Effect of metformin on survival rate in experimental sepsis

ORIG

I N A L

A R T I C L E

Effect of metformin on survival rate in experimental sepsis V Gras1, B Bouffandeau2, Ph Montravers3, JD Lalau4

SUMMARY

RÉSUMÉ

Aim: Because “metformin-associated lactic acidosis” refers to metformin and concurrent pathologies as co-precipitating factors, the respective impact in the outcome of metformin therapy, metformin accumulation, and general diseases should be determined. We therefore constructed a model of sepsis in mice treated with metformin at a dose corresponding to clinical practice, or to accumulation. Methods: 460 mice were separated in 3 groups: no metformin therapy, a 7-day metformin therapy at 50 mg.kg-1.day-1 (MET50) or 500 mg.kg-1.day-1 (MET500). Blood was drawn on day 7 in 40 metformin-treated animals for determining metformin concentrations. The 420 other mice were divided in 14 subgroups according to the amount of an intra-peritoneal inoculum of E. coli ranging from 5.103 to 1010 CFU/ml in order to construct a lethal dose curve. The survival rate was assessed at 7, 13, 24, 36, 60 and 120 hours thereafter. Results: Plasma metformin concentrations were 0.26 ± 0.13 mg/l in MET50, and 4.63 ± 1.92 mg/l in MET500. The comparative analysis of the survival rates at 120 hours showed no difference of mortality, always occurring for an inoculum amount > 108 CFU/ml. Comparing the survival rates from time 0 to 120 hours using Kaplan-Meyer curves and the Logrank test, there was no difference between the different groups. Conclusion: Metformin, even at a dose mimicking accumulation, does not aggravate the mortality rate in this model of sepsis. Consequently, metformin can not be considered as toxic in such a condition.

Effet de la metformine sur la survie dans un modèle expérimental de sepsis

Key-words: Type 2 diabetes · Metformin · Lactic acidosis · Sepsis. Gras V, Bouffandeau B, Montravers Ph, Lalau JD. Effect of metformin on survival rate in experimental sepsis Diabetes Metab 2006;32:147-150

Objectif : Parce que la dénomination « acidose lactique associée à la metformine » implique la metformine autant que des pathologies associées dans les facteurs précipitants, la part respective dans le pronostic du traitement par metformine, de l’accumulation de metformine, et des affections générales doit être déterminée. Nous avons ainsi construit un modèle de sepsis chez la souris traitée par metformine à une dose correspondant soit à la clinique, soit à une accumulation. Méthodes : 460 souris ont été réparties en 3 groupes : pas de traitement, un traitement par metformine de 7 jours à la dose de 50 mg.kg-1.jour-1 (MET50) ou de 500 mg.kg-1.jour-1 (MET500). Du sang a été prélevé au septième jour chez 40 animaux traités pour mesurer la metforminémie. Les 420 autres animaux ont été divisés en 14 sous-groupes en fonction du taux d’un inoculum intra-péritonéal de E. coli variant de 5.103 to 1010 CFU/ml afin de construire une courbe dose-léthale. Le taux de survie a été apprécié à 7, 13, 24, 36, 60 et 120 heures après cette injection. Résultats : Les metforminémies étaient à 0,26 ± 0,13 mg/l pour MET50, et 4,63 ± 1,92 mg/l pour MET500. L’analyse comparative du taux de survie à 120 heures n’a pas montré de différence de mortalité, survenant dans chaque groupe pour un taux d’inoculum > 108 CFU/ml. La comparaison des taux de survie des temps 0 à 120 heures avec les courbes de Kaplan-Meyer et le test de Logrank n’a pas fait apparaître de différence entre les différents groupes. Conclusion : La metformine, même à une dose simulant une accumulation, n’aggrave pas la mortalité dans ce modèle de sepsis. En conséquence, la metformine ne peut pas être considérée comme toxique dans une telle situation.

Mots-clés : Diabète de type 2 · Metformine · Acidose lactique · Sepsis.

1

Service de Pharmacologie Clinique, Hôpital Sud, Amiens. 2 Service de Réanimation Médicale, Centre hospitalier, Elbeuf. 3 Service de Réanimation Médicale, Hôpital Bichat, Paris. 4 Service d’Endocrinologie-Nutrition, Hôpital Sud, Amiens.

