REVIEW ARTICLE

Anaesthetic management of patients with severe sepsisD. Eissa, E. G. Carton and D. J. Buggy*

Division of Anaesthesia, Mater Misericordiae University Hospital, Eccles Street, Dublin 7, Ireland

* Corresponding author. E-mail: donal.buggy@nbsp.ie, dbuggy@mater.ie

Key points

† Patients with sepsis oftenrequire surgicalinterventions.

† Anaesthesia can behazardous in thesecardiovascularly unstablepatients.

† Preoperative optimizationand intraoperative andpostoperative care needto be planned beforestarting.

† Techniques that preservecardiovascular andrespiratory function arerequired.

Summary. Severe sepsis, a syndrome characterized by systemic inflammation and acuteorgan dysfunction in response to infection, is a major healthcare problem affecting all agegroups throughout the world. Anaesthetists play a central role in the multidisciplinarymanagement of patients with severe sepsis from their initial deterioration at ward level,transfer to the diagnostic imaging suite, and intraoperative management for emergencysurgery. The timely administration of appropriate i.v. antimicrobial therapy is a crucial stepin the care of patients with severe sepsis who may require surgery to control the source ofsepsis. Preoperative resuscitation, aimed at optimizing major organ perfusion, is based onjudicious use of fluids, vasopressors, and inotropes. Intraoperative anaesthesiamanagement requires careful induction and maintenance of anaesthesia, optimizingintravascular volume status, avoidance of lung injury during mechanical ventilation, andongoing monitoring of arterial blood gases, lactate concentration, haematological andrenal indices, and electrolyte levels. Postoperative care overlaps with ongoingmanagement of the severe sepsis syndrome patient in the intensive care unit. Thesepatients are by definition, high risk, already requiring multiple supports, and requireexperienced and skilful decision-making to optimize their chances of a favourableoutcome. Similar to acute myocardial infarction, stroke, or acute trauma, the initial hours(golden hours) of clinical management of severe sepsis represent an importantopportunity to reduce morbidity and mortality. Rapid clinical assessment, resuscitationand surgical management by a focused multidisciplinary team, and early effectiveantimicrobial therapy are the key components to improved patient outcome.

Keywords: anaesthesia; emergency service; anaesthesia, general; infection; surgery;perioperative period

EpidemiologySevere sepsis and septic shock are major healthcare prob-lems with a reported incidence of 66–132 per 100 000population in the USA and UK, respectively.1 2 In 2001, a con-sensus conference (Society of Critical Care Medicine, Euro-pean Society of Intensive Care Medicine, American Collegeof Chest Physicians, American Thoracic Society, and SurgicalInfection Society) concluded that the basic definitions of sys-temic inflammatory response syndrome (SIRS), as originallydescribed in 1992 by the American College of Chest Phys-icians and the Society of Critical Care Medicine,3 shouldremain largely unchanged4 (Table 1). Because of the limit-ations of the definitions of SIRS and infection, the 2001 con-sensus conference suggested an expanded list of possiblesigns of systemic inflammation that may be observed in‘septic-looking’ patients (Table 2).

Severe sepsis occurs in 1–2% of all hospitalizations andaccounts for as much as 25% of intensive care unit (ICU)bed utilization. It is common in elderly, immune-compromised, and critically ill patients and is a major

cause of death in ICUs worldwide.5 Sepsis is the secondleading cause of death in non-coronary ICU patients. Mor-tality remains high at 30–50% despite improved care inthe past 10–15 yr.1 5 6

Causes of sepsisSevere sepsis may have infective and non-infective causes(Table 3). Infections are common and amenable to treat-ment; therefore, in patients presenting with clinical signs ofsystemic inflammation (SIRS), an infective cause should beactively sought. Community-acquired infections in previouslywell patients are easier to recognize than nosocomial infec-tions in debilitated hospitalized patients. Infections leadingto sepsis include central nervous system (CNS) infections,for example, meningitis or encephalitis, cardiovascular infec-tions (e.g. infective endocarditis), respiratory infections (e.g.pneumonia), gastrointestinal infections (e.g. peritonitis), orurinary tract infections (e.g. pyelonephritis).1 7 Although bac-terial infections are the most common infective cause,viruses and fungi can also cause septic shock. Non-infective

British Journal of Anaesthesia 105 (6): 734–43 (2010)Advance Access publication 27 October 2010 . doi:10.1093/bja/aeq305

& The Author [2010]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved.For Permissions, please email: journals.permissions@oxfordjournal.org

causes include severe trauma or haemorrhage and acutesystemic disease, including myocardial infarction, pulmonaryembolus, and acute pancreatitis. Table 4 summarizes thepresentation of severe sepsis syndrome, the pathophysiologyunderpinning the symptoms and signs, and the organismsmost commonly implicated.

Pathophysiology of sepsisA detailed discussion of the physiology of sepsis is beyondthe scope of this review, but has itself been recently reviewedcomprehensively.8 This review concentrates on anaestheticmanagement of patients with severe sepsis syndrome.

Anaesthetic managementAnaesthetists are frequently involved in the care of severelyseptic patients in the emergency department, operatingtheatre, or ICU. Infection source control, involving surgicaldrainage of an abscess or debridement of necrotic tissuecoupled with early effective antimicrobial therapy, is centralto the successful treatment of a patient with severe sepsis.In high-risk surgical or trauma patients with sepsis, earlyhaemodynamic optimization before the development oforgan failure reduced mortality by 23% in comparison withthose who were optimized after the development of organfailure.9 10

Preoperative assessment

Although not all patients with severe sepsis have an infectivefocus, it is prudent to examine patients systematicallylooking for a source of infection (Table 4). The primarysource may be self-evident (e.g. trauma, burns, recentsurgery) or may be more difficult to identify (e.g. empyemaof the gall bladder, pancreatitis, gynaecological sepsis, softtissue, and bony infections), particularly in agitatedun-cooperative patients. The examination should focus onthe severity of SIRS, the state of intravascular hydration,the presence of shock or multi-organ dysfunction, and theadequacy of haemodynamic resuscitation.

