Additional file 1
Detailed materials and methods

Animal preparation
	For the experiment, animals were premedicated with intramuscular injection of ketamine (KŽtalar¨, Virbac, France, 2.5 mg/kg of body weight) and xylazine (SŽdaxylan¨, CEVA SantŽ Animale, France, 2.5 mg/kg of body weight). Then, we used Isoflurane (Aerrane¨, Baxter, France) for the intubation process, and maintenance of anaesthesia was performed with a continuous infusion of midazolam (Hypnovel¨, Roche, France, 1 mg/kg body weight/h) for the whole experiment. Animals were mechanically ventilated (Osiris 2¨, Taema, France) with a tidal volume of 8 mL/kg, a positive end-expiratory pressure set at 4 cmH2O to limit cardiovascular effects, FiO2 0.6 to prevent fatal hypoxemia during the study, and respiratory rate 20 to 24 breaths/min only adjusted to maintain normocapnia (40 Ð 45 mmHg) at baseline. We chose to maintain ventilation similar in all animals during the experiment. No recruitment manoeuvres were done. Muscle relaxation was obtained by a continuous intravenous infusion of cisatracurium besylate (Nimbex¨, Hospira, France, 2 mg/kg body weight/h). Analgesia was achieved by a subcutaneous injection of buprenorphine (Vetergesic¨, Sogeval, France, 0.1 mg/kg body weight). Body surface area was calculated by KelleyÕs formula (1).

Microcirculatory parameters
	Skin microvascular blood flow was measured continuously using a laser Doppler flowmeter probe and device (Periflux¨ PF407; Perimed, Jiirfalla, Sweden). The blood flow was measured in a volume of 1mm3 of solid tissue. The fiber optic probe was applied on the left hind paw of the animals and fixed with an adhesive tape. Laser Doppler signal was continuously registered on a personal computer. Readjusting the pen of the recorder to zero when the probe was fixed to a white non-moving surface performed flux zero calibration. Skin blood flow was measured at rest and during reactive hyperemia from T0 to T300, and values were expressed in arbitrary perfusion units (PU). Reactive hyperemia was produced by arrest of leg blood flow with a pneumatic cuff inflated to a suprasystolic pressure (50 mmHg above systolic pressure) for 3 min. On completion of the ischemic period, the occluding cuff was rapidly deflated to zero. Peak flow was defined as the highest flow signal during the post occlusive phase. Reactive hyperemia was further analysed (Perisoft¨ 2.5 software) according to its duration and initial reactive hyperemia uphill slope. 
      Images of the sublingual and rectal microcirculation were obtained with SDF videomicroscopy (Microscan¨; Microvision Medical, Amsterdam, The Netherlands). After gentle removal of saliva and other secretions with isotonic saline-drenched gauze, the device was applied to the sublingual region, avoiding pressure artifacts by establishing a threshold image. The microcirculation of rectum was investigated through the anus by insertion of an SDF objective to the depth of 6 to 10 cm from the edge. The SDF imaging device was fitted with an analogue video camera that needs to be digitalized by separate analogue to digital converter devices for offline image analysis. At every time, the optical probe was placed gently on the rectal mucosa and the sublingual mucosa. Two videos in 5 different fields were captured and videos of at least 10s were recorded, while precautions were taken to avoid pressure and moving artifacts when capturing movies. Every 10 videos were taken at the tongue and the rectum at each time point in total. Video clips were directly saved as digital files to a hard drive of a personal computer. Video clips were blindly analyzed offline by two investigators in random order to prevent coupling. Assessment of microcirculatory parameters of convective oxygen transport (microvascular flow index (MFI)) was done according to published recommendations. The semi quantitative MFI, ranging from 0 (no flow) to 3 (continuous flow), was based on the determination of the predominant type of flow in four quadrants. The MFI was the sum of these scores divided by the number of scored quadrants. 
      
Biological methods 
	Apparent strong ion difference (SID) was calculated as [(Na+ + K+ + Ca2+ + Mg2+) Ð (Cl- + Lactate-)] (2). Sodium and chloride balance was determined by calculating for each individual the difference between the cumulative amounts of electrolytes administered and the cumulative amounts of electrolytes collected in the urine. 

Metabolic methods 
	Lactate, pyruvate, beta-hydroxy-butyrate and acetoacetate, glycerol and non-esterified fatty acids (NEFA)] were analyzed. One ml of blood was immediately precipitated with ATCA 10% (v:v). Tubes containing acidified blood were frozen and kept frozen until analysis (lactate, pyruvate and ketone bodies). Glycerol and NEFA were measured on serum. Blood samples were rapidly addressed on ice to the lab, where they were centrifuged, decanted, aliquoted and frozen until analysis. Lactate, beta-hydroxy-butyrate, glycerol and NEFA were respectively performed using LACT2 kit (Roche Diagnostics GmbH, Mannheim, Germany), RANBUT, GLY and NEFA kits (Randox Laboratories, Crumlin, Unitd Kingdom) according to manufacturerÕs recommendations. Pyruvate and acetoacetate were quantified by home-developed spectrophotometric assays, respectively using LDH and 3HBDH.

References
1. 	Kelley KW, Curtis SE, Marzan GT, Karara HM, Anderson CR. Body surface area of female swine. J Anim Sci 1973;36:927?30. 
2. 	Stewart PA. Modern quantitative acid-base chemistry. Can J Physiol Pharmacol 1983;61:1444?61.