Address correspondence and reprint requests to: JD Lalau. Hôpital Sud, 80054 Amiens Cedex 1, France. Lalau.Jean-Daniel @chu-amiens.fr Received: July 23 th, 2005; accepted: November 7th, 2005.

Diabetes Metab 2006;32:147-150 • © 2006 Masson, all rights reserved

147

V Gras et al.

Introduction Metformin is recognised as a leading drug in the treatment of type 2 diabetes. Although rarely, metformin, when accumulated, has been associated with lactic acidosis, a condition still considered as having a poor prognosis [1-4]. However, given that metformin use but also, and more frequently, coincidental pathologies, such as myocardial infarction, sepsis, haemorrhage, etc, may lead to lactic acidosis, there is a great confusion about the respective impact of metformin and of concurrent pathologies in the vital outcome [5,6]. Furthermore, in the true “metformin-associated lactic acidosis”, referring to metformin and concurrent pathologies as co-precipitating factors of lactic acidosis, plasma metformin concentrations varied largely, according to whether metformin therapy was stopped or not, renal failure was present or not, and renal failure was long-standing enough to lead to metformin accumulation or not [6]. It would therefore seem important to determine whether use of metformin, with or without accumulation, contributes to outcome. We constructed for this purpose an experimental model with a common general cause of lactic acidosis, namely sepsis, in animals treated with metformin at two dosages, corresponding to either the clinical daily dose, or to accumulation.

Material and methods

Metformin administration Mice were treated with metformin via drinking water during 7 days. Knowing a daily drink intake of 5 ml per animal, metformin powder was added in the drinking solution in order to obtain the two required concentrations’ of 50 mg.kg-1.day-1 and 500 mg.kg-1.day-1.

Validation of the metformin dose Blood was drawn on day 7 by intracardiac puncture in twenty mice in the metformin-treated groups for measurement of plasma metformin concentrations.

Experimental septis Preparation of the microorganisms

A strain of E. coli, CB1496, was used. E. coli were grown in brain-heart infusion broth. The final inoculum was made when bacteria were in the log phase of growth according to the procedure previously described by one of us [7,8]. The bacteria were diluted to give the number of microorganisms required for bacterial challenge. Assessment of purity and validity counts were carried out immediately before inoculation. Fourteen different amounts of inoculum were selected, ranging from 5.103 to 1010 colony forming unit (CFU) per ml, i.e. to deliver an inoculum of, respectively, 5.103, 104, 5.104, 105, 5.105, 106, 5.106, 107, 5.107, 108, 5.108, 109, 5.109, 1010 CFU per ml.

Study design We studied three groups of mice according to three metformin treatments: no treatment (MET 0), treatment at a dose equivalent to clinical practice (≈50 mg.kg-1.day-1: MET 50) or to accumulation (≈500 mg.kg-1.day-1: MET 500). In a first experiment, blood was drawn in 20 animals in each group treated with metformin for the measurement of plasma metformin concentrations to ensure that the selection dose was appropriate. In a second experiment, a total number of 420 mice were divided into the three experimental groups, and further subdivided in 14 subgroups of 10 animals each according to the amount of an intraperitoneal inoculum of Escherichia coli (E. coli). The amounts ranging from 5.103 to 1010 CFU per ml were selected in order to construct lethal dose curves and, ultimately, to allow the comparison of the survival rate between the experimental groups.

Animals Experiments were performed on Swiss pathogen-free mice (Charles River, Paris, France), in the mean body weight of 25 g. They were fed on a standard laboratory chow until the end of the experiment. The procedures followed were in agreement with the guidelines of the European regulation for the conduct of animal experimentation. 148

Diabetes Metab 2006;32:147-150 • www.masson.fr/revues/dm

Intra-peritoneal injection and assessment of survival rate

Each animal received a 1 ml-intra-peritoneal inoculum of microorganisms. After the inoculation, the animals were returned to their cages. The survival rate was assessed at 7, 13, 24, 36, 60 and 120 hours after the injection. Lethal dose curves

Lethal dose curves were plotted and LD50 was calculated according to the method described by Ike et al [9].

Analytical technique Plasma metformin concentration was measured in duplicate in the same laboratory with high-performance liquid chromatography according to the method described by Lacroix [10].

