- Email: [email protected]
Intensity of peri-operative glycemic control and postoperative outcomes in patients with diabetes: a meta-analysis
diabetes research and clinical practice 102 (2013) 8–15
Contents available at Sciverse ScienceDirect
Diabetes Research and Clinical Practice jou rnal hom ep ag e: w ww.e l s e v i er . c om/ loca te / d i ab r es
Intensity of peri-operative glycemic control and postoperative outcomes in patients with diabetes: a meta-analysis Bharath Sathya a, Rebecca Davis b, Tracey Taveira c, Hilary Whitlatch d, Wen-Chih Wu d,* a
Department of Medicine, Brown University and Hospitalist, Rhode Island Hospital, United States College of Pharmacy, University of Rhode Island, United States c College of Pharmacy, University of Rhode Island and Clinical Pharmacist, Pharmacy Service, Providence Veterans Affairs Medical Center, United States d Providence Veterans Affairs Medical Center and Department of Medicine, Brown University, United States b
article info
abstract
Article history:
Aims: Peri-operative hyperglycemia is a risk factor for postoperative morbidity and mortal-
Received 25 February 2013
ity. However, the role of specific glycemic targets in reducing this risk has not been defined,
Received in revised form
particularly among patients with diabetes. Thus, our objective was to conduct a meta-
19 March 2013
analysis relating distinct peri-operative glycemic targets and postoperative outcomes in
Accepted 13 May 2013
patients with diabetes.
Available online 6 June 2013
Methods: A systematic review was performed by two authors utilizing pre-specified terms: ‘‘diabetes mellitus’’ and ‘‘perioperative’’ and ‘‘mortality’’ and ‘‘blood glucose’’ or ‘‘strict
Keywords:
glucose control’’ or ‘‘intensive insulin therapy’’ in PUBMED, CENTRAL and EMBASE. Gly-
Metanalysis
cemic control was considered strict when perioperative targets ranged between 100 and
Diabetes mellitus
150 mg/dL (5.6–8.3 mmol/l), moderate when the targets ranged between 150 and 200 mg/dL
Perioperative
8.3–11.1 mmol/l), and liberal when the target was >200 mg/dL (11.1 mmol/l). The data were
Mortality and blood glucose
combined utilizing the Dersimoan–Laird random-effects method. The primary endpoint was postoperative mortality with secondary endpoints of postoperative atrial fibrillation, wound infection, and stroke. Results: The literature search yielded 760 studies, of which only 6 met inclusion criteria. When compared with a liberal target, pooled data showed that a moderate glycemic target was associated with reduced postoperative mortality (OR = 0.48, 95% CI 0.24–0.76) and stroke (OR = 0.61, 95% CI 0.38–0.98), but no differences in atrial fibrillation or wound infection were found. There were no significant differences in postoperative outcomes between moderate versus strict perioperative glycemic target. Conclusions: Pooled results suggest that in patients with diabetes, a moderate peri-operative glycemic target (150–200 mg/dl [5.6–8.3 mmol/l]) is associated with reduction in postoperative mortality and stroke compared with a liberal target (>200 mg/dl [11.1 mmol/l]), whereas no significant additional benefit was found with more strict glycemic control (<150 mg/dl [5.6 mmol/l]). Published by Elsevier Ireland Ltd.
* Corresponding author at: 830 Chalkstone Avenue, Providence, RI 02908, United States. Tel.: +1 401 273 7100x6237; fax: +1 401 457 3311. E-mail address: [email protected] (W.-C. Wu). Abbreviations: SC, Subcutaneous; CII, Continuous insulin infusion; GIK, Glucose-insulin-potassium infusion. 0168-8227/$ – see front matter . Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.diabres.2013.05.003
diabetes research and clinical practice 102 (2013) 8–15
1.
Introduction
Peri-operative hyperglycemia has been associated with increased ventilator dependence, atrial fibrillation, wound infection and mortality [1,2]. Despite its clinical significance, the optimal perioperative glycemic targets for patients with diabetes are still uncertain. Most trials that informed the current American Diabetes Association guidelines for inpatient glycemic targets are based largely on critically ill patients who might or might not have undergone surgery [3–5]. Moreover, these trials included patients with and without diabetes [3,6]. Given the potential differences in mortality risk and hyperglycemia treatment strategies [6], it would be hard to assume that peri-operative glycemic management will have the same effect in the population with and without diabetes prior to a rigorous investigation [7,8]. In fact, studies have found insulin therapy to provide greater mortality reduction in patients without diabetes [7,8], while Szekely et al. noted that deleterious effects from hyperglycemia were not observed in patients with diabetes unless the blood glucose (BG) was >300 mg/dL [8]. Moreover, individual studies that have addressed the effects of hyperglycemia treatment in surgical patients with diabetes have small enrollments and have yielded conflicting results [9– 14]. Given the lack of well-powered trials in patients with diabetes undergoing surgery that support current guideline recommendations and the conflicting results of individual studies, we conducted the following meta-analysis. Our objective was to analyze the current available evidence and relate distinct strategies of perioperative glycemic control and postoperative outcomes in patients with diabetes undergoing surgery.
2.
Methods
2.1.
Search strategy
Fig. 1 – Diagram describing the search algorithm up to February 1st of 2012, inclusion and exclusion criteria.
and the presence of at least two distinct glycemic targets for comparison within the same evaluation period (historic controls excluded [2,16–18]) with associated postoperative outcomes. The authors independently extracted the following information from the trials: the study design, sample size, baseline patient characteristics, mean glycemic levels achieved in each group, type of surgery performed, glycemic targets, timing of intervention (intraoperative or postoperative), mortality data, rates of atrial fibrillation, stroke and infection. Disagreements over study selection were resolved by discussion with the senior author.
2.3. A systematic search was conducted (Fig. 1) utilizing the PUBMED, CENTRAL and EMBASE databases for studies examining the effects of intensive insulin therapy on perioperative outcomes for patients with diabetes. The search was not limited by date of publication, but was limited to those in the English language, and pertaining to human subjects, up to December 1st of 2012. The search was constructed using the medical subject heading (MeSH) terms and text words: ‘‘diabetes mellitus’’ and ‘‘perioperative’’ and ‘‘mortality’’ and ‘‘blood glucose’’ or ‘‘strict glucose control’’ or ‘‘intensive insulin therapy’’. A manual search was then conducted from the reference list of pertinent review articles or of works by authors well published in the topic.
