Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Exp Physiol. Author manuscript; available in PMC 2010 April 11.
Published in final edited form as:
PMCID: PMC2852632

Chronic Intermittent Hypoxia and Acetaminophen Induce Synergistic Liver Injury


Obstructive sleep apnea (OSA) leads to chronic intermittent hypoxia (CIH) during sleep. OSA has been associated with liver injury. Acetaminophen (APAP) is one of the most commonly used drugs, which has known hepatotoxicity. The goal of the present study was to examine whether CIH increases liver injury, hepatic oxidative stress and inflammation induced by chronic APAP treatment. C57BL/6J mice were exposed to CIH or intermittent air (IA) for 4 weeks. Mice in both groups were treated with intraperitoneal injections of either APAP (200 mg/kg) or normal saline daily. A combination of CIH and APAP caused liver injury with marked increases in serum alanine aminotransferase, aspartate aminotransferase (AST), gamma glutamyl transferase and total bilirubin levels, whereas CIH alone induced only elevation in serum AST levels. APAP alone did not affect serum levels of liver enzymes. Histopathology revealed hepatic necrosis and increased apoptosis in mice exposed to CIH and APAP, whereas the liver remained intact in all other groups. Mice exposed to CIH and APAP exhibited decreased hepatic glutathione in conjunction with a five-fold increase in nitrotyrosine levels, suggesting formation of toxic peroxynitrite in hepatocytes. APAP or CIH alone had no effect on either glutathione or nitrotyrosine. A combination of CIH and APAP caused marked increases in pro-inflammatory chemokines, monocyte chemoattractant protein-1 and macrophage inflammatory protein-2, which were not observed in mice exposed to CIH or APAP alone. We conclude that CIH and chronic APAP treatment lead to synergistic liver injury, which may have clinical implications for patients with OSA.

Keywords: Free radical, hypoxia, liver


Obstructive sleep apnea (OSA) is characterized by recurrent closure of the upper airway during sleep leading to chronic intermittent hypoxia (CIH) and sleep fragmentation (Gastaut et al., 1966). OSA is present in 2-9% of women and 4-24% of men in the US, but the prevalence may exceed 50% in the obese population (Bixler et al., 2001;Punjabi et al., 2002;Young et al., 1993). OSA has been independently associated with development of hypertension, cardiovascular disease, stroke, and diabetes mellitus type 2 (Marin et al., 2005;Shahar et al., 2001;Yaggi et al., 2005;Nieto et al., 2000;Peppard et al., 2000;Punjabi et al., 2003;Punjabi et al., 2004;Punjabi & Polotsky, 2005). There is also emerging evidence that OSA is associated with chronic liver injury and non-alcoholic steatohepatitis (Tanne et al., 2005;Tatsumi & Saibara, 2005;Zamora-Valdes & Mendez-Sanchez, 2007;Jouet et al., 2007;Singh et al., 2005;Kheirandish-Gozal et al., 2008;Kallwitz et al., 2007;Norman et al., 2008).

The CIH of OSA induces repetitive cycles of hypoxia and reoxygenation leading to generation of reactive oxygen species (ROS) and systemic oxidative stress (Lavie, 2003;Lavie et al., 2004;Row et al., 2002;Zhan et al., 2005;Chen et al., 2005;Prabhakar et al., 2007). We have previously developed a murine model of CIH (Polotsky et al., 2006) and have shown that CIH leads to both mild liver injury and elevations in biomarkers of oxidative stress (Li et al., 2007;Savransky et al., 2007c;Savransky et al., 2007d;Jun et al., 2008). We have also shown that CIH may act as ‘a second hit’ amplifying liver injury induced by a single high dose of acetaminophen (APAP) (Savransky et al., 2007d). However, effects of CIH on acute APAP overdose are of limited clinical significance given that APAP overdose is relatively rare. In contrast, APAP is among the most commonly used pain-relievers and an association between OSA and chronic intake of APAP can be frequently encountered..

APAP is metabolized in the liver. About 90% of a therapeutic dose of APAP is conjugated with glucuronic acid or sulfate and excreted into bile or blood (Jaeschke, 2005). The remaining 10% of APAP is metabolized by the hepatic cytochrome P450 system to N-acetyl-p-benzoquinone imine, which reacts with reduced glutathione (GSH) to form a GSH-adduct (Chen et al., 2003a;Jaeschke & Bajt, 2006). When APAP overdose occurs, N-acetyl-p-benzoquinone imine depletes mitochondrial GSH stores, and leads to oxidative stress with predominant generation of peroxynitrite, (Jaeschke & Bajt, 2006). Peroxynitrite is formed in the mitochondria from nitric oxide (NO) and superoxide and can be assessed by 3-nitrotyrosine (3-NT) levels. Inducible NO synthase (iNOS) may play a role in APAP liver toxicity (Jaeschke & Bajt, 2006;Michael et al., 2001). It is conceivable that CIH and APAP have a synergistic injurious effect on hepatocytes, causing severe oxidative stress with resultant tissue injury and inflammation.

The goal of the present study was to explore whether CIH exacerbates liver toxicity induced by chronic administration of APAP. We exposed C56BL/6J mice to CIH and APAP and examined: 1) serum activity of liver enzymes; 2) liver histopathology; 3) indices of oxidative stress, including GSH, 3-NT, and lipid peroxidation; 4) liver levels of pro-inflammatory cytokines.



A total of 38 wild-type, 6-8 week old male, lean C57BL/6J mice purchased from Jackson Laboratory (Bar Harbor, Maine) were used in the study. The study was approved by the Johns Hopkins University Animal Care and Use Committee and complied with the American Physiological Society Guidelines for Animal Studies. For all blood samples, injections and surgical procedures, anesthesia was induced and maintained with 1-2 % isoflurane administered through a facemask.

Experimental Design

A gas control delivery system was designed to regulate the flow of room air, nitrogen and oxygen into customized cages housing the mice as previously described (Polotsky et al., 2003). A maximum of three mice were housed continuously in a single customized cage (dimensions 27 × 17 × 17 cm) with constant access to food and water. During each period of intermittent hypoxia (IH), the FIO2 was reduced from 20.9 to 4.9 ± 0.1 % over a 30 second period and then rapidly reoxygenated to room air levels in the subsequent 30 second period resulting in 60 hypoxic events per hour. We have previously determined that an average PaO2 over a hypoxic cycle was 51.7 ± 4.2 mmHg (Savransky et al., 2007a). Thus, our murine model mimicked CIH of severe OSA.