Surviving Sepsis Campaign

Following an international process of consultation to stan-dardize the management of critically ill septic patients, theSurviving Sepsis Campaign suggested that therapies begrouped or ‘bundled’ for particular subsets of patients. Theconcept is not unlike that of Advanced Trauma Life Support(ATLS), where somewhat didactic therapies are proposed in

Table 2 Diagnostic criteria for sepsis

Documented or suspected infection with some of the followingclinical signs or laboratory data

1. Infection: documented or suspected infection

2. Signs of systemic inflammation

(a) General parameters

Fever (core temp. .38.88C)

Hypothermia (core temp. ,368C)

Tachycardia (.90 beats min21)

Tachypnoea (.30 bpm)

Altered mental status

Significant positive fluid balance (.20 ml kg21 over 24 h)

Hyperglycaemia (.7.7 mmol litre21) in non-diabeticpatients

(b) Inflammatory parameters

WCC ,4 or .12, .10% immature forms

C-reactive protein .2 SD above normal value

Plasma procalcitonin .2 SD above normal value

(c) Haemodynamic parameters

Arterial hypotension (SAP ,90 mm Hg)

SvO2.70%

CI.3.5

(d) Organ dysfunction parameters:

Hypoxic (PaO2/FIO2

,40)

Oliguria (,0.5 ml kg21 h21)

Creatinine increase (.0.5 mg dl21)

Coagulopathy (INR .1.5, aPPT.60 s, plt count,100)

Absent bowel sounds

Hyperbilirubinaemia

(e) Tissue hypoperfusion parameters

Lactate.3 mmol litre21

Decreased capillary refill

Mottling of skin

Table 1 2001 sepsis definitions by the American College of ChestPhysicians (ACCP) and the Society of Critical Care Medicine(SCCM)3 4

Pathological entity Definition

Bacteraemia Presence of bacteria in the bloodstream

Septicaemia The presence of large numbers of bacteriain the bloodstream often associated withsystemic signs and symptoms such asfever, rigors, and headache

SIRS The threshold definition is two or more ofthe following criteria:o temperature .388C or ,368Co heart rate .90 beats min21

o ventilatory frequency .20 bpm or PaCO2

,4.3 kPao WBC ,4×109 litre21 or .12×109 litre21

or .10% immature forms

Sepsis SIRS with clinical evidence of infection

Severe sepsis Sepsis associated with organ dysfunction,hypotension, or hypoperfusionabnormalities

Septic shock Sepsis-induced hypotension, despite fluidresuscitation, plus hypoperfusionabnormalities

Sepsis-inducedhypotension

A systolic arterial pressure ,90 mm Hg ora reduction of . 40 mm Hg from baselinein the absence of other causes forhypotension

Anaesthetic management of patients with severe sepsis BJA

735

given clinical situations. Although its detractors point outthat bundled therapies are not individualized to a particularpatient’s needs, and the lack of evidence-based medicineto underpin its guidelines, there is nonetheless some evi-dence that the process of care and outcomes improvedafter educational programmes were instituted based on theSurviving Sepsis Campaign.11 – 13

Antibiotic therapy

It is imperative that i.v. antibiotics should be started as earlyas possible after the diagnosis of severe sepsis and septicshock. There is no evidence that delaying until the start ofthe surgical procedure or until microbiology culture resultsare available is beneficial. Appropriate samples should beobtained for culture before giving first-line anti-microbialtherapy.14 Anti-microbial drugs are best given i.v. and in suf-ficient dosage to achieve therapeutic concentration. Thechoice of agents should be based on the clinical history,physical examination, likely pathogen(s), optimal penetrationof anti-microbial drugs into infected tissues, and the localpattern of sensitivity to anti-microbial agents. Broad-spectrum agents should be used initially with one or moreagents active against all likely bacterial/fungal pathogens.

Haemodynamic resuscitation

The objective of preoperative resuscitation measures is torapidly restore adequate oxygen delivery to peripheraltissues. If the patient is haemodynamically unstable, invasivearterial pressure monitoring, central venous access, and ICUor high dependency unit admission must be considered. Pla-cement of a central venous catheter (CVC) will allowmeasurement of central venous pressure (CVP), mixedvenous oxygen saturation (SvO2 ), administration of i.v.fluids, and vasopressor medication.15 – 17 Resuscitationmeasures begun in the emergency room can be continuedeven if the patient requires diagnostic imaging studies or

admission to the ICU before transfer to the operatingtheatre. The first 6 h of resuscitation of septic patients, theso-called ‘golden hours’, are crucial and frequently coincidewith the time for emergency surgery.11 18 There is little dis-agreement among clinicians that in the hypotensive septicpatient with lactate .3 mmol litre21, volume resuscitationusing crystalloids or colloids should be used initially, aimingto reach the following clinical endpoints: CVP 8–12 mm Hg,mean arterial pressure 65 mm Hg, urine output 0.5 ml kg21

h21, central venous oxygen saturation: .70% (Table 5).There is no evidence-based support for one type of i.v. fluidover another with regard to ICU stay, duration of mechanicalventilation, duration of renal replacement therapy, and 28day outcome.11 16 Colloid with pentastarch therapy wasassociated with higher rates of acute renal failure and renal-replacement therapy than Ringer’s lactate and its toxicity isincreased with accumulating doses.7

Vasopressor support with norepinephrine may be con-sidered even before optimal i.v. fluid loading has beenachieved. Low-dose vasopressin (0.03 units min21) may besubsequently added to reduce the requirement for high-dosenorepinephrine alone.10 18 19 Inotropes are added to volumeresuscitation and vasopressors, if there is evidence of contin-ued low cardiac output despite adequate cardiac filling andfluid resuscitation. The Surviving Sepsis Campaign rec-ommends that dobutamine is the first-line inotropetherapy to be added to vasopressors in septic patients.11

However, a study in septic patients showed no differencein efficacy and safety with epinephrine alone comparedwith norepinephrine plus dobutamine (28 day mortality:40% vs 34% respectively, P¼0.31) in the management ofseptic shock.19 There is no evidence to support the use ofdobutamine to achieve supernormal oxygen delivery interms of improving outcomes.16 – 18 Resuscitation effortsshould be continued as long as haemodynamic improve-ment accompanies each step in the process. Further i.v.fluid administration should be stopped when filling pressuresare high and no further improvement seen in tissue per-fusion is seen (e.g. serum lactate not decreasing). Transfu-sion of red blood cells may be considered if tissue oxygendelivery remains inadequate.20 21

Levosimendan may be a useful adjunct to conventionalinotropic therapy in cases of refractory myocardial dysfunc-tion in sepsis. Its inotropic effect is attributable to increasedcardiac troponin C sensitivity to calcium. The systemic andpulmonary vasodilator effect is attributable to its openingof ATP-dependent potassium channels.22 A single random-ized controlled trial in 28 patients with septic shock and ejec-tion fraction ,45% persisting .48 h after conventionaltreatment found that cardiac index and renal functionindices improved after levosimendan, compared with dobu-tamine.22 23 However, further larger clinical studies arerequired before levosimendan becomes a widely acceptedtherapy in septic shock.

