REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY

T rgG luD eH , StAn anCh aaH d JDe ogy,He hersOr DiviDa n ofCe of MSu ioloTex MedCa

TpatalfrenoinbidD

ofcarcothefolmaand without diabetes undergoing adult cardiac surgicalprocedures, derived from evidence-based recommen-dations (Table 1).

I.Pe

KeAsDoeff6,2

HianansplevtivcoopassSimwounlevdicdelevels were also reported in a retrospective, observa-tional study of 409 cardiac surgical patients by Gandhi

ThiBloline

ForBloothsecat w

AdSure-m

2Pu

MISCEL

LANEO

USDetrimental Effects of Hyperglycemia in therioperative Period

y Points: Poor Perioperative Glycemic Control issociated With Increased Morbidity and Mortalityenst and coworkers [4] retrospectively reviewed theects of hyperglycemia on the clinical outcomes of80 patients undergoing cardiac surgical procedures.

and coworkers [7]. Intraoperative hyperglycemia wasan independent risk factor for perioperative complica-tions, including death. Abnormal glucose values priorto surgery may also be predictors of decreased survivalafter surgery. Lastly, Anderson and coworkers [8] stud-ied the effect of elevated fasting blood glucose levelsprior to surgery in a group of 1,375 CABG patients.Patients with elevated fasting blood glucose had a1-year mortality that was twice as great as patients withnormal fasting values and equal to that of patients whowere suspected, or known to have diabetes mellitus.Collectively, these studies strongly suggest that in-

creased fasting glucose levels prior to surgery, and per-sistently elevated glucose levels during and immediatelyafter cardiac surgery, are predictive of increased periop-erative morbidity and mortality in patients with andwithout diabetes. The next section will review thosestudies showing that lowering perioperative glucose lev-els with insulin therapy will decrease morbidity andmortality in cardiac surgical patients.

s article was written by members of The Society of Thoracic Surgeonsod Glucose Guideline Task Force whose names appear in the author.

the full text of The Society of Thoracic Surgeons (STS) Guideline onod Glucose Management During Adult Cardiac Surgery, as well aser titles in STS Practice Guideline Series, visit http://www.sts.org/tions/aboutthesociety/practiceguidelines at the official website of STSww.sts.org.

dress correspondence to Dr Lazar, Department of Cardiothoracicgery, Boston Medical Center, 88 East Newton St, Boston, MA 02118;ail: harold.lazar@bmc.org.

009 by The Society of Thoracic Surgeons Ann Thorac Surg 2009;87:6639 0003-4975/09/$36.00blished by Elsevier Inc doi:10.1016/j.athoracsur.2008.11.011he Society of Thoracic Suuideline Series: Blood Guring Adult Cardiac Surg

arold L. Lazar, MD, Marie McDonnell, MDthony P. Furnary, MD, Richard M. Engelmarles R. Bridges, MD, ScD, Constance K. H

einrich Taegtmeyer, MD, DPhil, and Richarpartment of Cardiothoracic Surgery and Division of Endocrinolalth and Health Sciences, The University of Massachusetts, Amegon; The Baystate Medical Center, Springfield, Massachusetts;vid Geffen School of Medicine, Los Angeles, California; Divisionter, Philadelphia, Pennsylvania; University of Florida Collegergery, University Hospital, Linkoping, Sweden; Division of Cardas; and The Division of Cardiothoracic Surgery, Ronald Reganlifornia

he prevalence of diabetes mellitus in patients requir-ing cardiac surgery is rapidly increasing. These

tients have higher perioperative morbidity and mor-ity, significantly reduced long-term survival, and lessedom from recurrent episodes of angina [13]. There isw evidence to suggest that achieving glycemic controlpatients with diabetes decreases perioperative mor-ity and improves short-term and long-term survival.espite the emerging recognition of the importance

glycemic control, there are no specific guidelines fordiac surgeons as to what the optimal level of glu-se should be during the perioperative period, andbest method to achieve these target values. What

lows is an executive summary of guidelines for thenagement of hyperglycemia in both patients witheons Practicecose Managementryuart R. Chipkin, MD,, MD, Archana R. Sadhu, MD,n, MD, MS, Rolf Svedjeholm, MD, PhD,. Shemin, MDthe Boston Medical Center, Boston, The School of Publict, Massachusetts; The Starr-Wood Cardiac Group, Portland,sion of Endocrinology, Ronald Regan Medical Center,Cardiovascular Surgery, University of Pennsylvania Medicaledicine, Jacksonville, Florida; Department of Cardiothoracicgy, The University of Texas School of Medicine, Houston,ical Center, David Geffen School of Medicine, Los Angeles,

gher glucose levels during surgery were found to beindependent predictor of mortality in patients withd without diabetes. Fish and coworkers [5] retro-ectively reviewed the importance of blood glucoseels in the intraoperative and immediate postopera-e period to predict morbidity in 200 consecutiveronary artery bypass graft (CABG) patients. A post-erative serum glucose level ( 250 mg/dL) wasociated with a 10-fold increase in complications.ilar findings were reported by McAlister and co-rkers [6] in a retrospective study of 291 patientsdergoing CABG surgery. The average serum glucoseel on the first postoperative day significantly pre-ted the development of an adverse outcome. Thetrimental effects of elevated intraoperative glucose