Food intake and body weight were monitored daily for each animal. All animals were kept in a controlled environment (22-24 °C with a 12 hrs: 12 hrs light: dark cycle; lights on at 09.00) on a standard Purina chow diet with free access to food and water. Eighteen mice were placed in the IH chamber for four consecutive weeks. Twenty mice were exposed to intermittent room air (IA, control groups) for four weeks in identical chambers (Table 1). The CIH and IA states were induced during the light phase alternating with 12 hrs of constant room air during the dark phase. During the last two weeks of exposure, nine mice exposed to CIH and ten mice exposed to IA were treated with daily intraperitoneal injections of APAP (Sigma-Aldrich Corp., St. Louis, MO) at 200 mg/kg in 0.3 ml of 0.9% sodium chloride; while nine mice exposed to CIH and ten mice exposed to IA were treated with daily intraperitoneal injections of placebo (0.3 ml of 0.9% sodium chloride). The dose was selected based on our pilot data demonstrating that APAP in a dose up to 300 mg/kg did not induce significant elevation of liver enzymes in male C57BL/6J mice (not shown).

Basic Characteristics of C57BL/6J Mice Exposed to Chronic Intermittent Hypoxia and Acetaminophen

Sample Collection

Mice fasted for 5 hrs prior to bleeding and sacrifice. Arterial blood (1 ml) was obtained by direct cardiac puncture under 1-2 % isoflurane anesthesia. Serum was separated and frozen at −80° C. After blood withdrawal, the animals were euthanized with pentobarbital (60 mg, intraperitoneally). Liver was surgically removed and weighed. Liver tissue was either frozen in liquid nitrogen and kept at −80°C for further biochemical studies, or fixed in buffered 10% formalin for histological examination.

Biochemical Assays

Serum alanine aminotransferase (ALT), aspartate aminotransferase (AST), gamma glutamyl transferase (GGT) activity, serum total and direct bilirubin, total protein and albumin levels were measured by the Clinical Chemistry Laboratory of the Johns Hopkins Bayview Medical Center. Fasting blood glucose was measured with Accu-Chek® Comfort Curve TM kit from Roche Diagnostics, Inc. (Indianapolis, IN). Liver was homogenized using Omni EZ Connect Homogenizer (Omni International, Warrington, VA). GSH, 3-NT, and malondialdehyde (MDA) levels were measured in total liver lysate with kits from Oxis Research (Portland, OR). Tumor necrosis factor α (TNF-α), monocyte chemoattractant protein-1 (MCP-1), and macrophage-inflammatory protein-2 (MIP-2) levels in total liver lysate were determined with ELISA kits from R&D Systems, Inc. (Minneapolis, MN). Interleukin 6 (IL-6) levels in the liver were measured with an ELISA kit from RayBiotech, Inc.(Norcross, GA).


Paraffin-embedded tissue was sectioned at 3-5 μm thickness and stained with hematoxylin and eosin (H&E). Semi-quantitative analysis of the H&E sections was performed by an experienced hepatopathologist (MST) without knowledge of treatment conditions or laboratory findings. The scoring system developed and validated for steatohepatitis in humans has been adapted for the study with minor modifications (Kleiner et al., 2005). The following scoring criteria have been used: steatosis (0- none, 1-1-30% hepatocytes, 2- 31-60%, 3 > 60%), hepatocyte swelling (same as steatosis), lobular inflammation (0 - none, 1- average of one inflammatory infiltrate, 2 – average of 2-4 inflammatory infiltrates, 3 – 5-10 inflammatory infiltrates, 4 - > 10 inflammatory infiltrates per 10X field), and necrosis (1 – occasional necrotic cell; 2 – numerous necrotic cells in zone 3; 3 – confluent zone 3 necrosis; 4 – necrosis extending beyond zone 3). Terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling (TUNEL) was performed in liver sections to detect apoptosis using the Fluorescein FragEL DNA Fragmentation Detection Kit (Cat. No. QIA39) from Calbiochem (San Diego, CA). To determine the percentage of apoptotic cells, all cells on one section (10-50 mm2) were examined and TUNEL-positive cells were counted under fluorescent microscopy at an excitation wavelength of 330 – 380 nm. Results were expressed as a number of TUNEL-positive cells/total cells × 100%.

Gene Expression Analysis

Total RNA was extracted from liver using Trizol (Life Technologies, Rockville, MD) and cDNA was synthesized using Advantage RT for PCR kit from Clontech (Palo Alto, CA). Real-time reverse-transcriptase PCR (RT-PCR) was performed with primers and probe specific for iNOS (Mm00440488_m1), endothelial NOS (eNOS, Mm0113492_g1) and 18S (Li et al., 2005b;Li et al., 2005a) from Applied Biosystems (Foster City, CA). Of note, 18S is an optimal housekeeping gene for liver injury (Pelletier et al., 2000). The mRNA expression levels were normalized to 18S rRNA concentrations using the following formula: Gene of interest/18S = 2 Ct(18S)−Ct(Gene of Interest).

Statistical Analyses

All values are reported as mean ± standard error of the mean (SEM). Statistical comparisons between four groups of mice (CIH-APAP, IA-APAP, CIH-Placebo, IA-Placebo) were performed by a general linear model ANOVA across two independent variables, hypoxia and APAP, followed by the Tukey’s post-hoc test. Comparisons between Day 0, Day 14 and Day 28 of mouse exposure were performed using repeated-measures ANOVA. A p value of less than 0.05 was considered significant.


Mice exposed to CIH exhibited a decrease in food intake and 9-11% weight loss during first two weeks of the experiment, whereas IA control mice gained 8-9% of the starting body weight (Table 1). During the last two weeks of the experiment, mice receiving placebo showed adaptation to the CIH stimulus with restored food intake and weight gain (Table 1). APAP administration from Day 14 to Day 28 resulted in mild reduction in food intake in mice exposed to IA, while body weight did not differ significantly from the placebo group. In contrast, mice exposed to CIH and APAP showed a dramatic decrease in food intake and body weight by Day 28 (Table 1). CIH led to a decrease in liver weight in both placebo and APAP-injected mice.

