Thursday, April 9, 2009

Hepatorenal Syndrome

Initial reports by Frerichs (1861) and Flint (1863) [1], who had noted an association between advanced liver disease with ascites and acute oliguric renal failure in the absence of significant histological changes in the kidneys, led Heyd [2], and later Helwig and Schutz [3], to introduce the concept of the hepatorenal syndrome (HRS) to explain the increased frequency of acute renal failure after biliary surgery. However, because HRS could not be reproduced in animal models, pathophysiological concepts remained speculative and its clinical entity was not generally accepted. During the 1950s, HRS was more specifically characterised as a functional renal failure in patients with advanced liver disease, electrolyte disturbances and low urinary sodium concentrations [4]. Hecker and Sherlock [5] showed its temporal reversibility by norepinephrine administration. Over the next few decades, haemodynamic and perfusion studies by Epstein and other investigators [6] identified splanchnic and systemic vasodilatation and active renal vasoconstriction as the pathophysiological hallmarks of HRS. Improved models of ascites and circulatory dysfunction contributed to therapeutic advances, including the introduction of large-volume paracentesis, vasopressin analogues, and transjugular intrahepatic stent-shunt (TIPS), which in turn have led to an improved pathophysiological understanding of HRS [7]. Definition HRS is defined as the development of renal failure in patients with severe liver disease (acute or chronic) in the absence of any other identifiable cause of renal pathology. It is diagnosed following the exclusion of other causes of renal failure in patients with liver disease, such as hypovolaemia, drug nephrotoxicity, sepsis or glomerulonephritis. A similar syndrome can also occur in the setting of acute liver failure [8]. In the kidney there is marked renal vasoconstriction, resulting in a low glomerular filtration rate (GFR). In the extrarenal circulation arterial vasodilatation predominates, resulting in reduction of the total systemic vascular resistance and arterial hypotension [9].

Diagnostic Criteria
The International Ascites Club (1996) group has defined the diagnostic criteria for HRS, and these are listed in Table 1 [8].

Major criteria
Chronic or acute liver disease with advanced hepatic failure and portal hypertension
Low GFR, as indicated by serum creatinine > 1.5 mg/dl or 24-h creatinine clearance <> 500 g/d for several days in patients with ascites without peripheral oedema or > 1000 ml in patients with peripheral oedema)
No sustained improvement in renal function (decrease of serum creatinine to 1.5 mg/dl or less or increase in 24 h creatinine clearance to 40 ml/min or more) after withdrawal of diuretics and expansion of plasma volume with 1.5 l of isotonic saline
Proteinuria < style="font-weight: bold;">Additional criteria
Urine sodium <> plasma osmolality
Urine red blood cells <>

HRS occurs in about 4% of patients admitted to hospital with decompensated cirrhosis, the cumulative probability being 18% at 1 year, increasing to 39% at 5 years. Retrospective studies [12] indicate that HRS is present in approximately 17% of patients admitted to hospital with ascites and in more than 50% of cirrhotic patients dying of liver failure. The most frequent cause of renal failure in cirrhosis is spontaneous bacterial peritonitis (SBP). Approximately 30% of patients with SBP develop renal failure. Type 1 HRS is characterised by rapid and progressive renal impairment and is precipitated most commonly by SBP. Type 1 HRS occurs in approximately 25% of patients with SBP, even when rapid resolution of the infection is obtained with antibiotics. Without treatment, the median survival of patients with HRS type 1 is less than 2 weeks, and virtually all patients die within 10 weeks after the onset of renal failure. Type 2 HRS is characterised by a moderate and stable reduction in GFR and commonly occurs in patients with relatively well-preserved hepatic function. The median survival is 3–6 months. Although this is markedly longer than that in type 1 HRS, it is still shorter than that of patients with cirrhosis and ascites who do not have renal failure. People of all races who have chronic liver disease are at risk of HRS, and its frequency is equal in both sexes; most patients with chronic liver disease are in the 4th–8th decade of life.

In a prospective study published by Gines et al., once HRS had developed the median survival was only 1.7 weeks, and it was poorer particularly in patients with apparent precipitating factors. Overall survival at 4 and 10 weeks was 20% and 10%, respectively. Patients with low urinary sodium excretion (<>

Table 2. Risk factors for development of hepatorenal syndrome
Previous episodes of ascites
Absence of hepatomegaly
Poor nutritional status
Presence of oesophageal varices
Serum sodium <> 553 mosmol/l
Norepinephrine levels > 544pg/ml
Plasma renin activity > 3.5 ng /ml
Mean arterial pressure < class="MsoNormal">

The hallmark of HRS is renal vasoconstriction, although the pathogenesis is not fully understood.Multiple mechanisms are probably involved and include interplay between disturbances in systemic haemodynamics, activation of vasoconstrictor systems and a reduction in activity of the vasodilator systems [16–19]. The haemodynamic pattern of patients with HRS is characterised by increased cardiac output, low arterial pressure and reduced systemic vascular resistance. Renal vasoconstriction occurs in the absence of reduced cardiac output and blood volume, which is a point of contrast to most clinical conditions associated with renal hypoperfusion. Although the pattern of increased renal vascular resistance and decreased peripheral resistance is characteristic of HRS, it also occurs in other conditions, such as anaphylaxis and sepsis. Doppler studies of the brachial, middle cerebral and femoral arteries suggest that extrarenal resistance is increased in patients with HRS, while the splanchnic circulation is responsible for arterial vasodilatation and reduced total systemic vascular resistance. The renin-angiotensin-aldosterone system (RAAS) and the sympathetic nervous system (SNS) are the predominant systems responsible for renal vasoconstriction [20]. The activity of both systems is increased in patients with cirrhosis and ascites, and this effect is magnified in HRS. In contrast, an inverse relationship exists between the activity of these two systems and renal plasma flow (RPF) and the glomerular filtration rate (GFR). Endothelin is another renal vasoconstrictor that is present in increased concentration in HRS, although its role in the pathogenesis of this syndrome has yet to be identified. Adenosine is well known for its vasodilator properties, although it acts as a vasoconstrictor in the lungs and kidneys. Elevated levels of adenosine are more common in patients with heightened activity of the RAAS and may work synergistically with angiotensin II to produce renal vasoconstriction in HRS. This effect has also been described with the powerful renal vasoconstrictor, leukotriene E4. The vasoconstricting effect of these various systems is antagonised by local renal vasodilatory factors, the most important of which are the prostaglandins. Perhaps the strongest evidence supporting their role in renal perfusion is the marked decrease in RPF and the GFR when nonsteroidal medications known to bring about a sharp reduction in PG levels are administered. Nitrous oxide (NO) is another vasodilator that is believed to play an important part in renal perfusion. Preliminary studies, predominantly based on animal experiments, have demonstrated that NO production is increased in the presence of cirrhosis, although NO inhibition does not result in renal vasoconstriction owing to a compensatory increase in PG synthesis.However, when both NO and PG production are inhibited, marked renal vasoconstriction develops. These findings demonstrate that renal vasodilators have a critical role in maintaining renal perfusion, particularly in the presence of overactivity of renal vasoconstrictors. However, we do not yet know for certain whether vasoconstrictor activity becomes the predominant system in HRS and whether a reduction in the activity of the vasodilator system contributes to this [21–29].Various theories have been proposed to explain the development of HRS in cirrhosis. The two main ones are the arterial vasodilatation theory and the hepatorenal reflex theory. The first not only describes sodium and water retention in cirrhosis, but may also be the most rational hypothesis for the development of HRS. Splanchnic arteriolar vasodilatation in patients with compensated cirrhosis and portal hypertension may be mediated by several factors, the most important of which is probably NO. In the early phases of portal hypertension and compensated cirrhosis, this underfilling of the arterial bed causes a decrease in the effective arterial blood volume and results in homeostatic reflex activation of the endogenous vasoconstrictor systems. Activation of the RAAS and SNS occurs early with antidiuretic hormone secretion, a later event when a more marked derangement in circulatory function is present. This results in vasoconstriction not only of the renal vessels, but also in the vascular beds of the brain,muscle, spleen and extremities. The splanchnic circulation is resistant to these effects because of the continuous production of local vasodilators, such as NO. In the early phases of portal hypertension, renal perfusion is maintained within normal or near-normal limits as the vasodilatory systems antagonise the renal effects of the vasoconstrictor systems. However, as liver disease progress in severity, a critical level of vascular underfilling is achieved; renal vasodilatory systems are unable to counteract the maximal activation of the endogenous vasoconstrictors and/or intrarenal vasoconstrictors, which leads to uncontrolled renal vasoconstriction. Support for this hypothesis is provided by studies in which the administration of splanchnic vasoconstrictors in combination with volume expanders results in improvement in arterial pressure, RPF and GFR [30–34]. The alternative theory proposes that renal vasoconstriction in HRS is not related to systemic haemodynamics but is due either to a deficiency in the synthesis of a vasodilator factor or to a hepatorenal reflex that leads to renal vasoconstriction. Evidence points to the vasodilatation theory as a more tangible explanation for the development of HRS.