Perioperative glycemic targets
Given the lack of evidence to support current guideline recommendations for perioperative glycemic targets in patients with diabetes [4,5,19], we defined the intensity of perioperative glycemic control based on the glycemic targets set by the studies identified in this systematic review [9–14]. Thus, glycemic control were considered strict when the target BG ranged from 90–150 mg/dL, moderate when BG targets ranged from 150–200 mg/dL, and liberal when the BG target was >200 mg/dL. Given the spectrum of reported glycemic targets, studies were classified into strict, moderate, or liberal based on the upper limit of the intended glycemic target.
2.4. 2.2.
9
Study outcomes
Selection criteria
The identified articles and abstracts were then independently reviewed by two authors to ensure reliability of findings. Inclusion criteria for identified studies were: post-operative patients, adult population, outcome data specific for patients with diagnosed or undiagnosed diabetes based on 2011 American Diabetes Association criteria [15], presence of intra-operative or post-operative glycemic measurements,
The primary outcome was postoperative mortality. The secondary outcomes were postoperative incidence of atrial fibrillation, infection, or stroke. Most studies reported shortterm outcomes as 30-day event rates [9,11,12,14], with some studies reporting in-hospital events [10,13] and one study reporting 2-year postoperative mortality [11]. Authors of the reviewed studies were contacted if the stated information was not available in published materials.
10
2.5.
diabetes research and clinical practice 102 (2013) 8–15
Statistical analysis
A random effects model (Dersimonian–Laird) was employed to pool study results due to the assumed heterogeneity introduced by inclusion of both observational and randomizedcontrolled trials [20]. The results of summary effect estimates were presented as odds ratios with 95% confidence intervals, with 2-sided p values at 0.05 or less considered to be significant. Heterogeneity was estimated using I2 statistics (describes the percentage of variation across studies due to heterogeneity rather than chance) and tested using the corresponding Chi-square statistics. Significant heterogeneity was considered to be present when I2 >50% or a p value of < 0.10 in the chi-square statistic. The effect of publication bias and small study effects were assessed using the Begg’s test [21]. To determine whether any single trial substantially influenced our results, a sensitivity analysis was conducted for the primary outcome by recalculating the pooled estimates after systematically eliminating one different trial each time. The statistical analysis was performed using Stata SE 11.0# (Stata Corp., College Station, TX).
3.
Results
3.1.
Search results
The literature search yielded 754 citations (Fig. 1). Among the citations identified, 681 were excluded based on review of their abstracts and 3 studies were added from screening of reference lists of the potentially relevant studies. The remaining 76 studies underwent a full text review with 42 studies being excluded due to absence of a control group, and 15 studies excluded because of lack of outcome data specific for diabetes. Finally, 9 studies were excluded due to absence of postoperative adverse event data sought for in the current study, with remaining 6 studies included in the meta-analysis [9–14].
(200 patients) [13], and three were randomized trials (423 patients) [10–12]. Five out of six studies included exclusively patients with diagnosed diabetes [10–14]. One study included patients with diagnosed diabetes, and patients with undiagnosed diabetes who mostly met current guideline criteria for the diabetes diagnosis [22], such as preoperative glycosylated hemoglobin of at least 8% (definitively met criteria), or evidence of perioperative hyperglycemia, either (1) immediate serum glucose level greater than 126 mg/dL after coronary bypass surgery (likely met criteria given fasting state) or (2) 3day postoperative average serum glucose level greater than 126 mg/dL (uncertain whether criteria was met) [9]. The majority of the studies reported 30-day mortality with exception of those conducted by Kirdemir et al. and Ouattara et al. which measured in-hospital mortality [9–14]. A differential in the timing of intervention was noted among included studies, with four studies starting glycemic control postoperatively [9,11,12,14] and two studies starting glucose control intra-operatively [10,13]. It should be noted that both studies conducted by Lazar et al. reported zero mortality for the primary outcome, but the 2004 study reported 2-year postoperative outcomes, latter of which, was used in the pooled estimate calculations [11,12]. The glycemic targets are described in Table 2, which varied between trials. There were four trials that compared a moderate versus liberal glycemic control strategy and reported on the primary outcome [9,11,13,14] whereas there were only three trials that have outcome data comparing a strict versus moderate glycemic control strategy [9,10,14]. There was one study that also compared strict versus moderate glycemic control strategy but had no postoperative mortality in the enrolled subjects [12]. For retrospective studies, the grouping of glycemic control strategies were determined by the authors’ study design and chart abstraction strategy, which utilized mean post-operative blood glucose values in one instance [9] and the maximum blood glucose measured in the PACU in another instance [14].
3.3. 3.2.
Mortality
Included studies
The studies included were described in Table 1. Of the six included studies, two were retrospective analyses (5710 patients) [9,14], one was a non-randomized prospective study
There was a total of 150 deaths (74 in the liberal control group, 69 in moderate control group, and 7 in the strict control group). Pooled results suggest that a moderate glycemic control strategy was associated with a significant reduction in
Table 1 – Characteristics of included studies. Year of publication
Author name
Study type
Surgery type
Timing of study intervention
Primary endpoint
CABG CABG CABG Colon, spine, or joint surgery CABG
Intra + post operative Intra-operative Intra + post operative Stratification based on post operative glycemia) Stratification based on post operative glycemia Intra + post operative
30 day mortality In hospital mortality In hospital mortality 30 day mortality
2004 2005 2006 2009
Lazar et al. Ouattara et al. Kirdemir et al. Smith et al.
Prospective, randomized Prospective Non-randomized Prospective, randomized Retrospective cohort
2011
Bhamidipati et al.
Retrospective cohort
2011
Lazar et al.
Prospective randomized control trial
CABG, Coronary artery bypass graft surgery.