II.Cl

KePa

OglyFupaPayetarceisertinke199usbesiothainspa(4.pronodetiothetie200conthesu

incopsigwosiosynF

inLainspageglutwaginsan(ICachatemgmgtigindals13%0.0staKavaconrec0.00.0T

unstu1,5thiunhisranins215aninfmgredsivcarsureqcarwaglyinfurbe

TaRe

Cgb

Cao

Cu

Ce

Cau

Lr

Lr

Lc

664 REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL Ann Thorac SurgBLOOD GLUCOSE MANAGEMENT 2009;87:6639

MISC

ELLANEO

USBeneficial Effects of Glycemic Control oninical Outcomes During Cardiac Surgery

y Points: Glycemic Control ( 180 mg/dL) intients With Diabetes During Cardiac Surgery:

Reduces mortality Reduces morbidity Lowers the incidence of wound infections Reduces hospital length of stay Enhances long-term survival

ne of the earliest studies to examine the effects ofcemic control during cardiac surgery was reported byrnary and coworkers [9]. The study involved 3,554tients undergoing CABG surgery from 1987 to 2001.tients were divided into three groups based on thear of surgery, the method of glycemic control, and thegeted glucose levels. From 1987 to 1991, patients re-ved subcutaneous insulin, given every 4 hours to keepum glucose 200 mg/dL. From 1991 to 1998, a con-uous intravenous (IV) insulin infusion was used toep serum glucose between 150 and 200 mg/dL. From8 to 2001, the Portland protocol was instituted, whiched a continuous insulin drip to keep serum glucosetween 100 and 150 mg/dL. Continuous insulin infu-ns resulted in significantly lower mean glucose levelsn could be obtained with intermittent subcutaneousulin therapy. The perioperative mortality in CABGtients with diabetes was decreased by 50% after 19925% vs 1.9%; p 0.0001) when continuous insulintocols were instituted, and it was similar to that forndiabetic CABG patients. There was also a significantcrease in the incidence of deep sternal wound infec-ns (p 0.001). Furnary and coworkers [10] expandedir original series to include an additional 1,980 pa-nts managed with the Portland protocol from 2001 to5. They introduced a new method to assess glycemictrol called 3-blood glucose, or 3-BG, consisting ofaverage of all glucose values obtained on the day of

ble 1. Classification System Used for Evidence Basedcommendations

lass I: Conditions for which there is evidence for and/oreneral agreement that the procedure or treatment iseneficial, useful, and effectivelass II: Conditions for which there is conflicting evidencend/or a divergence of opinion about the usefulness/efficacyf a procedure or treatmentlass IIA: Weight of evidence/opinion is in favor ofsefulness/efficacylass IIB: Usefulness/efficacy is less well-established byvidence/opinion.lass III: Conditions for which there is evidence or generalgreement that the procedure/treatment is notseful/effective, or both, and in some cases may be harmfulevel of EvidenceA: Data derived from multipleandomized clinical trialsevel of EvidenceB: Data derived from a singleandomized trial or nonrandomized studiesevel of evidenceC: Only consensus opinion of experts,ase studies, or standard-of-carergery and the first and second postoperative days. An anrease in 3-BG was an independent predictor of peri-erative mortality (p 0.001). Mean 3-BG was alsonificantly related to the incidence of deep sternalund infections, hospital length of stay, blood transfu-ns, new onset atrial fibrillation, and low cardiac outputdrome.urther evidence to support the role of insulin therapythe CABG patient with diabetes was presented byzar and coworkers [11] using a modified glucose-ulinpotassium solution. In this trial involving 141tients with diabetes undergoing isolated CABG sur-ry, patients were prospectively randomized to receivecose-insulinpotassium to keep serum glucose be-een 120 and 180 mg/dL, or sliding scale insulin cover-e to maintain glucose 250 mg/dL. The glucose-ulinpotassium was started on induction of anesthesiad continued for 12 hours in the intensive care unitU). The glucose-insulinpotassium-treated patientsieved significantly better glycemic control immedi-ly prior to cardiopulmonary bypass (169 mg/dL vs 209/dL; p 0.0001), and after 12 hours in the ICU (134/dL vs 266 mg/dL; p 0.0001). Patients treated withht glycemic control had significantly higher cardiacices (p 0.0001) and less need for inotropic support (p0.05) and pacing (p 0.05). Tighter glycemic controlo resulted in a lower incidence of infections (0% vs; p 0.01) and atrial fibrillation (15% vs 60%; p

