http://surgerybook.net/

temporary double lumen dialysis catheter placed in the internal jugular, femoral, or subclavian vein (in

order of preference). Once the dialysis catheter is inserted, the blood flow through the filter is moved

by a peristaltic pump. SCUF is utilized exclusively for volume removal which can be achieved at up to 2

L/hr. No replacement fluid is used, thus, there is minimal clearance of toxins. Arteriovenous circuits,

employing the patient’s own perfusion pressure to drive the system, are rarely used currently. Rather a

pump is used to drive a continuous venovenous circuit that can provide hemofiltration, hemodialysis,

or, most often, both. Continuous venovenous hemofiltration (CVVH) uses replacement fluid in the same

direction as the blood flow to dilute toxins in plasma as the toxins are removed with the ultrafiltrate.

Clearance of toxins is through the process of convection based on particle size where small, medium,

and large molecules are sieved out across a pressure gradient. Thus, electrolytes are leeched out and

replacement fluid is warranted. Continuous venovenous hemodialysis (CVVHD) uses a dialysate that

flows through the filter in the opposite direction of blood flow (and no replacement fluid) to remove

toxic waste and fluid of small molecules (such as urea) through the process of diffusion across a

concentration gradient. Finally, continuous venovenous hemodiafiltration (CVVHDF) combines both

convection and diffusion through a high-permeability membrane, using both dialysate and replacement

fluids. In those at a high risk of bleeding, it is reasonable to consider citrate anticoagulation (in the

absence of liver failure) or no anticoagulation with predilution administration of replacement fluids. In

those with moderate bleeding risk and/or frequent filter clotting, low-dose unfractionated heparin or

LMWH should be administered or argatroban in those with heparin-induced thrombocytopenia. The use

of CRRT may be preferred in those with brain edema, persistent metabolic acidosis, large fluid removal

requirements, and severe hemodynamic instability although hybrid modes, described below, may play a

role. Most patients have optimal clearance and treatment with running CRRT at 30 mL/kg/hr.137 There

has been some discussion about using high flow CRRT (70 mL/kg/hr) for the treatment of sepsis. In

patients with severe sepsis or septic shock without renal failure, there is no role for high volume

hemofiltration as reinforced in a recent multicenter study, the high volume in intensive care (IVOIRE)

trial.138

Intermittent hemodialysis (IHD) is performed at high flow rates (>300 mL/min) in a short session

relying on diffusion to clear toxins. Although it is thought that IHD is contraindicated in the face of

hemodynamic instability, it is possible to utilize this mode. Sufficient data do not exist to determine

which dialysis mode is optimal in promoting eventual renal recovery. Analysis suggesting a lower rate

of recovery with IHD relies heavily on data from older observational trials.139,140

Hybrid modes include sustained low efficiency dialysis (SLED). This can be thought of as stretching

out an intermittent dialysis session to promote improved hemodynamic stability and take advantage of

the cost benefits of IHD. SLED uses a conventional hemodialysis machine over 8 to 12 hours with lower

blood and dialysate flow rates. Various strategies exist to improve hemodynamic tolerance.

ENDOCRINE ISSUES IN THE ICU

Glucose Control

11 Hyperglycemia is common in critically ill patients and associated with poor clinical outcomes.

Critically ill patients are hypermetabolic due to elevated cortisol, growth hormone and catecholamine

levels, leading to increased hepatic gluconeogenesis, insulin resistance and hyperglycemia. Since

critically ill patients with hyperglycemia have increased mortality and complications including infection

and worse neurologic outcomes, assiduous glucose control has been practiced in critically ill

hyperglycemic patients.141,142 However, the optimal blood glucose range is controversial. Intensive

insulin therapy (IIT) typically employs continuous insulin infusions to maintain glucose levels of 80 to

110 mg/dL. Conventional blood glucose therapy could utilize either bolus or continuous insulin to

maintain glucose levels at a higher range. The Leuven surgical trial of mostly cardiac surgery patients

provided the groundwork for the use of IIT. One thousand five hundred and forty-eight patients were

randomized to IIT with infusion insulin versus conventional glucose management of 180 to 200 mg/dL.

