Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Does metformin exposure before ICU stay have any impact on patients’ outcome? A retrospective cohort study of diabetic patients

Does metformin exposure before ICU stay have any impact on patients’ outcome? A retrospective... Background: Impact of metformin exposure before ICU stay remains controversial. Metformin is thought to induce lactic acidosis and haemodynamic instability but may reduce ICU mortality. We evaluated its influence on outcome in diabetic patients admitted in the ICU and then compared two different populations based on the presence of septic shock. Methods: We conducted a retrospective cohort study in a 24‑ bed French ICU between October 2010 and December 2013, including all ICU‑ admitted diabetic patients. Results: Among 635 diabetic patients admitted during the study period, 131 (21%) were admitted with septic shock. Multivariate analysis showed no difference in hospital mortality in all metformin users (OR 0.75 [95% CI 0.44–1.28]; p = 0.29), except in the septic shock subgroup (OR 0.61; 95% CI [0.37–0.99]; p = 0.04) despite higher vasopressor dosages in the first hours after shock onset. Blood lactate level was higher in metformin users than in non ‑ metformin users in all patients (p < 0.001), in septic shock patients (p < 0.001) and in patients without kidney injury (p < 0.001). Metformin users did not have more septic shock from unknown aetiology (p = 0.65) or unknown pathogen (p = 0.99). Conclusions: Metformin use before admission to ICU did not affect in ‑ hospital mortality. However, for patients with septic shock, mortality was lower, despite worse clinical presentation on admission. Blood lactate levels were always higher with or without septic shock and indifferent of kidney function. Keywords: Metformin, Septic shock, Diabetes, Lactic acidosis, ICU [2, 3]. It is thought to induce or worsen lactic acidosis, Background especially in acute renal or liver dysfunction [4]. But in Metformin is increasingly used as an oral antidiabetic a recent meta-analysis pooling 347 trials involving long- (OAD) agent, especially in patients with type 2 diabetes run metformin use, the authors found no case of met- mellitus. Metformin inhibits hepatic glucose production, formin-associated lactic acidosis (MALA), as well as no reduces intestinal glucose absorption and improves glu- difference in blood lactate level related to metformin use cose metabolism [1]. [5]. These results were confirmed in a large cohort of Its use is associated with a reduction in cardiovascu- diabetic patients treated with metformin despite various lar morbidity and mortality, in comparison with insu- metformin contraindications, in which no MALA has lin, other OADs or diet alone, in non-acutely ill patients been described by the authors [2]. In the ICU, MALA has been described in renal, liver, *Correspondence: sebastien.jochmans@gmail.com pulmonary or cardiovascular chronic failure [6], and Département de Médecine Intensive et Unité de Recherche Clinique, Groupe Hospitalier Sud Ile‑de‑France, Hôpital de Melun, 77000 Melun, several case reports described fatal or non-fatal MALA France in acute conditions. In contrast, a recent retrospective Full list of author information is available at the end of the article © The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 2 of 9 study in 17 Danish ICUs found that prior to admission 1. Chronic respiratory failure (previous pulmonary metformin use was associated with a reduction in 30-day function tests, history of acute respiratory decom- mortality [3]. pensation, oxygen or non-invasive ventilation at Our main objective in this study was to evaluate the home, sleep apnoea) and/or influence of pre-admission metformin use on outcome 2. Chronic cardiac failure (history of pulmonary in diabetic ICU patients and in a subgroup experiencing oedema, left ventricular ejection fraction < 45%) and/ septic shock (an acute condition known to induce lac- or tic acidosis [7, 8]). Secondary objectives were to assess 3. Chronic renal disease (calculated creatinine clear- MALA incidence and blood lactate levels in ICU patients ance with Modification of Diet in Renal Disease with diabetes, treated or not by metformin, with or with- [MDRD] < 60 mL/min/1.73 m ) and/or out septic shock. 4. Chronic liver disease (history of cirrhosis, previous INR > 1.2) and/or Methods 5. Myocardial infarction during the previous month We performed a retrospective cohort study in our Inten- sive Care Medicine Department between October 2010 Septic shock was defined according to the Surviving and December 2013. The study protocol was approved Sepsis Campaign definition [9]. Acute kidney injury was by the French Intensive Care Society (FICS)—Société de defined using Kidney Disease Improving Global Out - Réanimation de Langue Française (SRLF)—ethical review come (KDIGO) classification [10] and was considered for board. any stage of the classification. Patients Statistical analysis Inclusion and exclusion criteria Continuous variables were expressed as median [25th– All patients admitted within the study period with a 75th interquartile range] or mean  ±  standard deviation history of diabetes treated by insulin or oral antidiabet- [95% confidence interval] (after Shapiro–Wilks test) and ics were included. So-called diabetic patients treated compared using nonparametric Mann–Whitney (or only with diet were considered as unconfirmed diabetes Student’s t test) and linear regression tests. Categorical and were excluded. The other exclusion criteria were as variables were expressed as n (%) and compared using follows: Chi-square or Fisher’s exact tests. All tests were two- Unknown chronic antidiabetic treatment, modifica - tailed assuming alpha risk = 0.05. All collected data were tions of antidiabetic treatment during the month before analysed in univariate analysis regarding ICU and hospi- ICU admission and unavailable arterial blood gas sample tal survivals. We included in forward and backward step- within 4 h after ICU admission. wise multivariate regression models as covariates all data with p  <  0.1 in univariate analysis, with stratification by Data collection metformin use. We applied these models in ICU patients Collected clinical features were as follows: age, sex, and in the subgroups of septic shock and metformin users height, weight, Simplified Acute Physiology Score II with usual contraindication. We performed a post hoc (SAPS II), main admission cause, metformin contrain- validity assessment of the regression models by receiver dication (‘Definitions’ paragraph below), ICU admis - operating characteristic (ROC) curves, and we selected as sion biomarkers (leucocytes, platelets, haemoglobin, the result the model with the best area under the curve. creatinine, C-reactive protein, bilirubin and/or INR if Results of multivariate regression test were expressed by available), arterial blood gas samples at day 1, all bacte- odds ratio (95% confidence interval). Prognostic value riological tests, vasopressor dosages (close to the initia- of blood lactate level on mortality was tested with ROC tion even outside the ICU), urinary output and amount curves (results expressed by area under the curve [AUC] of intravascular input during the first 24  h, the use of % (95% confidence interval)), sensitivity and sensibility. invasive ventilation and renal replacement therapy, Statistical analysis and graphic representations were the presence of acute respiratory distress syndrome performed with SPSS Statistics V20 software (IBM , (ARDS), ICU and hospital length of stay and vital New York, NY, USA) and Prism 6 software (GraphPad status. Software Inc. , San Diego, CA, USA). Definitions Results Usual metformin contraindications (adapted from the Among the 3871 patients admitted in our ICU during instructions for the use of the medicinal product) were the study period, 635 (16.4%) were finally included (study defined as: flowchart is available in Additional file  1: Figure S1), Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 3 of 9 including 131 (20.6%) patients with septic shock at day 1 Table 1 Cohort of  ICU diabetics: main characteristics at  ICU admission, during  ICU stay and  ICU/hospital out- after ICU admission. come Metformin use before admission was found in 240 patients (37.8%) and was similar regarding occurrence ICU diabetics No metformin Metformin or non-occurrence of septic shock (p  =  0.69). Ratio of N 635 395 (62.2) 240 (37.8) metformin use in patients with one or more usual con- Age (y) 71 [61–79] 73 [62.5–80] 68 [60–78]* traindications was high (119 (49.6%)) with a similar rate Men 408 (64.3) 255 (64.6) 153 (63.8) in septic shock patients (p = 0.54). SAPS II 39 [31‑52] 40 [32‑52] 38 [29‑51] Usual metformin 387 (60.9) 268 (67.9) 119 (49.6)* ICU admission and hospital stay contraindication The main characteristics of ICU diabetics at admission Chronic respiratory 190 (29.9) 132 (33.4) 58 (24.2)* and during ICU or hospital stays are specified in Table  1 