantar-Zadeh  K, Kovesdy  CP.  Should restrictions be relaxed for metformin use in chronic kidney disease? no, we should never again compromise safety!  Diabetes Care. 2016;39(7):1281-1286.PubMedGoogle ScholarCrossref
    24.
    Bodmer  M, Meier  C, Krähenbühl  S, Jick  SS, Meier  CR.  Metformin, sulfonylureas, or other antidiabetes drugs and the risk of lactic acidosis or hypoglycemia: a nested case-control analysis.  Diabetes Care. 2008;31(11):2086-2091.PubMedGoogle ScholarCrossref
    25.
    Cucchiari  D, Podestà  MA, Merizzoli  E,  et al.  Dose-related effects of metformin on acid-base balance and renal function in patients with diabetes who develop acute renal failure: a cross-sectional study.  Acta Diabetol. 2016;53(4):551-558.PubMedGoogle ScholarCrossref
    26.
    Ekström  N, Schiöler  L, Svensson  AM,  et al.  Effectiveness and safety of metformin in 51 675 patients with type 2 diabetes and different levels of renal function: a cohort study from the Swedish National Diabetes Register.  BMJ Open. 2012;2(4):e001076.PubMedGoogle ScholarCrossref
    27.
    Eppenga  WL, Lalmohamed  A, Geerts  AF,  et al.  Risk of lactic acidosis or elevated lactate concentrations in metformin users with renal impairment: a population-based cohort study.  Diabetes Care. 2014;37(8):2218-2224.PubMedGoogle ScholarCrossref
    28.
    Kuo  CC, Yeh  HC, Chen  B, Tsai  CW, Lin  YS, Huang  CC.  Prevalence of metformin use and the associated risk of metabolic acidosis in US diabetic adults with CKD: a national cross-sectional study.  Medicine (Baltimore). 2015;94(51):e2175.PubMedGoogle ScholarCrossref
    29.
    Richy  FF, Sabidó-Espin  M, Guedes  S, Corvino  FA, Gottwald-Hostalek  U.  Incidence of lactic acidosis in patients with type 2 diabetes with and without renal impairment treated with metformin: a retrospective cohort study.  Diabetes Care. 2014;37(8):2291-2295.PubMedGoogle ScholarCrossref
    30.
    Salpeter  SR, Greyber  E, Pasternak  GA, Salpeter  EE.  Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus.  Cochrane Database Syst Rev. 2010;(4):CD002967.PubMedGoogle Scholar
    31.
    Stang  M, Wysowski  DK, Butler-Jones  D.  Incidence of lactic acidosis in metformin users.  Diabetes Care. 1999;22(6):925-927.PubMedGoogle ScholarCrossref
    32.
    National Committee for Quality Assurance. HEDIS and Quality Measurement: HEDIS Measures. 2017. http://www.ncqa.org/hedis-quality-measurement/hedis-measures. Accessed October 15, 2017.
    33.
    Levey  AS, Stevens  LA, Schmid  CH,  et al; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration).  A new equation to estimate glomerular filtration rate.  Ann Intern Med. 2009;150(9):604-612.PubMedGoogle ScholarCrossref
    34.
    Pongwecharak  J, Tengmeesri  N, Malanusorn  N, Panthong  M, Pawangkapin  N.  Prescribing metformin in type 2 diabetes with a contraindication: prevalence and outcome.  Pharm World Sci. 2009;31(4):481-486.PubMedGoogle ScholarCrossref
    35.
    Vasisht  KP, Chen  SC, Peng  Y, Bakris  GL.  Limitations of metformin use in patients with kidney disease: are they warranted?  Diabetes Obes Metab. 2010;12(12):1079-1083.PubMedGoogle ScholarCrossref
    36.
    Rachmani  R, Slavachevski  I, Levi  Z, Zadok  B, Kedar  Y, Ravid  M.  Metformin in patients with type 2 diabetes mellitus: reconsideration of traditional contraindications.  Eur J Intern Med. 2002;13(7):428.PubMedGoogle ScholarCrossref
    37.
    Marcum  ZA, Forsberg  CW, Moore  KP,  et al.  Mortality associated with metformin versus sulfonylurea initiation: a cohort study of veterans with diabetes and chronic kidney disease.  J Gen Intern Med. 2018;33(2):155-165.PubMedGoogle ScholarCrossref
    38.
    Holman  RR, Paul  SK, Bethel  MA, Matthews  DR, Neil  HA.  10-year follow-up of intensive glucose control in type 2 diabetes.  N Engl J Med. 2008;359(15):1577-1589.PubMedGoogle ScholarCrossref
    39.
    Roussel  R, Travert  F, Pasquet  B,  et al; Reduction of Atherothrombosis for Continued Health (REACH) Registry Investigators.  Metformin use and mortality among patients with diabetes and atherothrombosis.  Arch Intern Med. 2010;170(21):1892-1899.PubMedGoogle ScholarCrossref
    ×