07). This all contributed to a shorter hospital length ofy (6.5 days vs 9.2 days; p 0.0003). After 5 years, theplan-Meier curves showed a significant survival ad-ntage (p 0.04) for patients receiving better glycemictrol. They had a significantly lower incidence ofurrent ischemia (p 0.01) and wound infections (p 3), and were able to maintain a lower angina class (p3).he importance of tight glycemic control in patientsdergoing CABG surgery was also demonstrated in ady by Van den Berghe and coworkers [12] involving48 ventilated patients admitted to a surgical ICU. Ins prospective, randomized study, 62% of patients haddergone cardiac surgery, and only 13% had a priortory of diabetes. During their ICU stay, patients weredomized to a conventional therapeutic group in whichulin was administered only if serum glucose exceededmg/dL to maintain a target goal of 180 to 200 mg/dL,

d an intensive group that received a continuous insulinusion to maintain glucose levels between 80 and 110/dL. Intensive insulin therapy resulted in a significantuction in mortality (10% vs 20%; p 0.005), exclu-ely in those patients who required 5 days of ICUe with multiorgan failure and sepsis. Similarly, cardiacrgical mortality was only reduced in those patientsuiring 3 days of ICU care. Hospital mortality for alldiac surgical patients, irrespective of their ICU stay,s reduced from 5.1% to 2.1% (p 0.05). Intensivecemic control had no effect on morbidity and mortalitythose patients spending 3 days in the ICU. In ather attempt to identify those patients who mightnefit most from tight glycemic control, DAlessandro

d coworkers [13] sought to correlate the effect of tight

glyCAwi200200ICtesinsbahathaesp2.5engroEuimwoeffwoStewipativHrinstostesu

IIIDi

KeIntCaThIsunwo38CAtriintothouoplogweanbebeenorthehaopterpa

GintCAtobetarintincnogrodecardahoforintincboim

IVInRe

Iofallcocieprgluinsaneffvidabdeachavranas[19icaforsafhy

665Ann Thorac Surg REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL2009;87:6639 BLOOD GLUCOSE MANAGEMENT

MISCEL

LANEO

UScemic control with expected EuroScore outcomes inBG patients with diabetes. Three hundred patientsth diabetes undergoing CABG surgery from January3 to June 2004 receiving tight glycemic control (150 tomg/dL in the operating room; 140 mg/dL in the

U) were matched with 300 CABG patients with diabe-treated from March 2001 to September 2002, whenulin protocols were not present, using propensity-sed analyses. The group with tight glycemic controld an observed mortality that was significantly lowern expected (1.3% vs 4.3%; p 0.01). Mortality wasecially lower in the higher risk cohort (EuroScore 4;% vs 8.0%; p 0.03). In contrast, there was no differ-ce between observed and expected mortality in theup without tight glycemic control in patients withroScore 4. Two additional studies have shown theportance of glycemic control in lowering sternalund infections. Zerr and coworkers [14] studied theects of glycemic control on the incidence of sternalund infections in 1,585 CABG patients with diabetes.rnal wound infections increased from 1.3% in patientsth mean glucose values of 100 to 150 mg/dL to 6.7% intients with levels of 250 to 300 mg/dL. In a retrospec-e study involving CABG patients with diabetes,uska and coworkers [15] found that a continuousulin infusion maintaining glucose levels between 120160 mg/dL significantly decreased the incidence ofrnal wound infections compared with intermittentbcutaneous injections.

. Glycemic Control in Patients Withoutabetes During Cardiac Surgery

y Points: Intraoperative Glycemic Control Usingravenous Insulin Infusions is Not Necessary inrdiac Surgery Patients Without Diabetes Providedat Glucose Values Remain 180 mg/dLtight glycemic control necessary for all patientsdergoing cardiac surgery? Butterworth and co-rkers studied the effects of tight glycemic control in1 patients without diabetes undergoing isolatedBG surgery [16]. In this prospective, randomizedal, one group received an insulin infusion whenraoperative glucose levels exceeded 100 mg/dL. Theer group received no insulin coverage. The primarytcome was the incidence of new neurologic, neuro-hthalmologic, or neurobehavioral deficits, or neuro-ic-related deaths. Intraoperative glucose levelsre significantly lower in the patients who receivedinsulin infusion; however, there was no differencetween the incidences of neurologic complicationstween the groups. Furthermore, there was no differ-ce in operative mortality, need for inotropic support,length of hospital stay between the groups, despitefact that patients without intraoperative insulin