Histopathology of HRS
In previous definitions of HRS, changes in renal histology were reported to be absent or minimal, which reflected a rapid progression to death after development of HRS. Considering that many patients with HRS currently receive aggressive supportive treatment including renal replacement therapy to prolong survival until liver transplantation, it seems obvious that prolonged renal hypoperfusion, renal medullary hypoxia and the high frequency of infectious complications ultimately contribute to histologically detectable renal damage. However, it is increasingly recognised that structural renal damage may already be found even before renal dysfunction becomes manifest. In a series of cirrhotic patients [35] undergoing liver transplantation, 100% of renal biopsies showed glomerular abnormalities. Tubular function is usually well preserved at the time when HRS develops, but tubular abnormalities, including increased B2 microglobulin excretion, have been reported in deeply jaundiced patients with HRS [36–38]. With progressive circulatory dysfunction, prolonged renal hypoperfusion may eventually result in acute tubular necrosis by increasing the susceptibility to additional insults by radiographic contrast agents, aminoglycosides, haemorrhage, endotoxinaemia or any other cause of medullary hypoxia. The presence of acute tubular necrosis could partially explain the slow or absent renal recovery in HRS type 1 even after the initiation of vasopressor support. For instance, a recent case study reports full recovery of renal function in dialysis-dependent HRS after 7 weeks of treatment with ornipressin, dopamine and intravenous albumin [39].

The following measures may decrease the incidence of renal failure or HRS in patients with liver disease.

Prophylaxis Against
Bacterial Infections Bacterial infections occur in approximately 50% of patients with variceal haemorrhage, and antibiotic prophylaxis improves survival by approximately 10%. Patients who have had a previous episode of SBP have a 68% chance of recurrent infection at 1 year, and this carries a 33% chance of developing renal failure. As bacterial infections are an important cause of renal dysfunction in cirrhotic patients, prophylaxis with antibiotics is recommended in two clinical settings, namely variceal bleeding and a history of previous SBP [40, 41].

Volume Expansion
To prevent the development of renal failure in patients who develop SBP, it is now recommended that plasma volume expansion should be implemented in these patients by giving 20% albumin (1–1.5 g/kg over 1–3 days) at diagnosis to prevent circulatory dysfunction, renal impairment and mortality. Use of low-salt albumin as fluid replacement in patients undergoing large-volume paracentesis (8 g for each litre of ascitic fluid removed) is known to prevent paracentesis-induced circulatory dysfunction [42–45].

Judicious Use of Diuretics
It is important to identify the lowest effective dose of a diuretic for any individual patient, as diuretic-induced renal impairment is seen in approximately 20% of patients with ascites. It develops when the rate of diuresis exceeds the rate of ascites reabsorption, leading to intravascular volume depletion. Diuretic-induced renal impairment is usually moderate and rapidly reversible following diuretic withdrawal.

Avoidance of Nephrotoxic Drugs
Patients with cirrhosis and ascites are predisposed to the development of acute tubular necrosis during the use of aminoglycosides, with renal failure occurring in 33% of such patients as against 3–5% in the general population. Another important cause of renal failure is the use of nonsteroidal antiinflammatory drugs (NSAIDs) [46].

The ideal treatment for HRS is liver transplantation; however, because of the long waiting lists in the majority of transplant centres, most patient die before being offered a transplant. There is an urgent need for effective alternative therapies to increase survival chances for patients with HRS until transplantation can be performed. Treatment can be divided into initial management, pharmacological treatment and surgical manoeuvres.

Initial Management Optimise fluid management.
Renal function rarely recovers in the absence of liver recovery. The key goal in the management of these patients is to exclude reversible or treatable lesions (mainly hypovolaemia) and to support the patient until liver recovery or liver transplantation. The treatment of HRS is directed at reversing the haemodynamic changes induced by reduced renal perfusion pressure, stimulated sympathetic nervous system and increased synthesis of humoral and renal vasoconstrictor factors. In cirrhotic patients renal failure is frequently secondary to hypovolaemia (diuretics or gastrointestinal bleeding), NSAIDs or sepsis. Precipitating factors should be recognised and treated and nephrotoxic drugs, discontinued. All patients should be challenged with up 1.5 l of fluid, such as albumin solution or normal saline, to assess the renal response, as many patients with subclinical hypovolaemia will respond to this simple measure. This should be done with careful monitoring to avoid fluid overload. In practice, fluid overload is not usually a problem, as patients with severe liver disease function as ‘fluid sumps’ and their vasculature adapts to accommodate the extra fluid. This has been described by Hadengue et al., who reported increased venous compliance following fluid challenge in advanced cirrhosis [47, 48]. Monitor for sepsis. Evidence of sepsis should be sought in blood, ascitic, cannulae and urine cultures, and nonnephrotoxic broad-spectrum antibiotics should be started regardless of whether such evidence is found, as any delay in effective treatment of undiagnosed infection can increase mortality. In advanced cirrhosis, endotoxins and cytokines play important parts in fostering the hyperdynamic circulation and worsening renal function. Optimise blood pressure. If mean arterial pressure is low (<>

Pharmacological Treatment
All the drugs that have been investigated in HRS have one overriding aim: to increase renal blood flow. This has been achieved either indirectly, by splanchnic vasoconstriction, or directly, using renal vasodilators. One of the principal difficulties has been the lack of agents that act purely on the splanchnic circulation. Drugs that ‘spill over’ into the systemic circulation may actually exacerbate the intense renal vasoconstriction already present. Currently, there is significant enthusiasm for the use of vasoconstrictor agents in HRS. However, the numbers of patients studied have been small, mortality remains high and there have been no randomised placebo-controlled trials. This deficit clearly needs to be addressed but the possibilities are limited by the relative rarity of patients with ‘pure’ HRS without such confounding variables as sepsis and gastrointestinal bleeding. Important aspect of the situation mentioned in these reports are the need for a pressor response to the agents used and the recurrence of abnormal renal function after the cessation of vasoconstrictor therapy. HRS is effectively a marker of poor hepatic function, and these agents are probably best utilised as a bridge to further improvement in liver function following either cessation of alcohol abuse or liver transplantation. Thus, the decision to use vasoconstrictor agents for HRS should be based on whether the patient is a realistic transplant candidate and, if not, whether liver function might improve. Patients who do not satisfy these criteria will be tested unnecessarily, merely prolonging the process of dying when palliative care would be more appropriate. Dopamine. Nonpressor renal doses of dopamine [2–5 μg kg–1 min–1) are frequently prescribed to patients with acute deterioration of renal function. As shown by a recent, large scale, randomised trial, early renal dose dopamine has no role in the prevention of acute renal failure in critically ill patients and does not significantly improve renal function in patients with HRS.At higher doses, dopamine worsens the hyperdynamic circulation by exaggerating splanchnic hyperaemia and increasing portal pressure and may cause tachyarrhythmia. Thus, the use of dopamine monotherapy seems to offer no benefit in HRS. Combination therapy with dopamine and vasopressors has produced inconsistent results in HRS. Because beneficial renal effects have been reported only with vasopressor, and not with dopamine, monotherapy, it seems unlikely that dopamine contributed to renal improvement in these studies [51–54]. Misoprostol. Misoprostol, a synthetic prostaglandin E-1 analogue, has been used to reverse renal vasoconstriction in HRS. Low doses of misoprostol are vasodilatory, natriuretic and diuretic, whereas high-dose misoprostol increases renal vascular tone and inhibits sodium and water excretion. None of the five studies investigating misoprostol in HRS seems to indicate substantial benefit. Improvement of renal function occurred in 1 of these studies, but could also be explained by volume expansion [55]. N-Acetylcysteine. In 1999, the group at the Royal Free Hospital reported their experience with N-acetylcysteine (NAC) for the treatment of HRS [56]. This was based on experimental models of acute cholestasis, in which the administration of NAC resulted in an improvement in renal function. Twelve patients with HRS were treated with intravenous NAC, without any adverse effects, and the survival rates were 67% and 58% at 1 month and 3 months, respectively (this included 2 patients who received liver transplantation after improvement in renal function). The mechanism of action remains unknown, but this interesting study encourages further optimism for medical treatment of a condition that once carried a hopeless diagnosis without liver transplantation. Controlled studies with longer follow-up may help answer these pressing questions. Renal vasoconstrictor antagonists. Saralasin, an antagonist of angiotensin II receptors, was first used in 1979 in an attempt to reverse renal vasoconstriction. Because this drug inhibited the homeostatic response to hypotension commonly observed in patients with cirrhosis, it led to worsening hypotension and deterioration in renal function. Poor results were also observed with phentolamine, an alpha-adrenergic antagonist, highlighting the importance of the sympathetic nervous system in maintaining renal haemodynamics in patients with HRS. Antagonists of endothelin A receptor. A recent case series by Soper et al. reported an improvement in GFR in patients with cirrhosis, ascites and HRS who received an endothelin A receptor antagonist. All patients showed a dosedependent response in the form of improved inulin and para-aminohippurate excretion, RPF and GFR without changes in systemic haemodynamics. These patients were not candidates for liver transplantation and subsequently died. More work is needed to explore this therapeutic approach as a possible bridge to transplantation for patients with HRS [57-59]. Systemic vasoconstrictors. These medications have shown the most promise for treatment of HRS in recent years. Hecker and Sherlock used norepinephrine in 1956 to treat patients with cirrhosis who had HRS, and they were the first to describe an improvement in arterial pressure and urine output. However, no improvement was observed in the biochemical parameters of renal function, and all patients subsequently died. Octapressin, a synthetic vasopressin analogue, was first used in 1970 to treat HRS type 1. RPF and the GFR improved in all patients, all of whom subsequently died of sepsis, gastrointestinal bleeding or liver failure.Because of these discouraging results, the use of alternative vasopressin analogues, particularly ornipressin, attracted attention. Two important studies by Lenz et al. [60, 61] demonstrated that short term use of ornipressin resulted in an improvement in circulatory function and a significant increase in RPF and the GFR. The combination of ornipressin and albumin was subsequently tried by Guevera in patients with HRS [49]. This idea was based on data suggesting that the combination of plasma volume expansion and vasoconstrictors normalised renal sodium and water handling in patients who had cirrhosis with ascites. In this study, 8 patients were originally treated for 15 days with ornipressin and albumin. Treatment had to be discontinued in 4 patients after fewer than 9 days because of complications of ornipressin use that included ischaemic colitis, tongue ischaemia and glossitis. Although a marked improvement in the serum creatinine was observed during treatment, renal function deteriorated on treatment withdrawal. In the remaining 4 patients the improvement in RPF and the GFR was significant and was associated with a lowering of serum creatinine levels. These patients subsequently died, but no recurrence of HRS was observed. Owing to the high incidence of severe adverse effects with ornipressin, the same investigators used another vasopressin analogue with fewer adverse effects, namely terlipressin. In this study, nine patients were treated with terlipressin + albumin for 5–15 days. This treatment was associated with a marked fall in serum creatinine levels and an improvement in mean arterial pressure. Reversal of HRS was noted in seven of the nine patients, and HRS did not recur when treatment was discontinued. No adverse ischemic effects were reported: according to this study, terlipressin with albumin is a safe and effective treatment for HRS [59-62]. Alpha adrenergic agonists.Angeli et al. showed that long-term administration of midodrine (an alpha-adrenergic agonist) and octreotide improved renal function in patients with HRS type 1 [65]. All patients also received albumin, and the results obtained with this approach were compared against those observed with dopamine at nonpressor doses. None of the patients treated with dopamine showed any improvement in renal function, but in all the patients treated with midodrine, octreotide and volume expansion renal function did improve. No adverse effects were reported in these patients. Gulberg et al. treated seven patients who had cirrhosis and HRS type 1 with a combination of ornipressin and dopamine for infusion periods as long as 27 days, but only three of the seven patients survived [62]. This treatment can be used as a bridge to liver transplantation [61, 65]. Aquaretic agents. K-Opioid antagonists inhibit antidiuretic hormone secretion by the neurohypophysis and induce water excretion.Administration of niravoline at doses ranging from 0.5 to 2 mg induced a strong aquaretic response and was well tolerated in 18 cirrhotic patients with preserved renal function, but no data are available on the use of niravoline in patients with HRS.