insufficiency and Additional file  1: Table S1. In our study cohort, 588 Chronic cardiac 138 (21.7) 92 (23.3) 46 (19.2) (92.6%) patients were admitted for a medical cause, insufficiency mainly for acute respiratory failure (266 (41.9%)). There Chronic liver disease 75 (11.8) 49 (12.4) 26 (10.8) was no difference between metformin users (MET) Chronic kidney 144 (22.7) 128 (32.4) 16 (6.7)* failure and non-metformin users (NO-MET) in the reason for Recent myocardial 8 (1.3) 5 (1.3) 3 (1.3) admission. MET were younger with less chronic res- infarction piratory and renal failures. They had higher blood lac - pH 7.36 [7.28–7.42] 7.36 [7.29–7.43] 7.36 [7.27–7.42] tate level (p  <  0.001), lower bicarbonate (p  <  0.01) and PaCO (mmHg) 36 [29–43] 37 [30–44] 36 [28–43] also lower serum creatinine (p  <  0.001) with less acute HCO (mmHg) 21.3 [17–25.2] 21.9 [17.5–26] 20.4 [15.3–24]* kidney injury (p  <  0.001). Severity score (SAPS II) and Lactate (mmol/L) 1.4 [0.9–2.4] 1.2 [0.8–2.1] 1.8 [1.1–3.9]* need in organ support (i.e. invasive mechanical ventila- INR 1.25 [1.06–1.71] 1.26 [1.06–1.65] 1.24 [1.07–1.77] tion, vasopressor, renal replacement therapy) were simi- Bilirubin (µmol/L) 10 [7–16] 10 [7–16] 10 [7–16] lar. Among MET, there was no difference in lactate level C‑reactive protein 34 [8–115] 35 [8–115] 32 [8–115] between patients with or without usual contraindication (mg/L) (p = 0.86) (Additional file 1: Table S2). Haemoglobin (g/dL) 11.2 [9.6–13] 11.1 [9.6–12.7] 11.7 [9.7–13.4] The main characteristics for diabetics with septic shock Leucocytes (G/L) 11.2 [8.1–15.3] 10.8 [7.5–14.6] 11.7 [8.4–16.3] at admission and during ICU or hospital stays are speci- Platelets (G/L) 213 [155–277] 207 [155–271] 219 [157–293] fied in Table  2 and Additional file  1: Tables S3 and S4. Creatinine (µmol/L) 131 [85–238] 153 [90–285] 108 [80–174]* Aetiologies of shock are specified in Additional file  1: Acute kidney injury 392 (61.7) 268 (67.8) 124 (51.7)* Table S3. There was no difference between MET and Renal replacement 113 (17.8) 72 (18.2) 41 (17.1) NO-MET regarding unknown aetiology (p  =  0.65) and therapy unknown pathogen (p  =  0.99) (Additional file  1: Table Vasopressors 229 (36.1) 136 (34.4) 93 (38.8) S4). MET with septic shock had higher blood lactate Invasive ventilation 230 (36.2) 139 (35.2) 91 (37.9) than NO-MET at admission (p  <  0.001) and during the ICU length of stay 6 [3–10] 6 [3.5–10] 6 [3–9] (d) first 12 h (Fig.  1). Bicarbonate was lower (p < 0.01). They ICU death 117 (18.4) 75 (19) 42 (17.5) also received more renal replacement therapy (p = 0.02), Hospital length of 12 [6–23] 12 [6–23] 13 [7–23] while they had less chronic renal failure and there was no stay (d) significant difference in serum creatinine, pH, day 1 uri - Hospital death 140 (22) 92 (23.3) 48 (20) nary output or acute kidney injury occurrence. In MET, Values are n (%) or median [IQR 25th–75th] there was a linear correlation between blood lactate * p < 0.05 between metformin and no metformin and serum creatinine (ρ  =  0.36; p  <  0.01) in contrast to NO-MET (ρ =  0.09; p = 0.41) (Additional file  1: Figures S2 and S3). However, lactate was even higher in MET (p  <  0.001) with normal kidney function (MDRD creati- (p  =  0.09). Vasopressor dose was significantly higher nine clearance > 60 mL/min/1.73 m ). in MET the first hours after reaching criteria for septic Septic shock severity can also be evaluated by the shock (Fig. 2). amount of vascular filling and the dose of vasopressors. There was no difference in the number of patients with Mortality and length of stay intensive vascular filling (i.e. more than 50  mL/kg/day) ICU or hospital lengths of stay as well as ICU death between MET and NO-MET, but there was a statistical showed no statistically significant difference between trend for higher maximal dose of noradrenaline in MET MET and NO-MET in the cohort of diabetics and in the Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 4 of 9 Table 2 Subgroup of ICU diabetics with septic shock: main characteristics at ICU admission, during ICU stay and ICU/hos- pital outcome Septic shocks No metformin Metformin N 131 79 (60.3) 52 (39.7) Age (y) 70 [63–78] 71 [64–78] 66 [61–78] Men 89 (67.9) 56 (70.9) 33 (63.5) SAPS II 52 [42–69] 48 [40–68] 57 [46–68] Usual metformin contraindication 79 (60.3) 56 (70.9) 23 (44.2)* Chronic respiratory failure 30 (22.9) 21 (26.6) 9 (17.3) Chronic cardiac failure 27 (20.6) 20 (25.3) 7 (13.5) Chronic liver disease 26 (19.8) 18 (22.8) 8 (15.4) Chronic renal failure 19 (14.5) 16 (20.3) 3 (5.8)* Recent myocardial infarction 1 (0.8) 1 (1.3) 0 pH 7.32 [7.2–7.38] 7.32 [7.23–7.39] 7.26 [7.17–7.38] PaCO (mmHg) 34 [27–42] 35 [29–43] 34 [24–42] HCO (mmHg) 18.2 [13.3–22.2] 19.7 [14.7–24.1] 15.5 [10.1–19.9]* Lactate (mmol/L) 2.2 [1.1–5] 1.4 [1–2.8] 4.5 [2.1–8.7]* INR 1.5 [1.2–2.3] 1.6 [1.3–2.9] 1.4 [1.1–1.9] Bilirubin (µmol/L) 12 [8–24] 13 [8–26] 10 [8–19] C‑reactive protein (mg/L) 95 [24–224] 98 [30–225] 85 [14–212] Haemoglobin (g/dL) 10.6 [9.1–12.4] 10.7 [9.3–12.4] 10.5 [9.1–12.4] Leucocytes (G/L) 12.1 [8.3–19.6] 11.9 [8.5–19] 12.9 [8.4–21.6] Platelets (G/L) 185 [119–265] 199 [120–273] 173 [118–252] Creatinine (µmol/L) 167 [113–326] 163 [108–276] 176 [123–364] Urinary output day 1 (mL) 1200 [553–2200] 1200 [558–1925] 1425 [443–2400] Number of patients with vascular filling > 50 mL/kg ≥ 1 day 76 (60.3) 42 (58.4) 34 (69.4) Maximum dose of noradrenaline (mg/h) 2 [1–4.3] 2 [1–3.5] 3.5 [1.3–5]* (µg/kg/min) 0.43 [0.22–0.95] 0.4 [0.21–0.76] 0.61 [0.23–1.16] Maximum dose of adrenaline (mg/h) 2.5 [1.5–6] 3 [1.5–6.3] 2.5 [1.4–6] (µg/kg/min) 0.61 [0.25–1.22] 0.52 [0.22–1.3] 0.66 [0.27–0.98] Noradrenaline duration (h) 39 [18–64] 48 [19–71] 36 [15–59] Adrenaline duration (h) 36 [9–90] 36 [14–90] 30 [6–102] Vasopressor duration (h) 48 [24–96] 48 [24–97] 36 [23–72] Acute kidney injury 104 (79.4) 62 (78.5) 42 (80.8) ARDS 48 (36.6) 27 (34.2) 21 (40.4) Renal replacement therapy 51 (38.9) 24 (30.4) 27 (51.9)* Invasive ventilation 96 (73.3) 56 (70.9) 40 (76.9) ICU length of stay (d) 9 [5–16] 9 [6–19] 7 [4–13] Hospital length of stay (d) 15 [7–29] 15 [8–29] 16 [4–26] ICU death 51 (38.9) 31 (39.2) 20 (38.5) Hospital death 53 (40.5) 33 (41.8) 20 (38.5) Values are n (%) or median [IQR 25th–75th] * p < 0.05 between metformin and no metformin subgroup of septic shock patients. Hospital death was not multivariate analysis with odds ratio 0.61 [95% CI 0.36– significantly different in multivariate regression model 0.99]; p = 0.049 (Table 3). analysis (OR 0.75 [0.44–1.28]; p  =  0.29) (Additional Blood lactate levels showed a prognostic value in MET file  1: Table S5). In the subgroup of septic shock patients, (AUC 67.3% (95% CI 58.3–76.4); p  =  0.001) and NO- metformin was associated with a lower mortality after MET (AUC 68.6% (61.5–75.8); p  <  0.001) of the cohort Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 5 of 9 Fig. 1 Initial evolution of lactate level in ICU diabetics sustaining septic shock with or without pre‑admission metformin treatment. T0: time of septic shock diagnosis. Abscissa axis is log 10 scale. *p < 0.05 Fig. 2 Initial evolution of vasopressor dosage in ICU diabetics sustaining septic shock with or without pre‑admission metformin treatment. T0: time of septic shock diagnosis. Abscissa axis is log 10 scale. *p < 0.05 and also in MET (AUC 66.7% (51.5–81.9); p = 0.05) and NO-MET (AUC 65.5% (53–78.1); p  =  0.02) of shocked Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 6 of 9 Table 3 Hospital death among  septic shock patients: univariate analysis and  conditional forward stepwise multivariate analysis with metformin as analysis factor Survivor Non-survivor P-univariate Odds ratio P-multivariate N 80 (61.1) 51 (38.9) – – – Men 49 (61.3) 40 (78.4) 0.055 NS NS SAPS II 49 [40–61] 65 [46–80] 0.001 1.05 (1.04–1.07) < 0.001 Metformin (n) 32 (40) 20 (39.2) 1 0.61 (0.37–0.99) 0.049 Lactate (mmol/L) 1.7 [1–3.9] 3.2 [1.4–7.1] 0.003 1.21 (1.1–1.34) < 0.001 ARDS (n) 23 (28.8) 25 (49) 0.03 NS NS RRT (n) 24 (30) 27 (52.9) 0.011 NS NS Invasive ventilation (n) 48 (60) 48 (94.1) < 0.001 NS NS Urinary output day 1 (mL) 1400 [675–2400] 1030 [65–1900] 0.03 NS NS α = 0.05. Area under the curve of the multivariate model = 0.786 RRT renal replacement therapy, NS not significant patients. But prognostic cut-off value for lactate with Propensity-score-matched analyses yielded the same the highest sensitivity and specificity was higher in MET results. In our work, more than 90% were medical admis- (2.15  mmol/L, sensitivity 65%, specificity 61.6%) than sions, whereas two-thirds of the 7404 ICU patients with in NO-MET (1.35, sensitivity 66.2%, specificity 61.3%). type 2 diabetes in Christiansen et al.’s study were surgical Likewise in the subgroup of septic shock patients, cut-off admissions. However, no data were available concerning values were 4.45  mmol/L (sensitivity 57.9%, specificity septic shocks, vasopressor dosages or even blood lac- 56.7%) versus 1.45  mmol/L (sensitivity 58.1%, specificity tate levels. Mechanisms of this beneficial effect