Supplemental oxygen therapy is valuable in severelyseptic patients even if they do not have signs of respiratorydistress. Immediate tracheal intubation and mechanical

Table 3 Aetiology of severe sepsis

Infective causes Non-infective causes

CNS infections Severe trauma

CVS infections Haemorrhage

Respiratory infections Complication of surgery

Renal infections Complicated aortic aneurysm

GIT infections Myocardial infarction

Skin and soft tissue infections Pulmonary embolism

Bone and joint infections Cardiac tamponade

Subarachnoid haemorrhage

Burns

Acute pancreatitis

Drug overdose/toxicity

Diabetic ketoacidosis

Adrenal insufficiency

Anaphylaxis

BJA Eissa et al.

736

ventilation of the lungs can be considered if the patient’slevel of consciousness is low or if there is progressive distressand hypoxia.24 If there is an inadequate response to theseresuscitation measures, it is important to consider the pres-ence of an alternative diagnosis. The non-infective causes

of SIRS or an iatrogenic complication, for example, tensionpneumothorax after CVC placement, should also be con-sidered (Table 3).

Diagnostic imaging

Diagnostic imaging studies are increasingly important in con-firming the site of infection, excluding alternative pathologyand guiding radiological or surgical source control pro-cedures. If diagnostic imaging studies are considered appro-priate, it is important that all other therapeutic measures(e.g. i.v. fluid resuscitation, antimicrobial therapy, mechanicalventilation) are continued in a comprehensive manner. Com-puterized tomography is the most useful imaging modalityfor complex soft-tissue infections and deep-seated infectionsin the abdomen and thorax. Ultrasound imaging of the biliaryand urinary tract may also be considered. Expert interpret-ation of all imaging studies should be sought to assist inplanning the optimal management strategy.

Table 4 Common presentation, pathophysiology, and pathogens in severe sepsis

System affected Signs and symptoms Mechanism Common system pathogens

Central nervoussystem

Confusion, drowsiness, irritability comaheadache, neck stiffness, photophobia

Alteration in: blood–brain barrierneurotransmitter levels; receptorfunction energy availability

Community-acquired pathogens:Streptococcus pneumoniae; Neiserriameningitides; Listeria monocytogenes

Nosocomial pathogens: Pseudomonasaeruginosa; Escherichia coli

Cardiovascularsystem

Hypovolaemia, impaired myocardialcontractility, tachycardia, increasedcardiac output, decreased systemicvascular resistance (SVR), impairedresponsiveness to vasopressor agents,short of breath, orthopnoea, raisedvenous pressure

(a) Poor intake, inadequatereplacement, excessive insensiblelosses

Major community-acquired pathogens:Enterococcus; Streptococcus bovis;Streptococcus spp.; coagulase-negativestaphylococci; Coxiella burnetii; Staphylococcusaureus; Campylobacter; E. coli; fungus

(b) Increase in microvascularpermeability andhypoalbuminaemia

Major nosocomial pathogens: Staphylococcusspp.; methicillin-resistant S. aureus;methicillin-resistant Staphylococcusepidermidis; methicillin-resistant coagulasenegative Staphylococcus

(c) Myocardial depression

(d) Down-regulation of adrenergicreceptors heart valve dysfunction

Respiratorysystem

Hypoxaemia, cyanosis, tachypnoea, useof accessory muscles, change insputum: volume, purulance

(a) Increase in capillarypermeability; alveolar flooding

Major community-acquired pathogens:S. pneumoniae, Haemophilus influenzae,Legionella sp.

(b) Neutrophils recruited to thelung

Major nosocomial pathogens: aerobicGram-negative bacilli

(c) Pulmonary microemboliplatelet aggregates

Gastrointestinalsystem

Vomiting, diarrhoea, abdominal pain,tenderness, liver failure, cholestasis

Major community-acquired pathogens: E. coli; Bacteroides fragilis

Major nosocomial pathogens: aerobic Gram-negative bacilli anaerobes

GU system (a) Frequency, dysuria, haematuria,flank pain, renal failure

Mechanisms? Major community-acquired pathogens: any ofthe above-mentioned organisms as a result ofbacteraemia

Major nosocomial pathogens: any of theabove-mentioned organisms as a result ofbacteraemia

Table 5 Goal-directed therapy: a summary of clinical targets

Clinical parameter Goal

Central venouspressure

8–12 mm Hg (≥8 mm Hg inspontaneously breathing patient, ≥12mm Hg in ventilated patients)

Mean arterial pressure Between 65 and 90 mm Hg

Central venousoxygen saturation

≥70 mm Hg

Urine output ≥0.5 ml kg21 h21

Haematocrit ≥30%

Anaesthetic management of patients with severe sepsis BJA

737

Source control

Source control measures include drainage or debridementprocedures and definitive correction of anatomical abnorm-alities leading to ongoing contamination of previouslysterile tissue. Drainage procedures apply to well-circumscribed infections that can be drained either percuta-neously under image-guidance or by an open surgicalapproach. Debridement refers to the physical removal ofnon-viable solid tissue usually by an open surgical approach.Definitive surgical interventions are indicated to correct ana-tomical abnormalities and prevent further contamination.