Surgical Manoeuvres
Transjugular intrahepatic portosystemic shunting. It is well documented that portal hypertension plays a central role in the development of refractory ascites and HRS. Earlier studies showed improved renal function after sideto- side portocaval shunting, but at the cost of a high surgical mortality in advanced cirrhosis. The transjugular intrahepatic portosystemic shunt (TIPS) was introduced as a less invasive method of reducing increased portal pressure. Guevarra et al. have investigated hepatic and renal haemodynamic changes after placement of TIPS in patients with HRS. One month after placement of TIPS a marked improvement in renal function was observed, as indicated by a significant reduction in serum creatinine and blood urea nitrogen and increased urine volume, RPF and GFR. These improvements were associated with a reduction in plasma rennin, aldosterone and norepinephrine activity. These changes were statistically significant, albeit less pronounced than observed in a similar group of patients receiving ornipressin and albumin infusions. Renal improvements were more pronounced at 30 days than at 7 days, possibly because of the deleterious effects of contrast media or the resolution of concomitant problems. After TIPS, GFR improved significantly but did not reach normal values, suggesting that TIPS does not correct all mechanisms contributing to HRS. Brensing et al. [65] found a sustained improvement of renal function after TIPS in 31 patients with type 1 or 2 HRS, allowing the discontinuation of haemodialysis in four of seven patients.After TIPS 3-, 6-, 12- and 18-month survival rates were 81%, 71%, 48% and 35%, respectively, in the total patient cohort, with survival in HRS type 1 patients being significantly worse than in the others. The use of TIPS to prolong survival until liver transplantation seems promising [65-68]. Other surgical shunts. Despite the theoretical benefit of improving portal hypertension and thus HRS by means of a portosystemic shunt, only a few scattered case reports have shown any benefit. Currently, particularly with the recent introduction of TIPS, portocaval shunts are not indicated in this setting. Renal replacement therapy. Many clinicians are reluctant to institute renal replacement therapy in advanced cirrhosis, because the outcome is poor unless liver transplantation is a realistic option. Intermittent haemodialysis can be a problem because patients with HRS are prone to develop circulatory and coagulation problems, and biocompatibility is also a problematic issue [69]. In an early study in the United Kingdom 100% mortality was observed in cirrhotic patients with HRS despite early institution of renal support [70]. However, modern renal replacement therapies such as continuous endogenous haemofiltration (CVVH) are certainly capable of prolonging life in patients with type 1 HRS who have not responded to medical therapies or TIPS. Because the underlying hepatic problem persists, the long-term prognosis is grim and treatment should be confined to patients who are candidates for liver transplantation or have a realistic chance of hepatic recovery. The molecular adsorbent recirculating system (MARS) is a modified dialysis method that uses albumin-containing dialysate in a closed-loop secondary circuit for adsorptive removal of albumin-bound toxins. In a randomised study, short-term survival of eight HRS patients treated with MARS was superior to that of five other HRS patients treated with CVVH [71]. In contrast to previous reports on haemodialysis, treatment was well tolerated. Unfortunately, the study was terminated after enrolment of only 13 patients, which makes evaluation of any influence on mortality difficult.Moreover, the control group seems to have received a smaller dialytic dose: creatinine levels were decreased in the MARS group only.Nonetheless, the favourable effects of this system deserve evaluation in a prospective study of adequate power. Liver transplantation. Liver transplantation is the ideal treatment for HRS, but is completely dependent on the availability of Donors. Patients with HRS have a higher risk of postoperative morbidity, early mortality and longer hospitalisation than other transplant recipients.Gonwa et al. [72] reported that at least one third of such patients require haemodialysis postoperatively, with a smaller proportion (5%) requiring long-term dialysis. Because renal dysfunction is common in the first few days after transplantation, avoidance nephrotoxic immunosuppressants is generally recommended until renal function is recovered. However, the GFR gradually improves to an average of 40–50 ml/min by the 6th postoperative week. The systemic and neurohumoral abnormalities associated with HRS also resolve in the 1st postoperative month. Long-term survival rates are excellent, with the survival rate at 3 years approaching approximately 60%. This is only slightly lower than the 70–80% survival rate of transplant recipients without HRS and is markedly better than the survival rate of patients with HRS who do not receive transplants, which is virtually nil at 3 years [73, 74].

Thursday, August 14, 2008



He who works with his hands is a laborer.
He who works with his head and his hands is a craftsman.
St. Francis of Assisi

The care of critically ill patients requires one or more pipelines to the vascular system, for both monitoring and interventions. This chapter presents some guidelines for the insertion of vascular catheters, including a brief description of the common percutaneous access routes (1-3). The emphasis here is on the craft of establishing vascular access. The labor of vascular cannulation is a skill learned at the bedside.



Handwashing is mandatory (and often overlooked) before the insertion of vascular devices. Scrubbing with antimicrobial cleansing solutions does not reduce the incidence of catheter-related sepsis (4), so a simple soap-and-water scrub is sufficient.


In 1985, the Centers for Disease Control introduced a strategy for blood and body fluid precautions known as universal precautions (5). This strategy is based on the assumption that all patients are potential sources of human immunodeficiency virus (HIV) and other blood-borne pathogens (e.g., hepatitis viruses) until proven otherwise. The following recommendations apply to the insertion of vascular catheters.

Use protective gloves for all vascular cannulations.

Use sterile gloves for all cannulations except those involving the introduction of a short catheter into a peripheral vein.

Caps, gowns, masks, and protective eyewear are not necessary unless splashes of blood are anticipated (e.g., in a trauma victim). These measures do not reduce the incidence of catheter-related sepsis (6).

Avoid needlestick injuries. Do not recap needles or manually remove needles from syringes. Place all sharp instruments in puncture-resistant containers immediately after use.

If a needlestick injury is sustained during the procedure, follow the recommendations in Table 4.1.

Needlestick injuries are reported in up to 80% of medical students and interns (9). Therefore, in patients who are known risks for transmitting HIV or viral hepatitis, vascular cannulation should be performed only by an experienced senior-level resident or fellow.


The increased use of rubber gloves (made of latex or vinyl) as protection against HIV infections has resulted in an increased recognition of allergic reactions to latex (10). These reactions can be manifest as a contact dermatitis (urticarial lesions of the hands and face), or as a conjunctivitis, rhinitis, or asthma. The latter three manifestations are reactions to airborne latex particles, and they do not require direct physical contact with gloves. They often appear when the affected individual enters an area where latex gloves are being used. Therefore, a latex allergy should be suspected in any ICU team member who develops atopic symptoms when in the ICU. When this occurs, a switch to vinyl gloves will eliminate the problem. Latex allergy can be manifest as anaphylaxis (10), so the transition to vinyl gloves for suspected latex allergy should not be delayed.


Agents that reduce skin microflora are called antiseptics, whereas agents that reduce the microflora on inanimate objects are called disinfectants. Common antiseptic agents are listed in Table 4.2 (11,12). The most widely used antiseptic agents are alcohol and iodine, both of which have a broad spectrum of antimicrobial activity. Alcohol (commonly used as a 70% solution) may not work well on dirty skin (that is, it does not have a detergent action), so it is often used in combination with another antiseptic agent. The most popular antiseptic solution currently in use is a povidone-iodine preparation (e.g., Betadine), also known as an iodophor, a water-soluble complex of iodine and a carrier molecule. The iodine is released slowly from the carrier molecule, and this reduces the irritating effects of iodine on the skin. This preparation should be left in contact with the skin for at least 2 minutes to allow sufficient time for iodine to be released from the carrier molecule.