remain 56.2%), respectively. unclear: in ICU patients, metformin may supply higher Among MET, there was no significant difference in amounts of lactate serving as an energetic carbon source hospital death between patients with or without usual and therefore is available for ischaemic tissues with glu- contraindication (OR 1.24 [0.48–3.2]; p  =  0.66) (Addi- cose preservation. Metformin may also decrease cellular tional file 1: Table S6). hypoxia of less perfused tissues by decreasing oxygen consumption. Discussion However, clinical severity seems higher in MET. Lac- In our large cohort on critically ill diabetic patients, met- tate levels are significantly higher in ICU diabetics with formin use before admission to ICU did not affect in- or without septic shock (Additional file  1: Figure S4). This hospital mortality; however, pre-admission metformin issue still remains controversial with studies finding no treatment was independently associated with a decrease effect of metformin on lactate rate [5, 15, 16] or, on the in hospital mortality in the group of septic shock contrary, finding an increased lactate [17–27]. One rea - patients, even with an initial clinical presentation appear- son for this discrepancy may be that ICU patients, unlike ing more severe. Indeed, independent of kidney func- other patients, suffer acute stress with endogenous cat - tion, vasopressor dosages and serum lactate levels were echolamine release leading to increased lactate levels higher during the first hours after shock onset in MET. through adrenergic receptor stimulation. Physiological Nevertheless, metformin did not seem to induce shock studies showed that metformin enhances lactate pro- per se because there was no more septic shock from duction and decreases oxygen consumption [23–25] by unknown aetiology or unknown pathogens in MET than inhibiting mitochondrial chain complexes [19, 22–24, in NO-MET. 27]. Therefore, in our study, prognostic cut-off values are A beneficial association between metformin and mor - higher in MET, especially when there is a septic shock, tality has been already described both in selected patients as previously found [28]. It is usually admitted that lac- with chronic heart failure [11], liver disease [12, 13], tic acidosis in metformin users is due to a reduced renal mild-to-moderate kidney failure [14] which are usual drug clearance. Lactate and creatinine levels (and cre- contraindications, and in ICU patients [3]. In this lat- atinine clearance) are linearly correlated in our study ter study, based on retrospective analysis of Northern as previously shown [17, 18, 21, 26, 29–31]. But lactate Denmark database, 30-day mortality was lower in met- levels remain higher in patients without kidney injury formin users than in non-metformin users with adjusted with metformin than without. This last issue was only hazard ratio  =  0.8 (95% confidence interval 0.71–0.95). previously described in case reports and one cohort Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 7 of 9 study [26], although another study failed to find hyper - same prognostic value in patients previously untreated lactatemia when kidney function was normal [29]. MET by metformin. Similarly, elevated doses of vasopressors, probably received more haemodialysis for the purpose which are used as a criterion for poor outcome for exam- of either correcting deeper hypobasemia or eliminating ple in the SOFA score, may not carry the same prog- plasma metformin. nostic significance. Metformin blood dosage has never been performed. However, it seems linearly correlated to lactate concentration [18, 21, 31]. Lastly, comparison • Vasopressor dosages are higher in septic shock diabet- between MET treated or non-treated by renal replace- ics with pre-admission metformin. This increase in ment therapy was unfeasible because analysis would catecholamines need, which has not been previously lack power and be statistically unreliable. If current sci- described, is not due to acidosis per se because pH val- entific opinion suggests its use in metformin overdose, ues are similar with or without metformin. Recent data there is no strong proof. There is indeed a contradiction suggest that metformin decreases adenylate cyclase between studies finding a beneficial association between activity and therefore cyclic AMP concentration [32]. sepsis and metformin and in contrast the desire to elimi- The effects of vasopressors are mediated by adrenergic nate metformin by haemodialysis. Therefore, we suggest receptors, G protein and adenylate cyclase stimulations that future studies should seek to answer two ques- leading to an increase in cyclic AMP concentration. It tions: Is there a benefit in giving metformin during the is assumed that it is necessary to increase vasopressor first hours of septic shock in diabetic patients previously dosages in order to obtain the same haemodynamic untreated by metformin? Is there really a benefit in the effect and compensate decreased adenylate cyclase early elimination of metformin by haemodialysis in dia- activity induced by metformin. Indeed, metformin betic patients with septic shock and without acute kid- does not seem to produce sepsis-like shock because ney injury? there is as much septic shock of unknown aetiology or germ in MET than in NO-MET. However, metformin Conclusions actually seems to worsen the criteria usually used to Metformin use before admission to ICU is associated assess the severity of septic shocks. with a decrease in mortality in septic shock patients despite a worse clinical presentation on admission. Met- Finally, in our study, patients treated with metformin formin users have higher lactate levels independent of despite the presence of the usual contraindications do kidney function and need higher vasopressor dosages not have higher lactate levels. The mortality rate is not during the first hours of septic shock. Metformin does increased either. These contraindications have been chal - not seem to induce shock per se. The presence or absence lenged for several years so that metformin seems delete- of one of the usual contraindications to taking metformin rious only in terminal kidney disease [33]. Our collected does not alter lactate levels or hospital mortality. data did not allow us to evaluate outcome according to the intensity of each organ failure. It is possible that our Additional file patients had mainly mild-to-moderate lung, liver, heart or kidney injury that would be insufficient to worsen out - Additional file 1: Figure S1. Study flowchart. Figure S2. Linear regres‑ come or lactate level. sion between blood creatinine and lactate levels in metformin users Our study is subject to certain limitations. First, it is patients. Figure S3. Linear regression between blood creatinine and lactate levels in non‑metformin users patients. Figure S4. Lactate levels. a retrospective study, avoiding observation bias, but Table S1. Main admission pattern of ICU‑admitted diabetics. Table S2. with selection bias due to non-inclusion of patients ICU‑admitted diabetics with preadmission metformin treatment with or with missing data. Thus, we cannot determine whether without usual metformin contraindication. Table S3. Aetiologies and germs responsible for septic shocks in ICU‑ diabetics. Table S4. Septic metformin users are more likely to be admitted to ICU shocks without aetiology at the end of hospital stay. Table S5. Hospital than other antidiabetics’ takers, and also whether the death among ICU‑admitted diabetic patients: univariate analysis and con‑ presence of a contraindication for its use is linked to a ditional forward stepwise multivariate analysis with metformin as analysis factor. Table S6. Hospital death among metformin patients: univariate higher rate of hospitalization. The lack of randomiza - analysis and conditional forward stepwise multivariate analysis with usual tion of metformin therapy does not indicate whether the contraindication as analysis factor. improvement in observed survival is due to metformin itself or whether the clinical presentation and biological Abbreviations characteristics of patients taking metformin appear to AMP: adenosine monophosphate; ARDS: acute respiratory distress syn‑ be ‘falsely’ more severe. We have included in our logis- drome; AUC: area under the curve; ICU: intensive care unit; KDIGO: Kidney tic regression model certain parameters such as lactate Disease Improving Global Outcome; MALA: metformin‑associated lactic acidosis; MDRD: Modification of Diet in Renal Disease; MET: metformin users; and bicarbonate levels, which are both influenced by the NO‑MET: non‑metformin users; OAD: oral antidiabetic; ROC: receiver operating presence of metformin and most likely do not have the Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 8 of 9 characteristic; RRT: renal replacement therapy; SAPS II: Simplified Acute Physi‑ 4. Arroyo D, Melero R, Panizo N, Goicoechea M, Rodríguez‑Benítez P, Vinuesa ology Score II; SOFA: Sepsis‑Related Organ Failure Assessment score. SG, et al. Metformin‑associated acute kidney injury and lactic acidosis. Int J Nephrol. 