d glucose levels 200 mg/dL. In this study, intra-erative glycemic control failed to improve short-m or long-term clinical outcomes in a group of

tients without diabetes. shandhi and coworkers [17] looked at the effects ofensive intraoperative insulin therapy in 400 electiveBG patients. Patients were prospectively randomizeda continuous insulin group to maintain serum glucosetween 80 and 100 mg/dL, or a conventional groupgeted to keep serum glucose 200 mg/dL usingermittent boluses of intravenous (IV) insulin. Theidence of diabetes was 20% in both groups. There wasdifference in the primary outcome between theups, which consisted of the composite incidence ofath, sternal wound infections, prolonged ventilation,diac arrhythmias, strokes, and renal failure within 30ys of surgery. There was also no difference in ICU orspital stay between the groups. There was a tendencymore deaths (p 0.06) and strokes (p 0.02) in theensive insulin group. This study was limited in that itluded patients both with and without diabetes, andth groups received intensive insulin therapy in themediate postoperative period.

. Management of Hyperglycemia Usingsulin Protocols in the Perioperative Periodcommendations: Class I

Glycemic control is best achieved with continuousinsulin infusions rather than intermittent subcu-taneous insulin injections or intermittent IV insu-lin boluses (level of evidence A).

All patients with diabetes undergoing cardiacsurgical procedures should receive an insulin in-fusion in the operating room, and for at least 24hours postoperatively to maintain serum glucoselevels 180 mg/dL (level of evidence B).

ntravenous insulin therapy is the preferred methodinsulin delivery during the perioperative period. Itows for rapid titration, which facilitates glycemicntrol during periods of malabsorption, insulin defi-ncy, and resistance [18]. Table 2 describes variousotocols that are readily available for use and targetcose values that can be achieved. Choosing anulin infusion protocol is dependent on the needsd resources of the institution. To ensure safe andective implementation of any protocol, those indi-uals involved in the patients care must be comfort-le using it. The success of any protocol can betermined by outcomes such as the time needed toieve the target value, specific BG concentrations,

erage BG control, percentage of values in the desiredge, or an area under-the-curve calculation reportedthe percentage of time spent in a determined range]. This issue is addressed specifically on the Amer-n Association of Clinical Endocrinologists websitehospital management of hyperglycemia [20, 21]. Forety tracking, the number of episodes (or percent) ofpoglycemic events and any clinical consequences

ould be monitored.

V.foCl

Cl

Epriconincinf11]wi

suorthewhinsglainsinsNPpriT

Ta -Rat

G cripti

Ma atedfor

Leu of inYa se vPo ose r

e ava

DIG ose rUn lated

dete

Atl ed rafor

Glu ates

Cla ates

DIG ction;

666 REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL Ann Thorac SurgBLOOD GLUCOSE MANAGEMENT 2009;87:6639

MISC

ELLANEO

USPreoperative Management and Assessmentr Patients With Diabetes Recommendations:ass I

Patients taking insulin should hold their nutri-tional insulin (lispro, aspart, glulisine, or regular)after dinner the evening prior to surgery (level ofevidence B).

Scheduled insulin therapy, using a combinationof long-acting and short-acting subcutaneousinsulin, or an insulin infusion protocol, shouldbe initiated to achieve glycemic control forin-hospital patients awaiting surgery (level ofevidence C).

All oral hypoglycemic agents and noninsulin dia-betes medications should be held for 24 hoursprior to surgery (level of evidence C).

The hemaglobin A1c (HbA1c) level should beobtained prior to surgery in patients with diabetesor those patients at risk for postoperative hyper-glycemia to characterize the level of preoperativeglycemic control (level of evidence C).

ass IIA

Prior to surgery, it is reasonable to maintain bloodglucose concentration 180 mg/dL (level of evi-dence B).

