Intravenous Almitrine Bismesylate Reversibly Induces Lactic Acidosis and Hepatic Dysfunction in Patients with Acute Lung Injury 

This survey was approved by the human ethics committee of the institute, and written consent was obtained from patients or the next of kin. Twenty-five consecutive tracheally intubated and ventilated (volume-cycled) patients with severe ALI, defined as a Pa^O2:FI^O2ratio <or= to 150 mmHg, and treated with almitrine after therapeutic optimization (see therapeutic strategy) were included prospectively. Among these, eight had ARDS (Murray score >2.5) based on classic criteria, including bilateral lung infiltrates shown on a chest roentgenograph and a thoracic computed tomograph and an absence of cardiac failure assessed by echocardiography, a pulmonary artery occlusion pressure <18 mmHg, or both. [12]The 17 other patients had focal lung lesions that involved <50% of the total lung surface, as judged by four independent, blinded observers by chest roentgenography and computed tomographic assessment of parenchymal consolidation.

Twelve patients had an oximetric thermodilution pulmonary artery catheter (Abbott Laboratories, Chicago, IL) placed that allowed measurement of mean pulmonary arterial, right atrial, and pulmonary artery occlusion pressures. The other 13 patients were monitored using central venous catheters and transesophageal descending aortic flow velocity (ODM 2; Abbott Laboratories, Rungis, France). [13]In all patients, arterial pressure was measured via arterial catheter (Seldicath catheter; Plastimed Laboratories, St. Leu la Foret, France). Arterial blood sampling allowed determination of arterial blood gases (ABL 300; Radiometer, Copenhagen, Denmark), prothrombin time (%), the coagulation factors II, V, and VII, aspartate aminotransferase (AST, n < 35 IU/l), alanine aminotransferase (ALT, n < 56 IU/l), total bilirubin (n < 17 [micro sign]M), alkaline phosphatase (n < 115 IU/l), glucose (n, 4 - 5.5 mM) and lactate (n < 2 mM, +/− 0.06 mM at the precision test; Ektachem 700, Kodak, Rochester, NY). The cumulative dose of almitrine, the Logistic Organ Dysfunction System score, [14]and the Simplified Acute Physiology Score II [15](both scores performed at the time of almitrine administration) were calculated also.

Patients with compromised arterial oxygenation were administered intravenous almitrine after preliminary therapeutic optimization that consisted of adequate fluid resuscitation (based on filling pressures and mixed venous oxygen saturation), appropriate use of vasoactive drugs to maintain blood pressure, determination of the best positive end-expiratory pressure level based on oxygen transport when it was available, and administration of inhaled NO at a dose of 4 - 15 ppm, as previously described. [7,12]Inhaled NO was administered continuously via fenestrated silicone catheter (1.3 mm diameter; Vygon, Paris, France) positioned within the endotracheal tube, at a flow rate adapted to provide a therapeutic fraction of 4 - 15 ppm. Potential NO toxicity was assessed by (1) the methemoglobin level determined daily by spectrophotometry (OSM2 Hemoximeter, Radiometer), which was consistently <1.5%, and by (2) NO and nitrogen dioxide (NO²) levels in both the inspired and expired limbs of the ventilator, measured twice a day by chemiluminescence (EcoPhysic, Massy, France). The NO (²) level was always <100 ppb.

When the Pa^O2:FI^O2ratio remained <or= to 150 mmHg after these modalities were used, patients were administered a continuous infusion of intravenous almitrine (Servier, Suresnes, France) through a central catheter. The dose of almitrine was titrated from 2 - 8 [micro sign]g [middle dot] kg^-1[middle dot] min^-1to achieve the best Pa^O2:FI^O2ratio with the minimal amount of almitrine. [1]Hemodynamic and biologic parameters were recorded before the administration of almitrine (T0) and every 12 h for 60 h in all patients (T12 to T60).

For the patients whose lactate concentration increased by more than +2 mM during the first 24 h of almitrine administration, almitrine was discontinued at 24 h (T24). For the other patients, in whom the plasma lactate concentration remained stable between T0 and T24, almitrine administration was continued at the same infusion rate for at least 36 h more (until T60). In all patients, NO inhalation was maintained at least until T60.

Mean values between groups were compared using two-way analysis of variance for subsequent measures. Intragroup comparisons were made using the Newman-Keuls test. Data are expressed as mean +/− SD, and P < 0.05 was considered significant.

Potential univariate correlates of lactic acidosis were identified by chi-squared analysis for qualitative parameters. and by the Mann-Whitney U test for quantitative parameters. Considering the small sample size, a logistic exact procedure was used for multivariate analysis (LogXact, Cytel Software).

(Table 1) summarizes selected clinical characteristics of the 25 patients at the time of intravenous almitrine administration. The mean Logistic Organ Dysfunction System score was 5 +/− 3, the Simplified Acute Physiology Score II was 34 +/− 10, and the Pa^O2: FI^O2ratio was 93 +/− 33 mmHg for the 25 patients. The causes of ALI were related mainly to primary lung lesions (21 of 25 patients), including pneumonia (n = 7), postcardiosurgical atelectesia (n = 6), and lung contusions after thoracic trauma (n = 4).