2011;2011:749653. Authors’ contributions 5. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal and nonfatal SJ, JEA, CV and MM had the original idea; SJ dealt with the administrative files lactic acidosis with metformin use in type 2 diabetes mellitus. Cochrane and authorizations; SJ, JEA, JC, LVPV, OE, OS and NR collected data; SJ and MM Database Syst Rev. 2010;4:CD002967. performed the statistical analysis; and SJ, JC and CV wrote the manuscript. All 6. Cicero AFG, Tartagni E, Ertek S. Metformin and its clinical use: new insights authors read and approved the final manuscript. for an old drug in clinical practice. Arch Med Sci. 2012;8(5):907–17. 7. Kjelland CB, Djogovic D. The role of serum lactate in the acute care set‑ Authors’ information ting. J Intensive Care Med. 2010;25(5):286–300. First results of this study have been presented by JEA during the 44th Annual 8. Rishu AH, Khan R, Al‑Dorzi HM, Tamim HM, Al‑ Qahtani S, Al‑ Ghamdi G, Congress of the French Intensive Care Society in Paris 2016. et al. Even mild hyperlactatemia is associated with increased mortality in critically ill patients. Crit Care. 2013;17(5):R197. Author details 9. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, Département de Médecine Intensive et Unité de Recherche Clinique, Groupe et al. Surviving sepsis campaign: international guidelines for manage‑ Hospitalier Sud Ile‑de‑France, Hôpital de Melun, 77000 Melun, France. Ser‑ ment of severe sepsis and septic shock, 2012. Intensive Care Med. vice de Réanimation Médicale, AP‑HP, Hôpital Bicêtre, 94270 Le Kremlin‑Bicê ‑ 2013;39(2):165–228. tre, France. Département de Médecine Intensive, Groupe Hospitalier Sud 10. Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Ile‑de‑France, Hôpital de Melun, 77000 Melun, France. Service de Réanima‑ Work Group. KDIGO clinical practice guideline for acute kidney injury. tion Polyvalente, Hôpital de Bethune, 62408 Bethune, France. Kidney Int Suppl. 2012;2:1–138. 11. Romero SP, Andrey JL, Garcia‑Egido A, Escobar MA, Perez V, Corzo R, et al. Acknowledgements Metformin therapy and prognosis of patients with heart failure and new‑ We are indebted to Charles Timoney and Sean A. Freeman for manuscript onset diabetes mellitus. A propensity‑matched study in the community. corrections. Int J Cardiol. 2013;166(2):404–12. 12. Zhang X, Harmsen WS, Mettler TA, Kim WR, Roberts RO, Therneau Competing interests TM, et al. Continuation of metformin use after a diagnosis of cirrhosis SJ received fees from ResMed. JC received fees from Hamilton Medical. CV significantly improves survival of patients with diabetes. Hepatology. received fees from Astute Medical. The remaining authors have disclosed that 2014;60(6):2008–16. they do not have any competing interest. None of these competing interests 13. Harris K, Smith L. Safety and efficacy of metformin in patients with are related to the present manuscript. type 2 diabetes mellitus and chronic hepatitis C. Ann Pharmacother. 2013;47(10):1348–52. Availability of data and materials 14. Inzucchi SE, Lipska KJ, Mayo H, Bailey CJ, McGuire DK. Metformin in The study data will not be shared except in the case of a new study and after patients with type 2 diabetes and kidney disease: a systematic review. agreement of an ethics committee. If needed, please contact corresponding JAMA. 2014;312(24):2668–75. author. 15. Chang C‑H, Sakaguchi M, Dolin P. Epidemiology of lactic acidosis in type 2 diabetes patients with metformin in Japan. Pharmacoepidemiol Drug Consent for publication Saf. 2016;25(10):1196–203. Not applicable. 16. Kamber N, Davis WA, Bruce DG, Davis TME. Metformin and lactic acidosis in an Australian community setting: the Fremantle Diabetes Study. Med J Ethics approval and consent to participate Aust. 2008;188(8):446–9. The study protocol was approved by the French Intensive Care Society 17. Sipahi S, Solak Y, Acikgoz SB, Genc AB, Yildirim M, Yilmaz U, et al. Retro‑ (FICS)—Société de Réanimation de Langue Française (SRLF)—ethical review spective analysis of lactic acidosis‑related parameters upon and after board (Reference Number CE SRLF 15‑29). In this retrospective study, the metformin discontinuation in patients with diabetes and chronic kidney requirement for patients’ consent was waived. disease. Int Urol Nephrol. 2016;48(8):1305–12. 18. Cucchiari D, Podestà MA, Merizzoli E, Calvetta A, Morenghi E, Angelini C, Funding et al. Dose‑related effects of metformin on acid‑base balance and renal No external source of funding. function in patients with diabetes who develop acute renal failure: a cross‑sectional study. Acta Diabetol. 2016;53(4):551–8. 19. DeFronzo R, Fleming GA, Chen K, Bicsak TA. Metformin‑associated Publisher’s Note lactic acidosis: current perspectives on causes and risk. Metabolism. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 2016;65(2):20–9. lished maps and institutional affiliations. 20. Hitchings AW, Archer JRH, Srivastava SA, Baker EH. Safety of metformin in patients with chronic obstructive pulmonary disease and type 2 diabetes Received: 28 February 2017 Accepted: 4 November 2017 mellitus. COPD. 2015;12(2):126–31. 21. Adam WR, O’Brien RC. A justification for less restrictive guidelines on the use of metformin in stable chronic renal failure. Diabet Med J Br Diabet Assoc. 2014;31(9):1032–8. 22. Piel S, Ehinger JK, Elmér E, Hansson MJ. Metformin induces lactate production in peripheral blood mononuclear cells and platelets through References specific mitochondrial complex I inhibition. Acta Physiol Oxf Engl. 1. Gong L, Goswami S, Giacomini KM, Altman RB, Klein TE. Metformin 2015;213(1):171–80. pathways: pharmacokinetics and pharmacodynamics. Pharmacogenet 23. Protti A, Lecchi A, Fortunato F, Artoni A, Greppi N, Vecchio S, et al. Met‑ Genomics. 2012;22(11):820–7. formin overdose causes platelet mitochondrial dysfunction in humans. 2. Eurich DT, Weir DL, Majumdar SR, Tsuyuki RT, Johnson JA, Tjosvold L, Crit Care. 2012;16(5):R180. et al. Comparative safety and effectiveness of metformin in patients with 24. Protti A, Fortunato F, Monti M, Vecchio S, Gatti S, Comi GP, et al. Metformin diabetes mellitus and heart failure: systematic review of observational overdose, but not lactic acidosis per se, inhibits oxygen consumption in studies involving 34,000 patients. Circ Heart Fail. 2013;6(3):395–402. pigs. Crit Care. 2012;16(3):R75. 3. Christiansen CF, Johansen MB, Christensen S, O Brien JM, Tønnesen E, 25. Protti A, Russo R, Tagliabue P, Vecchio S, Singer M, Rudiger A, et al. Sørensen HT. Preadmission metformin use and mortality among inten‑ Oxygen consumption is depressed in patients with lactic acidosis due to sive care patients with diabetes: a cohort study. Crit Care. 2013;17(5):R192. biguanide intoxication. Crit Care. 2010;14(1):R22. Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 9 of 9 26. Liu F, Lu J, Tang J, Li L, Lu H, Hou X, et al. Relationship of plasma creatinine 30. Eppenga WL, Lalmohamed A, Geerts AF, Derijks HJ, Wensing M, Egberts and lactic acid in type 2 diabetic patients without renal dysfunction. Chin A, et al. Risk of lactic acidosis or elevated lactate concentrations in met‑ Med J (Engl). 2009;122(21):2547–53. formin users with renal impairment: a population‑based cohort study. 27. Dykens JA, Jamieson J, Marroquin L, Nadanaciva S, Billis PA, Will Diabetes Care. 2014;37(8):2218–24. Y. Biguanide‑induced mitochondrial dysfunction yields increased 31. Duong JK, Furlong TJ, Roberts DM, Graham GG, Greenfield JR, Williams lactate production and cytotoxicity of aerobically‑poised HepG2 KM, et al. The role of metformin in metformin‑associated lactic acidosis cells and human hepatocytes in vitro. Toxicol Appl Pharmacol. (MALA): case series and formulation of a model of pathogenesis. Drug 2008;233(2):203–10. Saf. 2013;36(9):733–46. 28. Filho RR, Rocha LL, Corrêa TD, Pessoa CMS, Colombo G, Assun‑ 32. Miller RA, Chu Q, Xie J, Foretz M, Viollet B, Birnbaum MJ. Biguanides sup‑ cao MSC. Blood lactate levels cutoff and mortality prediction in press hepatic glucagon signaling by decreasing production of cyclic AMP. sepsis‑time for a reappraisal? A retrospective cohort study. Shock. Nature. 2013;494(7436):256–60. 2016;46(5):480–5. 33. Hung S‑ C, Chang Y‑K, Liu J‑S, Kuo K ‑L, Chen Y ‑H, Hsu C‑ C, et al. Metformin 29. Lepelley M, Giai J, Yahiaoui N, Chanoine S, Villier C. Lactic acidosis in use and mortality in patients with advanced chronic kidney disease: diabetic population: is metformin implicated? Results of a matched case‑ national, retrospective, observational, cohort study. Lancet Diabetes control study performed on the type 2 diabetes population of Grenoble Endocrinol. 2015;3(8):605–14. Hospital University. J Diabetes Res. 2016;2016:3545914. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Intensive Care Springer Journals

Does metformin exposure before ICU stay have any impact on patients’ outcome? A retrospective cohort study of diabetic patients

Loading next page...