In placebo-treated mice, CIH doubled serum AST levels and slightly increased serum total protein levels, whereas serum ALT, GGT, total bilirubin, albumin and fasting blood glucose levels were not altered (Figure 1, Table 1). Mice treated with APAP and exposed to IA exhibited normal levels of liver enzymes and a mild increase in serum total bilirubin level (Figure 1). CIH significantly increased liver injury induced by APAP, as evidenced by a three-fold increase in serum ALT, a decline in fasting blood glucose, a marked increase in serum total bilirubin, and appearance of direct bilirubin (0.06 ± 0.02 mg/dl), which was absent in all other animal groups (Figure 1, Table 1).

Figure 1
The effects of chronic intermittent hypoxia (CIH) and acetaminophen (APAP) on serum enzymatic activity of (A) alanine aminotransferase (ALT), (B) aspartate aminotransferase (AST), (C) gamma glutamyl transferase (GGT) and (D) serum levels of total bilirubin. ...

Histological examination of the livers revealed normal hepatic architecture with no significant lobular or portal inflammation, steatosis, or necrosis and mild hepatocyte swelling in mice exposed to IA and placebo (Figure 2A), IA and APAP (Figure 2B), and CIH and placebo (Figure 2C). In contrast, animals exposed to CIH and APAP showed striking liver injury with discernable swelling of hepatocytes (the average score of 3 ± 0 vs 1.3 ± 0.2 in all other groups, p < 0.001) and necrosis with the average score of 1.5 ± 0.6 vs 0 in other groups of animals, p = 0.01. The size of liver necrosis varied from occasional hepatocytes to large confluent areas in zone 3 around the central veins (Figure 2D). Combined CIH-APAP injury led to a 2-fold increase in apoptosis in the liver, whereas CIH or APAP alone had no effect (Figure 3).

Figure 2
Liver histology in mice exposed to (A) intermittent air (IA) and placebo, (B) intermittent air (IA) and acetaminophen (APAP), (C) chronic intermittent hypoxia (CIH) and placebo, (D) chronic intermittent hypoxia (CIH) and acetaminophen (APAP). The livers ...
Figure 3
TUNEL (Terminal deoxynucleotidyl Transferase Biotin-dUTP Nick End Labeling) in liver tissue of mice exposed to (A) intermittent air and placebo, (B) chronic intermittent hypoxia (CIH) and acetaminophen (APAP). White arrows point to apoptotic cells. ( ...

Liver tissue was assessed for markers of oxidative stress and inflammation. In mice injected with placebo, CIH did not have a significant effect on GSH and 3-NT levels, although there was a trend toward increased MDA (p = 0.09, Fig. 4). In mice exposed to IA and APAP, there was no evidence of oxidative stress in the liver. Simultaneous exposure to CIH and APAP depleted GSH stores by 26% and induced a five-fold increase in 3-NT levels in the liver, compared to APAP alone, while MDA levels were not affected (Figure 4). An increase in 3-NT could be related to up-regulation of NO biosynthesis, therefore NOS expression was examined. iNOS mRNA was undetectable, whereas eNOS mRNA was expressed at identical levels (eNOS/18S mRNA ratio = 0.00017 ± 0.00004) in all groups of mice. Taken together, these results suggest that CIH significantly exacerbates oxidative stress induced by APAP preferentially affecting reactive nitrogen species.

Figure 4
The effects of chronic intermittent hypoxia (CIH) and acetaminophen (APAP) on levels of (A) reduced glutathione (GSH), (B) 3-nitrotyrosine (3-NT), and (C) malondialdehyde (MDA) in the liver. * denotes p < 0.05 for the difference between CIH and ...

CIH alone did not induce noticeable inflammatory changes in liver parenchyma. Control IA animals exposed to APAP showed increases in liver IL-6 and TNF-α levels, whereas levels of chemokines MCP-1 and MIP-2 were similar to those of mice treated with placebo (Figure 5). Animals exposed to both CIH and APAP exhibited a 6-fold increase in hepatic MCP-1 and a nearly 40% increase in hepatic MIP-2 (Figure 5). Levels of TNF-α and IL-6 were similar in hypoxic and normoxic mice treated with APAP.

Figure 5
The effects of chronic intermittent hypoxia (CIH) and acetaminophen (APAP) on levels of (A) tumor necrosis factor alpha (TNF-α), (B) interleukin 6 (IL-6), (C) monocyte chemoattractant protein-1 (MCP-1), and (D) macrophage inflammatory protein-2 ...


We have previously shown that CIH leads to mild liver injury in C57BL/6J mice and that CIH can greatly exacerbate toxicity of APAP administered as a single high dose (600 mg/kg). The purpose of this study was to determine whether a combination of CIH and chronic treatment with APAP can cause liver dysfunction and to characterize mechanisms involved. The study resulted in several novel findings. First, mice exposed to CIH and chronic APAP developed liver injury with elevation of serum ALT and histopathologic evidence of increased apoptosis and hepatocellular necrosis, which was not observed in mice exposed to CIH alone or APAP alone. Second, mice exposed to CIH and treated with APAP suffered oxidative injury to the liver, with accumulation of 3-NT and decreased GSH, which was not observed in the other groups of mice. Third, mice exposed to CIH and treated with APAP exhibited increased levels of chemokines MCP-1 and MIP-2 in the liver, compared to other groups of mice. In the discussion below we will examine the possible mechanisms of CIH-APAP interaction and discuss clinical implications of our work.