fforts should be made to optimize glucose controlor to surgery, because poor preoperative glycemictrol has been associated with increased morbidity,luding a higher incidence of deep sternal woundections and prolonged postoperative length of stay [10,. In general, all oral diabetes medications should be

ble 2. List of Published and Commercially Available Variable

lucose Target Protocol Brief Des

rkovitz Five algorithms with pre-calculinfusion rates are determined

ven General guidelines on titrationle Calculated rates based on glucortland Specified infusion rates by gluc

needed; five target ranges arfloor

AMI Specified infusion rates by gluciversity of Washington Four algorithms with pre-calcu

multiplier; infusion rates arerange

anta Medical Center 10 algorithms with pre-calculatinfusion rates are determined

commander Computer calculated infusion ralgorithms

rian Computer calculated infusion rprogrammed algorithms

AMI diabetes and insulin-glucose infusion in acute myocardial infarthheld within the 24 hours prior to surgery, especially tolfonylureas (eg, glipizide) and glinides (eg, nategliniderepaglinide). These drugs can induce hypoglycemia inabsence of food. Patients who are taking insulin ando are admitted on the day of surgery should betructed to continue their basal insulin dose (eg,rgine, detemir, or NPH) and hold their nutritionalulin (eg, lispro, aspart, glulisine, or regular) unlesstructed otherwise by their primary physician. TheH insulin may be reduced by one half or one thirdor to surgery to avoid hypoglycemia.o achieve rapid control in a hospitalized patient withperglycemia (glucose persistently 180 mg/dL for hours before surgery), insulin therapy either withravenous variable-rate continuous infusion or subcu-eous basal plus rapid-acting insulin should be usedpending on the availability of either therapy. For thetient noted to be hyperglycemic in the preoperativea on the day of surgery, IV insulin therapy is anective way to achieve rapid control. Patients with aown history of diabetes (either type 1 or type 2) can berted immediately on IV therapy in the preoperativea. All preoperative medications should be reviewed totermine the potential for insulin resistance. Theselude steroids, protease inhibitors, and anti-psychoticgs. Finally, patients with renal insufficiency should bentified, because insulin clearance is impaired and thek for hypoglycemia is increased.he hemaglobin A1c (HbA1c), a glycosylated hemoglo-, is an accurate indicator of glycemic control for aonth to 3-month period. The American Diabetes

sociation has reported that adequate glycemic controlassociated with an HbA1c 7% [21]. Obtaining anA1c prior to surgery in diabetic patients or thosetients at risk for postoperative hyperglycemia will help

e Insulin Infusion Protocols

on mg/dL Reference

rates using a multiplier;each glucose range

120199 27

sulin drip 80110 12alue and rate of change 90120 20ange with IV bolus asilable for both ICU and

70110 1080120100150125175150200

ange 126180 28rates using armined for each glucose

80180 29

tes using a multiplier;each glucose range

80110 30

based on programmed 80120 31

based on glucostabilizer 80110 32

ICU intensive care unit; IV intravenous.hy12inttandepaareeffknstaaredeincdruiderisT

bin2-mAsisHbpaoptimize glycemic control in those patients with ele-

vawhup

VICl

PtheMoshsenplewiperooP

maduhasingluinibethehyhavaconthemethe

VIRe

PperecgluwhverenshblorecshfreimanW

ICinsbereqpa

VIan

667Ann Thorac Surg REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL2009;87:6639 BLOOD GLUCOSE MANAGEMENT

MISCEL

LANEO

USted HbA1c levels. It will also identify those patientso might require more aggressive glycemic controlon hospital discharge.

. Intraoperative Control Recommendations:ass I

Glucose levels 180 mg/dL that occur in patientswithout diabetes only during cardiopulmonarybypass may be treated initially with a single orintermittent dose of IV insulin as long as levelsremain 180 mg/dL. However, in those patientswith persistently elevated serum glucose ( 180mg/dL) after cardiopulmonary bypass, a continu-ous insulin drip should be instituted, and anendocrinology consult should be obtained (levelof evidence B).

If an intravenous insulin infusion is initiated inthe preoperative period, it should be continuedthroughout the intraoperative and early postop-erative period according to institutional protocolsto maintain serum glucose 180 mg/dL (level ofevidence C).

atients receiving IV infusions of insulin should haveir blood glucose monitored every 30 to 60 minutes.re frequent monitoring, as often as every 15 minutes,ould be made during periods of rapidly fluctuatingsitivity, such as during the administration of cardio-gia and systemic cooling and rewarming. Patientsth IV insulin infusions initiated in the preoperativeriod should have them continued in the operatingm (OR) to maintain serum glucose 180 mg/dL.atients with no history of diabetes prior to surgery,y exhibit transient elevation of BG 180 mg/dLring cardiopulmonary bypass. These patients mayve insulin resistance and should be treated with agle or intermittent dose of IV insulin to maintaincose 180 mg/dL. Caution should be exercised intiating a continuous IV insulin drip in these patients,cause insulin requirements may decrease rapidly inimmediate postoperative period resulting in serious

poglycemia [22]. However, those patients not known tove diabetes who have persistently elevated glucoselues ( 180 mg/dL) during surgery should receive atinuous IV insulin drip. Because a large percentage ofse patients may ultimately be found to have diabetesllitus, an endocrinology consult should be obtained inpostoperative period.