 
/lp/springer_journal/does-metformin-exposure-before-icu-stay-have-any-impact-on-patients-jbXFjgyl0X

References (36)

Publisher
Springer Journals
Copyright
Copyright © 2017 by The Author(s)
Subject
Medicine & Public Health; Intensive / Critical Care Medicine; Emergency Medicine; Anesthesiology
eISSN
2110-5820
DOI
10.1186/s13613-017-0336-8
Publisher site
See Article on Publisher Site

Abstract

Background: Impact of metformin exposure before ICU stay remains controversial. Metformin is thought to induce lactic acidosis and haemodynamic instability but may reduce ICU mortality. We evaluated its influence on outcome in diabetic patients admitted in the ICU and then compared two different populations based on the presence of septic shock. Methods: We conducted a retrospective cohort study in a 24‑ bed French ICU between October 2010 and December 2013, including all ICU‑ admitted diabetic patients. Results: Among 635 diabetic patients admitted during the study period, 131 (21%) were admitted with septic shock. Multivariate analysis showed no difference in hospital mortality in all metformin users (OR 0.75 [95% CI 0.44–1.28]; p = 0.29), except in the septic shock subgroup (OR 0.61; 95% CI [0.37–0.99]; p = 0.04) despite higher vasopressor dosages in the first hours after shock onset. Blood lactate level was higher in metformin users than in non ‑ metformin users in all patients (p < 0.001), in septic shock patients (p < 0.001) and in patients without kidney injury (p < 0.001). Metformin users did not have more septic shock from unknown aetiology (p = 0.65) or unknown pathogen (p = 0.99). Conclusions: Metformin use before admission to ICU did not affect in ‑ hospital mortality. However, for patients with septic shock, mortality was lower, despite worse clinical presentation on admission. Blood lactate levels were always higher with or without septic shock and indifferent of kidney function. Keywords: Metformin, Septic shock, Diabetes, Lactic acidosis, ICU [2, 3]. It is thought to induce or worsen lactic acidosis, Background especially in acute renal or liver dysfunction [4]. But in Metformin is increasingly used as an oral antidiabetic a recent meta-analysis pooling 347 trials involving long- (OAD) agent, especially in patients with type 2 diabetes run metformin use, the authors found no case of met- mellitus. Metformin inhibits hepatic glucose production, formin-associated lactic acidosis (MALA), as well as no reduces intestinal glucose absorption and improves glu- difference in blood lactate level related to metformin use cose metabolism [1]. [5]. These results were confirmed in a large cohort of Its use is associated with a reduction in cardiovascu- diabetic patients treated with metformin despite various lar morbidity and mortality, in comparison with insu- metformin contraindications, in which no MALA has lin, other OADs or diet alone, in non-acutely ill patients been described by the authors [2]. In the ICU, MALA has been described in renal, liver, *Correspondence: sebastien.jochmans@gmail.com pulmonary or cardiovascular chronic failure [6], and Département de Médecine Intensive et Unité de Recherche Clinique, Groupe Hospitalier Sud Ile‑de‑France, Hôpital de Melun, 77000 Melun, several case reports described fatal or non-fatal MALA France in acute conditions. In contrast, a recent retrospective Full list of author information is available at the end of the article © The Author(s) 2017. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 2 of 9 study in 17 Danish ICUs found that prior to admission 1. Chronic respiratory failure (previous pulmonary metformin use was associated with a reduction in 30-day function tests, history of acute respiratory decom- mortality [3]. pensation, oxygen or non-invasive ventilation at Our main objective in this study was to evaluate the home, sleep apnoea) and/or influence of pre-admission metformin use on outcome 2. Chronic cardiac failure (history of pulmonary in diabetic ICU patients and in a subgroup experiencing oedema, left ventricular ejection fraction < 45%) and/ septic shock (an acute condition known to induce lac- or tic acidosis [7, 8]). Secondary objectives were to assess 3. Chronic renal disease (calculated creatinine clear- MALA incidence and blood lactate levels in ICU patients ance with Modification of Diet in Renal Disease with diabetes, treated or not by metformin, with or with- [MDRD] < 60 mL/min/1.73 m ) and/or out septic shock. 4. Chronic liver disease (history of cirrhosis, previous INR > 1.2) and/or Methods 5. Myocardial infarction during the previous month We performed a retrospective cohort study in our Inten- sive Care Medicine Department between October 2010 Septic shock was defined according to the Surviving and December 2013. The study protocol was approved Sepsis Campaign definition [9]. Acute kidney injury was by the French Intensive Care Society (FICS)—Société de defined using Kidney Disease Improving Global Out - Réanimation de Langue Française (SRLF)—ethical review come (KDIGO) classification [10] and was considered for board. any stage of the classification. Patients Statistical analysis Inclusion and exclusion criteria Continuous variables were expressed as median [25th– All patients admitted within the study period with a 75th interquartile range] or mean  ±  standard deviation history of diabetes treated by insulin or oral antidiabet- [95% confidence interval] (after Shapiro–Wilks test) and ics were included. So-called diabetic patients treated compared using nonparametric Mann–Whitney (or only with diet were considered as unconfirmed diabetes Student’s t test) and linear regression tests. Categorical and were excluded. The other exclusion criteria were as variables were expressed as n (%) and compared using follows: Chi-square or Fisher’s exact tests. All tests were two- Unknown chronic antidiabetic treatment, modifica - tailed assuming alpha risk = 0.05. All collected data were tions of antidiabetic treatment during the month before analysed in univariate analysis regarding ICU and hospi- ICU admission and unavailable arterial blood gas sample tal survivals. We included in forward and backward step- within 4 h after ICU admission. wise multivariate regression models as covariates all data with p  <  0.1 in univariate analysis, with stratification by Data collection metformin use. We applied these models in ICU patients Collected clinical features were as follows: age, sex, and in the subgroups of septic shock and metformin users height, weight, Simplified Acute Physiology Score II with usual contraindication. We performed a post hoc (SAPS II), main admission cause, metformin contrain- validity assessment of the regression models by receiver dication (‘Definitions’ paragraph below), ICU admis - operating characteristic (ROC) curves, and we selected as sion biomarkers (leucocytes, platelets, haemoglobin, the result the model with the best area under the curve. creatinine, C-reactive protein, bilirubin and/or INR if Results of multivariate regression test were expressed by available), arterial blood gas samples at day 1, all bacte- odds ratio (95% confidence interval). Prognostic value riological tests, vasopressor dosages (close to the initia- of blood lactate level on mortality was tested with ROC tion even outside the ICU), urinary output and amount curves (results expressed by area under the curve [AUC] of intravascular input during the first 24  h, the use of % (95% confidence interval)), sensitivity and sensibility. invasive ventilation and renal replacement therapy, Statistical analysis and graphic representations were the presence of acute respiratory distress syndrome performed with SPSS Statistics V20 software (IBM , (ARDS), ICU and hospital length of stay and vital New York, NY, USA) and Prism 6 software (GraphPad status. Software Inc. , San Diego, CA, USA). Definitions Results Usual metformin contraindications (adapted from the Among the 3871 patients admitted in our ICU during instructions for the use of the medicinal product) were the study period, 635 (16.4%) were finally included (study defined as: flowchart is available in Additional file  1: Figure S1), Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 3 of 9 including 131 (20.6%) patients with septic shock at day 1 Table 1 Cohort of  ICU diabetics: main characteristics at  ICU admission, during  ICU stay and  ICU/hospital out- after ICU admission. come Metformin use before admission was found in 240 patients (37.8%) and was similar regarding occurrence ICU diabetics No metformin Metformin or non-occurrence of septic shock (p  =  0.69). Ratio of N 635 395 (62.2) 240 (37.8) metformin use in patients with one or more usual con- Age (y) 71 [61–79] 73 [62.5–80] 68 [60–78]* traindications was high (119 (49.6%)) with a similar rate Men 408 (64.3) 255 (64.6) 153 (63.8) in septic shock patients (p = 0.54). SAPS II 39 [31‑52] 40 [32‑52] 38 [29‑51] Usual metformin 387 (60.9) 268 (67.9) 119 (49.6)* ICU admission and hospital stay contraindication The main characteristics of ICU diabetics at