Increased Nitric Oxide in the Exhaled Air of Patients with Decompensated Liver Cirrhosis

Methods

Patients

Fifty-six patients were consecutively selected from those hospitalized in our department. All had biopsy-proven chronic hepatitis or liver cirrhosis; none had primary lung disease, hypertension, or infection. They could walk in the ward unaided and did not need intensive care. Physical examination findings and blood data were analyzed to classify hepatocellular function in liver cirrhosis according to the Child criteria. The clinical background of these patients is summarized in Table 1. Healthy volunteers served as controls (15 men; 34 ±2 years of age; body surface area, 1.84 ±0.03 m²). All medications were discontinued 24 hours before each study began. No antihypertensives or vasodilators, including nitrates and angiotensin-converting enzyme inhibitors, were used in these patients. The study was approved by the hospital ethics committee, and informed consent was obtained from each study participant.

Nitric Oxide Measurement

The nitric oxide concentration in exhaled air was determined at least 3 hours after meals while each participant was at rest in the sitting position, as previously described [7]. Each participant was asked to inhale synthetic air (Taiyo Sanso Co., Osaka, Japan) free of nitric oxide (< 3 parts per billion [ppb]) through a mask and a T-valve, and to exhale the air into a wide-bore Teflon tube (internal diameter, 25 mm; length, 600 mm). Exhaled air was continuously drawn from this tube with a vacuum pump and was introduced into a chemiluminescence analyzer (APNE-350E, Horiba Co., Kyoto, Japan). Measurement of nitric oxide concentration was based on the reaction of nitric oxide with ozone. The sensitivity of the analyzer to nitric oxide ranged from 2 to 1000 ppb. The system was calibrated with dilutions of certified nitric oxide gas (450 ppb in nitrogen; Taiyo Sanso Co.) using mass flowmeters (Estec Co., Kyoto, Japan). Expired volume was measured with a hot-wire flow meter connected to the T-valve on the expiratory side, and minute ventilation was calculated using a breath-by-breath respirometer (RM-280, Minato Medical Science Co., Tokyo, Japan). The nitric oxide concentration and minute ventilation were recorded with a computer-assisted data recorder (DS1100, Fukuda Denshi Co., Tokyo, Japan), and the output of nitric oxide was calculated as follows: nitric oxide output = (nitric oxide_ex −nitric oxide_in) x minute ventilation/body surface area, where nitric oxide_ex was the nitric oxide concentration in exhaled air, and nitric oxide_in was the nitric oxide concentration in inhaled air. Nitric oxide concentration and minute ventilation were monitored simultaneously for 10 minutes, and the data obtained during the last 3 minutes was averaged. During the study period, the ambient levels of nitric oxide concentration were less than 5 ppb. Nitric oxide output was reproducible in patients with cirrhosis and in controls (coefficient of variation, 10.8% [n = 5] for patients and 9.3% [n = 5] for controls) on separate days, and there was no significant time-course change in nitric oxide output at rest.

Echocardiographic Measurement

To examine the relation between systemic hemodynamics and nitric oxide production, we measured cardiac output using transthoracic two-dimensional echocardiography (SSD-2200, Aloka Co., Tokyo, Japan) in 19 patients with liver cirrhosis and in 6 controls. This was done on the same day that nitric oxide concentrations were measured. A physician, who was blinded to the patient characteristics and the exhaled nitric oxide output values, obtained echocardiographic views and recorded them on videotape. Another physician, who was also blinded to these data, measured cardiac output using the echocardiographic images. Left ventricular dimension was measured in the long-axis view of the left ventricle while the patient was in the left lateral decubitus position. Left ventricular volume and cardiac index were obtained by the following formulae according to the Teichholz equation [8]: left ventricular volume = 7.0 x dimension^3/(2.4 + dimension); cardiac index = (left ventricular end-diastolic volume -end-systolic volume) x heart rate/body surface area. Blood pressure was measured with a sphygmomanometer. Total peripheral resistance index was calculated as mean blood pressure × 80/cardiac index. The cardiac index obtained by this method on separate days was reproducible in patients with cirrhosis and in controls (coefficient of variation, 9.1% [n = 6] in patients and 8.6% [n = 5] in controls).

Statistical Analysis

Values are expressed as the mean ±SE. Differences between patients and controls were compared using one-way analysis of variance (ANOVA) followed by the Fisher test. The correlation coefficient was calculated using the least-squares method. Statistical significance was set at P < 0.05.

Results

Patients with decompensated liver cirrhosis had markedly depressed liver function but normal serum creatinine levels (Table 1). There were no intergroup differences in minute ventilation per m² body surface area (patients with chronic hepatitis, 5.1 ±0.3 L/min; Child A patients, 5.6 ±0.3 L/min; Child B patients, 5.4 ±0.2 L/min; Child C patients, 6.2 ±0.3 L/min; and controls, 5.4 ±0.2 L/min; P = 0.12). The level of exhaled nitric oxide output per m² body surface area was significantly greater in patients with Child C (190 ±11 nL/min; P < 0.001) or Child B liver cirrhosis (166 ±12 nL/min; P < 0.001) than in controls (97 ±8 nL/min) (Figure 1). In patients with Child A liver cirrhosis (119 ±10 nL/min; P = 0.17) or chronic hepatitis (129 ±19 nL/min; P = 0.13), the level of nitric oxide output per m² body surface area was similar to that in controls.