Intravenous almitrine enhanced the Pa^O2:FI^O2ratio by more than two times in the 25 patients included in our survey (Table 2, Figure 1). The improvement in the Pa^O2:FI^O2ratio was associated with an increased systolic pulmonary arterial pressure in the 12 patients with pulmonary artery catheters. The other measured hemodynamic parameters, including cardiac output, remained unchanged.

Eight patients had increased plasma lactate concentration during the first 24 h of intravenous almitrine administration-a +3.5 +/− 1.8 mM increase compared with T0; these patients were included in a group labeled “high lactate”(HL group, Figure 1and Figure 2). Figure 1and Figure 2show that plasma lactate increased at 12 h and peaked at T24. Intravenous almitrine discontinuation in the HL group at T24 was followed by a decrease in plasma lactate level that returned to normal in all patients within 36 h. In parallel, the pH and the bicarbonate concentration decreased during intravenous almitrine infusion and returned to normal at T60. In contrast, intravenous almitrine did not alter the pH, the bicarbonate concentration, or the plasma lactate level between T0 and T60 in the 17 other patients who were included in the low-lactate group (LL group).

Impaired hepatic function also was observed in the HL group 24 h after intravenous administration of almitrine (Figure 2, Table 3). The AST and ALT serum concentrations increased, whereas prothrombin time and factor VII decreased in the HL group. When intravenous almitrine was discontinued, factor VII returned to normal within 36 h, whereas the AST serum concentration and the prothrombin time remained altered, slowly improving in the next 3 or 4 days (data not shown). Total bilirubin, glucose, and alkaline phosphatase concentrations remained stable after almitrine administration in both groups, despite a greater baseline bilirubin level in the HL group. Intravenous almitrine did not alter hepatic function tests throughout the 60 h of observation in the LL group.

The amount of almitrine administered was greater in the HL group than in the LL group during the first 24 h (459 +/− 155 mg for the HL group and 370 +/− 113 mg for the LL group; P < 0.05) but was less over the 60 h because almitrine was discontinued in the HL group (459 +/− 55 mg for the HL group and 911 +/− 361 mg for the LL group, P < 0.005).

(Figure 2) shows that despite the difference in plasma lactate concentrations and the hepatic function tests between the HL and LL groups, intravenous almitrine induced a similar improvement in the Pa^O2: FI (^O)²ratio in both groups. Although almitrine was discontinued in the HL group at T24, the improvement in the Pa^O2: FI^O2ratio lasted until T60 in the two groups.

Univariate analysis revealed that the level of bilirubin before almitrine administration and patient gender were predictive factors for the occurrence of lactic acidosis after almitrine administration, whereas age, catecholamine treatment, Logistic Organ Dysfunction System score, Simplified Acute Physiology Score II, and cause of ALI (ARDS vs. non-ARDS) were not. Therefore, as shown in Figure 3, female patients were at high risk for the occurrence of lactic acidosis; five of eight were in the HL group. In contrast, all male patients with normal initial levels of bilirubin were in the LL group (10 of 10). Initial bilirubin level and gender both were also found to be predictive factors when used in the multiple logistic exact model. In this case, the exact odds ratio for gender was 2.46 (P = 0.038), and the [Greek small letter beta] parameter for bilirubin (exact assumption) was 0.056 (P = 0.01).

Intravenous almitrine consistently improved the Pa^O2: FI (^O)²ratio in patients with ALI. Intravenous almitrine was also associated with lactic acidosis and impaired hepatic function in 30%(n = 8) of our patients. Lactic acidosis and hepatic dysfunction appeared within 12 h of almitrine administration and were reversed after drug discontinuation. In contrast, in the other 70%(n = 17) of patients, intravenous almitrine induced no change in any of the measured biologic parameters.

Acute lung injury is characterized by impaired gas exchange caused by maldistribution of ventilation:perfusion ratios. The recent introduction of intravenous almitrine and inhaled NO dramatically modified the therapeutic management of hypoxemia in patients with ALI. [1,8,9]In our patients with ALI, the introduction of intravenous almitrine after therapeutic optimization, which included inhaled NO, increased the Pa^O2:FI^O2ratio within 12 h from 93 +/− 7 mmHg to 207 +/− 23 mmHg (P < 0.001). Improvement of the Pa^O2:FI^O2ratio was observed in every patient observed, regardless of plasma lactate concentration and hepatic dysfunction.

Interestingly, the Pa^O2:FI^O2ratio remained greater at T24 and T60 than at T0 in the HL group, even after almitrine was discontinued (Figure 2). This may be explained by the pharmacologic properties of almitrine:[1,2](1) almitrine is highly bound (>99%) to low-density lipoprotein and its half-life is >24 h in humans;(2) hydroxy metabolite of almitrine show almitrine-like activity; and (3) because almitrine is mainly metabolized by the liver and excreted in bile, hepatic dysfunction in the HL group further prolonged almitrine and the half-lives of its metabolites, as previously described. [16] 

Disturbances of acid-base equilibrium and particularly lactic acidosis are among the most frequent metabolic disorders seen in the intensive care unit. Hyperlactatemia may result from decreased lactate clearance, from increased lactate production related to any cause of local anaerobic metabolism or cellular dysfunction, or from both.