admission Chronic respiratory 190 (29.9) 132 (33.4) 58 (24.2)* and during ICU or hospital stays are specified in Table  1 insufficiency and Additional file  1: Table S1. In our study cohort, 588 Chronic cardiac 138 (21.7) 92 (23.3) 46 (19.2) (92.6%) patients were admitted for a medical cause, insufficiency mainly for acute respiratory failure (266 (41.9%)). There Chronic liver disease 75 (11.8) 49 (12.4) 26 (10.8) was no difference between metformin users (MET) Chronic kidney 144 (22.7) 128 (32.4) 16 (6.7)* failure and non-metformin users (NO-MET) in the reason for Recent myocardial 8 (1.3) 5 (1.3) 3 (1.3) admission. MET were younger with less chronic res- infarction piratory and renal failures. They had higher blood lac - pH 7.36 [7.28–7.42] 7.36 [7.29–7.43] 7.36 [7.27–7.42] tate level (p  <  0.001), lower bicarbonate (p  <  0.01) and PaCO (mmHg) 36 [29–43] 37 [30–44] 36 [28–43] also lower serum creatinine (p  <  0.001) with less acute HCO (mmHg) 21.3 [17–25.2] 21.9 [17.5–26] 20.4 [15.3–24]* kidney injury (p  <  0.001). Severity score (SAPS II) and Lactate (mmol/L) 1.4 [0.9–2.4] 1.2 [0.8–2.1] 1.8 [1.1–3.9]* need in organ support (i.e. invasive mechanical ventila- INR 1.25 [1.06–1.71] 1.26 [1.06–1.65] 1.24 [1.07–1.77] tion, vasopressor, renal replacement therapy) were simi- Bilirubin (µmol/L) 10 [7–16] 10 [7–16] 10 [7–16] lar. Among MET, there was no difference in lactate level C‑reactive protein 34 [8–115] 35 [8–115] 32 [8–115] between patients with or without usual contraindication (mg/L) (p = 0.86) (Additional file 1: Table S2). Haemoglobin (g/dL) 11.2 [9.6–13] 11.1 [9.6–12.7] 11.7 [9.7–13.4] The main characteristics for diabetics with septic shock Leucocytes (G/L) 11.2 [8.1–15.3] 10.8 [7.5–14.6] 11.7 [8.4–16.3] at admission and during ICU or hospital stays are speci- Platelets (G/L) 213 [155–277] 207 [155–271] 219 [157–293] fied in Table  2 and Additional file  1: Tables S3 and S4. Creatinine (µmol/L) 131 [85–238] 153 [90–285] 108 [80–174]* Aetiologies of shock are specified in Additional file  1: Acute kidney injury 392 (61.7) 268 (67.8) 124 (51.7)* Table S3. There was no difference between MET and Renal replacement 113 (17.8) 72 (18.2) 41 (17.1) NO-MET regarding unknown aetiology (p  =  0.65) and therapy unknown pathogen (p  =  0.99) (Additional file  1: Table Vasopressors 229 (36.1) 136 (34.4) 93 (38.8) S4). MET with septic shock had higher blood lactate Invasive ventilation 230 (36.2) 139 (35.2) 91 (37.9) than NO-MET at admission (p  <  0.001) and during the ICU length of stay 6 [3–10] 6 [3.5–10] 6 [3–9] (d) first 12 h (Fig.  1). Bicarbonate was lower (p < 0.01). They ICU death 117 (18.4) 75 (19) 42 (17.5) also received more renal replacement therapy (p = 0.02), Hospital length of 12 [6–23] 12 [6–23] 13 [7–23] while they had less chronic renal failure and there was no stay (d) significant difference in serum creatinine, pH, day 1 uri - Hospital death 140 (22) 92 (23.3) 48 (20) nary output or acute kidney injury occurrence. In MET, Values are n (%) or median [IQR 25th–75th] there was a linear correlation between blood lactate * p < 0.05 between metformin and no metformin and serum creatinine (ρ  =  0.36; p  <  0.01) in contrast to NO-MET (ρ =  0.09; p = 0.41) (Additional file  1: Figures S2 and S3). However, lactate was even higher in MET (p  <  0.001) with normal kidney function (MDRD creati- (p  =  0.09). Vasopressor dose was significantly higher nine clearance > 60 mL/min/1.73 m ). in MET the first hours after reaching criteria for septic Septic shock severity can also be evaluated by the shock (Fig. 2). amount of vascular filling and the dose of vasopressors. There was no difference in the number of patients with Mortality and length of stay intensive vascular filling (i.e. more than 50  mL/kg/day) ICU or hospital lengths of stay as well as ICU death between MET and NO-MET, but there was a statistical showed no statistically significant difference between trend for higher maximal dose of noradrenaline in MET MET and NO-MET in the cohort of diabetics and in the Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 4 of 9 Table 2 Subgroup of ICU diabetics with septic shock: main characteristics at ICU admission, during ICU stay and ICU/hos- pital outcome Septic shocks No metformin Metformin N 131 79 (60.3) 52 (39.7) Age (y) 70 [63–78] 71 [64–78] 66 [61–78] Men 89 (67.9) 56 (70.9) 33 (63.5) SAPS II 52 [42–69] 48 [40–68] 57 [46–68] Usual metformin contraindication 79 (60.3) 56 (70.9) 23 (44.2)* Chronic respiratory failure 30 (22.9) 21 (26.6) 9 (17.3) Chronic cardiac failure 27 (20.6) 20 (25.3) 7 (13.5) Chronic liver disease 26 (19.8) 18 (22.8) 8 (15.4) Chronic renal failure 19 (14.5) 16 (20.3) 3 (5.8)* Recent myocardial infarction 1 (0.8) 1 (1.3) 0 pH 7.32 [7.2–7.38] 7.32 [7.23–7.39] 7.26 [7.17–7.38] PaCO (mmHg) 34 [27–42] 35 [29–43] 34 [24–42] HCO (mmHg) 18.2 [13.3–22.2] 19.7 [14.7–24.1] 15.5 [10.1–19.9]* Lactate (mmol/L) 2.2 [1.1–5] 1.4 [1–2.8] 4.5 [2.1–8.7]* INR 1.5 [1.2–2.3] 1.6 [1.3–2.9] 1.4 [1.1–1.9] Bilirubin (µmol/L) 12 [8–24] 13 [8–26] 10 [8–19] C‑reactive protein (mg/L) 95 [24–224] 98 [30–225] 85 [14–212] Haemoglobin (g/dL) 10.6 [9.1–12.4] 10.7 [9.3–12.4] 10.5 [9.1–12.4] Leucocytes (G/L) 12.1 [8.3–19.6] 11.9 [8.5–19] 12.9 [8.4–21.6] Platelets (G/L) 185 [119–265] 199 [120–273] 173 [118–252] Creatinine (µmol/L) 167 [113–326] 163 [108–276] 176 [123–364] Urinary output day 1 (mL) 1200 [553–2200] 1200 [558–1925] 1425 [443–2400] Number of patients with vascular filling > 50 mL/kg ≥ 1 day 76 (60.3) 42 (58.4) 34 (69.4) Maximum dose of noradrenaline (mg/h) 2 [1–4.3] 2 [1–3.5] 3.5 [1.3–5]* (µg/kg/min) 0.43 [0.22–0.95] 0.4 [0.21–0.76] 0.61 [0.23–1.16] Maximum dose of adrenaline (mg/h) 2.5 [1.5–6] 3 [1.5–6.3] 2.5 [1.4–6] (µg/kg/min) 0.61 [0.25–1.22] 0.52 [0.22–1.3] 0.66 [0.27–0.98] Noradrenaline duration (h) 39 [18–64] 48 [19–71] 36 [15–59] Adrenaline duration (h) 36 [9–90] 36 [14–90] 30 [6–102] Vasopressor duration (h) 48 [24–96] 48 [24–97] 36 [23–72] Acute kidney injury 104 (79.4) 62 (78.5) 42 (80.8) ARDS 48 (36.6) 27 (34.2) 21 (40.4) Renal replacement therapy 51 (38.9) 24 (30.4) 27 (51.9)* Invasive ventilation 96 (73.3) 56 (70.9) 40 (76.9) ICU length of stay (d) 9 [5–16] 9 [6–19] 7 [4–13] Hospital length of stay (d) 15 [7–29] 15 [8–29] 16 [4–26] ICU death 51 (38.9) 31 (39.2) 20 (38.5) Hospital death 53 (40.5) 33 (41.8) 20 (38.5) Values are n (%) or median [IQR 25th–75th] * p < 0.05 between metformin and no metformin subgroup of septic shock patients. Hospital death was not multivariate analysis with odds ratio 0.61 [95% CI 0.36– significantly different in multivariate regression model 0.99]; p = 0.049 (Table 3). analysis (OR 0.75 [0.44–1.28]; p  =  0.29) (Additional Blood lactate levels showed a prognostic value in MET file  1: Table S5). In the subgroup of septic shock patients, (AUC 67.3% (95% CI 58.3–76.4); p  =  0.001) and NO- metformin was associated with a lower mortality after MET (AUC 68.6% (61.5–75.8); p  <  0.001) of the cohort Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 5 of 9 Fig. 1 Initial evolution of lactate level in ICU diabetics sustaining septic shock with or without pre‑admission metformin treatment. T0: time of septic shock diagnosis. Abscissa axis is log 10 scale. *p < 0.05 Fig. 2 Initial evolution of vasopressor dosage in ICU diabetics sustaining septic shock with or without pre‑admission metformin treatment. T0: time of septic shock diagnosis. Abscissa axis is log 10 scale. *p < 0.05 and also in MET (AUC 66.7% (51.5–81.9); p = 0.05) and NO-MET (AUC 65.5% (53–78.1); p  =  0.02) of shocked Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 6 of 9 Table 3 Hospital death among  septic shock patients: univariate analysis and  conditional forward stepwise multivariate analysis with metformin as analysis factor Survivor Non-survivor P-univariate Odds ratio P-multivariate N 80 (61.1) 51 (38.9) – – – Men 49 (61.3) 40 (78.4) 0.055 NS NS SAPS II 49 [40–61] 65 [46–80] 0.001 1.05 (1.04–1.07) < 0.001 Metformin (n) 32 (40) 20 (39.2) 1 0.61 (0.37–0.99) 0.049 Lactate (mmol/L) 1.7 [1–3.9] 3.2 [1.4–7.1] 0.003 1.21 (1.1–1.34) < 0.001 ARDS (n) 23 (28.8) 25 (49) 0.03 NS NS RRT (n) 24 (30) 27 (52.9) 0.011 NS NS Invasive ventilation (n) 48 (60) 48 (94.1) < 0.001 NS NS Urinary output day 1 (mL) 1400 [675–2400] 1030 [65–1900] 0.03 NS NS α = 0.05. Area under the curve of the multivariate model = 0.786 RRT renal replacement therapy, NS not significant patients. But prognostic cut-off value for lactate with Propensity-score-matched analyses yielded the same the highest sensitivity and specificity was higher in MET results. In our work, more than 90% were medical admis- (2.15  mmol/L, sensitivity 65%, specificity 61.6%) than sions, whereas two-thirds of the 7404 ICU patients with in NO-MET (1.35, sensitivity 66.2%, specificity 61.3%). type 2 diabetes in Christiansen et al.’s study were surgical Likewise in the subgroup of septic shock patients, cut-off admissions. However, no data were available concerning values were 4.45  mmol/L (sensitivity 57.9%, specificity septic shocks, vasopressor dosages or even blood