Statistical analysis Results are presented as means ± SEM. Data by combined Kaplan-Meyer curves and the Logrank test in the experimental groups with an amount of inoculum of E. coli ≥108 CFU/ml (i.e. ≥ the threshold of mortality) were compared.

Metformin and survival in experimental sepsis

Results

Survival (percentage)

Plasma metformin concentrations All the animals were surviving at day 7, when plasma metformin concentrations were determined. The mean concentrations were 0.26 ± 0.13 mg/l (2 ± 1 µmol/l) in the MET 50 group and 4.63 ± 1.92 mg/l (36 ± 15 µmol/l) in the MET 500 group.

Survival rates after sepsis The comparative analysis of the overall survival rates at 120 hours between the three experimental groups showed no difference of mortality: mortality occurred in each group only for an inoculum amount of E. coli superior to 108 CFU/ml, with almost no survival at a ten-fold higher amount of inoculum (figure 1). It was decided to perform a subsequent analysis, comparing combined data by Kaplan-Meyer curves and the Logrank test in the experimental groups with an amount of inoculum of E. coli ≥ 108 CFU/ml (i.e. ≥ the threshold of mortality; figure 2). After 13 hours, about 50% of the mice were dead in each group. This level remained stable until the 60th hour. After 120 hours, almost all animals died. Considering all the times after the intraperitoneal injection, there was no statistical difference in survival rates between the three groups.

Discussion This is the first time that results of this nature, obtained in rats with experimental sepsis, have been reported. In the present experiment, the mice were given metformin in drinking water to deliver a dose of either 50 or 500 mg.kg1 .day-1. The dose of 50 mg.kg-1.day-1 corresponds, on a weight-related basis, to the maximal clinical daily dose in

Survival (percentage) 100

80 60

MET 0 MET 50 MET 500

40

20 0 1.00E+03 1.00E+04 1.00E+05 1.00E+06 1.00E+07 1.00E+08 1.00E+09 1.00E+10 Inoculum size (CFU/ml)

Figure 1 Survival rate in the three experimental groups at 120 hours after the inoculum injection.

100

80

MET 0 MET 50 MET 500

P = N.S. 60

40

20 0 0

20

40

60

80

100

120

140

Time after inoculation (hours) Figure 2 Combined data by Kaplan-Meyer curves and the Logrank test in the experimental groups with an amount of inoculum of E. coli ≥ 108 CFU/ml (i.e. ≥ the threshold of mortality).

man. However, since rodents are less sensitive to the metabolic effects of metformin than man [11] and since metformin clearance is 4 to 5 times that of creatinine [12,13] this selection of the metformin dose was considered appropriate for reproducing the clinical practice. Comparatively, the dose of 500 mg.kg-1.day-1 was intentionally delivered to provoke an accumulation of metformin, which is indeed evidenced by the far higher concentrations of plasma metformin at residual rate [14]. Escherichia coli was used as infectious agent because of its frequency as offending pathogen in diabetic patients. In a study published recently, Muller and coworkers reported an increased risk of urinary tract infection in diabetic patients [15]. Similarly, Laupland and coworkers reported an increased risk of septic shock in diabetic patients [16]. Finally, this E. coli strain was commonly used in the laboratory of one of us and its pathogenicity had been previously assessed [7,8]. The striking finding is that there was no difference in survival rates between the experimental groups, whether metformin was given or not, and whether metformin was given to reproduce a therapeutic dose or a pharmacological one. In other words, the survival rate was only determined by the amount of the infectious agent. The present finding questions metformin as being potentially toxic. That metformin accumulation may lead to lactic acidosis does not necessarily imply that metformin aggravates the course of a concurrent disease. Effectively, if metformin were toxic in the true “metformin-associated lactic acidosis”, i.e. when metformin and concurrent pathologies are co-precipitating factors of lactic acidosis, plasma metformin concentrations would have been higher Diabetes Metab 2006;32:147-150 • © 2006 Masson, all rights reserved

149

V Gras et al.