IIT resulted in lower ICU and hospital mortality (4.6% and 7.2% vs. 8% and 10.9%, respectively), renal

failure, transfusion requirement, blood stream infections, and critical illness neuropathy, with more

frequent hypoglycemia.143 Other trials that included surgical patients reached opposite conclusions. The

Volume Substitution and Insulin Therapy in Severe Sepsis (VISEP) trial was a multicenter trial that was

stopped early because of the increased rate of hypoglycemia and serious adverse events in the IIT group

without a difference in mortality.144 Similarly, the multicenter Glucontrol trial was terminated early

324

http://surgerybook.net/

because of the high rate of hypoglycemia in the IIT group.145 The largest trial, Normoglycemic in

Intensive Care Evaluation Survival Using Glucose Algorithm Regulation (NICE – SUGAR) included over

6,000 medical and surgical ICU patients randomized to IIT with blood glucose targets of 81 to 108

mg/dL and conventional blood glucose targets of <180 mg/dL. The IIT group had decreased mean

blood glucose levels, significantly higher mortality and more incidence of severe hypoglycemia.146

Current guidelines recommend moderate blood glucose control of 144 to 180 mg/dL for critically ill

patients to avoid both marked hyperglycemia and iatrogenic hypoglycemia.147 This can be achieved

predominantly by avoiding the use of glucose-containing intravenous fluids to avoid the need for insulin

administration. If insulin is needed, continuous infusions and intermittent short-acting agents may be

safer in ICU patients.

Adrenal Insufficiency

Absolute adrenal insufficiency is rare in critically ill patients, with an incidence estimated at less than

5%.148 Relative adrenal insufficiency, manifested by hypotension and fever (and rarely with electrolyte

abnormalities), is common in ICU patients and associated with worse outcomes. In general, most

(absolute) adrenal insufficiency is secondary to abrupt withdrawal of therapeutic steroids. In the ICU,

however, although total levels of cortisol may be normal, they may not respond appropriately to stress

situations. Further, the typical diurnal variation of cortisol appreciated in normal volunteers is not seen

in ICU patients, as stressors are not eliminated at night, making cortisol assay an easier task. In the past,

relative adrenal insufficiency was defined by a cortisol level of less than 11 to 25 mcg/dL and/or failure

to raise the level by more than 9 mcg/dL after stimulation by high-dose (250 mcg) ACTH.149 However,

there is much debate as to what cortisol level is appropriate in septic shock, what constitutes an

adequate response to ACTH stimulation and what dose of ACTH should be used. After careful study, it

appears that ACTH, or its synthetic analogue, cosyntropin, stimulation is not helpful in distinguishing

those with relative adrenal insufficiency who would derive benefit from steroid administration. Prior

work in the 1980 s and 90 s did not show benefit to high-dose steroid administration in septic shock.

The more recent Annane French trial demonstrated a benefit to administration of low-dose steroid

administration with concomitant mineralocorticoids in septic shock in terms of mortality.150 However,

the Corticosteroid Therapy of Septic Shock (CORTICUS) trial, a large multicenter study, did not

demonstrate a mortality benefit to administration of steroids in septic shock.151 However, patients

received therapy later than in the French trial and were not also administered mineralocorticoids.

Subsequent meta-analyses of the role of steroids in septic shock reveal no difference in mortality but

improved shock reversal.152,153 Current recommendations advise administration of low-dose,

“replacement” steroids (glucocorticoids) to septic patients who remain hypotensive despite adequate

volume resuscitation, particularly those whose serum cortisol levels are below 20 mcg/dL.20 Still

controversial is whether or not to administer a mineralocorticoid concomitantly to those felt adrenally

insufficient. If fludrocortisone is not administered, then it is important to remember that glucocorticoids

have varied mineralocorticoid potency – none seen with dexamethasone and half seen with

methylprednisolone compared to hydrocortisone. Typically 200 to 300 mg hydrocortisone daily is

administered in intermittent doses, every 8 hours, for 5 to 7 days and then tapered as guided by clinical

response. A parallel to the relative insufficiency of cortisol seen in ICU patients with septic shock is that

of arginine vasopressin. Therapy with low, replacement doses of vasopressin might be considered

analogous to that with glucocorticoids and may, in fact, be complimentary to this therapy.154 A final

point to mention is that the induction agent etomidate is associated with relative adrenal insufficiency.

Etomidate is laudatory because it is not associated with hemodynamic instability and, thus, had been

favored in shock. However, etomidate is a reversible inhibitor of 11-beta-hydroxylase, and thus may

cause transient adrenal insufficiency which theoretically can worsen outcome in sepsis. Studies have not

definitively demonstrated worse outcome between patients intubated using etomidate versus other

induction agents such as ketamine. Practitioners should balance the risks and benefits of etomidate

administration and consider the transient administration of steroids to blunt adverse effects in the

airway management of those with sepsis.155

Thyroid Abnormalities

Thyroid abnormalities occur infrequently in ICU patients. Although the half-life of commonly

administered thyroid medications is long, it is possible that chronically ill patients who eventually

present to the ICU could develop severe hypothyroidism. The typical patient with myxedema coma has

a long-standing need for replacement thyroid medications and is ill after infection or trauma. Patients

325

http://surgerybook.net/

are hypothermic, bradycardic, and obtunded with low serum sodium and glucose levels. The treatment

includes securing an airway, hydration, warming, and administration of thryoxine and hydrocortisone.

