Intervention Review

You have free access to this content

Peri-operative glycaemic control regimens for preventing surgical site infections in adults

  1. Lillian S Kao1,*,
  2. Derek Meeks1,
  3. Virginia A Moyer2,
  4. Kevin P Lally3

Editorial Group: Cochrane Wounds Group

Published Online: 8 JUL 2009

Assessed as up-to-date: 24 MAR 2009

DOI: 10.1002/14651858.CD006806.pub2


How to Cite

Kao LS, Meeks D, Moyer VA, Lally KP. Peri-operative glycaemic control regimens for preventing surgical site infections in adults. Cochrane Database of Systematic Reviews 2009, Issue 3. Art. No.: CD006806. DOI: 10.1002/14651858.CD006806.pub2.

Author Information

  1. 1

    University of Texas Health Science Center at Houston, Department of Surgery, Houston, Texas, USA

  2. 2

    Baylor College of Medicine and Texas Children's Hospital, Academic General Pediatrics, Houston, Texas, USA

  3. 3

    University of Texas at Houston, Houston, Texas, USA

*Lillian S Kao, Department of Surgery, University of Texas Health Science Center at Houston, 5656 Kelley Street, Suite 30S 62008, Houston, Texas, TX 77026, USA. Lillian.S.Kao@uth.tmc.edu.

Publication History

  1. Publication Status: New
  2. Published Online: 8 JUL 2009

SEARCH

 
Characteristics of included studies [ordered by study ID]
Bilotta 2007

MethodsRandomised controlled trial


ParticipantsPatients with subarachnoid haemorrhage undergoing emergency cerebral aneurysm clipping.
Inclusion criteria: Prehospital Rankin score 1 to 7 and not endotracheally intubated


InterventionsIntra- and post-operatively until discharge from the ICU or upon postoperative day 14.
Conventional: insulin infusion, target range 80 to 220 mg/dL
Strict: insulin infusion, target range 80 to 120 mg/dL.
Insulin - Actrapid HM, NovoNordisk, Copenhagen, Denmark.
No intravenous dextrose was administered in the first 24 hours. Subsequently, a parenteral/ enteral nutrition per protocol was administered.


OutcomesPrimary outcome: infection (pneumonia, sepsis, urinary, wound infections) as defined by NNIS.
Secondary outcomes: Vasospasm, neurologic outcome, 6-month mortality.


Notes


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?YesComputer-generated random sequence, stratified by Hunt-Hess score (classifies patients with subarachnoid haemorrhage).

Allocation concealment?YesSealed envelopes, not reported whether or not they were sequentially numbered or opaque.
Randomisation and allocation concealment appeared adequate. Important baseline characteristics such as body mass index and diabetes were similar between the groups.

Blinding?
All outcomes
YesOutcome assessment of infection and neurologic outcome were assessed by blinded personnel. Blinding of caregivers was not specified, but is unlikely given different insulin algorithms. The outcomes are unlikely to have been influenced by lack of blinding of the caregivers, although no standardized protocol for initiating a workup for infection was detailed.

Incomplete outcome data addressed?
All outcomes
YesFollow-up was 100%. No missing outcome data reported.

Gandhi 2007

MethodsRandomised controlled trial


ParticipantsAdults with and without diabetes undergoing elective cardiac surgery who had intra-operative hyperglycaemia defined as glucose levels ≥ 100 mg/dL.
Patients undergoing off-pump cardiopulmonary bypass procedures were excluded.
Mean age of patients was 63 years. 69% male. ˜20% of patients had diabetes, among whom the mean Hemoglobin A1c was 7 (SD 2).


InterventionsIntraoperatively followed by strict (intensive) glycaemic control post-operatively for 24 hours then individualized recommendations by the hospital diabetes consulting service.
Conventional: regular insulin IV bolus/infusion, target range < 200 mg/dL.
Strict: regular insulin infusion, target range 80 to 100 mg/dL.