I. Glycemic Control in the ICUcommendation: Class I

Patients with and without diabetes with persis-tently elevated serum glucose ( 180 mg/dL)should receive IV insulin infusions to maintainserum glucose 180 mg/dL for the duration oftheir ICU care (level of evidence A).

All patients who require 3 days in the ICU

because of ventilatory dependency or requiring theneed for inotropes, intra-aortic balloon pump, or leftventricular assist device support, anti-arrhythmics,dialysis, or continuous veno-venous hemofiltrationshould have a continuous insulin infusion to keepblood glucose 150 mg/dL, regardless of diabeticstatus (level of evidence B).

Before intravenous insulin infusions are discon-tinued, patients should be transitioned to a sub-cutaneous insulin schedule using institutionalprotocols (level of evidence B).

atients with or without diabetes mellitus who haversistently elevated serum glucose 180 mg/dL shouldeive intravenous insulin infusions to maintain serumcose 180 mg/dL [912]. Furthermore, those patientso require 3 days of ICU care due to prolongedntilatory support, inotropic or mechanical support,al insufficiency, or need for anti-arrhythmic therapyould have continuous IV insulin infusions to keepod glucose 150 mg/dL [10, 12]. When patients areeiving IV insulin infusions in the ICU, glucose levelsould be monitored at least hourly until stable. Thisquency avoids fluctuations in glucose levels and min-izes the risk of hypoglycemia, which is fortunately rared has resulted in minimal morbidity [1012].hen patients are ready to be discharged from the

U, patients should be transitioned to a subcutaneousulin-dosing schedule. Daily insulin requirements canestimated by extrapolating the amount of insulinuired in the preceding 24 hours and considering thetients current nutritional intake [23].

II. Glycemic Control in the Stepdown Unitsd on the Floor Recommendations: Class I

A target blood glucose level 180 mg/dL shouldbe achieved in the peak postprandial state (levelof evidence B).

A target blood glucose level 110 mg/dL shouldbe achieved in the fasting and pre-meal statesafter transfer to the floor (level of evidence C).

Oral hypoglycemic medications should be re-started in patients who have achieved targetblood glucose levels if there are no contraindica-tions. Insulin dosages should be reduced accord-ingly (level of evidence C).

According to the American Association of ClinicalEndocrinologists, a reasonable goal for a noncriti-cally ill patient on a regular hospital ward is 110mg/dL pre-meal and 180 mg/dL postprandial orrandomly [24]. The best method to achieve thiscontrol is with scheduled subcutaneous basal and,or bolus insulin therapy, such as glargine ordetermir (basal) and lispro, aspart, or glulisine(bolus). Patients with type 2 diabetes who haveused oral diabetes medications preoperatively canbe restarted on those medications once they havereached their targeted glucose goals and are eat-ing a regular diet. Metformin should not be re-

started until stable renal function has been docu-

IXRe

AshdeWthemithediseddiestatecmetiowirinin

IX

ImdustucontivconimimsuThglymiconpa

Re

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

668 REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL Ann Thorac SurgBLOOD GLUCOSE MANAGEMENT 2009;87:6639

MISC

ELLANEO

USmented. Thiazolidineadiones (eg, pioglitazone,rosiglitazone) can be re-started in patients with-out congestive heart failure or liver dysfunction.

. Preparation for Hospital Dischargecommendations: Class I

Prior to discharge, all patients with diabetes andthose who have started a new glycemic controlregimen, should receive in-patient education re-garding glucose monitoring, medication adminis-tration (including subcutaneous insulin injectionif necessary), nutrition, and lifestyle modification(level of evidence C).

Upon discharge, changes in therapy for glycemiccontrol should be communicated to primary carephysicians, and follow-up appointments shouldbe arranged with an endocrinologist when appro-priate (level of evidence C).

ll patients with hyperglycemia after cardiac surgeryould be assessed by an inpatient diabetes team tocide on a glycemic control program after discharge.hen hyperglycemia is discovered for the first time inperioperative period, or if insulin is first being ad-

nistered, or when a new insulin protocol is instituted,patient should receive specialized education prior tocharge. This can be provided by a certified diabetesucator, and supplemented by nurses or registeredticians with expertise in diabetes. Education should berted at least 2 days prior to discharge, includinghniques of glucose monitoring, administration ofdications, nutrition, exercise, and lifestyle modifica-n [25, 26]. Appropriate follow-up should be arrangedth primary care physicians prior to discharge. Refer-g physicians should be informed of any changes madethe diabetes management plan.