CABG
30 day mortality 30 day mortality
Comparison group Study name
N
Mean age
% Male
Studies of moderate glycemic control (target 150–200 mg/dL) Lazar et al., 2004 69 63.5 66.7 Ouattara et al., 2005
no data no data available available Studies of strict glycemic control (target 100–150 mg/dL) Kirdemir et al., 2006 100 57 65.0 Lazar et al., 2011 42 65 26.0 Studies of both moderate and strict glycemic control Smith et al., 2009 447 63.7 45.4
Bhamidipati et al., 2011
a
35
1739
63.7
71.9
Moderate or strict glycemic control groups
Type of insulin
Mean blood glucose (mg/dL)
Blood glucose target (mg/dL)
no data available no data available
266.8
<250
208
SC CII
Mean age
% Male
Type of insulin
72
63.7
58.3
GIK
134.3
125–200
<200
165
no data available
no data available
no data available
148
150 – 200
195.5 135
<200 120–180
100 40
57 64.8
59.0 38.0
CII CII
172 103
100–150 90–120
no data available
N/A
200a
739
65.1
47.4
no data available
N/A
no data available
214.6
>180a
2785
64.2
76.3
no data available
Moderate: 152.4, strict: 119.0
Moderate: 140–199 strict: <140 Moderate: 127–179 strict: 126
Retrospective study, glycemic target chosen by author’s study design and chart abstraction strategy.
N
Mean blood glucose (mg/dL)
Blood glucose target (mg/dL)
diabetes research and clinical practice 102 (2013) 8–15
Table 2 – Characteristics of glycemic control groups.
11
12
diabetes research and clinical practice 102 (2013) 8–15
Fig. 2 – Forest plot of studies comparing moderate (150–200 mg/dL) versus liberal (>200 mg/dL) peri-operative glycemic control strategies and postoperative mortality.
mortality versus a liberal control strategy (Odds ratio = 0.48, 95% CI 0.24–0.76, p = 0.004) (Fig. 2), without significant heterogeneity found in the pooled estimates (I2 = 20.7%, p = 0.29). Sensitivity analysis recalculating the pooled estimates eliminating one study each time showed consistency of the results (Supplemental Figure). In addition, Begg’s test did not reveal signficant publication bias ( p = 0.31).
Supplementary material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.diabres.2013. 05.003. On the other hand, pooled results did not show a significant difference between the effect of moderate versus strict glycemic control strategies on postoperative mortality (Odds ratio = 0.94, 95% CI 0. 40–2.19, p = 0.88) (Fig. 3).
Fig. 3 – Forest plot of studies comparing strict (<150 mg/dL) versus moderate (150–200 mg/dL) peri-operative glycemic control strategies and postoperative mortality.
diabetes research and clinical practice 102 (2013) 8–15
3.4.
Stroke
Pooled estimates from 2 studies that compared moderate versus liberal glycemic strategy showed a significant reduction in the incidence of stroke (Odds ratio = 0.61, 95% CI 0.38– 0.98, p = 0.04) without significant heterogeneity (I2 = 0%, p = 0.68) [9,11]. Meanwhile, pooled results from 3 studies showed no significant difference between the effect of moderate versus strict glycemic control strategies on postoperative stroke (Odds ratio 1.85, 95% CI 0.72–4.74, p = 0.20) [9,10,12].
3.5.
Atrial fibrillation
The relationship between glycemic control and postpoperative incident atrial fibrillation was not significant throughout all glycemic management strategies in the current analysis. Pooled estimates from 2 studies for moderate versus liberal peri-operative glycemic control strategies did not show a significant association with incidence of post-operative atrial fibrillation (Odds ratio 0.54, 95% CI 0.17–1.76, p = 0.31) [9,11]. Similarly, 3 studies that compared strict versus moderate perioperative glycemic control did not demonstrate a significant difference between the effect of moderate versus strict glycemic control strategies on postoperative incident atrial fibrillation (Odds ratio 0.71, 95% CI 0.39–1.30, p = 0.27) [9,10,12].
3.6.
Wound infection
The relationship between glycemic control and wound infection was not significant throughout all glycemic management strategies. The odds ratios were 0.25 (95% CI 0.01– 5.20, p = 0.37) for the moderate versus liberal glycemic control strategies from 2 studies [9,11], and 0.52 (95% CI 0.01–32.1, p = 0.75) for the strict versus moderate glycemic control strategies from 3 studies [9,10,12].
4.
Discussion
In the current meta-analysis, we have demonstrated that when compared to a liberal glycemic control strategy (BG >200 mg/dL), moderate control (BG 150–200 mg/dL), during or immediately after surgery, was associated with a significantly lower risk of mortality and stroke in patients with diabetes. However, we found no significant difference between strict (BG < 150 mg/dL) versus moderate glycemic control with respect to postoperative mortality or stroke. The mortality benefits of moderate glycemic control versus liberal strategy in patients with diabetes found in our study are comparable to those found by the metanalysis of mostly patients without diabetes undergoing cardiac surgery by Haga et al. [23], with an effect size that is similar to ours (48% reduction in mortality versus 52% in our study). Although our conclusions differed in the glycemic targets (<200 versus <180 mg/dL), it is unclear the basis for the target BG of <180 mg/dL reported by the Haga and coauthors when only one out of 7 studies analyzed used 180 mg/dL as the target. The remaining 6 studies had glycemic targets in the treatment arm that ranged from 100 to 200 mg/dL. Therefore, we believe that
13
the BG target of <200 mg/dL may be more suited for the patients with diabetes given our findings, despite the recommended perioperative BG target of <180 mg/dL by the Society of Thoracic Surgeon guidelines for patients with and without diabetes [19]. The decreased risk of perioperative stroke in favor of a moderate perioperative glycemic control as opposed to a liberal strategy helped to shed light on the potential clinical processes associated with a decreased mortality in our study. Although our analyses in postoperative atrial fibrillation and wound infection did not yield statistically significant results, possibly due to small sample size, other metanalyses that combined patients with and without diabetes undergoing surgery have suggested that perioperative glycemic levels <150–200 mg/dL may also be associated with a decreased incidence of postoperative atrial fibrillation and infection [23,24]. We did not find significant outcome differences between a strict (BG < 150 mg/dL) versus moderate perioperative glycemic control. These results are mirrored by a metanalysis by Hua and co-authors in a mixed population of patients with and without diabetes undergoing cardiac surgery [24]. Similar to our study, the authors found no significant mortality benefits between a treatment arm that targeted a BG level of <130 mg/dL and a control arm with BG targets of <200 mg/dL. However, these investigators found a significantly lower infection rate favoring the strict BG control arm. On the other hand, a metanalysis in the critically ill population of patients with and without diabetes [25], showed no difference in mortality or infection outcomes between an intervention arm that achieved mean BG levels of 103 to 124 mg/dL and a control arm with mean BG levels of 139– 171 mg/dL. In addition, the intervention arm had 7 times higher incidence of hypoglycemic events than the control arm. Although it is still controversial, we speculate that any potential benefits of a strict glycemic control may have been offset by the potential deleterious effects of perioperative hypoglycemia, found more commonly in the strict glycemic control arm [12]. It is not surprising then, that the current guidelines support a moderate perioperative glycemic target of <180 mg/dL [4,5,19], although the BG target of <180 mg/dL is highly debatable. The current results have to be interpreted in the context of potential limitations, such as inclusion of studies with distinct design (randomized versus retrospective), glycemic targets and types of surgeries. In addition, the number of available studies that met criteria were small, there were differences in the timing of the intervention (intra- versus postoperative versus both), and the length of postoperative mortality followup was variable (ranged from in-hospital only to 2-years). In spite of these differences, our results did not show significant heterogeneity in the pooled estimates for the primary endpoint, with consistent results in the sensitivity analyses excluding one different trial each time.