The treatment of thyroid storm is discussed elsewhere in more detail. It occurs in patients with Graves

disease or toxic nodular goiters who become infected, pregnant, traumatized, or undergo surgery.

Supportive treatment includes ICU admission to monitor hemodynamic status and arrhythmias, cooling,

hydration, and beta blockade. Propylthiouracil (PTU) or methimazole is administered both to inhibit the

release of T4 and to block the conversion of T4 to T3. Thereafter, iodine as SSKI or Lugol solution is

given, once hemodynamics have been stabilized to inhibit T4 release. Glucocorticoids may also be

indicated. Finally, there is an entity of pseudohypothyroidism, the so-called sick syndrome, that occurs

in ICU patients who have normal thyroid function but low levels of T3, freeT4, and T4 as T3 is

converted to its isomer reverse T3. Several mechanisms can contribute to the inhibition of 5′-

monodeiodination and therefore to the low serum T3 concentrations in patients with nonthyroidal

illness. These include elevated cortisol levels with glucocorticoid therapy, free fatty acids that inhibit

deiodinase, cytokines, and treatment with drugs that inhibit 5′-monodeiodinase activity such as

amiodarone. No treatment is required. If there is additional evidence to suggest a diagnosis of

hypothyroidism, critically ill patients, should receive cautious thyroid hormone replacement beginning

with approximately half the expected full replacement dose of levothyroxine.

INFECTIOUS DISEASE ISSUES IN THE ICU

Sepsis Guidelines

Evidence-based sepsis resuscitation and global care guidelines are set forth in the Surviving Sepsis

Campaign, now in its third version.20 These guidelines prioritize care for those with the most acute

forms of sepsis – that is, severe sepsis and septic shock. Severe sepsis refers to those infected patients

with hypotension and/or hypoperfusion with organ dysfunction (manifested by oliguria, acute mental

status change, new coagulopathy or thrombocytopenia, or acute lung injury). Septic shock refers to

those who require administration of pressors despite initial volume resuscitation. Care of the septic

patient with severe sepsis and septic shock is suggested to adhere to bundles that optimally should be

completed within 3 and 6 hours (Table 10-2). Resuscitation targets a MAP of >65 mm Hg, a central

venous pressure (CVP) of >8 mm Hg, and a central venous oxygen saturation (ScvO2

) of >70%.

Pressors are administered if the initial fluid bolus does not achieve the targets above and inotropes

and/or blood transfusion are given to maintain a hemoglobin >10 g/dL if the ScvO2 goal is not met.

The principal evidence underlying the resuscitation strategy is derived from Emanuel Rivers’ early goaldirected therapy study (EGDT) published in 2001.156 This single center trial allocated patients to usual

care or CVP-directed intravenous fluid administration with MAP and ScvO2 goals achieved by

administration of dopamine, dobutamine, and packed red blood cells. In this study, Rivers and

colleagues demonstrated a significant difference in mortality between the groups – 30.5% in the study

patients versus 46.5% in the controls. Further, the current surviving sepsis guidelines render several

specific recommendations regarding fluid administration. They advocate for the use of crystalloids as

the initial fluid therapy of choice and against the use of hydroethyl starches and suggest that albumin

may be beneficial if large volume crystalloid resuscitation is required.

12 In the interim, three trials have been conducted that call into question Rivers’ resuscitation

paradigm. The ProCESS trial, a large, multicenter controlled trial randomized patients into three groups

with excellent protocol compliance (90%) and equivalent usage of early antibiotics. The three groups

were the EGDT group (439 patients) exactly as described by Rivers with CVP and ScvO2 monitoring and

administration of dobutamine and packed cells. The second (protocol-based standard care) group (446

patients) was resuscitated to clinical euvolemia within 6 hours. Insertion of a CVC or and ScvO2 catheter

was not mandatory. Finally, a third (usual care) group (456 patients) was resuscitated according to the

clinical judgment of the provider. Approximately half of the protocol and usual group patients received

CVCs. Fluid administration was not clinically different between the protocol- and goal-directed therapy

groups (2.3 L vs. 3.3 L, respectively). However, dobutamine use and packed cell transfusion was

roughly eight times higher and double, respectively, in the EGDT group. Most importantly, there was

no difference in 60-day, 90-day, or 1-year mortality, length of stay, or organ failure.36 The second, the