Shaving is not recommended for hair removal because it abrades the skin and promotes bacterial colonization. If hair removal is necessary, the hair can be clipped or a depilatory can be applied (6).


Vascular cannulation can be performed by advancing the catheter over a needle or guidewire that is in contact with the lumen of a blood vessel.


A catheter-over-needle device is shown in Figure 4.1. The catheter fits snugly over the insertion needle, and has a tapered end to minimize damage to the catheter tip and soft tissues during insertion. This device can be held like a pencil (i.e., between the thumb and forefinger) as it is inserted through the skin and directed to the target vessel. When the tip of the needle enters the lumen of a patent blood vessel, blood moves up the needle by capillary action and enters the flashback chamber. When this occurs, the catheter is threaded over the needle and into the lumen of the blood vessel as the needle is withdrawn.

The advantage of a catheter-over-needle device is the ability to cannulate vessels in a simple one-step procedure. The disadvantage is the tendency for the catheter tip to become frayed as it passes through the skin and soft tissues, and to subsequently damage the vascular endothelium and promote phlebitis and thrombosis. To minimize this risk, the catheter-over-needle device is usually reserved for cannulation of superficial vessels.


Guidewire-assisted vascular cannulation was introduced in the early 1950s and is often called the Seldinger technique, after its inventor. This technique is illustrated in Figure 4.2. A small-bore needle (usually 20 gauge) is used to probe for the target vessel. When the tip of the needle enters the vessel, a thin wire with a flexible tip (called a J-tip because of its shape) is passed through the needle and into the vessel lumen. The needle is then removed, leaving the wire in place to serve as a guide for cannulation of the vessel. When cannulating deep vessels, a rigid dilator catheter is first threaded over the guidewire and removed; this creates a tract that facilitates insertion of the vascular catheter.

The guidewire technique has the presumed advantage of minimizing damage to soft tissues and blood vessels by using a small-bore probe needle. However, the use of a rigid dilator catheter (as explained above) seems to nullify this advantage. Nevertheless, the guidewire technique is currently the preferred method for central venous and arterial cannulation (1,2).


Vascular catheters are composed of plastic polymers impregnated with barium or tungsten salts to enhance radiopacity. Catheters intended for short-term cannulation (days) are usually made of polyurethane; catheters used for long-term venous access (weeks to months) are composed of a more flexible and less thrombogenic derivative of silicone. The silicone catheters (e.g., Hickman and Broviac catheters) are too flexible for routine percutaneous insertion, and are not appropriate for use in the ICU.


Some vascular catheters are impregnated or coated with heparin to reduce thrombogenicity. However, this measure has not proven effective in reducing the incidence of catheter-associated thrombosis (13). Because heparin-coated catheters can be a source of heparin-induced thrombocytopenia (see Chapter 45), catheters used in the ICU should not be impregnated or coated with heparin.


The size of vascular catheters is commonly expressed in terms of the outside diameter, and the units of measurement are shown in Table 4.3. The French size is a metric derivative equivalent to the outside diameter in millimeters multiplied by 3; that is, French size = outside diameter (in mm) ´ 3. The gauge system was developed for wires and needles, and has been adopted for catheters. There is no simple mathematical relationship between gauge size and other units of measurement, and a table of reference values such as Table 4.3 is needed to make the appropriate conversions (14). Gauge sizes usually range from 14 (largest diameter) to 27 (smallest diameter).

As described in Chapter 1, the steady or laminar flow through a rigid tube is influenced most by the radius of the tube (see the Hagen-Poisseuille equation in the section on peripheral blood flow in Chapter 1). The influence of tube diameter on flow rate is demonstrated in Table 4.3 for gravity flow of one unit of packed red blood cells diluted with 250 mL normal saline flowing through catheters of equal length (15). Note that a little more than a doubling of tube diameter (from 0.7 mm to 1.65 mm) is associated with almost a quadrupling of flow rate (from 24.7 to 96.3 mL/minute). Thus, catheter size (diameter) is an important consideration if rapid flow rates are desired.


Multilumen catheters were introduced for clinical use in the early 1980s, and are now used routinely for central venous cannulation. The design of a triple-lumen catheter is shown in Figure 4.3. These catheters have an outside diameter of 2.3 mm (6.9 French) and may have three channels of equal diameter (usually 18 gauge) or may have one larger channel (16 gauge) and two smaller channels of equal diameter (18 gauge). The distal opening of each channel is separated from the other by at least 1 cm to help minimize mixing of infusate solutions.

Multilumen catheters have proven to be valuable aids because they minimize the number of venipunctures needed for monitoring and infusion therapy, yet they do not increase the risk of thrombosis or infection when compared with single-lumen catheters (13).


Another valuable addition to the family of vascular catheters is the introducer catheter, shown in Figure 4.3. These large-bore catheters (8 to 9 French) can be used as conduits for insertion and removal of smaller vascular catheters (including multilumen catheters and pulmonary artery catheters) through a single venipuncture. The side-arm infusion port on the catheter provides an additional infusion line, and allows a continuous flush to prevent thrombus formation around smaller catheters that sit in the lumen of the introducer catheter. This side-arm infusion port also allows the introducer catheter to be used as a stand-alone infusion device (a rubber membrane on the hub of the catheter provides an effective seal when fluids are infusing through the side-arm port of the catheter). The large diameter of introducer catheters makes them particularly valuable as infusion devices when rapid infusion rates are necessary (e.g., in the resuscitation of massive hemorrhage).


The following is a brief description of common vascular access routes in the arm (antecubital veins and radial artery), the thoracic inlet (subclavian and jugular veins), and the groin (femoral artery and vein).


The veins in the antecubital fossa provide rapid and safe vascular access for acute resuscitative therapy. Although long catheters can be inserted into the antecubital veins and advanced into the superior vena cava, such peripherally inserted central venous catheters (PICC devices) are more appropriate for home infusion therapy than for treating critically ill patients (16). Short catheters (5 to 7 cm) are preferred for acute resuscitation via the antecubital veins because they are more easily inserted and allow more rapid infusion rates than the longer PICC catheters.


The surface anatomy of the antecubital veins is shown in Figure 4.4. The basilic vein runs along the medial aspect of the antecubital fossa, and the cephalic vein is situated on the opposite side. The basilic vein is preferred for cannulation because it runs a straighter and less variable course up the arm than the cephalic vein.

Insertion Technique

The patient need not be supine, but the arm should be straight and abducted. The antecubital veins can be distended by tourniquet or by inflating a blood pressure cuff to just above the diastolic pressure (this allows arterial inflow while impeding venous outflow). Once the veins are visible or palpable, a catheter-over-needle device is used to insert a short 16- or 18-gauge catheter into the basilic or cephalic vein.

Blind Insertion

If the antecubital veins are neither visible nor palpable, palpate the brachial artery pulse at a point 1 inch above the antecubital crease. The basilic vein (or brachial vein) should lie just medial to the palpated pulse at this point, and can be entered by inserting the catheter-over-needle device through the skin at a 35° to 45° angle and advancing the needle until blood return is noted. This approach has a reported success rate of 80% (17). Injury to the median nerve (which is also medial to the artery, but deep to the veins) can occur with excessive movement of the probe needle.


Cannulation of the antecubital veins is recommended (18,19).

For rapid venous access (e.g., cardiopulmonary resuscitation)

For thrombolytic therapy in acute myocardial infarction

For trauma victims who require thoracotomy

Remember that the shorter the catheter, the more rapid the flow rate through the catheter (see Chapter 1). Thus, insertion of short catheters into the antecubital veins permits more rapid volume resuscitation than insertion of the longer central venous catheters.


The radial artery is a favored site for arterial cannulation because the vessel is superficial and accessible and the insertion site is easy to keep clean. The major disadvantage of the radial artery is its small size, which limits the success rate of cannulation and promotes vascular occlusion.


The surface anatomy of the radial and ulnar arteries is shown in Figure 4.4. The radial artery is usually palpable at a point just medial to the styloid process of the radius. The ulnar artery is on the opposite (medial) side of the wrist, just lateral to the pisiform bone. Although the radial artery is preferred, the ulnar artery is the larger of the two arteries and should be easier to cannulate (20).

The Allen Test

The Allen test evaluates the capacity of the ulnar artery to supply blood to the digits when the radial artery is occluded. The test is performed by first occluding the radial and ulnar arteries with the thumb and index finger. The patient is then instructed to raise the wrist above the head and to make a fist repeatedly until the fingers turn white. The ulnar artery is then released, and the time required for return of the normal color to the fingers is recorded. A normal response time is 7 seconds or less, and a delay of 14 seconds or greater is evidence of insufficient flow in the ulnar artery.

Although a positive Allen test (i.e., 14 seconds or longer for return of color to the digits) is often stated as a contraindication to radial artery cannulation, in numerous instances the Allen test has indicated inadequate flow in the ulnar artery, yet subsequent radial artery cannulation has been uneventful (2,21). Thus, a positive Allen test is not a contraindication to radial artery cannulation. Another limitation is the need for patient cooperation to perform the test. Therefore, this test is not worth the time it takes.

Insertion Technique

The wrist should be hyperextended to bring the artery closer to the surface. A short 20-gauge catheter is appropriate, and can be inserted by a catheter-over-needle device or by the guidewire technique. When using a catheter-over-needle device, the following through-and-through technique is recommended: When the needle tip first punctures the artery (and blood appears in the flashback chamber), the tip of the catheter is just outside the vessel. To position the catheter tip in the lumen of the vessel, the needle is passed completely through the artery and then withdrawn until blood returns again through the needle. At this point, the catheter tip should be in the lumen of the artery, and the catheter can be advanced while the needle is retracted. If two attempts at cannulation are unsuccessful, switch to an alternative site (to reduce trauma to the vessel).