5.
Conclusions
This meta-analysis showed that in patients with diabetes, a moderate perioperative glycemic target of 150–200 mg/dL is associated with reduction in postoperative mortality and stroke versus a more liberal target, whereas no additional
14
diabetes research and clinical practice 102 (2013) 8–15
benefits were found with a more strict control of glycemia. Although the current data represent the best available evidence to guide clinical practice, larger randomized-controlled trials should be conducted to confirm and expand on these results. The design of such a study should ideally answer the questions of the timing of the intervention as well as the optimal glycemic targets of the intervention. In that case, effect estimates in the current study could help on the planning and power calculations of such a study.
[6]
[7]
[8]
Author contributions [9]
BS, researched data, wrote manuscript. RD, researched data, contributed analysis data. TT, analyzed data and reviewed/ edited manuscript. HW, contributed to discussion and reviewed/edited manuscript. WW, contributed in researched data, data analysis, discussion and reviewed/edited manuscript.
Conflict of interest
[10]
[11]
The authors declare that they have no conflict of interest.
Financial support Providence VA Medical Center, for Dr. Wu and Dr. Whitlatch’s time and College of Pharmacy – University of Rhode Island, for Drs. Davis and Taveira’s time.
Disclosure The views expressed in this article are those of the authors and do not necessarily reflect those of the Department of Veterans Affairs.
[12]
[13]
[14]
[15] [16]
references
[1] Smiley DD, Umpierrez GE. Perioperative glucose control in the diabetic or nondiabetic patient. South Med J 2006;99:580–9. [2] Furnary AP. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. Cardiopulm Support Physiol 2003;125:1007–21. [3] Grisedale DEG, de Sourza RJ, van Dam RM, Heyland DK, Cook DJ, Malhotra A, et al. Intensive insulin therapy and mortality among critically ill patients: a meta-analysis including NICE–SUGAR study data. CMAJ 2009;180: 821–7. [4] Moghissi ES, Korytkowski MT, DiNardo M, Einhorn D, Hellman R, Hirsch IB, et al. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Endocr Pract 2009;15:353–69. [5] Moghissi ES, Korytkowski MT, DiNardo M, Einhorn D, Hellman R, Hirsch IB, et al. American Association of Clinical Endocrinologists and American Diabetes Association
[17]
[18]
[19]
[20] [21]
consensus statement on inpatient glycemic control. Diabetes Care 2009;32:1119–31. Egi M, Bellomo R, Stachowski E, French CJ, Hart GK, Hegarty C, et al. Blood glucose concentration and outcome of critical illness: the impact of diabetes. Crit Care Med 2008;36(1039):2249–55. 2210.1097/CCM. 2240b2013e31818. Ascione R, Rogers CA, Rajakaruna C, Angelini GD. Inadequate blood glucose control is associated with inhospital mortality and morbidity in diabetic and nondiabetic patients undergoing cardiac surgery. Circulation 2008;113–23. Szekely A. Impact of hyperglycemia on perioperative mortality after coronary artery bypass graft surgery. J Thorac Cardiovasc Surg 2011;142:430–7. Bhamidipati CM, LaPar D, Stukenborg GJ, Morrison CC, Kern JA, Korn IL, et al. Superiority of moderate control of hyperglycemia to tight control in patients undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2011;141:543–51. Kirdemir P, Yildirim V, Kiris I, Gulmen S, Kuralay E, Ibrisim E, et al. Does continuous insulin therapy reduce postoperative supraventricular tachycardia incidence after coronary artery bypass operations in diabetic patients. J Cardiothorac Vasc Anesth 2006;22:383–7. Lazar HL, Chipkin SR, Fitzgerald CA, Bao Y, Cabral H, Apstein CS. Tight glycemic control in diabetic coronary artery bypass graft patients improves perioperative outcomes and decrease recurrent ischemic events. Circulation 2004;109:1497–502. Lazar HL, McDonnell MM, Chipkin S, Fitzgerlad C, Bliss C, Cabral H. Effects of aggressive versus moderate glycemic control on clinical outcomes in diabetic coronary artery bypass graft patients. Ann Surg 2011;254:458–64. Ouattara A, Lecomte P, Le Manach Y, Landi M, Jacqueminet S, Platonov I, et al. Poor intraoperative blood glucose control is associated with a worsened hospital outcome after cardiac surgery in diabetic patients. Anesthesiology 2005;103:687–94. Smith DK, Bowen J, Bucher LHT, Jurkovitz C, Reed III J, et al. A study of perioperative hyperglycemia in patients with diabetes having colon, spine, and joint surgery. J Perianesth Nurs 2009;24:362–9. Standards of medical care in diabetes – 2011. Diabetes Care 2011;34(Suppl 1):S11–61. D’Alessandro C, Leprince P, Golmard JL, Ouattara A, Aubert S, Pavie A, et al. Strict glycemic control reduces EuroSCORE expected mortality in diabetic patients undergoing myocardial revascularization. J Thorac Cardiovasc Surg 2007;134:29–37. Furnary AP, Gao G, Grunkemeier GL, Wu Y, Zerr KJ, Bookin SO, et al. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2003;125:1007–21. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A. Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetic patients after cardiac surgical procedures. Ann Thorac Surg 1999;67:352–60. discussion 360-352. Lazar HL, McDonnell M, Chipkin SR, Furnary AP, Engelman RM, Sadhu AR, et al. The society of thoracic surgeons practice guidelines series: blood glucose management during adult cardiac surgery. Ann Thorac Surg 2009;87: 663–9. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101.