lac- 56.7%) versus 1.45  mmol/L (sensitivity 58.1%, specificity tate levels. Mechanisms of this beneficial effect remain 56.2%), respectively. unclear: in ICU patients, metformin may supply higher Among MET, there was no significant difference in amounts of lactate serving as an energetic carbon source hospital death between patients with or without usual and therefore is available for ischaemic tissues with glu- contraindication (OR 1.24 [0.48–3.2]; p  =  0.66) (Addi- cose preservation. Metformin may also decrease cellular tional file 1: Table S6). hypoxia of less perfused tissues by decreasing oxygen consumption. Discussion However, clinical severity seems higher in MET. Lac- In our large cohort on critically ill diabetic patients, met- tate levels are significantly higher in ICU diabetics with formin use before admission to ICU did not affect in- or without septic shock (Additional file  1: Figure S4). This hospital mortality; however, pre-admission metformin issue still remains controversial with studies finding no treatment was independently associated with a decrease effect of metformin on lactate rate [5, 15, 16] or, on the in hospital mortality in the group of septic shock contrary, finding an increased lactate [17–27]. One rea - patients, even with an initial clinical presentation appear- son for this discrepancy may be that ICU patients, unlike ing more severe. Indeed, independent of kidney func- other patients, suffer acute stress with endogenous cat - tion, vasopressor dosages and serum lactate levels were echolamine release leading to increased lactate levels higher during the first hours after shock onset in MET. through adrenergic receptor stimulation. Physiological Nevertheless, metformin did not seem to induce shock studies showed that metformin enhances lactate pro- per se because there was no more septic shock from duction and decreases oxygen consumption [23–25] by unknown aetiology or unknown pathogens in MET than inhibiting mitochondrial chain complexes [19, 22–24, in NO-MET. 27]. Therefore, in our study, prognostic cut-off values are A beneficial association between metformin and mor - higher in MET, especially when there is a septic shock, tality has been already described both in selected patients as previously found [28]. It is usually admitted that lac- with chronic heart failure [11], liver disease [12, 13], tic acidosis in metformin users is due to a reduced renal mild-to-moderate kidney failure [14] which are usual drug clearance. Lactate and creatinine levels (and cre- contraindications, and in ICU patients [3]. In this lat- atinine clearance) are linearly correlated in our study ter study, based on retrospective analysis of Northern as previously shown [17, 18, 21, 26, 29–31]. But lactate Denmark database, 30-day mortality was lower in met- levels remain higher in patients without kidney injury formin users than in non-metformin users with adjusted with metformin than without. This last issue was only hazard ratio  =  0.8 (95% confidence interval 0.71–0.95). previously described in case reports and one cohort Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 7 of 9 study [26], although another study failed to find hyper - same prognostic value in patients previously untreated lactatemia when kidney function was normal [29]. MET by metformin. Similarly, elevated doses of vasopressors, probably received more haemodialysis for the purpose which are used as a criterion for poor outcome for exam- of either correcting deeper hypobasemia or eliminating ple in the SOFA score, may not carry the same prog- plasma metformin. nostic significance. Metformin blood dosage has never been performed. However, it seems linearly correlated to lactate concentration [18, 21, 31]. Lastly, comparison • Vasopressor dosages are higher in septic shock diabet- between MET treated or non-treated by renal replace- ics with pre-admission metformin. This increase in ment therapy was unfeasible because analysis would catecholamines need, which has not been previously lack power and be statistically unreliable. If current sci- described, is not due to acidosis per se because pH val- entific opinion suggests its use in metformin overdose, ues are similar with or without metformin. Recent data there is no strong proof. There is indeed a contradiction suggest that metformin decreases adenylate cyclase between studies finding a beneficial association between activity and therefore cyclic AMP concentration [32]. sepsis and metformin and in contrast the desire to elimi- The effects of vasopressors are mediated by adrenergic nate metformin by haemodialysis. Therefore, we suggest receptors, G protein and adenylate cyclase stimulations that future studies should seek to answer two ques- leading to an increase in cyclic AMP concentration. It tions: Is there a benefit in giving metformin during the is assumed that it is necessary to increase vasopressor first hours of septic shock in diabetic patients previously dosages in order to obtain the same haemodynamic untreated by metformin? Is there really a benefit in the effect and compensate decreased adenylate cyclase early elimination of metformin by haemodialysis in dia- activity induced by metformin. Indeed, metformin betic patients with septic shock and without acute kid- does not seem to produce sepsis-like shock because ney injury? there is as much septic shock of unknown aetiology or germ in MET than in NO-MET. However, metformin Conclusions actually seems to worsen the criteria usually used to Metformin use before admission to ICU is associated assess the severity of septic shocks. with a decrease in mortality in septic shock patients despite a worse clinical presentation on admission. Met- Finally, in our study, patients treated with metformin formin users have higher lactate levels independent of despite the presence of the usual contraindications do kidney function and need higher vasopressor dosages not have higher lactate levels. The mortality rate is not during the first hours of septic shock. Metformin does increased either. These contraindications have been chal - not seem to induce shock per se. The presence or absence lenged for several years so that metformin seems delete- of one of the usual contraindications to taking metformin rious only in terminal kidney disease [33]. Our collected does not alter lactate levels or hospital mortality. data did not allow us to evaluate outcome according to the intensity of each organ failure. It is possible that our Additional file patients had mainly mild-to-moderate lung, liver, heart or kidney injury that would be insufficient to worsen out - Additional file 1: Figure S1. Study flowchart. Figure S2. Linear regres‑ come or lactate level. sion between blood creatinine and lactate levels in metformin users Our study is subject to certain limitations. First, it is patients. Figure S3. Linear regression between blood creatinine and lactate levels in non‑metformin users patients. Figure S4. Lactate levels. a retrospective study, avoiding observation bias, but Table S1. Main admission pattern of ICU‑admitted diabetics. Table S2. with selection bias due to non-inclusion of patients ICU‑admitted diabetics with preadmission metformin treatment with or with missing data. Thus, we cannot determine whether without usual metformin contraindication. Table S3. Aetiologies and germs responsible for septic shocks in ICU‑ diabetics. Table S4. Septic metformin users are more likely to be admitted to ICU shocks without aetiology at the end of hospital stay. Table S5. Hospital than other antidiabetics’ takers, and also whether the death among ICU‑admitted diabetic patients: univariate analysis and con‑ presence of a contraindication for its use is linked to a ditional forward stepwise multivariate analysis with metformin as analysis factor. Table S6. Hospital death among metformin patients: univariate higher rate of hospitalization. The lack of randomiza - analysis and conditional forward stepwise multivariate analysis with usual tion of metformin therapy does not indicate whether the contraindication as analysis factor. improvement in observed survival is due to metformin itself or whether the clinical presentation and biological Abbreviations characteristics of patients taking metformin appear to AMP: adenosine monophosphate; ARDS: acute respiratory distress syn‑ be ‘falsely’ more severe. We have included in our logis- drome; AUC: area under the curve; ICU: intensive care unit; KDIGO: Kidney tic regression model certain parameters such as lactate Disease Improving Global Outcome; MALA: metformin‑associated lactic acidosis; MDRD: Modification of Diet in Renal Disease; MET: metformin users; and bicarbonate levels, which are both influenced by the NO‑MET: non‑metformin users; OAD: oral antidiabetic; ROC: receiver operating presence of metformin and most likely do not have the Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 8 of 9 characteristic; RRT: renal replacement therapy; SAPS II: Simplified Acute Physi‑ 4. Arroyo D, Melero R, Panizo N, Goicoechea M, Rodríguez‑Benítez P, Vinuesa ology Score II; SOFA: Sepsis‑Related Organ Failure Assessment score. SG, et al. Metformin‑associated acute kidney injury and lactic acidosis. Int J Nephrol. 