Arterial occlusion occurs in as many as 25% of radial artery cannulations, but digital ischemia is rare (2,22). Despite being well tolerated in most patients, cannulation of the radial artery (or any artery) should be reserved for monitoring blood pressure, and is not to be used as a convenience measure for monitoring blood gases or other blood components (23).


More than 3 million central venous cannulations are performed yearly in the United States (24), and a majority of these procedures are performed via the subclavian vein (25). The subclavian vein is well suited for cannulation because it is a large vessel (about 20 mm in diameter) and is prevented from collapsing by its surrounding structures. The immediate risks of subclavian vein cannulation include pneumothorax (1% to 2%) and hemothorax ( less than 1%) (25). The incidence of bleeding is no different in the presence or absence of a coagulopathy (26); that is, the presence of a coagulation disorder is not a contraindication to subclavian vein cannulation.


The subclavian vein is a continuation of the axillary vein as it passes over the first rib, and the apical pleura lies about 5 mm deep to the vein at its point of origin. As shown in Figure 4.5, the subclavian vein runs most of its course along the underside of the clavicle. The vein runs along the outer surface of the anterior scalene muscle, which separates the vein from its companion artery on the underbelly of the muscle. At the thoracic inlet, the subclavian vein meets the internal jugular vein to form the brachiocephalic vein. The convergence of the right and left brachiocephalic veins forms the superior vena cava.

Anatomic Distances

The lengths of the vascular segments involved in subclavian (and internal jugular) vein cannulation are shown in Table 4.4. The average distance from venipuncture site to the right atrium is 14.5 cm and 18.5 cm for right-sided and left-sided cannulations, respectively. These distances are far shorter than catheter lengths recommended for right-sided (20 cm) and left-sided (30 cm) central venous cannulations, and are more consistent with a recent report showing that the average distance to the right atrium is 16.5 cm in central venous cannulation from either side in adults (27). Therefore, to avoid placing catheter tips in the right atrium (which can lead to cardiac perforation and fatal cardiac tamponade), all central venous catheters should be no longer than 15 or 16 cm in length (27).


The patient is placed supine, with arms at the sides and head faced away from the insertion site. A towel roll can be placed between the shoulder blades, but this is uncomfortable and unnecessary. Identify the clavicular insertion of the sternocleidomastoid muscle. The subclavian vein lies just underneath the clavicle where the muscle inserts onto the clavicle. The vein can be entered from either side of the clavicle.

Infraclavicular Approach (Insertion Site 1 in Figure 4.5). Identify the lateral margin of the sternocleidomastoid muscle as it inserts on the clavicle. The catheter is inserted in line with this margin, but below the clavicle. Insert the probing needle (18 or 20 gauge) with the bevel pointing upward (toward the ceiling) and advance the needle along the underside of the clavicle and toward the suprasternal notch. The path of the needle should be parallel to the patient's back. When the vein is entered, turn the bevel of the needle to 3 o'clock so the guidewire threads in the direction of the superior vena cava.

Supraclavicular Approach (Insertion Site 2 in Figure 4.5). Identify the angle formed by the lateral margin of the sternocleidomastoid muscle and the clavicle. The probe needle is inserted so that it bisects this angle. Keep the bevel of the needle facing upward and direct the needle under the clavicle in the direction of the opposite nipple. The vein should be entered at a distance of 1 to 2 cm from the skin surface (the subclavian vein is more superficial in the supraclavicular approach). When the vein is entered, turn the bevel of the needle to 9 o'clock so the guidewire threads in the direction of the superior vena cava.


Patient comfort and ease of insertion are the most compelling reasons to select the subclavian vein for central venous access. Selection of the infraclavicular versus supraclavicular approach is largely a matter of personal preference. Some recommend avoiding the subclavian vein in ventilator-dependent patients because of the risk of pneumothorax. However, the risk of pneumothorax is too small to justify abandoning the subclavian vein in patients with respiratory failure.


Cannulation of the internal jugular vein reduces (but does not eliminate) the risk of pneumothorax, but introduces new risks (e.g., carotid artery puncture and thoracic duct injury).


The internal jugular vein is located under the sternocleidomastoid muscle in the neck and, as shown in Figure 4.5, the vein follows an oblique course as it runs down the neck. When the head is turned to the opposite side, the vein forms a straight line from the pinna of the ear to the sternoclavicular joint. Near the base of the neck, the internal jugular vein becomes the most lateral structure in the carotid sheath (which contains the carotid artery sandwiched between the vein laterally and the vagus nerve medially).

Insertion Technique

The right side is preferred because the vessels run a straighter course to the right atrium. The patient is placed in a supine or Trendelenburg position, with the head turned to the opposite side. The internal jugular vein can be entered from an anterior or posterior approach.

The Anterior Approach (Insertion Site 4 in Figure 4.5). The anterior approach is through a triangular region created by two heads of the sternocleidomastoid muscle. The carotid artery is palpated in the triangle and retracted medially. The probe needle is inserted at the apex of the triangle with the bevel facing up, and the needle is advanced toward the ipsilateral nipple, at a 45° angle with the skin surface. If the vein is not encountered by a depth of 5 cm, the needle is withdrawn 4 cm and advanced again in a more lateral direction. When a vessel is entered, look for pulsations. If the blood is red and pulsating, you have entered the carotid artery. In this situation, remove the needle and tamponade the area for 5 to 10 minutes. When the carotid artery has been punctured, no further attempts should be made on either side because puncture of both arteries can have serious consequences.

The Posterior Approach (Insertion Site 3 in Figure 4.5). The insertion site for this approach is 1 centimeter superior to the point where the external jugular vein crosses over the lateral edge of the sternocleidomastoid muscle. The probe needle is inserted with the bevel positioned at 3 o'clock. The needle is advanced along the underbelly of the muscle in a direction pointing to the suprasternal notch. The internal jugular vein should be encountered 5 to 6 cm from the skin surface with this approach (28).

Carotid Artery Puncture. If the carotid artery has been punctured with a probing needle, the needle should be removed and pressure should be applied to the site for at least 5 minutes (10 minutes is recommended for patients with a coagulopathy). No further attempts should be made to cannulate the internal jugular vein on either side, to avoid puncture of both carotid arteries. If the carotid artery has been inadvertently cannulated, the catheter should not be removed, as this could provoke serious hemorrhage. In this situation, a vascular surgeon should be consulted immediately.


As with the subclavian vein, cannulation of the internal jugular vein is safe and effective when performed by skilled operators. However, several disadvantages of internal jugular cannulation deserve mention. (a) Accidental puncture of the carotid artery is reported in 2 to 10% of attempted cannulations (28). (b) Awake patients often complain of the limited neck mobility when the internal jugular vein is cannulated. (c) In agitated patients, inappropriate neck flexion can result in thrombotic occlusion of the catheter and vein. (d) In patients with tracheostomies, the insertion site can be exposed to infected secretions that drain from the tracheal stoma.


Cannulation of the external jugular vein has two advantages: (a) There is no risk of pneumothorax, and (b) hemorrhage is easily controlled. The major drawback is difficulty advancing the catheter.


The external jugular vein runs along a line extending from the angle of the jaw to a point midway along the clavicle. The vein runs obliquely across the surface of the sternocleidomastoid muscle and joins the subclavian vein at an acute angle. This acute angle is the major impediment to advancing catheters that have been inserted into the external jugular vein.

Insertion Technique

The patient is placed in the supine or Trendelenburg position, with the head turned away from the insertion site. If necessary, the vein can be occluded just above the clavicle (with the forefinger of the nondominant hand) to engorge the entry site. As many as 15% of patients so not have an identifiable external jugular vein, even under optimal conditions of vein engorgement (28).

The external jugular vein has little support from surrounding structures, so the vein should be anchored between the thumb and forefinger when the needle is inserted. The bevel of the needle should be pointing upward when it enters the vein. The recommended insertion point is midway between the angle of the jaw and the clavicle (see Fig. 4.5). Use a 16-gauge single-lumen catheter that is 10 to 15 cm in length. If the catheter does not advance easily, do not force it, as this may result in vascular perforation at the junction between the external jugular and subclavian veins.


This approach is best reserved for temporary access in patients with a severe coagulopathy, particularly when the operator is inexperienced and does not feel comfortable cannulating the subclavian or internal jugular veins. Contrary to popular belief, cannulation of the external jugular is not always easier to accomplish than central venous cannulation because of the difficulty in advancing catheters past the acute angle at the junction of the subclavian vein.


The femoral vein is the easiest of the large veins to cannulate and also does not carry a risk of pneumothorax. The disadvantages associated with this route are venous thrombosis (10%), femoral artery puncture (5%), and limited ability to flex the hip (which can be bothersome for awake patients). Contrary to popular belief, the infection rate with femoral vein catheters is no different from that of subclavian or internal jugular vein catheters (28).


The anatomy of the femoral sheath is shown in Figure 4.6. The femoral vein is the most medial structure in the femoral sheath and is situated just medial to the femoral artery. At the inguinal ligament, the femoral vessels are just a few centimeters below the skin surface.

Insertion Technique

Palpate the femoral artery just below the inguinal crease and insert the needle (bevel up) 1 to 2 cm medial to the palpated pulse. Advance the needle at a 45° angle to the skin surface, entering the vein at a depth of 2 to 4 cm. Once in the vessel, if the catheter or guidewire will not pass beyond the tip of the needle, tilt the needle so that it is more parallel to the skin surface (this may move the needle tip away from the far side of the vessel wall and into more direct contact with the lumen of the vessel). Femoral vein catheters should be at least 15 cm long.