diabetes research and clinical practice 102 (2013) 8–15
[22] American Diabetes Association. Standard of medical care in diabetes – 2012. Diabetes Care 2012;35. [23] Haga KK, McClymont KL, Clarke S, Grounds RS, Ng KY, Glyde DW, et al. The effect of tight glycaemic control, during and after cardiac surgery, on patient mortality and morbidity: a systematic review and meta-analysis. J Cardiothorac Surg 2011;6:3.
15
[24] Hua J, Chen G, Li H, Fu S, Zhang L -M, Scott M, et al. Intensive intraoperative insulin therapy versus conventional insulin therapy during cardiac surgery: a meta-analysis. J Cardiothorac Vasc Anesth 2012. [25] Marik PE, Preiser JC. Toward understanding tight glycemic control in the ICU: a systematic review and metaanalysis. Chest 2010;137:544–51.
Contents available at Sciverse ScienceDirect
Diabetes Research and Clinical Practice jou rnal hom ep ag e: w ww.e l s e v i er . c om/ loca te / d i ab r es
Intensity of peri-operative glycemic control and postoperative outcomes in patients with diabetes: a meta-analysis Bharath Sathya a, Rebecca Davis b, Tracey Taveira c, Hilary Whitlatch d, Wen-Chih Wu d,* a
Department of Medicine, Brown University and Hospitalist, Rhode Island Hospital, United States College of Pharmacy, University of Rhode Island, United States c College of Pharmacy, University of Rhode Island and Clinical Pharmacist, Pharmacy Service, Providence Veterans Affairs Medical Center, United States d Providence Veterans Affairs Medical Center and Department of Medicine, Brown University, United States b
article info
abstract
Article history:
Aims: Peri-operative hyperglycemia is a risk factor for postoperative morbidity and mortal-
Received 25 February 2013
ity. However, the role of specific glycemic targets in reducing this risk has not been defined,
Received in revised form
particularly among patients with diabetes. Thus, our objective was to conduct a meta-
19 March 2013
analysis relating distinct peri-operative glycemic targets and postoperative outcomes in
Accepted 13 May 2013
patients with diabetes.
Available online 6 June 2013
Methods: A systematic review was performed by two authors utilizing pre-specified terms: ‘‘diabetes mellitus’’ and ‘‘perioperative’’ and ‘‘mortality’’ and ‘‘blood glucose’’ or ‘‘strict
Keywords:
glucose control’’ or ‘‘intensive insulin therapy’’ in PUBMED, CENTRAL and EMBASE. Gly-
Metanalysis
cemic control was considered strict when perioperative targets ranged between 100 and
Diabetes mellitus
150 mg/dL (5.6–8.3 mmol/l), moderate when the targets ranged between 150 and 200 mg/dL
Perioperative
8.3–11.1 mmol/l), and liberal when the target was >200 mg/dL (11.1 mmol/l). The data were
Mortality and blood glucose
combined utilizing the Dersimoan–Laird random-effects method. The primary endpoint was postoperative mortality with secondary endpoints of postoperative atrial fibrillation, wound infection, and stroke. Results: The literature search yielded 760 studies, of which only 6 met inclusion criteria. When compared with a liberal target, pooled data showed that a moderate glycemic target was associated with reduced postoperative mortality (OR = 0.48, 95% CI 0.24–0.76) and stroke (OR = 0.61, 95% CI 0.38–0.98), but no differences in atrial fibrillation or wound infection were found. There were no significant differences in postoperative outcomes between moderate versus strict perioperative glycemic target. Conclusions: Pooled results suggest that in patients with diabetes, a moderate peri-operative glycemic target (150–200 mg/dl [5.6–8.3 mmol/l]) is associated with reduction in postoperative mortality and stroke compared with a liberal target (>200 mg/dl [11.1 mmol/l]), whereas no significant additional benefit was found with more strict glycemic control (<150 mg/dl [5.6 mmol/l]). Published by Elsevier Ireland Ltd.
* Corresponding author at: 830 Chalkstone Avenue, Providence, RI 02908, United States. Tel.: +1 401 273 7100x6237; fax: +1 401 457 3311. E-mail address: [email protected] (W.-C. Wu). Abbreviations: SC, Subcutaneous; CII, Continuous insulin infusion; GIK, Glucose-insulin-potassium infusion. 0168-8227/$ – see front matter . Published by Elsevier Ireland Ltd. http://dx.doi.org/10.1016/j.diabres.2013.05.003
diabetes research and clinical practice 102 (2013) 8–15
1.
Introduction
Peri-operative hyperglycemia has been associated with increased ventilator dependence, atrial fibrillation, wound infection and mortality [1,2]. Despite its clinical significance, the optimal perioperative glycemic targets for patients with diabetes are still uncertain. Most trials that informed the current American Diabetes Association guidelines for inpatient glycemic targets are based largely on critically ill patients who might or might not have undergone surgery [3–5]. Moreover, these trials included patients with and without diabetes [3,6]. Given the potential differences in mortality risk and hyperglycemia treatment strategies [6], it would be hard to assume that peri-operative glycemic management will have the same effect in the population with and without diabetes prior to a rigorous investigation [7,8]. In fact, studies have found insulin therapy to provide greater mortality reduction in patients without diabetes [7,8], while Szekely et al. noted that deleterious effects from hyperglycemia were not observed in patients with diabetes unless the blood glucose (BG) was >300 mg/dL [8]. Moreover, individual studies that have addressed the effects of hyperglycemia treatment in surgical patients with diabetes have small enrollments and have yielded conflicting results [9– 14]. Given the lack of well-powered trials in patients with diabetes undergoing surgery that support current guideline recommendations and the conflicting results of individual studies, we conducted the following meta-analysis. Our objective was to analyze the current available evidence and relate distinct strategies of perioperative glycemic control and postoperative outcomes in patients with diabetes undergoing surgery.