CIH, APAP and Oxidative Stress

OSA is associated with increased serum levels of malondialdehyde (MDA) and 8-isoprostanes (Lavie, 2003;Lavie et al., 2004). Experimentally induced CIH in rodents causes lipid peroxidation in the brain, carotid bodies, adrenal glands and myocardium (Row et al., 2002;Zhan et al., 2005;Chen et al., 2005;Kumar et al., 2006;Peng et al., 2006a;Peng et al., 2006b). We have shown that CIH leads to increased lipid peroxidation in the liver (Li et al., 2007;Jun et al., 2008) and that liver injury in CIH is associated with increased levels of lipid peroxidation products (Savransky et al., 2007d). Repetitive cycles of hypoxia and reoxygenation increase ROS generation by the mitochondrial electron-transport chain, which is the main source of ROS during normal metabolism (Chen et al., 2003b;Lahiri et al., 2006;Lavie, 2003;Chandel & Schumacker, 2000;Prabhakar & Kumar, 2004). Microsomal enzyme systems such as CYP2E1, CYP4A, xanthine oxidase, nitric oxide synthases, and NADPH oxidase can also participate in the formation of ROS in the liver (Peng et al., 2006a;Lavie, 2003;Choi et al., 2006;Robertson et al., 2001;Jun et al., 2008). Our present data suggest that, although CIH alone tends to increase lipid peroxidation, APAP alone did not have a similar effect. Moreover, APAP abolished a CIH-induced increase in lipid peroxidation. We speculate that this unexpected finding is a result of perturbations in superoxide metabolism. Indeed, superoxide has to be converted to hydrogen peroxide and hydroxyl radicals to induce lipid peroxidation (Halliwell & Gutteridge, 2007). Several reports indicate that hydrogen peroxide and hydroxyl radicals as well as lipid peroxidation are of limited relevance in APAP toxicity, because superoxide is rapidly utilized in the peroxynitrite reaction: NO + O2•- → ONOO- that is several times faster than the superoxide dismutase (SOD) reaction (Halliwell & Gutteridge, 2007;Jaeschke & Bajt, 2006). Of note, conflicting reports indicate that SOD1 is necessary for APAP-induced liver injury (Lei et al., 2006). Nevertheless, it is conceivable that APAP shunts CIH-induced ROS away from lipid peroxidation to the mitochondrial peroxynitrite reaction.

Peroxynitrite leads to rapid protonation, depletion of antioxidant defense, and nitration and oxidation of amino acids (Beckman & Koppenol, 1996;Alvarez & Radi, 2003;Halliwell & Gutteridge, 2007). Nitration of tyrosine resulting in 3-NT is considered a reliable biomarker of peroxynitrite generation. Peroxynitrite and 3-NT damage the mitochondria and cytoskeleton, leading to liver injury (Halliwell & Gutteridge, 2007;Greenacre & Ischiropoulos, 2001;Cover et al., 2005). Exposure of isolated and perfused rat liver to hypoxia resulted in 3-NT accumulation in pericentral regions (Arteel et al., 1999), but the effects of CIH and OSA on hepatic 3-NT levels were not previously studied. In contrast, the role of 3-NT in APAP toxicity is well known (Cover et al., 2005;Hinson et al., 1998;Jaeschke & Bajt, 2006;Knight et al., 2001). We have shown that (a) CIH alone induces mild liver injury without an increase in 3-NT, (b) chronic administration of APAP does not induce liver injury nor increases 3-NT, (c) a combination of CIH and APAP leads to overwhelming liver injury with necrosis in association with a 4-5 fold increase in 3-NT. Since peroxynitrite is generated predominantly in mitochondria (Cover et al., 2005), our findings suggest that CIH and APAP interact to induce mitochondrial dysfunction promoting liver injury. One possibility is that CIH inhibits APAP metabolism resulting in higher APAP levels, which would be a simple and elegant explanation of the CIH-APAP synergy. However, APAP levels have not been measured, which is a limitation of the study.

What are the mechanisms of an increase in hepatic peroxynitrite levels during CIH and APAP treatment? One mechanism would be accelerated generation of peroxynitrite due to an increase in NO production. There is no consensus in the literature on the effects of CIH on NOS. In rodents, CIH increases levels of eNOS in pulmonary vasculature (Snow et al., 2008), decreases eNOS expression in penile tissue (Soukhova-O’Hare et al., 2008), and decreases neuronal NOS expression in the hippocampus (Tjong et al., 2008). Patients with OSA exhibit decreased levels of eNOS and increased levels of iNOS in venous endothelial cells and these changes are reversible by CPAP treatment (Jelic et al., 2008). The effect of CIH on NOS isoforms in the liver has not been previously assessed. APAP appears to increase hepatic NO acting via mechanisms other than iNOS (Jaeschke & Bajt, 2006). Our study of the role of NO was limited, because NO levels and NOS activities have not been measured. Nevertheless, undetectable iNOS and very low levels of eNOS mRNA in all experimental groups may suggest that excessive NO production is not a main mechanism of high peroxynitrite levels in our model.

Another potential mechanism of an increase in peroxynitrite is GSH depletion. Specific depletion of mitochondrial GSH causes accumulation of ROS in hepatocytes exposed to hypoxia in vitro, leading to cell death (Lluis et al., 2005). In the present study we have shown that, while CIH and APAP alone do not affect hepatic GSH levels, a combination of both insults leads to GSH depletion. GSH scavenges peroxynitrite decomposing it to NO (Halliwell & Gutteridge, 2007). Therefore, depletion of GSH by a combination of CIH and APAP can selectively increase peroxynitrite levels, causing mitochondrial damage, liver injury and hepatic necrosis.

CIH, APAP and Hepatic Inflammation

In the absence of APAP, exposure to CIH for four weeks caused an increase in serum AST, whereas elevations in serum ALT and hepatocellular swelling in histological sections previously observed after exposure to CIH for 12 weeks (Savransky et al., 2007d), were not detected. Similar to the previous report (Savransky et al., 2007d), CIH alone did not induce significant inflammation in the liver of lean C57BL/6J mice. In contrast, in mice with diet-induced hepatic steatosis, CIH caused lobular inflammation with increased levels of TNF-α, IL-6, interleukin-1β, and a chemokine MIP-2 (Savransky et al., 2007b). These data suggest that CIH acts as ‘a second hit’, causing hepatitis only in the presence of the primary insult. In the present study, CIH again acts as a powerful ‘second hit’ in APAP-induced liver injury. APAP alone significantly increased levels of IL-6 and TNF-α, consistent with previous reports (Jaeschke, 2005), but MCP-1 and MIP-2 levels rose only in the presence of CIH (Figure 5). TNF-α induces the transcription of chemokines (Bataller & Brenner, 2005;Laurens et al., 2005), but ROS may be required for induction to occur (Chen et al., 2004). We speculate that peroxynitrite up-regulated hepatic MCP-1 and MIP-2 synthesis during combined exposure to CIH and APAP. An increase in these chemotactic factors would then promote liver inflammation and injury.