A surgeon with experience in dealing with complex infec-tions in critically ill patients is best placed to be involved inthe decision-making process regarding a particular sourcecontrol procedure.25 The immediate goal is to achieve ade-quate control of the source of infection with the least phys-iological embarrassment. Source control intervention maycause further complications such as bleeding, fistulas, orinadvertent organ injury. The optimal timing of any surgicalintervention depends on the diagnosis and the clinicalcourse of the patient. In some patients, immediate surgeryor within 1–2 h of presentation (e.g. upper airway infectionsleading to airway compromise, necrotizing fasciitis) is life-saving.25 26 There are also a number of commonly occurringsevere infections (intra-abdominal abscess, infections associ-ated with intravascular or prosthetic device, infective endo-carditis with structural heart damage leading tocardiogenic shock) which may require urgent surgical inter-vention. The exception to this rule is peripancreatic necrosisassociated with acute pancreatitis, where percutaneous drai-nage and full supportive therapy facilitate delayed surgicalintervention, which is associated with improved outcome.27

Clear and timely communication between the anaesthe-tist, surgeon, microbiologist-infectious disease physician,and radiologist is essential for rapid implementation of aneffective treatment plan, which can be discussed with thepatient and their family. It is vital that the anaesthetistassumes a central role in the multidisciplinary team.

Intraoperative managementThe primary goal of the anaesthetist during the intraopera-tive period is to provide safe and optimal care for criticallyill septic patients so that they may benefit maximally fromthe surgical or radiological source control procedure. Themajority of surgical source control procedures are optimallycarried out in the operating theatre under general anaesthesia.

Before induction

Many source control procedures are done out of hours, so it isimportant that the anaesthetist has appropriate help avail-able in the operating theatre. Some thought should begiven early to whether the patient may require ICU manage-ment after operation. Awareness of the microbiologicalsamples sent for culture, the anti-microbial agents whichwere started, and timing of the next scheduled dose isimportant to optimize type and timing of intraoperative

antimicrobial therapy.28 Therapeutic concentrations of effec-tive antimicrobial agents should be maintained throughoutthe perioperative period as the procedure itself may causefurther bacteraemia and clinical deterioration. Invasivehaemodynamic monitoring is likely to be indicated inaddition to standard intraoperative monitoring. Serialmeasurements of arterial blood gases and lactate concen-tration should be readily available from near-patient testingequipment. If large volume loss is anticipated during the sur-gical procedure, it is worth considering placement of anappropriate volume resuscitation intravascular device.

Induction of anaesthesia and initiation ofmechanical ventilation

Patients undergoing source control procedures are in aninherently unstable cardiovascular state due to the com-bined effects of sepsis, anaesthesia, intravascular volumeloss, bleeding, and surgical stress. De-nitrogenation of thelungs, breathing 100% O2 through a tightly fitted facemaskfor up to 3 min, may be considered before induction ofanaesthesia. Because many surgical procedures on severelyseptic patients occur on an emergency basis, a modifiedrapid sequence induction, perhaps using rocuronium ratherthan succinylcholine to facilitate tracheal intubation, maybe required. Options for the induction technique are many,including ketamine, etomidate, and slow administration ofmore commonly used agents such as propofol. Most i.v. orinhalation anaesthetic agents cause vasodilation or impairedventricular contractility. Induction of anaesthesia is ideally adeliberate step-wise process, using small doses of i.v. anaes-thetic agents, titrated to clinical response. The choice ofinduction agent or narcotic is less important than the carewith which they are administered. Ketamine or midazolammay provide a degree of haemodynamic stability and short-acting opioids such as fentanyl or alfentanil will enable areduction in the dose of anaesthetic induction agent. Withthe exception of remifentanil, the effects and duration ofaction of i.v. opioids may be increased by impaired hepaticand renal perfusion. Remifentanil infusion, either as aprimary agent or as a background adjunct to another induc-tion drug, has much to recommend it in the setting of induc-tion of anaesthesia in the septic, unstable patient. Althoughit can cause bradycardia, many of these patients are tachy-cardic, and its effects on myocardial contractility areminimal. Further, remifentanil avoids sudden reductions insystemic vascular resistance.29 Placement of a cuffed tra-cheal tube is facilitated by the use of neuromuscular blockingagents (preferably non-histamine releasing agents).

Continued volume resuscitation and incremental doses ofvasopressors are helpful to counteract the hypotensive effectof anaesthetic agents and positive pressure mechanical ven-tilation. Options for the use of vasopressors include ephe-drine, phenylephrine, and metaraminol, but there is noevidence base to support the use of any of these in prefer-ence to another. Norepinephrine infusion may be used fora more prolonged effect.10 18 The goal of mechanically

BJA Eissa et al.

738

ventilating patients with severe sepsis is to use sufficientlyhigh fractional inspired oxygen concentration (FIO2 ) to main-tain adequate oxygenation (PaO2

.12 kPa). There is nowstrong evidence supporting a low tidal volume ventilatorystrategy, to minimize the impact of positive pressure venti-lation on the lung tissue itself, and also on venous returnand cardiac output.30 Shear forces caused by high tidalvolumes or high inspiratory pressures will exacerbate lunginjury. Therefore, where oxygenation is adequate, theconcept of ‘permissive hypercapnia’ has arisen, where lowalveolar minute ventilation to minimize ventilatory lungdamage inevitably results in a degree of hypercapnia (typi-cally PaCO2 .8–9 kPa), which is tolerated and appears rela-tively safe in the short term (i.e. more than 3–4 days).31

Maintenance of anaesthesia

There is no evidence to suggest an outcome benefit whenanaesthesia is maintained by the inhalation or i.v. route.Options for maintaining anaesthesia include inhalationagents, i.v. agents, and opioids, for example, remifentanilinfusion using 0.25–0.5 mg kg21 min21. The anaesthetistshould choose the technique which they believe best fitswith their assessment of the individual patient’s risk factorsand co-morbidities, and their own experience and expertise.The MAC of inhalation anaesthetic agents is reduced insevere sepsis.32 In patients with significant lung dysfunction,maintenance of stable concentrations of anaesthetic agentsin the brain may be more reliably achieved when using i.v.rather than inhalation agents. Whatever technique is used,the depth of anaesthesia achieved can be estimated usingbispectral index monitoring. During surgery, the haemo-dynamic state may be further complicated by blood loss orsystemic release of bacteria or endotoxins. Transfusion ofblood products should proceed without delay if the surgicalprocedure is complicated by excessive blood loss.