2.
Methods
2.1.
Search strategy
Fig. 1 – Diagram describing the search algorithm up to February 1st of 2012, inclusion and exclusion criteria.
and the presence of at least two distinct glycemic targets for comparison within the same evaluation period (historic controls excluded [2,16–18]) with associated postoperative outcomes. The authors independently extracted the following information from the trials: the study design, sample size, baseline patient characteristics, mean glycemic levels achieved in each group, type of surgery performed, glycemic targets, timing of intervention (intraoperative or postoperative), mortality data, rates of atrial fibrillation, stroke and infection. Disagreements over study selection were resolved by discussion with the senior author.
2.3. A systematic search was conducted (Fig. 1) utilizing the PUBMED, CENTRAL and EMBASE databases for studies examining the effects of intensive insulin therapy on perioperative outcomes for patients with diabetes. The search was not limited by date of publication, but was limited to those in the English language, and pertaining to human subjects, up to December 1st of 2012. The search was constructed using the medical subject heading (MeSH) terms and text words: ‘‘diabetes mellitus’’ and ‘‘perioperative’’ and ‘‘mortality’’ and ‘‘blood glucose’’ or ‘‘strict glucose control’’ or ‘‘intensive insulin therapy’’. A manual search was then conducted from the reference list of pertinent review articles or of works by authors well published in the topic.
Perioperative glycemic targets
Given the lack of evidence to support current guideline recommendations for perioperative glycemic targets in patients with diabetes [4,5,19], we defined the intensity of perioperative glycemic control based on the glycemic targets set by the studies identified in this systematic review [9–14]. Thus, glycemic control were considered strict when the target BG ranged from 90–150 mg/dL, moderate when BG targets ranged from 150–200 mg/dL, and liberal when the BG target was >200 mg/dL. Given the spectrum of reported glycemic targets, studies were classified into strict, moderate, or liberal based on the upper limit of the intended glycemic target.
2.4. 2.2.
9
Study outcomes
Selection criteria
The identified articles and abstracts were then independently reviewed by two authors to ensure reliability of findings. Inclusion criteria for identified studies were: post-operative patients, adult population, outcome data specific for patients with diagnosed or undiagnosed diabetes based on 2011 American Diabetes Association criteria [15], presence of intra-operative or post-operative glycemic measurements,
The primary outcome was postoperative mortality. The secondary outcomes were postoperative incidence of atrial fibrillation, infection, or stroke. Most studies reported shortterm outcomes as 30-day event rates [9,11,12,14], with some studies reporting in-hospital events [10,13] and one study reporting 2-year postoperative mortality [11]. Authors of the reviewed studies were contacted if the stated information was not available in published materials.
10
2.5.
diabetes research and clinical practice 102 (2013) 8–15
Statistical analysis
A random effects model (Dersimonian–Laird) was employed to pool study results due to the assumed heterogeneity introduced by inclusion of both observational and randomizedcontrolled trials [20]. The results of summary effect estimates were presented as odds ratios with 95% confidence intervals, with 2-sided p values at 0.05 or less considered to be significant. Heterogeneity was estimated using I2 statistics (describes the percentage of variation across studies due to heterogeneity rather than chance) and tested using the corresponding Chi-square statistics. Significant heterogeneity was considered to be present when I2 >50% or a p value of < 0.10 in the chi-square statistic. The effect of publication bias and small study effects were assessed using the Begg’s test [21]. To determine whether any single trial substantially influenced our results, a sensitivity analysis was conducted for the primary outcome by recalculating the pooled estimates after systematically eliminating one different trial each time. The statistical analysis was performed using Stata SE 11.0# (Stata Corp., College Station, TX).
3.
Results
3.1.
Search results
The literature search yielded 754 citations (Fig. 1). Among the citations identified, 681 were excluded based on review of their abstracts and 3 studies were added from screening of reference lists of the potentially relevant studies. The remaining 76 studies underwent a full text review with 42 studies being excluded due to absence of a control group, and 15 studies excluded because of lack of outcome data specific for diabetes. Finally, 9 studies were excluded due to absence of postoperative adverse event data sought for in the current study, with remaining 6 studies included in the meta-analysis [9–14].
(200 patients) [13], and three were randomized trials (423 patients) [10–12]. Five out of six studies included exclusively patients with diagnosed diabetes [10–14]. One study included patients with diagnosed diabetes, and patients with undiagnosed diabetes who mostly met current guideline criteria for the diabetes diagnosis [22], such as preoperative glycosylated hemoglobin of at least 8% (definitively met criteria), or evidence of perioperative hyperglycemia, either (1) immediate serum glucose level greater than 126 mg/dL after coronary bypass surgery (likely met criteria given fasting state) or (2) 3day postoperative average serum glucose level greater than 126 mg/dL (uncertain whether criteria was met) [9]. The majority of the studies reported 30-day mortality with exception of those conducted by Kirdemir et al. and Ouattara et al. which measured in-hospital mortality [9–14]. A differential in the timing of intervention was noted among included studies, with four studies starting glycemic control postoperatively [9,11,12,14] and two studies starting glucose control intra-operatively [10,13]. It should be noted that both studies conducted by Lazar et al. reported zero mortality for the primary outcome, but the 2004 study reported 2-year postoperative outcomes, latter of which, was used in the pooled estimate calculations [11,12]. The glycemic targets are described in Table 2, which varied between trials. There were four trials that compared a moderate versus liberal glycemic control strategy and reported on the primary outcome [9,11,13,14] whereas there were only three trials that have outcome data comparing a strict versus moderate glycemic control strategy [9,10,14]. There was one study that also compared strict versus moderate glycemic control strategy but had no postoperative mortality in the enrolled subjects [12]. For retrospective studies, the grouping of glycemic control strategies were determined by the authors’ study design and chart abstraction strategy, which utilized mean post-operative blood glucose values in one instance [9] and the maximum blood glucose measured in the PACU in another instance [14].
3.3. 3.2.