OutcomesPrimary outcome: Composite of death, sternal wound infections, prolonged pulmonary ventilation, cardiac arrhythmias, stroke, and acute renal failure within 30 days after surgery.
Secondary outcomes: length of stay in the ICU and hospital.


Notes


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?YesComputer generated, permuted blocks of 4, stratified by surgeon, surgical procedure, and diabetes.

Allocation concealment?YesSequential, opaque, sealed, tamper-proof envelopes.
Randomisation and allocation concealment appear adequate - baseline characteristics appear similar between groups.

Blinding?
All outcomes
YesOutcome assessment was performed by trained study personnel, blinded to treatment assignment, using confirmable objective criteria based on the Society of Thoracic Surgeons database. Nurses not associated with the study performed the insulin titration per assigned protocols.

Incomplete outcome data addressed?
All outcomes
UnclearA per protocol analysis was performed in that patients who did not have intra-operative hyperglycaemia ≥ 100 mg/dL were excluded from analysis. A total of 15 patients were excluded, evenly distributed between the groups. Three patients in each group were excluded due to cancellation of surgery. No intention to treat analysis was performed. Eight patients were lost to follow-up.

Grey 2004

MethodsRandomised controlled trial.


ParticipantsCritically ill surgery patients. Patients predicted to have a brief stay or not expected to survive beyond 48 hours were excluded from the study. Also excluded: patients with active infections, disseminated cancer, or receiving chemotherapy, radiation, or corticosteroids.
Mean age ˜55-56 years. 69% of patients male. 11% diabetic. 25% received corticosteroids and 54% received vasopressors.


InterventionsPost-operatively, throughout their ICU stay.
Conventional: regular insulin infusion, target range 180 to 220.
Strict: regular insulin infusion, target range 80 to 120.
Nutritional support was managed by a critical care nutritional support team using standardized guidelines (not described in detail).


OutcomesSerum glucose values and nosocomial infections (intravascular device infection, pneumonia, urinary tract infection per 1,000 ICU days) as defined by the Centers for Disease Control.


Notes


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?YesCoin toss.

Allocation concealment?UnclearAllocation concealment not described.

Blinding?
All outcomes
UnclearAssessment of outcomes was not stated to have been performed by blinded personnel.

Incomplete outcome data addressed?
All outcomes
YesDid not report on missing data.

Lazar 2004

MethodsRandomised controlled trial.


ParticipantsAdults with diabetes undergoing primary or reoperative coronary artery bypass surgery.
Mean age ˜ 63 years. 62% male.


InterventionsIntra- and post-operatively (for 12 hours), patients were randomised to GIK (glucose-insulin-potassium) or no-GIK groups.
GIK: IV regular insulin bolus/insulin infusion (with glucose and potassium), target range 126 to 200 mg/dL.
No GIK: Subcutaneous regular insulin, target range 80 to 249 mg/dL.


OutcomesMortality, ischemic events, and wound complications.


Notes


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearThe method of sequence generation was not described.

Allocation concealment?UnclearNo method of allocation concealment was described.
Although neither the method of sequence generation or allocation concealment were described, important baseline variables such as age, sex, type of diabetes, or urgency of surgery were comparable between groups per the authors.

Blinding?
All outcomes
UnclearAssessment of outcomes was not stated to have been performed by blinded personnel.

Incomplete outcome data addressed?
All outcomes
No30-day and in hospital data appeared to be complete. Follow-up data (over 5 years) was only available for 60/72 (83%) of the GIK group and 60/69 (87%) of the no-GIK group. No method was discussed for addressing the incomplete data.

Li 2006

MethodsRandomised controlled trial, but method not stated.


ParticipantsDiabetic adults undergoing coronary artery bypass grafting.
Mean age ˜ 63 years. 62% male.
Mean Body Mass Index ˜ 27 kg/m2.