Blind Insertion

If the femoral artery pulse is not palpable, draw an imaginary line from the anterior superior iliac crest to the pubic tubercle, and divide the line into three equal segments. The femoral artery lies at the junction between the middle and most medial segment, and the femoral vein is 1 to 2 cm medial to this point. This method of locating the femoral vein has a reported success rate of over 90% (29).


Femoral vein cannulation is usually reserved for patients who are paralyzed or comatose and immobile. This approach is not recommended for cardiopulmonary resuscitation (because of the delayed transit times for bolus drug injections) (18) or in patients with bleeding disorders (28).


Cannulation of the femoral artery is usually reserved for situations where radial artery cannulation is unsuccessful or contraindicated. Despite its reserve status, the femoral artery is larger than the radial artery, and is easier to cannulate. The complications of femoral artery cannulation are the same as for radial artery cannulation (thrombosis, bleeding, and infection). The incidence of infection is the same with radial and femoral artery catheters, and the incidence of thrombosis is lower with femoral artery cannulation (2). Thrombosis of the femoral artery, like that in the radial artery, only rarely results in troublesome ischemia in the distal extremity (2).

Localization and cannulation of the femoral artery proceeds as described in the section on femoral vein cannulation. The Seldinger technique is preferred for catheter insertion, and catheters should be 18 gauge in diameter and 15 to 20 cm long.


Femoral artery cannulation is a viable alternative and may be preferable to radial artery cannulation in patients who are paralyzed or otherwise immobile, unless they have a significant coagulopathy (in which case the radial artery is preferred). The incidence of thrombotic complications is lower in femoral artery cannulations, and the pressure in the femoral artery more closely approximates the pressure in the aorta than does the pressure in the radial artery (see Chapter 8).



Inadvertent air entry is one of the most feared complications of central venous cannulation. The importance of maintaining a closed system during insertion is highlighted by the following statement:

A pressure gradient of 4 mm Hg along a 14-gauge catheter can entrain air at a rate of 90 mL/second and can produce a fatal air embolus in 1 second (30).

Preventive Measures

Prevention is the hallmark of reducing the morbidity and mortality of venous air embolism. The most effective method of preventing air entry is to keep the venous pressure more positive than atmospheric pressure. This is facilitated by placing the patient in the Trendelenburg position with the head 15° below the horizontal plane. Remember that the Trendelenburg position does not prevent venous air entry because patients still generate negative intrathoracic pressures while in the Trendelenburg position. When changing connections in a central venous line, a temporary positive pressure can be created by having the patient hum audibly. This not only produces a positive intrathoracic pressure, but allows clinicians to hear when the intrathoracic pressure is positive. In ventilator-dependent patients, the nurse or respiratory therapist should initiate a mechanical lung inflation when changing connections.

Clinical Presentation

The usual presentation is acute onset of dyspnea that occurs during the procedure. Hypotension and cardiac arrest can develop rapidly. Air can pass across a patent foramen ovale and obstruct the cerebral circulation, producing an acute ischemic stroke. A characteristic "mill wheel" murmur can be heard over the right heart, but this murmur may be fleeting.

Therapeutic Maneuvers

If a venous air embolism is suspected, immediately place the patient with the left side down, and attempt to aspirate air directly from the venous line. In dire circumstances, a needle should be inserted through the chest wall and into the right ventricle to aspirate the air. Unfortunately, the mortality in severe cases of venous air embolism remains high despite these maneuvers.


Pneumothorax is a concern primarily with subclavian vein cannulation but can also complicate jugular vein cannulation (2,30). This is one reason that postinsertion chest films are recommended after all central venous cannulations (or attempts). If possible, postinsertion films should be obtained in the upright position and during expiration. Expiratory films facilitate the detection of small pneumothoraxes because expiration decreases the volume of air in the lungs, but not the volume of air in the pleural space. Thus, during expiration, the volume of air in the pleural space is a larger fraction of the total volume of the hemithorax, thereby magnifying the radiographic appearance of the pneumothorax (31).

Upright films are not always possible in ICU patients. When supine films are necessary, remember that pleural air does not often collect at the apex of the lung when the patient is in the supine position (32,33). In this situation, pleural air tends to collect in the subpulmonic recess and along the anteromedial border of the mediastinum (see Chapter 28).

Delayed Pneumothorax

Pneumothoraxes may not be radiographically evident until 24 to 48 hours after central venous cannulation (31,33). Therefore, the absence of a pneumothorax on an immediate postinsertion chest film does not absolutely exclude the possibility of a catheter-induced pneumothorax. This is an important consideration in patients who develop dyspnea or other signs of pneumothorax in the first few days after central venous cannulation. In the absence of signs and symptoms, there is little justification for serial chest films following central venous catheter placement.


The properly placed central venous catheter should run parallel to the superior vena cava, and the tip of the catheter should be positioned above the junction of the superior vena cava and right atrium. The following conditions warrant corrective measures.

Tip Against the Wall of the Vena Cava

Catheters inserted from the left side must make an acute turn downward when they reach the superior vena cava. If they fail to make this turn, the catheters can end up in a position like the one shown in Figure 4.7. The tip of the catheter is up against the lateral wall of the vena cava, and in this position, the catheter tip can stab the vessel wall and perforate the vena cava. Therefore, catheters that abut the wall of the vena cava should be repositioned as soon as possible. (The problem of vascular perforation is discussed in more detail in Chapter 5.)

Tip in the Right Atrium

The Food and Drug Administration has issued a strong warning about the risk of cardiac perforation from catheter tips that are advanced into the heart (24). However, cardiac perforation is a rare complication of central venous cannulation (27), even though over half of central venous catheters may be misplaced in the right atrium (27). Nevertheless, tamponade is often fatal, so cardiac placement of catheters should be avoided. A few measures help to minimize the risk of cardiac perforation. The most effective measure is to use shorter catheters, as recommended earlier. The tip of indwelling catheters should be above the third right costal cartilage (this is the level where the vena cava meets the right atrium). If the anterior portion of the third rib cannot be visualized, keep the catheter tip at or above the tracheal carina.



All human things are subject to decay.
John Dryden

The treatment of critically ill patients is dominated by the notion that promoting the supply of oxygen to the vital organs is a necessary and life-sustaining measure. Oxygen is provided in a liberal and unregulated fashion, while the tendency for oxygen to degrade and decompose organic (carbon-based) matter is either overlooked or underestimated. In contrast to the notion that oxygen protects cells from injury in the critically ill patient, the accumulated evidence over the past 15 years suggests that oxygen is responsible for the cell injury that accompanies critical illness. Oxygen's ability to act as a lethal toxin has monumental implications for the way we treat critically ill patients.


An oxidation reaction is a chemical reaction between oxygen and another chemical species. Because oxygen removes electrons from other atoms and molecules, oxidation is also described as the loss of electrons by an atom or molecule. The chemical species that removes the electrons is called an oxidizing agent or oxidant. The companion process (i.e., the gain of electrons by an atom or molecule) is called a reduction reaction, and the chemical species that donates the electrons is called a reducing agent. Because oxidation of one atom or molecule must be accompanied by reduction of another atom or molecule, the overall reaction is often called a redox reaction.

When an organic molecule (a molecule with a carbon skeleton) reacts with oxygen, electrons are removed from carbon atoms in the molecule. This disrupts one or more covalent bonds, and as each bond ruptures, energy is released in the form of heat and light (and sometimes sound). The organic molecule then breaks into smaller fragments. When oxidation is complete, the parent molecule is broken down into the smallest molecules capable of independent existence. Because organic matter is composed mainly of carbon and hydrogen, the end-products of oxidation are simple combinations of oxygen with carbon and hydrogen: carbon dioxide and water.


Oxygen is a weak oxidizing agent, but some of its metabolites are potent oxidants capable of producing widespread and lethal cell injury (1). The mechanism whereby oxygen metabolism can produce more powerful oxidants than the parent molecule is related to the atomic structure of the oxygen molecule, which is described below.


Oxygen in its natural state is a diatomic molecule, as shown by the familiar O2 symbol at the top of Figure 3.1. The orbital diagram to the right of the O2 symbol shows how the outer electrons of the oxygen molecule are arranged. The circles in the diagram represent orbitals. (An orbital is an energy field that can be occupied by electrons. It is distinct from an orbit, which is a path that represents a specific point in space and time.) The arrows in the diagram represent electrons that are spinning in the same or opposite directions (indicated by the direction of the arrows). Note that one of oxygen's orbitals contains two electrons with opposing spins, and the other two orbitals each contain a single electron spinning in the same direction. The orbital with the paired electrons is obeying one of the basic rules of the quantum atom: An electron orbital can be occupied by two electrons if they have opposing spins. Thus, the two outermost orbitals that contain single electrons are only half full, and their electrons are unpaired. An atom or molecule that has one or more unpaired electrons in its outer orbitals is called a free radical (2). (The term free indicates that the atom or molecule is capable of independent existence—it is free-living.)

Free radicals tend to be highly reactive chemical species by virtue of their unpaired electrons. However, not all free radicals are highly reactive. This is the case with oxygen, which is not a highly reactive molecule despite having two unpaired electrons.

The reason for oxygen's sluggish reactivity is the directional spin of its two unpaired electrons. No two electrons can occupy the same orbital if they have the same directional spin. Thus, an electron pair cannot be added to oxygen because one orbital would have two electrons with the same directional spin, which is a quantum impossibility. This spin restriction limits oxygen to single electron additions, and this not only increases the number of reactions needed to reduce molecular oxygen to water, but it also produces more highly reactive intermediates.