View larger version:

• In this window
• In a new window

Figure 1. Bars indicate mean with 95% CI. Nitric oxide (NO) output in exhaled air in controls, patients with chronic hepatitis (CH), and patients with liver cirrhosis.

The results of hemodynamic measurements showed that patients with Child C liver cirrhosis had a greater cardiac index per m² body surface area (4.3 ±0.3 L/min compared with 2.9 ±0.2 L/min; P < 0.001) and a smaller total peripheral resistance per m² body surface area (1732 ±125 dyne/s x cm⁵ compared with 2680 ±235 dyne/s x cm⁵; P = 0.004) than controls. There was a positive correlation between the level of nitric oxide output and cardiac index (r = 0.621; P < 0.001) (Figure 2).

View larger version:

• In this window
• In a new window

Figure 2. Relation between nitric oxide (NO) output and cardiac index in patients with liver cirrhosis and controls.

Discussion

We have shown that nitric oxide output is increased in the air exhaled by patients with cirrhosis, especially patients with decompensated cirrhosis. Although we did not identify the origin of the increased synthesis of nitric oxide, several potential sources can be considered. Patients with liver cirrhosis often have endotoxemia even when they have no signs of infection [4], and elevated concentrations of cytokines, such as tumor necrosis factor-α, have been shown in patients with liver diseases [9, 10]. The liver may produce large amounts of nitric oxide in these patients: Hepatocytes and Kupffer cells are known to produce nitric oxide in vitro in response to lipopolysaccharide and several cytokines [11, 12]. The plasma levels of cytokines, including tumor necrosis factor, are much lower in patients with liver cirrhosis than in these in vitro studies [10-15]. However, in vitro studies have also shown that endotoxin and cytokines also induce nitric oxide synthase in other tissues, including vascular endothelium, smooth muscle, and bronchial epithelium [13-15]. Thus, it is possible that vascular and bronchial tissues in the lungs of patients with liver cirrhosis produce nitric oxide as a result of continuous stimulation by the lower concentrations of cytokines, because the plasma levels of cytokines in patients with cirrhosis are similar to those in normal persons who have become hypotensive through the administration of endotoxin [10, 16].

Because most nitric oxide is inactivated by hemoglobin or rapidly metabolized to nitrite and nitrate [3], nitric oxide in exhaled air may be the residual of excessive local production of nitric oxide by the lung rather than a product of the liver. Because plasma nitrite and nitrate levels reflect the sum of nitric oxide production in the entire body, including the liver, the significance of the increase in exhaled nitric oxide output may differ from that of the increased plasma concentrations of nitrite and nitrate in patients with cirrhosis [17]. We did not measure the plasma levels of endotoxin, cytokines, and nitrite and nitrate, but these levels have been reported to vary widely, even among patients with similar degrees of liver damage [10, 17]. This is compatible with our finding of a large spread within each group and a significant overlap among the groups in exhaled nitric oxide levels.

Increased production of nitric oxide in the lungs may exert a vasodilatory effect on the pulmonary circulation. It has been suggested that such increased production may play a significant role in patients with the hepatopulmonary syndrome, because the administration of methylene blue, an inhibitor of guanylate cyclase, restores hypoxia in these patients [18]. If other vascular beds behave in a similar way, induction of nitric oxide synthase in the whole body will result in the hyperdynamic state associated with the decrease in total peripheral resistance and the increase in total venous capacitance found in patients with cirrhosis [1]. In our study, patients with liver cirrhosis were in a hyperdynamic state that was associated with an increase in nitric oxide output. Although it is not necessarily true that the increase in nitric oxide caused the hyperdynamic state, previous experimental and human studies have suggested that nitric oxide may play such a causative role. Rats with cirrhosis have enhanced sensitivity to the pressor effect of nitric oxide inhibition, indicating that increased endogenous nitric oxide contributes to arterial hypotension [19]. In addition, hyporesponsiveness to pressor stimuli is reversed by the inhibition of nitric oxide in animal models of portal hypertension [20]. Thus, nitric oxide has been shown to play an important role in regulating the vasomotor tone in experimental liver cirrhosis. Furthermore, methylene blue reversed severe hypotension in a human patient with liver failure [21]. Taken together, these findings suggest that excessive production of nitric oxide may account for the hemodynamic abnormalities seen in patients with liver cirrhosis. The inhibition of nitric oxide synthesis might reverse these abnormalities, although the need for development of specific inhibitors of constitutive or inducible nitric oxide synthase remains.

Increased nitric oxide output in exhaled air is associated with systemic circulatory disturbances in patients with liver cirrhosis. Measurement of nitric oxide output in exhaled air is a noninvasive examination that may provide clues to understanding the hemodynamic disturbances seen in patients with liver cirrhosis.