2011;2011:749653. Authors’ contributions 5. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal and nonfatal SJ, JEA, CV and MM had the original idea; SJ dealt with the administrative files lactic acidosis with metformin use in type 2 diabetes mellitus. Cochrane and authorizations; SJ, JEA, JC, LVPV, OE, OS and NR collected data; SJ and MM Database Syst Rev. 2010;4:CD002967. performed the statistical analysis; and SJ, JC and CV wrote the manuscript. All 6. Cicero AFG, Tartagni E, Ertek S. Metformin and its clinical use: new insights authors read and approved the final manuscript. for an old drug in clinical practice. Arch Med Sci. 2012;8(5):907–17. 7. Kjelland CB, Djogovic D. The role of serum lactate in the acute care set‑ Authors’ information ting. J Intensive Care Med. 2010;25(5):286–300. First results of this study have been presented by JEA during the 44th Annual 8. Rishu AH, Khan R, Al‑Dorzi HM, Tamim HM, Al‑ Qahtani S, Al‑ Ghamdi G, Congress of the French Intensive Care Society in Paris 2016. et al. Even mild hyperlactatemia is associated with increased mortality in critically ill patients. Crit Care. 2013;17(5):R197. Author details 9. Dellinger RP, Levy MM, Rhodes A, Annane D, Gerlach H, Opal SM, Département de Médecine Intensive et Unité de Recherche Clinique, Groupe et al. Surviving sepsis campaign: international guidelines for manage‑ Hospitalier Sud Ile‑de‑France, Hôpital de Melun, 77000 Melun, France. Ser‑ ment of severe sepsis and septic shock, 2012. Intensive Care Med. vice de Réanimation Médicale, AP‑HP, Hôpital Bicêtre, 94270 Le Kremlin‑Bicê ‑ 2013;39(2):165–228. tre, France. Département de Médecine Intensive, Groupe Hospitalier Sud 10. Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Ile‑de‑France, Hôpital de Melun, 77000 Melun, France. Service de Réanima‑ Work Group. KDIGO clinical practice guideline for acute kidney injury. tion Polyvalente, Hôpital de Bethune, 62408 Bethune, France. Kidney Int Suppl. 2012;2:1–138. 11. Romero SP, Andrey JL, Garcia‑Egido A, Escobar MA, Perez V, Corzo R, et al. Acknowledgements Metformin therapy and prognosis of patients with heart failure and new‑ We are indebted to Charles Timoney and Sean A. Freeman for manuscript onset diabetes mellitus. A propensity‑matched study in the community. corrections. Int J Cardiol. 2013;166(2):404–12. 12. Zhang X, Harmsen WS, Mettler TA, Kim WR, Roberts RO, Therneau Competing interests TM, et al. Continuation of metformin use after a diagnosis of cirrhosis SJ received fees from ResMed. JC received fees from Hamilton Medical. CV significantly improves survival of patients with diabetes. Hepatology. received fees from Astute Medical. The remaining authors have disclosed that 2014;60(6):2008–16. they do not have any competing interest. None of these competing interests 13. Harris K, Smith L. Safety and efficacy of metformin in patients with are related to the present manuscript. type 2 diabetes mellitus and chronic hepatitis C. Ann Pharmacother. 2013;47(10):1348–52. Availability of data and materials 14. Inzucchi SE, Lipska KJ, Mayo H, Bailey CJ, McGuire DK. Metformin in The study data will not be shared except in the case of a new study and after patients with type 2 diabetes and kidney disease: a systematic review. agreement of an ethics committee. If needed, please contact corresponding JAMA. 2014;312(24):2668–75. author. 15. Chang C‑H, Sakaguchi M, Dolin P. Epidemiology of lactic acidosis in type 2 diabetes patients with metformin in Japan. Pharmacoepidemiol Drug Consent for publication Saf. 2016;25(10):1196–203. Not applicable. 16. Kamber N, Davis WA, Bruce DG, Davis TME. Metformin and lactic acidosis in an Australian community setting: the Fremantle Diabetes Study. Med J Ethics approval and consent to participate Aust. 2008;188(8):446–9. The study protocol was approved by the French Intensive Care Society 17. Sipahi S, Solak Y, Acikgoz SB, Genc AB, Yildirim M, Yilmaz U, et al. Retro‑ (FICS)—Société de Réanimation de Langue Française (SRLF)—ethical review spective analysis of lactic acidosis‑related parameters upon and after board (Reference Number CE SRLF 15‑29). In this retrospective study, the metformin discontinuation in patients with diabetes and chronic kidney requirement for patients’ consent was waived. disease. Int Urol Nephrol. 2016;48(8):1305–12. 18. Cucchiari D, Podestà MA, Merizzoli E, Calvetta A, Morenghi E, Angelini C, Funding et al. Dose‑related effects of metformin on acid‑base balance and renal No external source of funding. function in patients with diabetes who develop acute renal failure: a cross‑sectional study. Acta Diabetol. 2016;53(4):551–8. 19. DeFronzo R, Fleming GA, Chen K, Bicsak TA. Metformin‑associated Publisher’s Note lactic acidosis: current perspectives on causes and risk. Metabolism. Springer Nature remains neutral with regard to jurisdictional claims in pub‑ 2016;65(2):20–9. lished maps and institutional affiliations. 20. Hitchings AW, Archer JRH, Srivastava SA, Baker EH. Safety of metformin in patients with chronic obstructive pulmonary disease and type 2 diabetes Received: 28 February 2017 Accepted: 4 November 2017 mellitus. COPD. 2015;12(2):126–31. 21. Adam WR, O’Brien RC. A justification for less restrictive guidelines on the use of metformin in stable chronic renal failure. Diabet Med J Br Diabet Assoc. 2014;31(9):1032–8. 22. Piel S, Ehinger JK, Elmér E, Hansson MJ. Metformin induces lactate production in peripheral blood mononuclear cells and platelets through References specific mitochondrial complex I inhibition. Acta Physiol Oxf Engl. 1. Gong L, Goswami S, Giacomini KM, Altman RB, Klein TE. Metformin 2015;213(1):171–80. pathways: pharmacokinetics and pharmacodynamics. Pharmacogenet 23. Protti A, Lecchi A, Fortunato F, Artoni A, Greppi N, Vecchio S, et al. Met‑ Genomics. 2012;22(11):820–7. formin overdose causes platelet mitochondrial dysfunction in humans. 2. Eurich DT, Weir DL, Majumdar SR, Tsuyuki RT, Johnson JA, Tjosvold L, Crit Care. 2012;16(5):R180. et al. Comparative safety and effectiveness of metformin in patients with 24. Protti A, Fortunato F, Monti M, Vecchio S, Gatti S, Comi GP, et al. Metformin diabetes mellitus and heart failure: systematic review of observational overdose, but not lactic acidosis per se, inhibits oxygen consumption in studies involving 34,000 patients. Circ Heart Fail. 2013;6(3):395–402. pigs. Crit Care. 2012;16(3):R75. 3. Christiansen CF, Johansen MB, Christensen S, O Brien JM, Tønnesen E, 25. Protti A, Russo R, Tagliabue P, Vecchio S, Singer M, Rudiger A, et al. Sørensen HT. Preadmission metformin use and mortality among inten‑ Oxygen consumption is depressed in patients with lactic acidosis due to sive care patients with diabetes: a cohort study. Crit Care. 2013;17(5):R192. biguanide intoxication. Crit Care. 2010;14(1):R22. Jochmans et al. Ann. Intensive Care (2017) 7:116 Page 9 of 9 26. Liu F, Lu J, Tang J, Li L, Lu H, Hou X, et al. Relationship of plasma creatinine 30. Eppenga WL, Lalmohamed A, Geerts AF, Derijks HJ, Wensing M, Egberts and lactic acid in type 2 diabetic patients without renal dysfunction. Chin A, et al. Risk of lactic acidosis or elevated lactate concentrations in met‑ Med J (Engl). 2009;122(21):2547–53. formin users with renal impairment: a population‑based cohort study. 27. Dykens JA, Jamieson J, Marroquin L, Nadanaciva S, Billis PA, Will Diabetes Care. 2014;37(8):2218–24. Y. Biguanide‑induced mitochondrial dysfunction yields increased 31. Duong JK, Furlong TJ, Roberts DM, Graham GG, Greenfield JR, Williams lactate production and cytotoxicity of aerobically‑poised HepG2 KM, et al. The role of metformin in metformin‑associated lactic acidosis cells and human hepatocytes in vitro. Toxicol Appl Pharmacol. (MALA): case series and formulation of a model of pathogenesis. Drug 2008;233(2):203–10. Saf. 2013;36(9):733–46. 28. Filho RR, Rocha LL, Corrêa TD, Pessoa CMS, Colombo G, Assun‑ 32. Miller RA, Chu Q, Xie J, Foretz M, Viollet B, Birnbaum MJ. Biguanides sup‑ cao MSC. Blood lactate levels cutoff and mortality prediction in press hepatic glucagon signaling by decreasing production of cyclic AMP. sepsis‑time for a reappraisal? A retrospective cohort study. Shock. Nature. 2013;494(7436):256–60. 2016;46(5):480–5. 33. Hung S‑ C, Chang Y‑K, Liu J‑S, Kuo K ‑L, Chen Y ‑H, Hsu C‑ C, et al. Metformin 29. Lepelley M, Giai J, Yahiaoui N, Chanoine S, Villier C. Lactic acidosis in use and mortality in patients with advanced chronic kidney disease: diabetic population: is metformin implicated? Results of a matched case‑ national, retrospective, observational, cohort study. Lancet Diabetes control study performed on the type 2 diabetes population of Grenoble Endocrinol. 2015;3(8):605–14. Hospital University. J Diabetes Res. 2016;2016:3545914.

Journal

Annals of Intensive CareSpringer Journals

Published: Dec 2, 2017

There are no references for this article.