in the patients with poor prognosis. It appeared on the contrary that the majority of patients with plasma metformin at the therapeutic level or even lower had the poorest prognosis while the majority of patients with high plasma metformin levels survived. This is indeed not really surprising given the notion that lactate production and related metabolic acidosis might be an adapted protective response under both anaerobic and anaerobic conditions [17,18], and given the various vascular properties of metformin accounting for protection under ischaemic conditions (in sharp contrast with those of phenformin, an other biguanide) [19]. That metformin may have beneficial effects is reinforced by the impressive benefit on global mortality in the UKPDS [20]. It has recently been shown that, contrarily to a classical view, the lactate overproduction during septic shock should be regarded as a result of exaggerated aerobic glycolysis and, therefore, as a marker of disease, instead of a marker of hypoxia [18]. Considering this and also the effect of metformin on lactate metabolism, it would have been interesting to assess in our study the outcome according to lactate concentrations. However, lactate concentrations were unfortunately not available. It may be concluded that metformin accumulation does not aggravate the mortality rate in an experimental model of sepsis. In other terms, metformin, whether accumulated or not, can no longer be considered as toxic in such a condition.

References

150

systematic review and meta-analysis. Arch Intern Med 2003; 163:2594-602 5.

6. 7.

8.

9.

10.

11.

12.

13. 14. 15.

16.

1.

Sirtori C, Pasik C. Re-evaluation of a biguanide, metformin: mechanism of action and tolerability. Pharmacol Res 1994;30:187-228

17.

2.

Misbin R, Green L, Stadel B, Gueriguian J, Guffi A, Flemming G. Lactic acidosis in patients with diabetes treated with metformin. N Engl J Med 1998;338:265-6

18.

3.

Chang C, Chen YC, Fang JT, Huang CC. Metformin-associated lactic acidosis: case reports and literature review. J Nephrol 2000;15: 398-402

19.

4.

Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus:

Diabetes Metab 2006;32:147-150 • www.masson.fr/revues/dm

20.

Lalau JD, Lacroix C, Compagnon P, et al. Role of metformin accumulation in metformin-associated lactic acidosis. Diabetes Care 1995; 18:779-84 Lalau JD, Race JM. Metformin and lactic acidosis in diabetic humans. Diabetes Obesity Metab 2000;2:131-7 Montravers P, Andremont A, Massis L, Carbon C. Investigation of the potential role of Enterococcus faecalis in the pathophysiology of experimental peritonitis. J Infect Dis 1994;169:821-30 Dupont H, Montravers P, Mohler J, Carbon C. Disparate findings on the role of virulence factors of Enterococcus faecalis in mouse and rat models of peritonitis. Infect Immun 1998;169:2570-5 Ike Y, Hashimoto H, Clewell D. Hemolysin of Steptococcus faecalis subspecies zymogenes contributes to virulence in mice. Infect Immun 1984;15:528-30 Lacroix C, Danger P, Wojciechowski F. Microdosage de la metformine plasmatique et intra-érythrocytaire par chromatographie en phase liquide. Ann Clin (Paris) 1991;49:98-101 Sterne J. Pharmacology and mode of action of hypoglycaemic guanidine derivates. In: Campell IW, ed, Oral hypoglycaemic agents. London: Academic Press 1969,193-245 Sirtori C, Franceschini G, Galli-Kienle M, Cighette G, Bondioli A, Conti F. Disposition of metformin (N, N-dimethylbiguanide) in man. Clin Pharmacol Ther 1978;24:683-93 Scheen AJ. Clinical Pharmacokinetics of metformin. Clin Pharmacokinet 1996;30:359-71 Wilcock C, Balley CJ. Accumulation of metformin by tissues of the normal and diabetic mouse. Xenobiotica 1994;24:49-57. Muller M, Gorter KJ, Hak E, et al. Increased risk of common infections in patients with type 1 and type 2 diabetes mellitus. Clin Infect Dis 2005;41:281-8 Laupland KB, Gregson DB, Zygun DA, Doig CJ, Mortis G, Church DL. Severe bloodstream infections: a population-based assessment. Crit Care Med 2004;32:992-7 Leverve X. Lactic acidosis. A new insight? Minerva Anestesiol 1999; 65:205-9 Levy B, Gibot S, Cravoisy A, Bollaert PE. Relation between muscle Na+K+ATPase activity and raised lactate concentrations in septic shock: a prospective study. Lancet 2005;365:871-5 Wiernsperger NF. Metformin: intrinsic vasculoprotective properties. Diabetes Technol Ther 2000;2:259-72 UK Prospective Diabetes Study (UKPDS) group. Effect of intensive blood-glucose control with metformin on complication in overweight patients with type 2 diabetes (UKPDS 34). Lancet 1998;352:854-65