Oxygen is metabolized at the very end of the electron transport chain, where the electrons and protons that have completed the transport process are left to accumulate. The complete reduction of molecular oxygen to water requires the addition of four electrons and four protons, as shown in the reaction sequence in Figure 3.1. Each metabolite in this sequence is accompanied by an orbital diagram to demonstrate the changes occurring at each point in the pathway.

Superoxide Radical

The first reaction adds one electron to oxygen, and produces the superoxide radical.

O2 = e- --» O2.

Note the superscript dot on the superoxide symbol. This signifies an unpaired electron, and is the conventional symbol for a free radical. The superoxide radical has one unpaired electron, and thus is less of a free radical than oxygen. Superoxide is neither a highly reactive radical nor a potent oxidant (3). Nevertheless, it has been implicated in conditions associated with widespread tissue damage, such as the reperfusion injury that follows a period of ischemia (2). The toxicity of superoxide may be caused by the large daily production, which is estimated at 1 billion molecules per cell, or 1.75 kg (4 lb) for a 70-kg adult (4).

Hydrogen Peroxide

The addition of one electron to superoxide creates hydrogen peroxide, a strong oxidizing agent (and the source of acid rain in the atmosphere) (5).

O2. + e- + 2H+ -- H2O2

Hydrogen peroxide is very mobile, and crosses cell membranes easily. It is a powerful cytotoxin and is well known for its ability to damage endothelial cells. It is not a free radical, but it may have to generate a free radical (a hydroxyl radical) to express its toxicity.

Hydrogen peroxide is loosely held together by a weak oxygen-oxygen bond (this bond is represented by the lower orbital in the orbital diagram for hydrogen peroxide). This bond ruptures easily, producing two hydroxyl radicals, each with one unpaired electron. An electron is donated to one of the hydroxyl radicals, creating one hydroxyl ion (OH-) and one hydroxyl radical (•OH). The electron is donated by iron in its reduced form, Fe(II), which serves as a catalyst for the reaction. Iron is involved in many free radical reactions, and is considered a powerful pro-oxidant. The role of transition metals in free radical reactions is discussed again later in the chapter.

Hydroxyl Radical

The iron-catalyzed dissociation of hydrogen peroxide proceeds as follows:

H2O2 + FE(II) --» OH- + .OH + FE(III)

(Note that Roman numerals are used instead of plus signs to designate the oxidation state of iron, as recommended by the International Union of Chemistry.) The hydroxyl radical is the ace of free radicals. It is one of the most reactive molecules in biochemistry and often reacts with another chemical species within five molecular diameters from its point of origin (2). This high degree of reactivity limits the mobility of the hydroxyl radical, and this may serve as a protective device to limit the toxicity of the hydroxyl radical. However, the hydroxyl radical is always dangerous because it can oxidize any molecule in the human body.

Hypochlorous Acid

The metabolism of oxygen in neutrophils has an additional pathway (not shown in Figure 3.1) that uses a myeloperoxidase enzyme to chlorinate hydrogen peroxide, creating hypochlorous acid (hypochlorite).

H2O2 + 2Cl- --» 2HOCL

When neutrophils are activated, the conversion of oxygen to superoxide increases twentyfold. This is called the respiratory burst, which is an unfortunate term because the increased O2 consumption has nothing to do with energy metabolism. When the increased metabolic traffic reaches hydrogen peroxide, about 40% is diverted to hypochlorite production and the remainder forms hydroxyl radicals (6). Hypochlorite is the active ingredient in household bleach. It is a powerful germicidal agent and requires only milliseconds to produce lethal damage in bacteria (7).


The final reaction in oxygen metabolism adds an electron to the hydroxyl radical and produces two molecules of water.

.OH + OH- + e- + 2H+ -- 2H2O

Therefore, the metabolism of one molecule of oxygen requires four chemical reactions, each involving the addition of a single electron. This process, then, requires four reducing equivalents (electrons and protons).

Under normal conditions, about 98% of the oxygen metabolism is completed, and less than 2% of the metabolites escape into the cytoplasm (3). This is a tribute to cytochrome oxidase, which carries on the reactions in a deep recess that effectively blocks any radical escape. This degree of suppression is necessary because of the ability of free radicals to start chain reactions (see next section).

Proposed Scheme

The superoxide radical is mobile but not toxic, whereas the hydroxyl radical is toxic but not mobile. Combining the advantages of each oxidant yields a scheme that has the mobile oxidant serving as a transport vehicle that can reach distant places. Once at the desired location, this metabolite could then generate hydroxyl radicals to produce local damage (3). This scheme is intuitively satisfying, regardless of its validity.


The damaging effects of oxidation are largely the result of free radical reactions. This section describes the two basic types of free radical reactions: those involving free radicals and nonradicals and those involving two free radicals.


When a free radical reacts with a nonradical, the nonradical loses an electron and is transformed into a free radical. Therefore, the union of a radical and a nonradical begets another radical (thus illustrating the survival value of the free radical). Because free radicals are often highly reactive in nature. This type of radical-regenerating reaction tends to become repetitive, creating a series of self-sustaining reactions known collectively as a chain reaction (3). The tendency to produce chain reactions is one of the most characteristic features of free radical reactions. A fire is one example of a chain reaction involving free radicals, and fires illustrate a very important feature of chain reactions: the tendency to produce widespread damage. A chain reaction that is capable of producing widespread organ damage is described below.

Lipid Peroxidation

The rancidity that develops in decaying food is the result of oxidative changes in polyunsaturated fatty acids (8). This same process, called lipid peroxidation, is also responsible for the oxidative damage of membrane lipids. The lipophilic interior of cell membranes is rich in polyunsaturated fatty acids (e.g., arachidonic acid) and the characteristic low melting point of these fatty acids may be responsible for the fluidity of cell membranes. Oxidation increases the melting point of membrane fatty acids and reduces membrane fluidity. The membranes eventually lose their selective permeability and become leaky, predisposing cells to osmotic disruption (8).

The peroxidation of membrane lipids proceeds as shown in Figure 3.2. The reaction sequence is initiated by a strong oxidant such as the hydroxyl radical, which removes an entire hydrogen atom (proton and electron) from one of the carbon atoms in a polyunsaturated fatty acid. This creates a carbon-centered radical (C•), which is then transformed into an oxygen-centered peroxy radical (COO•) that can remove a hydrogen atom from an adjacent fatty acid and initiate a new series of reactions. The final propagation reaction creates a self-sustaining chain reaction that will continue until the substrate (i.e., fatty acid) is exhausted, or until something interferes with the propagation reaction. (The latter mechanism is the basis for the antioxidant action of vitamin E, which is described later.)


Free radical reactions have been implicated in the pathogenesis of more than 100 diseases (9), but it is not clear whether oxidant injury is a cause or a consequence of disease (9,10). However, a chain reaction is an independent process (i.e., independent of the initiating process), and if it causes tissue injury it becomes an independent pathologic process (a primary illness).


Two radicals can react by sharing electrons to form a covalent bond. This eliminates the free radicals but does not necessarily eliminate the risk of toxicity. In the example below, the product of a radical-radical reaction is much more destructive than both radicals combined.

Nitric Oxide Transformation

Nitric oxide has been placed in a category of its own as a free radical because of its beneficial actions as a vasodilator, neurotransmitter, and bactericidal agent (11). The regard for nitric oxide has been so favorable that it was named "Molecule of the Year" by Science Magazine in 1992. However, despite its favorable profile, nitric oxide can become a toxin in the presence of superoxide. The reaction of superoxide with nitric oxide generates a powerful oxidant called peroxynitrite, which is 2000 times more potent than hydrogen peroxide as an oxidizing agent (12). Peroxynitrite can either cause direct tissue damage or can decompose and produce hydroxyl radicals and nitrogen dioxide.

NO. + O2. --» ONOOO-(peroxynitrite)

ONOOOH --» .OH + NO2

The transformation of nitric oxide into a source of oxidant injury demonstrates how free radicals can promote oxidant damage indirectly.


Evidence for endogenous antioxidant protection is provided by the simple observation that accelerated decay begins at the moment of death. This section presents the substances that are believed to play a major role in this protection.

An antioxidant is defined as any chemical species that can reduce or delay the oxidation of an oxidizable substrate (2). The nonenzyme antioxidants are included in Table 3.1.


There are three enzymes that function as antioxidants, shown in Figure 3.3. Note that the reaction sequence in this figure is the same as in Figure 3.1.

Superoxide Dismutase

The discovery of superoxide dismutase (SOD) enzyme in 1969 was the first indication of free radical activity in humans, and this began the frenzy of interest in free radicals. The role of SOD as an antioxidant is not clear. In fact, SOD promotes the formation of hydrogen peroxide, which is an oxidant. How can an enzyme that promotes the formation of an oxidant be defined as an antioxidant? In fact, if SOD increases the metabolic traffic flowing through hydrogen peroxide, and the catalase and peroxidase reactions are unable to increase their activity proportionally, the hydrogen peroxide levels may rise, and in this situation SOD functions as a pro-oxidant (13). Thus, SOD is not an antioxidant at least some of the time.


Catalase is an iron-containing heme protein that reduces hydrogen peroxide to water. It is present in most cells, but is lowest in cardiac cells and neurones. Inhibition of the catalase enzyme does not enhance the toxicity of hydrogen peroxide for endothelial cells (14), so the role of this enzyme as an antioxidant is unclear.

Glutathione Peroxidase

The peroxidase enzyme reduces hydrogen peroxide to water by removing electrons from glutathione in its reduced form and then donating the electrons to hydrogen peroxide. Glutathione is returned to its reduced state by a reductase enzyme that transfers the reducing equivalents from NADPH. The total reaction can be written as follows:

Peroxidase reaction: H2O2 + 2GSH --» 2H2O + GSSG

Reductase reaction: NADPH + H + GSSG --» 2GSH + NADP

where GSSG and GSH are oxidized and reduced glutathione, respectively.