Intravascular volume resuscitation should continue asindicated throughout the surgical procedure. Although aCVP of 8–12 cm H2O is a commonly used haemodynamicgoal in the initial resuscitation of septic patients, intraopera-tive CVP values may be increased by raised intra-thoracic andintra-abdominal pressure. Changes in dynamic markers(pulse pressure variation, stroke volume variation) havebeen shown to predict volume responsiveness more accu-rately than pressure-based estimates (CVP or pulmonaryartery occlusion pressure). Changes in dynamic markers ofvolume responsiveness can be used intraoperatively toguide i.v. volume therapy, especially in patients withregular sinus heart rhythm and whose lungs are ventilatedby controlled mechanical ventilation. Concurrent transoeso-phageal echocardiography or oesophageal Doppler may beused to define changes in stroke volume variation.33 34

There are many devices available to monitor changes incardiac output either continuously (pulmonary artery cath-eter, oesophageal Doppler, impedance plethysmography) orat discrete time intervals (trans-thoracic or trans-oesophageal echocardiography, or serial measurement of

mixed-venous O2 saturation). Throughout the surgical pro-cedure, cardiovascular parameters (heart rate, cardiacfilling pressures, inotropic state, systemic arterial pressure)can be adjusted to optimize tissue oxygen delivery ratherthan to achieve set values of cardiac output or arterialpressure. The adequacy of global oxygen delivery may beassessed by serum lactate ,2 mmol litre21 and mixed-venous O2 saturation .70%.

Oxygenation may be impaired by non-cardiogenic pul-monary oedema, which is caused by the increased capillarypermeability in sepsis. Management options for hypoxaemiaduring maintenance of anaesthesia include increasing theinspired oxygen concentration and incrementally increasingPEEP. The inspired oxygen concentration can be increaseduntil SaO2

is at least 90% and the use of PEEP may be con-sidered during the surgical procedure. The PEEP may be cau-tiously increased in haemodynamically stable patients ifthere is still hypoxia despite increasing the FIO2

. Hypercarbiashould be avoided specifically in patients with raised intra-cranial pressure, compensated metabolic acidosis, or thelater stages of pregnancy. In all other circumstances, hyper-carbia may be well tolerated and there is some evidence thatpermissive hypercapnia may have inherent protectiveeffects.31 35

Protective lung strategies are advisable for mechanicalventilation of the lungs. The difference between the pressureinside and outside the alveolar air space at end-inspiration isthe transpulmonary pressure. Plateau airway pressure,measured during volume-control mechanical ventilationwhen an end-inspiratory pause has been applied, is an indi-cator of the maximal pressure applied inside the alveolar sac.The pressure outside the alveolar sac cannot be measureddirectly but is estimated clinically by assessing changes inpleural pressure. Extra-alveolar or pleural pressure can beabruptly increased by placing the patient in the Trendelen-berg position or by the increased intra-abdominal pressureassociated with inflation of a pneumoperitoneum for laparo-scopic surgery. Pulmonary gas exchange may deteriorate ifpleural pressure is increased and plateau pressure remainsconstant (i.e. a reduction in transpulmonary pressure). Onthe other hand, high transpulmonary pressures are associ-ated with lung injury. In patients with early acute lunginjury, the ventilatory strategy should aim to strike an expe-dient balance between significant reduction in transpulmon-ary airway pressure (e.g. ,20–25 cm H2O, with associatedreduction in alveolar ventilation), and excessive transpul-monary pressures (e.g. .25–30 cm H2O, and the associatedrisk of barotraumas).30 35 36 Recruitment of collapsed alveoliby manually ventilating the patient to a peak airway pressureof 30–40 mm Hg for short periods may reduce shunt andimprove intraoperative oxygenation. Caution is advisable inundertaking this manoeuvre in patients at risk of pneu-mothorax, such as patients with emphysematous bullae orsevere chronic obstructive pulmonary disease. During thesurgical procedure, regular near-patient testing of arterialblood gases, full blood count, coagulation screen, electro-lytes, lactate, and glucose concentration is advisable. Every

Anaesthetic management of patients with severe sepsis BJA

739

effort should be implemented to avoid intraoperativehypothermia as it is associated with impaired platelet andcoagulation factor dysfunction.37

Role of regional anaesthesia and nerve blocks inanaesthesia for septic patients

Peripheral nerve block may be effective at minimizing thesympathetic response to a painful stimulus, while avoidingthe systemic effects of opioid and may be used if an individ-ual assessment of the risk–benefit balance suggests that itmay be justified in their particular circumstances. However,the presence of coagulopathy, local or systemic spread ofinfection, and the fact that local anaesthetics may notwork properly in the presence of infection or acidosis maylimit the application of regional techniques in septic patients.Neuraxial block (spinal and epidural anaesthesia) should beundertaken with caution, since the haemodynamic effectsof these techniques in the setting of sepsis-induced cardio-vascular compromise may be difficult to reverse.38 39

Recent blood tests confirming normal coagulation areessential.

More than 700 000 central neuraxial blocks are conductedannually in the UK. The incidence of permanent injury fromCNB was 4.2 (95% CI 2.9–6.1) per 100 000 and that of para-plegia or death was 1.8 (95% CI 1.0–3.1) per 100 000 cases.Of the 52 cases which were the focus of follow-up for perma-nent injury from CNB, 22 made a complete recovery fromtheir serious complication within the follow-up period.40 48

Therefore, while epidural anaesthesia appears to have avery low risk of permanent neurological sequelae overall,severely septic patients may be at increased risk of this andother serious complications. Although there is no evidencethat placement of an epidural catheter in severely septicpatients increases the risk of epidural abscess or haematomaformation, a substantial proportion of clinical opinion wouldseem to believe that the risks associated with using it inthe context of severe sepsis is not justifiable.

End of surgical procedure

At the conclusion of the surgical procedure, administration offurther neuromuscular blocking agents to facilitate surgicalclosure of the abdomen or thorax may be considered. Therate of blood loss should be minimal before leaving the oper-ating theatre. Supplemental doses of antimicrobial agentsmay be considered. In patients who will require furthersurgery and in all severely ill patients, analgesia, sedation,and mechanical ventilation are maintained at the conclusionof the surgery. Safe transfer of the patient to the ICU isessential. A focused hand-over report is helpful for the ICUcolleagues which highlights the clinical presentation,response to resuscitation measures, antimicrobial agentsused, details of the surgical procedure preformed, blood pro-ducts used intraoperatively, and any specific problems thatshould be anticipated in the postoperative period.