Much of the lactate produced by the body is cleared by the liver by gluconeogenesis and oxidative phosphorylation. [17]In our study, the coexistence of hyperlactatemia and hepatic dysfunction, in the HL group, suggests an impaired hepatic lactate clearance as a mechanism of hyperlactatemia. The lung, which is another site of mammalian lactate metabolism, may account for the origin of hyperlactatemia. Because lactate is a preferential energetic substrate for rat and Wedded seal lungs, [18,19]almitrine might have reduced the lactate lung uptake in our patients. More realistically, inflamed lungs may produce more lactate in the presence of almitrine. Such an hypothesis is confirmed by recent data about lung lactate production in critically ill patients. [20] 

Other organs, such as the brain and the gut, and the erythrocytes are major lactate producers, [17]both in physiologic and pathophysiologic conditions. In our study, almitrine-induced hyperlactatemia was associated with a decrease in pH and bicarbonate concentration. This acidosis probably is related to mitochondrial dysfunction, which is known to induce an accumulation of Pi and H⁺. [17]Almitrine may affect mitochondrial function by reducing oxygen delivery to certain organs, but more likely by acting directly on mitochondrial adenosine triphosphatase activity, as previously suggested by several authors. [21-23]In mitochondria that are isolated from bovine heart and rat liver, almitrine decreased the adenosine triphosphate:adenosine diphosphate ratio, mainly by causing a large decrease in adenosine triphosphate synthesis. [23]Thus, by impairing the efficiency of mitochondrial adenosine triphosphate synthesis, almitrine may increase intracellular H⁺concentration, even during conditions of adequate cellular oxygen delivery. In addition, Leverve et al. [22]recently showed that almitrine may alter the regulatory enzymes of glycolysis. In cultures of hepatocytes almitrine inhibits gluconeogenesis and activates lactate and pyruvate kinase. These effects favor glycolysis and lactate, Pi, and H⁺production. Thus, lactic acidosis after almitrine administration may have been related, at least in part, to a direct intracellular effect of almitrine on glucose metabolism and mitochondrial function.

A preliminary report recently described an alteration in hepatic function tests in a patient with ARDS who was treated with almitrine. [11]It showed an increase in AST plasma concentrations to levels >1,000 IU/l that decreased as soon as almitrine was discontinued. In our study, similar alterations in results of hepatic function tests were observed. The AST and ALT levels increased and prothrombin time and factor VII decreased within 24 h of almitrine administration, and these values returned to normal after almitrine was discontinued. The decrease in factor VII is related to a true hepatic dysfunction rather than to a vitamin K deficiency, because the discontinuation of almitrine was followed by a quick recovery in factor VII level.

The mechanism of hepatic dysfunction after intravenous almitrine administration in patients in the HL group is uncertain. Liver injury is unlikely to be related to hemodynamic alterations because no difference was observed before and after almitrine administration, except for an increase in pulmonary artery systolic pressure (Table 1). In addition, no difference was observed between the HL and LL groups. Hepatic dysfunction may have been related to a local vasoconstrictor effect of the almitrine on hepatic vessels, to a direct, and potentially toxic, effect of almitrine on hepatocyte, or to both. [22]These effects may have been amplified by the impaired hepatic excretion of bilirubin seen before intravenous almitrine administration in the HL group.

In conclusion, our survey shows that a great improvement in the Pa (^O)²:FI^O2ration in all patients with ALI who were treated using intravenous almitrine. No alteration in plasma lactate concentration or in results of hepatic function tests were seen in most of these patients during the first 60 h of treatment. However, in some patients with abnormal plasma bilirubin concentrations, an almitrine improvement in arterial oxygenation was associated with lactic acidosis and an alteration in hepatic function test results. Although lactic acidosis and hepatic dysfunction returned to normal after intravenous almitrine was discontinued, arterial oxygenation remained improved for at least 36 h more. Therefore, we suggest that almitrine should be administered in short infusions of 12 - 24 h every 2 or 3 days. Consequently, the cumulative dose of almitrine would be reduced with a potential reduction in toxicity but with a sustained arterial oxygenation improvement. The intracellular mechanism of almitrine-induced hepatic dysfunction remains unclear, and more clinical and experimental studies are necessary. However, we recommend that the lactate concentration be monitored, that biologic hepatic function tests be performed every 12 h during the first 24 - 48 h of almitrine administration, and that the almitrine infusion be discontinued in cases of severe alteration. The bilirubin concentration before almitrine treatment and patient gender were predictive factors of the occurrence of lactic acidosis and hepatic dysfunction. This association should be confirmed by a larger study.

The authors thank Drs. Bernard Cholley and Randall C. Wetzel for help with revising the manuscript and Professor Eric Vicaut for help with the statistical analysis.