Nitrosative stress is not the only mechanism of hepatic inflammation in our model. APAP increases apoptosis of hepatocytes by up-regulating apoptosis signal-regulating kinase 1 and c-jun N-terminal kinase (Nakagawa et al., 2008;Gunawan et al., 2006), independently of peroxynitrite. CIH induces apoptosis in the brain and myocardium due to a direct effect of superoxide (Gozal et al., 2001;Shan et al., 2007;Chen et al., 2008). Our data reveal that both noxious stimuli do not affect hepatic apoptosis when administered separately. In contrast, a combination of CIH and APAP leads to a marked increase in apoptosis in the liver, which may contribute to hepatic inflammation and necrosis, independently of peroxynitrite.

Clinical Implications

Our data has significant clinical implications. OSA directly predisposes to several conditions associated with chronic pain (Loh et al., 1999;Goder et al., 2003;Gold et al., 2003;Gold et al., 2004). In addition, OSA is highly prevalent in the obese individuals (Punjabi et al., 2002), who frequently suffer from chronic musculoskeletal pain (Anandacoomarasamy et al., 2008;Janke et al., 2007). APAP is one of the most common analgesics, and chronic treatment with APAP is frequently recommended in chronic pain syndromes associated with OSA and obesity (Felson et al., 2000;Sarzi-Puttini et al., 2005;Bennett et al., 2003). Hence, CIH of OSA and chronic APAP use may frequently coexist in clinical practice. In addition, OSA and obesity are associated with fatty liver disease (Tanne et al., 2005;Tatsumi & Saibara, 2005;Bellentani et al., 2000;Nomura et al., 1988;Browning et al., 2004;Jouet et al., 2007;Singh et al., 2005), which may lead to increased generation of ROS, depletion of GSH, and augmented APAP toxicity (Laurent et al., 2004;Carmiel-Haggai et al., 2005). One limitation of our study is that the APAP dose administered to mice was 2-3-fold higher than recommended for humans. Nevertheless, our data suggest that patients with OSA may be uniquely vulnerable to the hepatic toxicity of APAP, inviting future clinical investigation.


CIH greatly enhances hepatic toxicity of chronic APAP administration resulting in liver necrosis and inflammation. The mechanism of injury appears to be related to the formation of peroxynitrite, depletion of GSH stores and apoptosis of hepatocytes.


This work was supported by National Institute of Health Grants HL68715 and HL80105, and American Heart Association 0765293U to Dr. Vsevolod Y. Polotsky, DA016370 to Dr. Michael S. Torbenson, Pilot and Feasibility Grant of the Clinical Nutrition Research Unit (DK72488) to Dr. Vladimir Savransky, NIH T32 HL07534 and American Lung Association-National Sleep Foundation Pickwick Fellowship SF-78568-N to Dr. Jonathan Jun, American Heart Association Mid-Atlantic Affiliate Postdoctoral Fellowship 0625514U to Dr. Jianguo Li, and Research Fellowship grant RE 2842/1-1 of the German Research Foundation (DFG) to Dr. Christian Reinke.


DISCLOSURE No conflicts of interest to report.