. Future Areas of Study

portant issues in the management of hyperglycemiaring cardiac surgery remain to be elucidated. Futuredies will determine: (1) the optimal level of glycemictrol and which, if any, specific time in the periopera-e period is most crucial for maintaining glycemictrol; (2) whether the level of glucose achieved is asportant as the amount of insulin delivered; and (3) theportance of preoperative HbA1c levels and whetherrgery should be delayed in patients with higher values.ese studies will increase our understanding of hyper-cemia during cardiac surgery and help us to deter-ne the most optimal methods to achieve glycemictrol and improve clinical outcomes in these high-risktients.

ferences

Szabo Z, Hakanson E, Svedjeholm R. Early postoperativeoutcome and medium-term survival in 540 diabetic and 2239

nondiabetic patients undergoing coronary artery bypassgrafting. Ann Thorac Surg 2002;74:7129.Cohen Y, Raz I, Merin G, Mozes B. Comparison of factorsassociated with 30-day mortality after coronary artery by-pass grafting in patients with versus without diabetes mel-litus. Israeli Coronary Artery Bypass (ISCAB) Study Consor-tium. Am J Cardiol 1998;81:711.Carson J, Scholz P, Chen A, Peterson E, Gold J, Schneider S.Diabetes mellitus increases short-term mortality and mor-bidity in patients undergoing coronary artery bypass graftsurgery. J Am Coll Cardiol 2002;40:41823.Doenst T, Wijeysundera D, Karkouti K, et al. Hyperglycemiaduring cardiopulmonary bypass is an independent riskfactor for mortality in patients undergoing cardiac surgery.J Thorac Cardiovasc Surg 2005;130:11449.Fish LH, Weaver TW, Moore AL, and Steel LG. Value ofpostoperative blood glucose in predicting complications andlength of stay after coronary artery bypass grafting. Am JCardiol 2003;92:746.McAlister FA, Man J, Bistritz L, Amad H, Tandon P. Diabetesand coronary artery bypass surgery: an examination ofperioperative glycemic control and outcomes. Diabetes Care2003;26:151824.Gandhi GY, Nuttall GA, Abel MD, et al. Intraoperativehyperglycemia and perioperative outcomes in cardiac sur-gery patients. Mayo Clin Proc 2005;80:8626.Anderson RE, Brismar K, Barr G, Ivert T. Effects of cardio-pulmonary bypass on glucose homeostasis after coronaryartery bypass surgery. Eur J Cardiothorac Surg 2005;28:42530.Furnary AP, Gao G, Grunkemeier GL, et al. Continuousinsulin infusion reduces mortality in patients with diabetesundergoing coronary artery bypass grafting. J Thorac Car-diovasc Surg 2003;125:100721.Furnary AP, Wu Y, Bookin SO. Effect of hyperglycemia andcontinuous intravenous insulin infusions on outcomes ofcardiac surgical procedures: the Portland Diabetic Project.Endocr Pract 2004;10(suppl)2:21-33.Lazar HL, Chipkin SR, Fitzgerald CA, Bao Y, Cabral H,Apstein CS. Tight glycemic control in diabetic coronaryartery bypass graft patients improves perioperative out-comes and decreases recurrent ischemic events. Circulation2004;109:1497502.Van den Berghe G, Wouters P, Weekers F, et al. Intensiveinsulin therapy in the critically ill patients. N Engl J Med2001;345:135967.DAlessandro C, Leprince P, Golmard JL, et al. Strict glyce-mic control reduces EuroSCORE expected mortality in dia-betic patients undergoing myocardial revascularization.J Thorac Cardiovasc Surg 2007;134:2937.Zerr KJ, Furnary AP, Grunkemeier GL, Bookin S, Kanhere V,Starr A. Glucose control lowers the risk of wound infectionin diabetics after open heart operations. Ann Thorac Surg1997;63:35661.Hruska LA, Smith JM, Hendy MP, Fritz VL, McAdams S.Continuous insulin infusion reduces infectious complica-tions in diabetics following coronary surgery. J Card Surg2005;20:4037.Butterworth J, Wagenknecht LE, Legault C, et al. Attemptedcontrol of hyperglycemia during cardiopulmonary bypassfails to improve neurologic or neurobehavioral outcomes inpatients without diabetes mellitus undergoing coronary ar-tery bypass grafting. J Thorac Cardiovasc Surg 2005;130:131923.Gandhi GY, Nuttall GA, Abel MD, et al. Intensive intraop-erative insulin therapy versus conventional glucose manage-ment during cardiac surgery: a randomized trial. Ann InternMed 2007;146:23343.Friedberg SJ, Lam YW, Blum JJ, Gregerman RI. Insulinabsorption: a major factor in apparent insulin resistance andthe control of type 2 diabetes mellitus. Metabolism 2006;55:6149.Braithwaite SS, Godara H, Song HJ, Rock P. No patient left

behind: evaluation and design of intravenous insulin infu-sion algorithms. Endocr Pract 2006;12(suppl 3):728.