Postoperative management of patientswith severe sepsisIt is important to note that pre-resuscitation measurementsshould be used to calculate the Intensive Care admissionAPACHE score and not those that have improved after resus-citation and the surgical procedure. Ongoing infusions ofvasopressor medication should be adjusted to match thepresent intravascular volume and the new mechanical ven-tilator settings. Having secured the patient’s airway, mech-anical ventilation settings can be decided, with theobjective of minimizing ventilation-induced volutraumaand barotraumas to the lungs. This is most likely to beachieved using low-pressure settings, a high fractionalinspired oxygen concentration (FIO2 ), and suitablyset alarm limits. Low tidal volumes (up to 6 ml kg21 of thepredicted body weight) and permissive hypercapnia maybe considered, provided that arterial pH does not decreasebelow 7.20.36 Pressure-controlled or volume-control modeof mechanical ventilation can be used. When an end-inspiratory pause is included in the respiratory cycle in thevolume-control mode, the achieved transpulmonarypressure (plateau pressure–pleural pressure) should belimited to 25–30 cm H2O to minimize lung parenchymalventilatory damage.41 The use of high PEEP (10–15 cmH2O) may be limited by the degree of associated haemo-dynamic instability. The FIO2 can be decreased (i.e. ,60%)to achieve an SpO2

of 93–95%.35 36 41

It is obviously important that antimicrobial therapy, whichwas started before operation, should be continued in the ICUand the time of the next scheduled dose was noted. Antimi-crobial regimens can be reassessed daily in light of microbio-logical results, and adjusted to ensure efficacy, preventresistance, and to avoid toxicity.

Duration of therapy should be limited to 7–10 days.14 28 Ithas been shown that patients who had a restrictive red bloodcell transfusion strategy (transfusion avoided unless Hb ,7 gdl21) had a significantly lower mortality rate (22% vs 28%)than those who were transfused at higher Hb levels, withthe possible exception of patients with acute myocardialinfarction and unstable angina.42 Fresh-frozen plasma maybe used to correct laboratory clotting abnormalities only ifthere is clinical bleeding or an invasive procedure isplanned.20 Platelets are transfused if counts are ≤5000mm23 regardless of bleeding, or if between 5000 and30 000 mm23 with significant bleeding risk.20 Deep venousthrombosis thromboprophylaxis should usually be consideredwhen concerns about coagulopathy have abated. Recombi-nant human activated protein C (rhAPC) may be consideredin adult patients with sepsis-induced organ dysfunctionwith clinical assessment of high risk of death (typicallyAPACHE score .25 or multiple organ failure) if there are nocontraindications to rhAPC. Adult patients with severesepsis and low risk of death (typically, APACHE II ,20 orone organ failure) should not receive rhAPC.43 44

Continuation of adequate glycaemic control (,8.5 mmollitre21) is important in the control of the septic process. In

BJA Eissa et al.

740

a large, international, randomized trial of ICU patients, therewas no significant difference between strict glycaemiccontrol (blood glucose 4–6 mmol litre21) and more liberalglycaemic control (blood glucose 6–10 mmol litre21) in therate of death or the mean organ failure score. However,the rate of severe hypoglycaemia (glucose level ≤2.2 mmollitre21) was higher in the intensive-therapy group than inthe conventional-therapy group (17% vs 4%, P,0.001), aswas the rate of serious adverse events (11% vs 5%,P¼0.01). Therefore, in severely septic patients, bloodglucose should be maintained in the range 6–10 mmollitre21.7

Nutrition is one of the cornerstones of management incritically ill septic patients. Enteral nutrition via a nasogastrictube is the best choice to maintain enterocyte integrity andnourish the patient. Gastrointestinal protective measures(stress ulcer prophylaxis) and antiemetic drugs are alsoprescribed. Total parenteral nutrition (TPN) should be con-sidered if there is a surgical contraindication to enteralnutrition or if nutritional requirements are not fully met byenteral nutrition alone. Patients may become rapidlyhypoglycaemic if TPN or enteral nutrition is stopped duringthe perioperative period.44

I.V. hydrocortisone may be considered when hypotensionresponds poorly to fluid resuscitation and vasopressors. A 7day trial treatment with low doses of hydrocortisone and flu-drocortisone significantly reduced the risk of death inpatients with septic shock and relative adrenal insufficiencywithout increasing adverse events (P,0.05).30 In this study,there were 81 deaths (70%) in the placebo group and 66deaths (58%) in the corticosteroid group at the end of ICUstay [relative risk (RR) 0.82; 95% CI 0.68–1.00; adjustedodds ratio (OR) 0.50; 95% CI 0.28–0.89; P¼0.02]. Althoughthis study was conducted in the ICU setting, it seemsprudent to extrapolate the finding to appropriately selectedpatients in the perioperative period.45

Hydrocortisone in a dose of 200 mg per day in four divideddoses or as a continuous infusion in a dose of 240 mg per day(10 mg h21) for 7 days is recommended for septic shock inthe ICU setting.10 45 Whether administration of low-dosesteroids during intraoperative management of the septicpatient would improve haemodynamic stability or outcomeis unknown and seems unlikely. The role of glucocorticoidsin the management of patients with severe sepsis requiresfurther investigation. In a multicentre, randomized, blinded,controlled trial of patients with septic shock who weretreated with corticosteroids, there was significantlydecreased mortality in patients who received vasopressincompared with norepinephrine (36% vs 45%, respectively,P¼0.03). In contrast, in septic patients who did not receivecorticosteroids, vasopressin use was associated withincreased mortality compared with norepinephrine (34% vs21%, respectively, P¼0.05).10 There appears to be a benefitto the use of low-dose glucocorticoids (e.g. hydrocortisone50 mg, four times daily, where normovolaemic septicpatients seem refractory to vasopressor therapy to maintainmajor organ perfusion and haemodynamic stability).

Acute renal failure occurs in 23% of patients with severesepsis. Renal replacement therapy may be initiated tocorrect acidosis, hyperkalaemia, or fluid overload and maybe continued until acute tubular necrosis has recovered.Sodium bicarbonate is not recommended for correctingacidosis unless pH ,7.1. Continuous veno-venous haemodia-filtration does not confer any survival benefit when com-pared with intermittent haemodialysis, the observedmortality being 67% for intermittent haemodialysis vs 65%for continuous haemodiafiltration, with an RR of 1.03 (95%CI 0.94–1.14), P¼0.54.46 However, continuous renal replace-ment may be more practical in hemodynamic unstablepatients. A study comparing daily with alternate-day haemo-dialysis found that daily haemodialysis resulted in bettercontrol of uraemia, fewer hypotensive episodes during hae-modialysis, and more rapid resolution of acute renal failure[mean (SD), 9 (2) vs 16 (6) days; P¼0.001] than did conven-tional intermittent haemodialysis on alternate days.47

Although the weight of current evidence suggests thathigher doses of renal replacement may be associated withimproved outcomes, these results may not apply specificallyto patients with severe sepsis.