Reference List

  • Alvarez B, Radi R. Peroxynitrite reactivity with amino acids and proteins. Amino Acids. 2003;25:295–311. [PubMed]
  • Anandacoomarasamy A, Caterson I, Sambrook P, Fransen M, March L. The impact of obesity on the musculoskeletal system. Int J Obes (Lond) 2008;32:211–222. [PubMed]
  • Arteel GE, Kadiiska MB, Rusyn I, Bradford BU, Mason RP, Raleigh JA, Thurman RG. Oxidative stress occurs in perfused rat liver at low oxygen tension by mechanisms involving peroxynitrite. Mol Pharmacol. 1999;55:708–715. [PubMed]
  • Bataller R, Brenner DA. Liver fibrosis. J Clin Invest. 2005;115:209–218. [PMC free article] [PubMed]
  • Beckman JS, Koppenol WH. Nitric oxide, superoxide, and peroxynitrite: the good, the bad, and ugly. Am J Physiol. 1996;271:C1424–C1437. [PubMed]
  • Bellentani S, Saccoccio G, Masutti F, Croce LS, Brandi G, Sasso F, Cristanini G, Tiribelli C. Prevalence of and risk factors for hepatic steatosis in Northern Italy. Ann Intern Med. 2000;132:112–117. [PubMed]
  • Bennett RM, Kamin M, Karim R, Rosenthal N. Tramadol and acetaminophen combination tablets in the treatment of fibromyalgia pain: a double-blind, randomized, placebo-controlled study. Am J Med. 2003;114:537–545. [PubMed]
  • Bixler EO, Vgontzas AN, Lin HM, Ten Have T, Rein J, Vela-Bueno A, Kales A. Prevalence of sleep-disordered breathing in women: effects of gender. Am J Respir Crit Care Med. 2001;163:608–613. [PubMed]
  • Browning JD, Szczepaniak LS, Dobbins R, Nuremberg P, Horton JD, Cohen JC, Grundy SM, Hobbs HH. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology. 2004;40:1387–1395. [PubMed]
  • Carmiel-Haggai M, Cederbaum AI, Nieto N. A high-fat diet leads to the progression of non-alcoholic fatty liver disease in obese rats. FASEB J. 2005;19:136–138. [PubMed]
  • Chandel NS, Schumacker PT. Cellular oxygen sensing by mitochondria: old questions, new insight. J Appl Physiol. 2000;88:1880–1889. [PubMed]
  • Chen C, Hennig GE, Manautou JE. Hepatobiliary excretion of acetaminophen glutathione conjugate and its derivatives in transport-deficient (TR-) hyperbilirubinemic rats. Drug Metab Dispos. 2003a;31:798–804. [PubMed]
  • Chen L, Einbinder E, Zhang Q, Hasday J, Balke CW, Scharf SM. Oxidative stress and left ventricular function with chronic intermittent hypoxia in rats. Am J Respir Crit Care Med. 2005;172:915–920. [PubMed]
  • Chen L, Zhang J, Gan TX, Chen-Izu Y, Hasday JD, Karmazyn M, Balke CW, Scharf SM. Left ventricular dysfunction and associated cellular injury in rats exposed to chronic intermittent hypoxia. J Appl Physiol. 2008;104:218–223. [PubMed]
  • Chen Q, Vazquez EJ, Moghaddas S, Hoppel CL, Lesnefsky EJ. Production of reactive oxygen species by mitochondria: central role of complex III. J Biol Chem. 2003b;278:36027–36031. [PubMed]
  • Chen XL, Zhang Q, Zhao R, Medford RM. Superoxide, H2O2, and iron are required for TNF-alpha-induced MCP-1 gene expression in endothelial cells: role of Rac1 and NADPH oxidase. Am J Physiol Heart Circ Physiol. 2004;286:H1001–H1007. [PubMed]
  • Choi SS, Sicklick JK, Ma Q, Yang L, Huang J, Qi Y, Chen W, Li YX, Goldschmidt-Clermont PJ, Diehl AM. Sustained activation of Rac1 in hepatic stellate cells promotes liver injury and fibrosis in mice. Hepatology. 2006;44:1267–1277. [PubMed]
  • Cover C, Mansouri A, Knight TR, Bajt ML, Lemasters JJ, Pessayre D, Jaeschke H. Peroxynitrite-induced mitochondrial and endonuclease-mediated nuclear DNA damage in acetaminophen hepatotoxicity. J Pharmacol Exp Ther. 2005;315:879–887. [PubMed]
  • Felson DT, Lawrence RC, Hochberg MC, McAlindon T, Dieppe PA, Minor MA, Blair SN, Berman BM, Fries JF, Weinberger M, Lorig KR, Jacobs JJ, Goldberg V. Osteoarthritis: new insights. Part 2: treatment approaches. Ann Intern Med. 2000;133:726–737. [PubMed]
  • Gastaut H, Tassinari CA, Duron B. Polygraphic study of the episodic diurnal and nocturnal (hypnic and respiratory) manifestations of the Pickwick syndrome. Brain Res. 1966;1:167–186. [PubMed]
  • Goder R, Friege L, Fritzer G, Strenge H, Aldenhoff JB, Hinze-Selch D. Morning headaches in patients with sleep disorders: a systematic polysomnographic study. Sleep Med. 2003;4:385–391. [PubMed]
  • Gold AR, Dipalo F, Gold MS, Broderick J. Inspiratory airflow dynamics during sleep in women with fibromyalgia. Sleep. 2004;27:459–466. [PubMed]
  • Gold AR, Dipalo F, Gold MS, O’Hearn D. The symptoms and signs of upper airway resistance syndrome: a link to the functional somatic syndromes. Chest. 2003;123:87–95. [PubMed]
  • Gozal D, Daniel JM, Dohanich GP. Behavioral and anatomical correlates of chronic episodic hypoxia during sleep in the rat. J Neurosci. 2001;21:2442–2450. [PubMed]
  • Greenacre SA, Ischiropoulos H. Tyrosine nitration: localisation, quantification, consequences for protein function and signal transduction. Free Radic Res. 2001;34:541–581. [PubMed]
  • Gunawan BK, Liu ZX, Han D, Hanawa N, Gaarde WA, Kaplowitz N. c-Jun N-terminal kinase plays a major role in murine acetaminophen hepatotoxicity. Gastroenterology. 2006;131:165–178. [PubMed]
  • Halliwell B, Gutteridge JMC. Free radicals in biology and medicine. 2007. Ref Type: Generic.
  • Hinson JA, Pike SL, Pumford NR, Mayeux PR. Nitrotyrosine-protein adducts in hepatic centrilobular areas following toxic doses of acetaminophen in mice. Chem Res Toxicol. 1998;11:604–607. [PubMed]
  • Jaeschke H. Role of inflammation in the mechanism of acetaminophen-induced hepatotoxicity. Expert Opin Drug Metab Toxicol. 2005;1:389–397. [PubMed]
  • Jaeschke H, Bajt ML. Intracellular signaling mechanisms of acetaminophen-induced liver cell death. Toxicol Sci. 2006;89:31–41. [PubMed]
  • Janke EA, Collins A, Kozak AT. Overview of the relationship between pain and obesity: What do we know? Where do we go next? J Rehabil Res Dev. 2007;44:245–262. [PubMed]