Mortality
Included studies
The studies included were described in Table 1. Of the six included studies, two were retrospective analyses (5710 patients) [9,14], one was a non-randomized prospective study
There was a total of 150 deaths (74 in the liberal control group, 69 in moderate control group, and 7 in the strict control group). Pooled results suggest that a moderate glycemic control strategy was associated with a significant reduction in
Table 1 – Characteristics of included studies. Year of publication
Author name
Study type
Surgery type
Timing of study intervention
Primary endpoint
CABG CABG CABG Colon, spine, or joint surgery CABG
Intra + post operative Intra-operative Intra + post operative Stratification based on post operative glycemia) Stratification based on post operative glycemia Intra + post operative
30 day mortality In hospital mortality In hospital mortality 30 day mortality
2004 2005 2006 2009
Lazar et al. Ouattara et al. Kirdemir et al. Smith et al.
Prospective, randomized Prospective Non-randomized Prospective, randomized Retrospective cohort
2011
Bhamidipati et al.
Retrospective cohort
2011
Lazar et al.
Prospective randomized control trial
CABG, Coronary artery bypass graft surgery.
CABG
30 day mortality 30 day mortality
Comparison group Study name
N
Mean age
% Male
Studies of moderate glycemic control (target 150–200 mg/dL) Lazar et al., 2004 69 63.5 66.7 Ouattara et al., 2005
no data no data available available Studies of strict glycemic control (target 100–150 mg/dL) Kirdemir et al., 2006 100 57 65.0 Lazar et al., 2011 42 65 26.0 Studies of both moderate and strict glycemic control Smith et al., 2009 447 63.7 45.4
Bhamidipati et al., 2011
a
35
1739
63.7
71.9
Moderate or strict glycemic control groups
Type of insulin
Mean blood glucose (mg/dL)
Blood glucose target (mg/dL)
no data available no data available
266.8
<250
208
SC CII
Mean age
% Male
Type of insulin
72
63.7
58.3
GIK
134.3
125–200
<200
165
no data available
no data available
no data available
148
150 – 200
195.5 135
<200 120–180
100 40
57 64.8
59.0 38.0
CII CII
172 103
100–150 90–120
no data available
N/A
200a
739
65.1
47.4
no data available
N/A
no data available
214.6
>180a
2785
64.2
76.3
no data available
Moderate: 152.4, strict: 119.0
Moderate: 140–199 strict: <140 Moderate: 127–179 strict: 126
Retrospective study, glycemic target chosen by author’s study design and chart abstraction strategy.
N
Mean blood glucose (mg/dL)
Blood glucose target (mg/dL)
diabetes research and clinical practice 102 (2013) 8–15
Table 2 – Characteristics of glycemic control groups.
11
12
diabetes research and clinical practice 102 (2013) 8–15
Fig. 2 – Forest plot of studies comparing moderate (150–200 mg/dL) versus liberal (>200 mg/dL) peri-operative glycemic control strategies and postoperative mortality.
mortality versus a liberal control strategy (Odds ratio = 0.48, 95% CI 0.24–0.76, p = 0.004) (Fig. 2), without significant heterogeneity found in the pooled estimates (I2 = 20.7%, p = 0.29). Sensitivity analysis recalculating the pooled estimates eliminating one study each time showed consistency of the results (Supplemental Figure). In addition, Begg’s test did not reveal signficant publication bias ( p = 0.31).
Supplementary material related to this article found, in the online version, at http://dx.doi.org/10.1016/j.diabres.2013. 05.003. On the other hand, pooled results did not show a significant difference between the effect of moderate versus strict glycemic control strategies on postoperative mortality (Odds ratio = 0.94, 95% CI 0. 40–2.19, p = 0.88) (Fig. 3).
Fig. 3 – Forest plot of studies comparing strict (<150 mg/dL) versus moderate (150–200 mg/dL) peri-operative glycemic control strategies and postoperative mortality.
diabetes research and clinical practice 102 (2013) 8–15
3.4.
Stroke
Pooled estimates from 2 studies that compared moderate versus liberal glycemic strategy showed a significant reduction in the incidence of stroke (Odds ratio = 0.61, 95% CI 0.38– 0.98, p = 0.04) without significant heterogeneity (I2 = 0%, p = 0.68) [9,11]. Meanwhile, pooled results from 3 studies showed no significant difference between the effect of moderate versus strict glycemic control strategies on postoperative stroke (Odds ratio 1.85, 95% CI 0.72–4.74, p = 0.20) [9,10,12].
3.5.
Atrial fibrillation
The relationship between glycemic control and postpoperative incident atrial fibrillation was not significant throughout all glycemic management strategies in the current analysis. Pooled estimates from 2 studies for moderate versus liberal peri-operative glycemic control strategies did not show a significant association with incidence of post-operative atrial fibrillation (Odds ratio 0.54, 95% CI 0.17–1.76, p = 0.31) [9,11]. Similarly, 3 studies that compared strict versus moderate perioperative glycemic control did not demonstrate a significant difference between the effect of moderate versus strict glycemic control strategies on postoperative incident atrial fibrillation (Odds ratio 0.71, 95% CI 0.39–1.30, p = 0.27) [9,10,12].
3.6.
Wound infection
The relationship between glycemic control and wound infection was not significant throughout all glycemic management strategies. The odds ratios were 0.25 (95% CI 0.01– 5.20, p = 0.37) for the moderate versus liberal glycemic control strategies from 2 studies [9,11], and 0.52 (95% CI 0.01–32.1, p = 0.75) for the strict versus moderate glycemic control strategies from 3 studies [9,10,12].
4.