The activity of the glutathione peroxidase enzyme in humans depends on the trace element selenium. Selenium is an essential nutrient with a recommended dietary allowance of 70 ug daily for men and 55 ug daily for women (15). Despite this recommendation, selenium is not included in most total parenteral nutrition regimens. Because the absence of dietary selenium produces measurable differences in glutathione peroxidase activity after just 1 week (16), the routine administration of selenium seems justified. However, selenium, has no clear-cut deficiency syndrome in humans, so there is little impetus to provide selenium on a routine basis.

Selenium status can be monitored with whole blood selenium levels. The normal range is 0.5 to 2.5 mg/L. Selenium can be provided intravenously as sodium selenite (17). The highest daily dose that is considered safe is 200 ug, given in divided doses (50 ug intravenously every 6 hours).



One of the major antioxidants in the human body is a sulfur-containing tripeptide glutathione (glycine-cysteine-glutamine), which is present in molar concentrations (0.5 to 10 mM/L) in most cells (18,19). Glutathione is a reducing agent by virtue of a sulfhydryl group in the cysteine residue of the molecule. It is normally in the reduced state (GSH), and the ratio of reduced to oxidized forms is 10:1. The major antioxidant action of glutathione is to reduce hydrogen peroxide directly to water, which diverts hydrogen peroxide from producing hydroxyl radicals. Glutathione is found in all organs, but is particularly prevalent in the lung, liver, endothelium, and intestinal mucosa. It is primarily an intracellular antioxidant, and plasma levels of glutathione are three orders of magnitude lower than intracellular levels.

Glutathione does not cross cell membranes directly, but is broken down first into its constituent amino acids and then reconstituted on the other side of the membrane. It is synthesized in every cell of the body, and largely remains sequestered within cells. Exogenous glutathione has little effect on intracellular levels (20), which limits the therapeutic value of this agent.


N-Acetylcysteine, a popular mucolytic agent (Mucomyst), is a sulfhydryl-containing glutathione analog capable of passing readily across cell membranes. N-Acetylcysteine is effective as a glutathione analog in acetaminophen toxicity, which is the result of an overwhelmed glutathione detoxification pathway (see Chapter 53). Therefore, N-acetylcysteine has a proven track record as an exogenous glutathione analog.

N-Acetylcysteine may prove to be a valuable antioxidant for therapeutic use. It protects the myocardium from ischemic injury, and has been successful in reducing the incidence of reperfusion injury during cardiac catheterization (21). It has also been used with some success in treating critically ill patients with acute respiratory distress syndrome and inflammatory shock syndromes (22,23).

Vitamin E

Vitamin E (a-tocopherol) is a lipid-soluble vitamin that functions primarily as an antioxidant that antagonizes the peroxidative injury of membrane lipids. It is the only antioxidant capable of halting the propagation of lipid peroxidation. The mechanism for this action is shown in Figure 3.4. Vitamin E inhabits the lipophilic interior of cell membranes, where the polyunsaturated fatty acids are also located. When a propagating wave of lipid peroxidation reaches vitamin E, it is oxidized to a free radical, thereby sparing any adjacent polyunsaturated fatty acids from oxidation. The vitamin E radical is poorly reactive, and this halts the propagation of the peroxidation reactions. This action has earned vitamin E the title of a chain-breaking antioxidant. The vitamin E radical is transformed back to vitamin E, and vitamin C can act as the electron donor in this reaction.

Vitamin E deficiency may be common in critically ill patients (24). The normal vitamin E level in plasma is 1 mg/dL, and a level below 0.5 mg/dL is evidence of deficiency (25). Considering the important role of vitamin E as an antioxidant, it seems wise to check the vitamin E status in patients who are at risk for oxidant injury (see Table 3.2).

Vitamin C

Vitamin C (ascorbic acid) is a reducing agent that can donate electrons to free radicals and fill their electron orbitals. It is a water-soluble antioxidant, and operates primarily in the extracellular space. Vitamin C is found in abundance in the lung, where it may play a protective role in inactivating pollutants that enter the airways.

The problem with vitamin C is its tendency to promote (rather than retard) the formation of oxidants in the presence of iron and copper (26-28). Vitamin C reduces iron to the Fe(II) state, and this normally aids in the absorption of iron from the intestinal tract. However, Fe(II) can promote the production of hydroxyl radicals, as described earlier. Thus, vitamin C can function as a pro-oxidant by maintaining iron in its reduced or Fe(II) state. The reactions involved are as follows:

Ascorbate + Fe(III) --» Fe(II) + Dehydroascorbate

Fe(II) + H2O2 --» .OH + OH- + Fe(III)

Several conditions that are common in ICU patients can promote an increase in free iron. Among these are inflammation, blood transfusions, and reductions in binding proteins. The prevalence of these conditions raises serious concerns about the use of vitamin C as an exogenous antioxidant in the ICU patient population.

Plasma Antioxidants

The plasma components with antioxidant activity are listed at the very bottom of Table 3.2. Most of the antioxidant activity in plasma can be traced to two proteins that make up only 4% of total plasma protein pool (27): ceruloplasmin (the copper transport or storage protein) and transferrin. Transferrin binds iron in the Fe(III) state, and ceruloplasmin oxidizes iron from the Fe(II) to Fe(III) state. Therefore, ceruloplasmin helps transferrin to bind iron, and both proteins then act to limit free iron in the plasma. For this reason, iron sequestration has been proposed as the major antioxidant activity in plasma (24). This is consistent with the actions of Fe(II) to promote free radical production, as shown in Figure 3.1.


The risk and severity of oxidation-induced tissue injury are determined by the balance between oxidant and antioxidant activities. When oxidant activity exceeds the neutralizing capacity of the antioxidants, the excess or unopposed oxidant activity can promote tissue injury. This condition of unopposed biological oxidation is known as oxidant stress (29).


Any imbalance in the activities of oxidants and antioxidants can result in unopposed oxidation. The box plots in Figure 3.5 show the effects of two conditions that promote oxidant-antioxidant imbalance on the level of oxidant stress in humans. The index of oxidant stress in this study is the activity of lipid hydroperoxides in urine, measured as spontaneous chemiluminescence and reported in counts per minute (CPM). Healthy, nonsmoking adults (the control group) show the lowest level of oxidant activity. The effects of an increase in oxidant burden is shown in a group of heavy smokers (one puff of a cigarette contains roughly one billion free radicals). The effects of a deficiency in antioxidant protection are shown for a group of patients with human immunodeficiency virus (HIV) infection (glutathione deficiency is common in HIV infections) (30). Each of the predisposing conditions is accompanied by a significant increase in oxidant activity in comparison to the activity in the healthy, control subjects. Note also that the HIV patients have ongoing oxidant stress when they are symptom-free. This supports the notion that oxidant stress is a cause and not a consequence of pathologic organ injury.


As mentioned earlier, oxidants have been implicated in the pathogenesis of more than 100 clinical diseases (9); the ones most likely to be encountered in the ICU are listed in Table 3.2. Unopposed biological oxidation has been documented in each of these clinical conditions (9,10,30-35). This does not establish a causal role for oxidation (this will require evidence that antioxidant therapy can improve outcome), but the tendency for oxidation to cause independent and progressive tissue damage (e.g., in chain reactions) is reason enough to suspect that oxidant-induced injury plays a role in these illnesses.


Most of the clinical conditions in Table 3.2 are accompanied by inflammation, and the conditions with multiorgan involvement are often associated with a progressive, systemic inflammatory response. As a result, inflammation has been proposed as a principal offender in pathologic forms of oxidant injury. The release of free radicals from activated neutrophils and macrophages creates an oxidant-intense environment, and the ability of host cells to withstand this oxidative assault may be the important factor in determining the clinical course of inflammatory conditions. In the desirable world, leukocyte-derived oxidants would annihilate all invading microbes, but would not affect the host cells. In the undesirable world, the inflammatory oxidants would destroy both the invader and the host. This proposed scheme is intuitively appealing, and emphasizes the value of providing antioxidants routinely in inflammatory illnesses.

Antioxidant Therapy

The value of maintaining antioxidant protection is not a debatable issue because loss of antioxidant defenses is a known cause of tissue destruction; the best example of this is the accelerated decay that occurs after death. Therefore, antioxidant supplements should be considered as a routine measure in patients who spend more than a few days in the ICU. In patients with any of the clinical conditions in Table 3.2, it would be wise to monitor some of the endogenous antioxidants (vitamin E, vitamin C, and selenium). A reduced antioxidant level in blood (or any body fluid) may not indicate a deficiency state (it may indicate that the antioxidant is being used), but would certainly be an indication to provide supplements. The real value of antioxidant therapy will be determined by clinical studies, some of which are currently in progress.


The tendency for aerobic metabolism to generate toxins has significant implications for the approach to metabolic support in critically ill patients. When metabolically-generated oxidants overwhelm the body's antioxidant defenses, the common practice of supporting metabolism by promoting the availability of oxygen and nutrients serves only to generate more toxic metabolites. The proper maneuver here is to support the antioxidant defenses. This approach adds another dimension to the concept of metabolic support by considering the output side of metabolism. Remember that metabolism is an engine (i.e., an energy converter) and like all engines, it has an exhaust that contains noxious byproducts of combustion. The exhaust from an automobile engine adds pollutants to the atmosphere; likewise, the exhaust from a metabolic engine adds pollutants to the "biosphere."