ARISE trial, randomized patients with early septic shock to receive EGDT (796 patients) or usual care

(804 patients) and corroborated the results of the ProCESS trial. There was no difference in 90-day

mortality, but the EGDT group was twice as likely to receive red blood cell transfusion and was six

326

http://surgerybook.net/

times as likely to receive dobutamine.37 The third trial – ProMISe from the United Kingdom, still awaits

publication. It is apparent that these studies will result in a de-emphasis on CVP and ScvO2

-guided

resuscitation in future iterations. Finally, the guidelines provide an in-depth evidence-based review of

the components of global sepsis care, the evidence for which is described throughout this chapter.

Use of Procalcitonin to Guide Antibiotic Therapy

Procalcitonin, a precursor of calcitonin with a half-life of about 24 hours, is released in response to

bacterial toxins and interleukin-1b and has been identified as a marker of sepsis and bacterial infection

for nearly a decade. Procalcitonin increases within 4 to 6 hours or bacterial infection, is attenuated in

the presence of viral infection and not affected in the presence of fungal infections. Far more important

clinically is the observation that procalcitonin levels halve once bacterial infection is controlled by

antibacterials. Protocols have been developed to stop empiric or therapeutic antibiotics once

procalcitonin levels decrease by 80% or to an absolute level of 0 to 0.5 ng/mL. Seven randomized trials

in ICU patients were subjected to a systematic review and meta-analysis suggesting that procalcitonin

assay reduces antibiotic use by an average of two days per patient without otherwise affecting

outcome.157 An in-depth economic analysis suggests that use of procalcitonin is effective in the ICU

setting.158 More recent studies support the notion that use of procalcitonin is useful in determining

when to stop antibiotics in ICU patients and that these shorter therapeutic courses are not associated

with a difference in outcome.159,160

Central Line Associated Blood Stream Infections

Central line associated blood stream infections (CLABSI) can be an important cause of morbidity,

resource utilization, and even attributable mortality in ICU patients. Infection control measures have

been successful in dramatically lowering the incidence of this infection, defined as 15 colony forming

units (cfu)/mL by semiquantitative or 103 by quantitative means. Perhaps the most successful

intervention to preventing CLABSI is to remove (or not insert) central catheters if they are not needed;

this is the cornerstone of care bundles or checklists (Table 10-3).161 Antiseptic or antibiotic-coated

catheters should be considered in ICUs with a high baseline rate of infection after employing less costly

measures. Chlorhexidine-impregnated dressings lower CLABSI rates as well.162 Routine wire changing of

catheters does not decrease CLABSI rate. In the past the subclavian site was felt to be the least prone to

infection; however, recent studies indicate that internal jugular (and perhaps even femoral) catheters

are not much worse. Further, no randomized controlled trial compares subclavian versus internal

jugular versus femoral insertion in a single study.

Catheter Associated Urinary Tract Infections

Infection control measures have also been undertaken to prevent catheter associated urinary tract

infections (CAUTI), a common cause of morbidity, but rarely of mortality, in surgical ICU patients.

CAUTI are the leading cause of secondary healthcare-associated bacteremia and may be associated with

symptomatic or asymptomatic bacteriuria of >10 cfu/mL.5 Most strategies focus on prompt removal of

indwelling catheters. There may be some role for use of antiseptic- (silver preparations) or antibioticcoated catheters (minocycline-rifampin) in high-risk patients, although clinical trials have not been

definitive.164,165

The Use of Selective Oral and Selective Digestive Decontamination (SOD and SDD) and

Other Strategies to Address Resistant Organisms in the ICU

Antibiotic-resistant bacteria are discussed more in detail in this text. Those relevant to ICU patients

include methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococcus, and

multi-drug resistant gram negative pathogens, including carbapenem-resistant enterobacteriaceae.

Various infection control strategies have been promulgated in ICU patients to address these virulent

organisms. We will discuss the use of selective decontamination and MRSA decontamination strategies

here. The practice of selective decontamination was developed many decades ago and is used

extensively in surgical and trauma patients in Europe. Its goal is to minimize ventilator-associated

pneumonia to lessen multiple organ dysfunction and improve outcome in high-risk populations. The

cornerstone of the therapy is the administration of a short course (4 day) of parenteral antibiotics to

address respiratory gram negative pathogens, typically a second generation cephalosporin. SOD

strategies add use of a chlorhexidine mouthwash. SDD employs a nonabsorbable enteral antibiotic (such

327