  • Jelic S, Padeletti M, Kawut SM, Higgins C, Canfield SM, Onat D, Colombo PC, Basner RC, Factor P, LeJemtel TH. Inflammation, oxidative stress, and repair capacity of the vascular endothelium in obstructive sleep apnea. Circulation. 2008;117:2270–2278. [PMC free article] [PubMed]
  • Jouet P, Sabate JM, Maillard D, Msika S, Mechler C, Ledoux S, Harnois F, Coffin B. Relationship between obstructive sleep apnea and liver abnormalities in morbidly obese patients: a prospective study. Obes Surg. 2007;17:478–485. [PubMed]
  • Jun JC, Savransky V, Nanayakkara A, Bevans S, Li J, Smith PL, Polotsky VY. Intermittent Hypoxia has Organ-Specific Effects on Oxidative Stress. Am J Physiol Regul Integr Comp Physiol. 2008 [PubMed]
  • Kallwitz ER, Herdegen J, Madura J, Jakate S, Cotler SJ. Liver enzymes and histology in obese patients with obstructive sleep apnea. J Clin Gastroenterol. 2007;41:918–921. [PubMed]
  • Kheirandish-Gozal L, Sans CO, Kheirandish E, Gozal D. Elevated serum aminotransferase levels in children at risk for obstructive sleep apnea. Chest. 2008;133:92–99. [PubMed]
  • Kleiner DE, Brunt EM, Van NM, Behling C, Contos MJ, Cummings OW, Ferrell LD, Liu YC, Torbenson MS, Unalp-Arida A, Yeh M, McCullough AJ, Sanyal AJ. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology. 2005;41:1313–1321. [PubMed]
  • Knight TR, Kurtz A, Bajt ML, Hinson JA, Jaeschke H. Vascular and hepatocellular peroxynitrite formation during acetaminophen toxicity: role of mitochondrial oxidant stress. Toxicol Sci. 2001;62:212–220. [PubMed]
  • Kumar GK, Rai V, Sharma SD, Ramakrishnan DP, Peng YJ, Souvannakitti D, Prabhakar NR. Chronic intermittent hypoxia induces hypoxia-evoked catecholamine efflux in adult rat adrenal medulla via oxidative stress. J Physiol. 2006;575:229–239. [PubMed]
  • Lahiri S, Roy A, Baby SM, Hoshi T, Semenza GL, Prabhakar NR. Oxygen sensing in the body. Prog Biophys Mol Biol. 2006;91:249–286. [PubMed]
  • Laurens M, Defamie V, Scozzari G, Schmid-Alliana A, Gugenheim J, Crenesse D. Hypoxia-reoxygenation-induced chemokine transcription is not prevented by preconditioning or intermittent hypoxia, in mice hepatocytes. Transpl Int. 2005;18:444–452. [PubMed]
  • Laurent A, Nicco C, Tran VN, Borderie D, Chereau C, Conti F, Jaffray P, Soubrane O, Calmus Y, Weill B, Batteux F. Pivotal role of superoxide anion and beneficial effect of antioxidant molecules in murine steatohepatitis. Hepatology. 2004;39:1277–1285. [PubMed]
  • Lavie L. Obstructive sleep apnoea syndrome - an oxidative stress disorder. Sleep Med Rev. 2003;7:35–51. [PubMed]
  • Lavie L, Vishnevsky A, Lavie P. Evidence for lipid peroxidation in obstructive sleep apnea. Sleep. 2004;27:123–128. [PubMed]
  • Lei XG, Zhu JH, McClung JP, Aregullin M, Roneker CA. Mice deficient in Cu,Zn-superoxide dismutase are resistant to acetaminophen toxicity. Biochem J. 2006;399:455–461. [PubMed]
  • Li J, Grigoryev DN, Ye SQ, Thorne L, Schwartz AR, Smith PL, O’Donnell CP, Polotsky VY. Chronic intermittent hypoxia upregulates genes of lipid biosynthesis in obese mice. J Appl Physiol. 2005a;99:1643–1648. [PubMed]
  • Li J, Savransky V, Nanayakkara A, Smith PL, O’Donnell CP, Polotsky VY. Hyperlipidemia and lipid peroxidation are dependent on the severity of chronic intermittent hypoxia. J Appl Physiol. 2007;102:557–563. [PubMed]
  • Li J, Thorne LN, Punjabi NM, Sun CK, Schwartz AR, Smith PL, Marino RL, Rodriguez A, Hubbard WC, O’Donnell CP, Polotsky VY. Intermittent hypoxia induces hyperlipidemia in lean mice. Circ Res. 2005b;97:698–706. [PubMed]
  • Lluis JM, Morales A, Blasco C, Colell A, Mari M, Garcia-Ruiz C, Fernandez-Checa JC. Critical role of mitochondrial glutathione in the survival of hepatocytes during hypoxia. J Biol Chem. 2005;280:3224–3232. [PubMed]
  • Loh NK, Dinner DS, Foldvary N, Skobieranda F, Yew WW. Do patients with obstructive sleep apnea wake up with headaches? Arch Intern Med. 1999;159:1765–1768. [PubMed]
  • Marin JM, Carrizo SJ, Vicente E, Agusti AG. Long-term cardiovascular outcomes in men with obstructive sleep apnoea-hypopnoea with or without treatment with continuous positive airway pressure: an observational study. Lancet. 2005;365:1046–1053. [PubMed]
  • Michael SL, Mayeux PR, Bucci TJ, Warbritton AR, Irwin LK, Pumford NR, Hinson JA. Acetaminophen-induced hepatotoxicity in mice lacking inducible nitric oxide synthase activity. Nitric Oxide. 2001;5:432–441. [PubMed]
  • Nakagawa H, Maeda S, Hikiba Y, Ohmae T, Shibata W, Yanai A, Sakamoto K, Ogura K, Noguchi T, Karin M, Ichijo H, Omata M. Deletion of Apoptosis Signal-Regulating Kinase 1 Attenuates Acetaminophen-Induced Liver Injury by Inhibiting c-Jun N-Terminal Kinase Activation. Gastroenterology. 2008 [PubMed]
  • Nieto FJ, Young TB, Lind BK, Shahar E, Samet JM, Redline S, D’Agostino RB, Newman AB, Lebowitz MD, Pickering TG. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study. JAMA. 2000;283:1829–1836. [see comments] [PubMed]
  • Nomura H, Kashiwagi S, Hayashi J, Kajiyama W, Tani S, Goto M. Prevalence of fatty liver in a general population of Okinawa, Japan. Jpn J Med. 1988;27:142–149. [PubMed]
  • Norman D, Bardwell WA, Arosemena F, Nelesen R, Mills PJ, Loredo JS, Lavine JE, Dimsdale JE. Serum aminotransferase levels are associated with markers of hypoxia in patients with obstructive sleep apnea. Sleep. 2008;31:121–126. [PubMed]
  • Pelletier SJ, Raymond DP, Crabtree TD, Berg CL, Iezzoni JC, Hahn YS, Sawyer RG, Pruett TL. Hepatitis C-induced hepatic allograft injury is associated with a pretransplantation elevated viral replication rate. Hepatology. 2000;32:418–426. [PubMed]
  • Peng YJ, Yuan G, Jacono FJ, Kumar GK, Prabhakar NR. 5-HT evokes sensory long-term facilitation of rodent carotid body via activation of NADPH oxidase. J Physiol. 2006a;576:289–295. [PubMed]