20. Goldberg PA, Bozzo JE, Thomas PG, et al. Glucometricsassessing the quality of inpatient glucose management.Diabetes Technol Ther 2006;8:5609.

21. Available at: http://www.aace.com/resources/igcre/.22. Chaney MA, Nikolov MP, Blakeman BP, Bakhos M. At-

tempting to maintain normoglycemia during cardiopulmo-nary bypass with insulin may initiate postoperative hypo-glycemia. Anesth Analg 1999;89:10915.

23. Schmeltz LR, DeSantis AJ, Schmidt K, et al. Conversion ofintravenous insulin infusions to subcutaneously adminis-tered insulin glargine in patients with hyperglycemia. En-docr Pract 2006;12:64150.

24. American Association of Clinical Endocrinologists medicalguidelines for clinical practice for the management of dia-betes mellitus. Endocr Pract 2007;13(suppl 1):168.

25. Ko SH, Song KH, Kim SR, et al. Long-term effects of astructured intensive diabetes education program (SIDEP) inpatients with type 2 diabetes mellitusa 4-year follow-upstudy. Diabet Med 2007;24:5562.

26. Fritsche A, Stumvoll M, Goebbel S, Reinauer KM, Schmull-ing RM, Haring HU. Long term effect of a structured inpa-tient diabetes teaching and treatment program in type 2diabetic patients: influence of mode of follow-up. DiabetesRes Clin Pract 1999;46:13541.

27. Markovitz LJ, Wiechmann RJ, Harris N, et al. Descriptionand evaluation of a glycemic management protocol forpatients with diabetes undergoing heart surgery. EndocrPract 2002;8:108.

28. Malmberg K, Ryden L, Efendic S, et al. Randomized trial ofinsulin-glucose infusion followed by subcutaneous insulintreatment in diabetic patients with acute myocardial infarc-tion (DIGAMI study): effects on mortality at 1 year. J Am CollCardiol 1995;26:5765.

29. Trence DL, Kelly JL, Hirsch IB. The rationale and managementof hyperglycemia for in-patients with cardiovascular disease:time for change. J Clin Endocrinol Metab 2003;88:24307.

30. Osburne RC, Cook CB, Stockton L, et al. Improving hyper-glycemia management in the intensive care unit: prelimi-nary report of a nurse-driven quality improvement projectusing a redesigned insulin infusion algorithm. DiabetesEduc 2006;32:394403.

31. Davidson PC, Steed RD, Bode BW. Glucommander: a com-puter-directed intravenous insulin system shown to be safe,simple, and effective in 120,618 h of operation. Diabetes Care2005;28:241823.

32. Juneja R, Roudebush C, Kumar N, et al. Utilization of acomputerized intravenous insulin infusion program to con-trol blood glucose in the intensive care unit. Diabetes Tech-nol Ther. 2007;9:23240.

669Ann Thorac Surg REPORT FROM STS WORKFORCE ON EVIDENCE BASED SURGERY LAZAR ET AL2009;87:6639 BLOOD GLUCOSE MANAGEMENT

MISCEL

LANEO

US

The Society of Thoracic Surgeons Practice Guideline Series: Blood Glucose Management During Adult Cardiac SurgeryI. Detrimental Effects of Hyperglycemia in thePerioperative PeriodKey Points: Poor Perioperative Glycemic Control is Associated With Increased Morbidity and Mortality

II. Beneficial Effects of Glycemic Control onClinical Outcomes During Cardiac SurgeryKey Points: Glycemic Control (< 180 mg/dL) in Patients With Diabetes During Cardiac Surgery

III. Glycemic Control in Patients WithoutDiabetes During Cardiac SurgeryKey Points: Intraoperative Glycemic Control UsingIntravenous Insulin Infusions is Not Necessary inCardiac Surgery Patients Without Diabetes ProvidedThat Glucose Values Remain 180 mg/dL

IV. Management of Hyperglycemia UsingInsulin Protocols in the Perioperative PeriodRecommendations: Class IV. Preoperative Management and Assessmentfor Patients With Diabetes Recommendations:Class IClass IIAVI. Intraoperative Control Recommendations:Class IVII. Glycemic Control in the ICURecommendation: Class IVIII. Glycemic Control in the Stepdown Unitsand on the Floor Recommendations: Class IIX. Preparation for Hospital DischargeRecommendations: Class IIX. Future Areas of StudyReferences