Analgesia and sedative medication is continued by infu-sion, but excessive use of sedation or neuromuscular block-ing agents is not recommended. Finally where applicable, itis wise to raise the subject of advanced care planning withthe patient and his family, and realistic expectations andoutcomes targeted.

In conclusion, severe sepsis is a major healthcare issue,with a persistently high mortality. Patients with severesepsis syndrome often require surgery for source of infectioncontrol. The anaesthetist has a crucially important role incoordinating and delivering resuscitation and therapeuticstrategies to optimize patient survival outcome. Early i.v.administration of effective antimicrobial therapy is essential.Preoperative resuscitation, aimed at optimizing major organperfusion, is based on judicious use of fluids, vasopressors,and inotropes. Intraoperative management requires carefulinduction of anaesthesia, using lowest effective doses of arange of agents. Maintenance of anaesthesia is challenging,requiring achievement of optimal volume status, avoidanceof lung injury during mechanical ventilation, and ongoingmonitoring of arterial blood gases, haematological andrenal indices, and electrolyte levels. Postoperative care over-laps with ongoing management of the severe sepsis syn-drome patient in the ICU. The care of critically ill septicpatients requiring anaesthesia and surgery will be furtherenhanced by testing promising therapeutic strategies, e.g.use of levosimendan for intraoperative inotropic support, inwell-designed clinical trials.

Conflict of interestD.J.B. is a member of the Editorial Board of BJA.

FundingD.J.B.’s time was supported by The Sisk Foundation.

Anaesthetic management of patients with severe sepsis BJA

741

References1 Dombrovskiy VY, Martin AA, Sunderram J, Paz HL. Rapid increase

in hospitalization and mortality rates for severe sepsis in theUnited States: a trend analysis from 1993 to 2003. Crit CareMed 2007; 35: 1414–5

2 Harrison DA, Welch CA, Eddleston JM. The epidemiology of severesepsis in England, Wales and Northern Ireland, 1996 to 2004: sec-ondary analysis of a high quality clinical database, the ICNARCCase Mix Programme Database. Crit Care 2006; 10: R42

3 Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organfailure and guidelines for the use of innovative therapies in sepsis.The ACCP/SCCM Consensus Conference Committee. AmericanCollege of Chest Physicians/Society of Critical Care Medicine.Chest 1992; 101: 1644–55

4 Levy MM, Fink MP, Marshall JC, et al. 2001 SCCM/ESICM/ACCP/ATS/SIS International Sepsis Definitions Conference. SCCM/ESICM/ACCP/ATS/SIS. Crit Care Med 2003; 31: 1250–6

5 Martin GS, Mannino DM, Eaton S, Moss M. The epidemiology ofsepsis in the United States from 1979 through 2000. N Engl JMed 2003; 348: 1546–54

6 Brun-Buisson C, Doyon F, Carlet J, et al. Incidence, risk factors,and outcome of severe sepsis and septic shock in adults. A multi-center prospective study in intensive care units. French ICU Groupfor Severe Sepsis. J Am Med Assoc 1995; 274: 968–74

7 Brunkhorst FM, Engel C, Bloos F, et al. German CompetenceNetwork Sepsis (SepNet). Intensive insulin therapy and penta-starch resuscitation in severe sepsis. N Engl J Med 2008; 358:125–39

8 Levi M, van der Poll T. Inflammation and coagulation. Crit CareMed 2010; 38(2 Suppl): S26–34

9 Kern JW, Shoemaker WC. Meta-analysis of hemodynamicoptimization in high-risk patients. Crit Care Med 2002; 30:1686–92

10 Russell JA, Walley KR, Gordon AC, et al. Dieter Ayers for the Vaso-pressin and Septic Shock Trial Investigators. Interaction of vaso-pressin infusion, corticosteroid treatment, and mortality ofseptic shock. Crit Care Med 2009; 37: 811–8

11 Dellinger RP, Levy MM, Carlet JM, et al. International SurvivingSepsis Campaign Guidelines Committee; American Associationof Critical-Care Nurses; American College of Chest Physicians;American College of Emergency Physicians; Canadian CriticalCare Society; European Society of Clinical Microbiology and Infec-tious Diseases; European Society of Intensive Care Medicine;European Respiratory Society; International Sepsis Forum; Japa-nese Association for Acute Medicine; Japanese Society of Inten-sive Care Medicine; Society of Critical Care Medicine; Society ofHospital Medicine; Surgical Infection Society; World Federationof Societies of Intensive and Critical Care Medicine. SurvivingSepsis Campaign: international guidelines for management ofsevere sepsis and septic shock: 2008. Crit Care Med 2008; 36:296–327. Erratum in: Crit Care Med 2008; 36: 1394–6

12 Ferrer R, Artigas A, Levy MM, et al. Edusepsis Study Group.Improvement in process of care and outcome after a multicentersevere sepsis educational program in Spain. J Am Med Assoc2008; 299: 2294–303

13 Levy MM, Dellinger RP, Townsend SR, et al. Surviving Sepsis Cam-paign. The Surviving Sepsis Campaign: results of an internationalguideline-based performance improvement program targetingsevere sepsis. Crit Care Med 2010; 38: 683–4

14 Handelsman J, Maki DG. Does combination antimicrobial therapyreduce mortality in Gram-negative bacteraemia? A meta-analy-sis. Lancet Infect Dis 2004; 4: 519–27

15 Finfer S, Bellomo R, Boyce N, et al. The SAFE study: a comparisonof albumin and saline for fluid resuscitation in the intensive careunit. N Engl J Med 2004; 350: 2247–56

16 Gattinoni L, Brazzi L, Pelosi P, et al. A trial of goal-oriented hemo-dynamic therapy in critically ill patients. N Engl J Med 1995; 333:1025–32