  • Peng YJ, Yuan G, Ramakrishnan D, Sharma SD, Bosch-Marce M, Kumar GK, Semenza GL, Prabhakar NR. Heterozygous HIF-1{alpha} deficiency impairs carotid body-mediated systemic responses and reactive oxygen species generation in mice exposed to intermittent hypoxia. J Physiol. 2006b;577:705–716. [PubMed]
  • Peppard PE, Young T, Palta M, Skatrud J. Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med. 2000;342:1378–1384. [PubMed]
  • Polotsky VY, Li J, Punjabi NM, Rubin AE, Smith PL, Schwartz AR, O’Donnell CP. Intermittent hypoxia increases insulin resistance in genetically obese mice. J Physiol. 2003;552:253–264. [PubMed]
  • Polotsky VY, Rubin AE, Balbir A, Dean T, Smith PL, Schwartz AR, O’Donnell CP. Intermittent hypoxia causes REM sleep deficits and decreases EEG delta power in NREM sleep in the C57BL/6J mouse. Sleep Med. 2006;7:7–16. [PubMed]
  • Prabhakar NR, Kumar GK. Oxidative stress in the systemic and cellular responses to intermittent hypoxia. Biol Chem. 2004;385:217–221. [PubMed]
  • Prabhakar NR, Kumar GK, Nanduri J, Semenza GL. ROS signaling in systemic and cellular responses to chronic intermittent hypoxia. Antioxid Redox Signal. 2007;9:1397–1403. [PubMed]
  • Punjabi NM, Ahmed MM, Polotsky VY, Beamer BA, O’Donnell CP. Sleep-disordered breathing, glucose intolerance, and insulin resistance. Respiration Physiology and Neurobiology. 2003 [PubMed]
  • Punjabi NM, Polotsky VY. Disorders of glucose metabolism in sleep apnea. J Appl Physiol. 2005;99:1998–2007. [PubMed]
  • Punjabi NM, Shahar E, Redline S, Gottlieb DJ, Givelber R, Resnick HE. Sleep-disordered breathing, glucose intolerance, and insulin resistance: the Sleep Heart Health Study. Am J Epidemiol. 2004;160:521–530. [PubMed]
  • Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP, Schwartz AR, Smith PL. Sleep-disordered breathing and insulin resistance in middle-aged and overweight men. Am J Respir Crit Care Med. 2002;165:677–682. [PubMed]
  • Robertson G, Leclercq I, Farrell GC. Nonalcoholic steatosis and steatohepatitis. II. Cytochrome P-450 enzymes and oxidative stress. Am J Physiol Gastrointest Liver Physiol. 2001;281:G1135–G1139. [PubMed]
  • Row BW, Kheirandish L, Neville JJ, Gozal D. Impaired spatial learning and hyperactivity in developing rats exposed to intermittent hypoxia. Pediatr Res. 2002;52:449–453. [PubMed]
  • Sarzi-Puttini P, Cimmino MA, Scarpa R, Caporali R, Parazzini F, Zaninelli A, Atzeni F, Canesi B. Osteoarthritis: an overview of the disease and its treatment strategies. Semin Arthritis Rheum. 2005;35:1–10. [PubMed]
  • Savransky V, Bevans S, Nanayakkara A, Li J, Smith PL, Torbenson MS, Polotsky VY. Chronic Intermittent Hypoxia Causes Hepatitis in a Mouse Model of Diet-Induced Fatty Liver. Am J Physiol Gastrointest Liver Physiol. 2007a [PubMed]
  • Savransky V, Bevans S, Nanayakkara A, Li J, Smith PL, Torbenson MS, Polotsky VY. Chronic intermittent hypoxia causes hepatitis in a mouse model of diet-induced fatty liver. Am J Physiol Gastrointest Liver Physiol. 2007b;293:G871–G877. [PubMed]
  • Savransky V, Nanayakkara A, Li J, Bevans S, Smith PL, Rodriguez A, Polotsky VY. Chronic intermittent hypoxia induces atherosclerosis. Am J Respir Crit Care Med. 2007c;175:1290–1297. [PMC free article] [PubMed]
  • Savransky V, Nanayakkara A, Vivero A, Li J, Bevans S, Smith PL, Torbenson MS, Polotsky VY. Chronic intermittent hypoxia predisposes to liver injury. Hepatology. 2007d;45:1007–1013. [PubMed]
  • Shahar E, Whitney CW, Redline S, Lee ET, Newman AB, Javier NF, O’Connor GT, Boland LL, Schwartz JE, Samet JM. Sleep-disordered breathing and cardiovascular disease: cross-sectional results of the Sleep Heart Health Study. Am J Respir Crit Care Med. 2001;163:19–25. [PubMed]
  • Shan X, Chi L, Ke Y, Luo C, Qian S, Gozal D, Liu R. Manganese superoxide dismutase protects mouse cortical neurons from chronic intermittent hypoxia-mediated oxidative damage. Neurobiol Dis. 2007;28:206–215. [PMC free article] [PubMed]
  • Singh H, Pollock R, Uhanova J, Kryger M, Hawkins K, Minuk GY. Symptoms of obstructive sleep apnea in patients with nonalcoholic fatty liver disease. Dig Dis Sci. 2005;50:2338–2343. [PubMed]
  • Snow JB, Kitzis V, Norton CE, Torres SN, Johnson KD, Kanagy NL, Walker BR, Resta TC. Differential effects of chronic hypoxia and intermittent hypocapnic and eucapnic hypoxia on pulmonary vasoreactivity. J Appl Physiol. 2008;104:110–118. [PubMed]
  • Soukhova-O’Hare GK, Shah ZA, Lei Z, Nozdrachev AD, Rao CV, Gozal D. Erectile dysfunction in a murine model of sleep apnea. Am J Respir Crit Care Med. 2008;178:644–650. [PMC free article] [PubMed]
  • Tanne F, Gagnadoux F, Chazouilleres O, Fleury B, Wendum D, Lasnier E, Lebeau B, Poupon R, Serfaty L. Chronic liver injury during obstructive sleep apnea. Hepatology. 2005;41:1290–1296. [PubMed]
  • Tatsumi K, Saibara T. Effects of obstructive sleep apnea syndrome on hepatic steatosis and nonalcoholic steatohepatitis. Hepatol Res. 2005 [PubMed]
  • Tjong YW, Li M, Hung MW, Wang K, Fung ML. Nitric oxide deficit in chronic intermittent hypoxia impairs large conductance calcium-activated potassium channel activity in rat hippocampal neurons. Free Radic Biol Med. 2008;44:547–557. [PubMed]
  • Yaggi HK, Concato J, Kernan WN, Lichtman JH, Brass LM, Mohsenin V. Obstructive sleep apnea as a risk factor for stroke and death. N Engl J Med. 2005;353:2034–2041. [PubMed]
  • Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med. 1993;328:1230–1235. [PubMed]
  • Zamora-Valdes D, Mendez-Sanchez N. Experimental evidence of obstructive sleep apnea syndrome as a second hit accomplice in nonalcoholic steatohepatitis pathogenesis. Ann Hepatol. 2007;6:281–283. [PubMed]
  • Zhan G, Serrano F, Fenik P, Hsu R, Kong L, Pratico D, Klann E, Veasey SC. NADPH oxidase mediates hypersomnolence and brain oxidative injury in a murine model of sleep apnea. Am J Respir Crit Care Med. 2005;172:921–929. [PMC free article] [PubMed]