InterventionsPost-operatively, the target glucose ranges were the same in both arms, but the route and duration differed.
Conventional: Subcutaneous insulin (Humulin® R, Eli Lilly and Company, Indianapolis, Indiana) for 5 days, target range 150 to 200 mg/dl.
Strict: intravenous insulin for 2 days, target range 150 to 200 mg/dL.
Post-operative nutritional support not described.


OutcomesThe primary outcomes were operative mortality and sternal wound infections. The secondary outcome was adequacy of glucose control as calculated by the composite average of daily mean glucose levels through the 5th post-operative day.


Notes


Risk of bias

ItemAuthors' judgementDescription

Adequate sequence generation?UnclearThe method of sequence generation was not described.

Allocation concealment?UnclearThe method of allocation concealment was not described.
Although neither the method of sequence generation or allocation concealment were described, important baseline variables such as age, sex, or underlying diseases were comparable between groups per the authors.

Blinding?
All outcomes
NoThe study was not blinded. Seven patients randomised to subcutaneous insulin were dropped from the treatment protocol due to physicians' concerns about their high glucose levels.

Incomplete outcome data addressed?
All outcomes
UnclearNo missing data were reported. An intention to treat analysis was not performed.

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Bilotta 2008No surgical procedures were performed.

De La Rosa 2008The RCT included both medical and surgical patients; randomisation was not stratified by type of patient. Surgical site infections were not measured.

He 2008No surgical patients were included.

Quinn 2006Both groups had intra-operative glycaemic control targeted between 180 to 270 mg/dL. Neither arm had strict glycaemic control.

Wang 2006No surgical patients were included.

 
Characteristics of studies awaiting assessment [ordered by study ID]
Arabi 2008

Methods

Participants

Interventions

Outcomes

Notesstudy being retrieved in full for evaluation.

Barcellos 2007

Methods

Participants

Interventions

Outcomes

Notesstudy being retrieved in full for evaluation.

Bilotta 2009

Methods

Participants

Interventions

Outcomes

Notesstudy being retrieved in full for evaluation.

He 2007

Methods

Participants

Interventions

Outcomes

Notesstudy being retrieved in full for evaluation.

 
Comparison 1. Intra-operative strict vs. conventional glycaemic control with insulin infusion

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Deep wound infection1371Risk Ratio (M-H, Fixed, 95% CI)0.86 [0.30, 2.52]

 2 Percentage of patients with intra-operative hypoglycemia below 60 mg/dL1371Risk Ratio (M-H, Fixed, 95% CI)1.01 [0.06, 15.95]

 3 Percentage of post-operative patients with hypoglycemia below 60 mg/dL1371Risk Ratio (M-H, Fixed, 95% CI)1.76 [0.76, 4.09]

 4 Mortality1371Risk Ratio (M-H, Fixed, 95% CI)9.05 [0.49, 166.88]

 
Comparison 2. intra- and post-operative strict versus conventional glycaemic control with intravenous insulin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Infection (pneumonia, sepsis, urinary, wound infection178Risk Ratio (M-H, Fixed, 95% CI)0.65 [0.35, 1.22]

 2 Wound infection178Risk Ratio (M-H, Fixed, 95% CI)0.48 [0.04, 5.03]

 3 Mortality178Risk Ratio (M-H, Fixed, 95% CI)0.81 [0.30, 2.20]

 
Comparison 3. Intra- and post-operative strict glycaemic control with intravenous glucose-insulin-potassium infusion (GIK) versus conventional glycemic control with subcutaneous insulin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Pneumonia and wound infections1140Risk Ratio (M-H, Fixed, 95% CI)0.05 [0.00, 0.84]

 
Comparison 5. Post-operative glycaemic control with intravenous insulin infusion versus subcutaneous insulin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Sternal wound infection193Risk Ratio (M-H, Fixed, 95% CI)0.82 [0.12, 5.60]

 2 Leg wound infection193Risk Ratio (M-H, Fixed, 95% CI)2.48 [0.10, 59.36]

 3 Mortality193Risk Ratio (M-H, Fixed, 95% CI)1.65 [0.15, 17.54]