17 Rivers E, Nguyen B, Havstad S, et al. Early Goal-Directed TherapyCollaborative Group. Early goal-directed therapy in the treatment ofsevere sepsis and septic shock. N Engl J Med 2001; 345: 1368–77

18 LeDoux D, Astiz ME, Carpati CM, et al. Effects of perfusion pressureon tissue perfusion in septic shock. Crit Care Med 2000; 28:2729–32

19 Djillali A, Vigno P, Renault A, et al. for the CATS Study Group.Norepinephrine plus dobutamine versus epinephrine alone formanagement of septic shock: a randomized trial. Lancet 2007;370: 676–84

20 American Society of Anaesthesiologists: Task Force on BloodComponent Therapy: practice guidelines for blood componenttherapy. Anesthesiology 1996; 84: 732–47

21 Fernandes CJ, Akamine N, DeMarco FVC, et al. Red blood celltransfusion does not increase oxygen consumption in criticallyill septic patients. Crit Care 2001; 5: 362–7

22 Hunter JD, Doddi M. Sepsis and the heart. Br J Anaesth 2010; 104:3–11

23 Morelli A, De CS, Teboul JL, et al. Effects of levosimendan on sys-temic and regional haemodynamics in septic myocardialdepression. Intensive Care Med 2005; 31: 638–44

24 Hayes MA, Timmins AC, Yau EHS, et al. Elevation of systemicoxygen delivery in the treatment of critically ill patients. N EnglJ Med 1994; 330: 1717–22

25 Marshall JC, Al Naqbi A. Principles of source control in the man-agement of sepsis. Crit Care Clin 2009; 25: 753–68, viii– ix

26 Natanson C, Shelhamer JH, Parrillo JE. Intubation of the tracheain the critical care setting. J Am Med Assoc 1985; 253: 1160–5

27 Mier J, Leon EL, Castillo A, Robledo F, Blanco R. Early versus latenecrosectomy in severe necrotizing pancreatitis. Am J Surg1997; 173: 71–5

28 Kumar A. Optimizing antimicrobial therapy in sepsis and septicshock. Crit Care Clin 2009; 25: 733–51, viii

29 Kollef MH, Levy NT, Ahrens TS, et al. The use of continuous IVsedation is associated with prolongation of mechanical venti-lation. Chest 1998; 114: 541–8

30 The Acute Respiratory Distress Syndrome Network: ventilationwith lower tidal volumes as compared with traditional tidalvolumes for acute lung injury and the acute respiratory distresssyndrome. N Engl J Med 2000; 342: 1301–8

31 Laffey JG, O’Croinin D, McLoughlin P, Kavanagh BP. Permissivehypercapnia—role in protective lung ventilatory strategies. Inten-sive Care Med 2004; 30: 347–56

32 Allaouchiche B, Duflo F, Tournadre JP, Debon R, Chassard D. Influ-ence of sepsis on sevoflurane minimum alveolar concentration ina porcine model. Br J Anaesth 2001; 86: 832–6

33 Biais M, Nouette-Gaulain K, Cottenceau V, Revel P, Sztark F.Uncalibrated pulse contour-derived stroke volume variationpredicts fluid responsiveness in mechanically ventilated patientsundergoing liver transplantation. Br J Anaesth 2008; 101: 761–8

34 Derichard A, Robin E, Tavernier B, et al. Automated pulse pressureand stroke volume variations from radial artery: evaluationduring major abdominal surgery. Br J Anaesth 2009; 103: 678–84

35 Grasso S, Stripoli T, De Michele M, et al. ARDSnet ventilatory pro-tocol and alveolar hyperinflation: role of positive end-expiratorypressure. Am J Respir Crit Care Med 2007; 176: 761–7

BJA Eissa et al.

742

36 Hager DN, Krishnan JA, Hayden DL, et al. Tidal volume reductionin patients with acute lung injury when plateau pressures are nothigh. Am J Respir Crit Care Med 2005; 172: 1241–5

37 Insler SR, Sessler DI. Perioperative thermoregulation and temp-erature monitoring. Anesthesiol Clin 2006; 24: 823–37

38 Guay J. Benefits of adding epidural analgesia to general anesthe-sia: a meta-analysis. Br J Anaesth 2006; 20: 335–40

39 Schuz-Stubner S, Pottinger JM, Coffin SA, Herwaldt LA, et al.Nosocomial infections and infection control in regional anaesthe-sia. Acta Anaesthesiol Scand 2008; 52: 1144–57

40 Buggy DJ. Central neuraxial block: defining risk more clearly. Br JAnaesth 2009; 102: 151–3

41 Ferguson ND, Frutos-Vivar F, Esteban A, et al. Airway pressures,tidal volumes, and mortality in patients with acute respiratorydistress syndrome. Crit Care Med 2005; 33: 21–30

42 Hebert PC, Wells G, Blajchman MA, et al. A multicenter, random-ized, controlled clinical trial of transfusion requirements in criticalcare. N Engl J Med 1999; 340: 409–17

43 Farimani MR, Bajestani NN. Comparison of early enteral feedingversus parenteral nutrition after resection of esophagealcancer. J Crit Care 2008; 23: 448–9

44 Laterre PF, Levy H, Clermont G, et al. Hospital mortality andresource use in subgroups of the Recombinant Human ActivatedProtein C Worldwide Evaluation in Severe Sepsis (PROWESS) trial.Crit Care Med 2004; 32: 2207–18

45 Annane D, Sebille V, Charpentier C, et al. Effect of treatmentwith low doses of hydrocortisone and fludrocortisone on mor-tality in patients with septic shock. J Am Med Assoc 2002; 288:862–71

46 Schiffl H, Lang SM, Fischer R. Daily hemodialysis and the outcomeof acute renal failure. N Engl J Med 2002; 346: 305–10

47 Tonelli M, Manns B, Feller-Kopman D. Acute renal failure in theintensive care unit: a systematic review of the impact of dialyticmodality on mortality and renal recovery. Am J Kidney Dis 2002;40: 875–85

48 Cook TM, Counsell D, Wildsmith JA. Major complications ofcentral neuraxial block: report on the Third National AuditProject of the Royal College of Anaesthetists. Royal College ofAnaesthetists Third National Audit Project. Br J Anaesth 2009;102: 179–90

Anaesthetic management of patients with severe sepsis BJA

743