Discussion
In the current meta-analysis, we have demonstrated that when compared to a liberal glycemic control strategy (BG >200 mg/dL), moderate control (BG 150–200 mg/dL), during or immediately after surgery, was associated with a significantly lower risk of mortality and stroke in patients with diabetes. However, we found no significant difference between strict (BG < 150 mg/dL) versus moderate glycemic control with respect to postoperative mortality or stroke. The mortality benefits of moderate glycemic control versus liberal strategy in patients with diabetes found in our study are comparable to those found by the metanalysis of mostly patients without diabetes undergoing cardiac surgery by Haga et al. [23], with an effect size that is similar to ours (48% reduction in mortality versus 52% in our study). Although our conclusions differed in the glycemic targets (<200 versus <180 mg/dL), it is unclear the basis for the target BG of <180 mg/dL reported by the Haga and coauthors when only one out of 7 studies analyzed used 180 mg/dL as the target. The remaining 6 studies had glycemic targets in the treatment arm that ranged from 100 to 200 mg/dL. Therefore, we believe that
13
the BG target of <200 mg/dL may be more suited for the patients with diabetes given our findings, despite the recommended perioperative BG target of <180 mg/dL by the Society of Thoracic Surgeon guidelines for patients with and without diabetes [19]. The decreased risk of perioperative stroke in favor of a moderate perioperative glycemic control as opposed to a liberal strategy helped to shed light on the potential clinical processes associated with a decreased mortality in our study. Although our analyses in postoperative atrial fibrillation and wound infection did not yield statistically significant results, possibly due to small sample size, other metanalyses that combined patients with and without diabetes undergoing surgery have suggested that perioperative glycemic levels <150–200 mg/dL may also be associated with a decreased incidence of postoperative atrial fibrillation and infection [23,24]. We did not find significant outcome differences between a strict (BG < 150 mg/dL) versus moderate perioperative glycemic control. These results are mirrored by a metanalysis by Hua and co-authors in a mixed population of patients with and without diabetes undergoing cardiac surgery [24]. Similar to our study, the authors found no significant mortality benefits between a treatment arm that targeted a BG level of <130 mg/dL and a control arm with BG targets of <200 mg/dL. However, these investigators found a significantly lower infection rate favoring the strict BG control arm. On the other hand, a metanalysis in the critically ill population of patients with and without diabetes [25], showed no difference in mortality or infection outcomes between an intervention arm that achieved mean BG levels of 103 to 124 mg/dL and a control arm with mean BG levels of 139– 171 mg/dL. In addition, the intervention arm had 7 times higher incidence of hypoglycemic events than the control arm. Although it is still controversial, we speculate that any potential benefits of a strict glycemic control may have been offset by the potential deleterious effects of perioperative hypoglycemia, found more commonly in the strict glycemic control arm [12]. It is not surprising then, that the current guidelines support a moderate perioperative glycemic target of <180 mg/dL [4,5,19], although the BG target of <180 mg/dL is highly debatable. The current results have to be interpreted in the context of potential limitations, such as inclusion of studies with distinct design (randomized versus retrospective), glycemic targets and types of surgeries. In addition, the number of available studies that met criteria were small, there were differences in the timing of the intervention (intra- versus postoperative versus both), and the length of postoperative mortality followup was variable (ranged from in-hospital only to 2-years). In spite of these differences, our results did not show significant heterogeneity in the pooled estimates for the primary endpoint, with consistent results in the sensitivity analyses excluding one different trial each time.
5.
Conclusions
This meta-analysis showed that in patients with diabetes, a moderate perioperative glycemic target of 150–200 mg/dL is associated with reduction in postoperative mortality and stroke versus a more liberal target, whereas no additional
14
diabetes research and clinical practice 102 (2013) 8–15
benefits were found with a more strict control of glycemia. Although the current data represent the best available evidence to guide clinical practice, larger randomized-controlled trials should be conducted to confirm and expand on these results. The design of such a study should ideally answer the questions of the timing of the intervention as well as the optimal glycemic targets of the intervention. In that case, effect estimates in the current study could help on the planning and power calculations of such a study.
[6]
[7]
[8]
Author contributions [9]
BS, researched data, wrote manuscript. RD, researched data, contributed analysis data. TT, analyzed data and reviewed/ edited manuscript. HW, contributed to discussion and reviewed/edited manuscript. WW, contributed in researched data, data analysis, discussion and reviewed/edited manuscript.
Conflict of interest
[10]
[11]
The authors declare that they have no conflict of interest.
Financial support Providence VA Medical Center, for Dr. Wu and Dr. Whitlatch’s time and College of Pharmacy – University of Rhode Island, for Drs. Davis and Taveira’s time.
Disclosure The views expressed in this article are those of the authors and do not necessarily reflect those of the Department of Veterans Affairs.
[12]
[13]
[14]
[15] [16]
references
[1] Smiley DD, Umpierrez GE. Perioperative glucose control in the diabetic or nondiabetic patient. South Med J 2006;99:580–9. [2] Furnary AP. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. Cardiopulm Support Physiol 2003;125:1007–21. [3] Grisedale DEG, de Sourza RJ, van Dam RM, Heyland DK, Cook DJ, Malhotra A, et al. Intensive insulin therapy and mortality among critically ill patients: a meta-analysis including NICE–SUGAR study data. CMAJ 2009;180: 821–7. [4] Moghissi ES, Korytkowski MT, DiNardo M, Einhorn D, Hellman R, Hirsch IB, et al. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Endocr Pract 2009;15:353–69. [5] Moghissi ES, Korytkowski MT, DiNardo M, Einhorn D, Hellman R, Hirsch IB, et al. American Association of Clinical Endocrinologists and American Diabetes Association
[17]
[18]
[19]
[20] [21]
consensus statement on inpatient glycemic control. Diabetes Care 2009;32:1119–31. Egi M, Bellomo R, Stachowski E, French CJ, Hart GK, Hegarty C, et al. Blood glucose concentration and outcome of critical illness: the impact of diabetes. Crit Care Med 2008;36(1039):2249–55. 2210.1097/CCM. 2240b2013e31818. Ascione R, Rogers CA, Rajakaruna C, Angelini GD. Inadequate blood glucose control is associated with inhospital mortality and morbidity in diabetic and nondiabetic patients undergoing cardiac surgery. Circulation 2008;113–23. Szekely A. Impact of hyperglycemia on perioperative mortality after coronary artery bypass graft surgery. J Thorac Cardiovasc Surg 2011;142:430–7. Bhamidipati CM, LaPar D, Stukenborg GJ, Morrison CC, Kern JA, Korn IL, et al. Superiority of moderate control of hyperglycemia to tight control in patients undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2011;141:543–51. Kirdemir P, Yildirim V, Kiris I, Gulmen S, Kuralay E, Ibrisim E, et al. Does continuous insulin therapy reduce postoperative supraventricular tachycardia incidence after coronary artery bypass operations in diabetic patients. J Cardiothorac Vasc Anesth 2006;22:383–7. Lazar HL, Chipkin SR, Fitzgerald CA, Bao Y, Cabral H, Apstein CS. Tight glycemic control in diabetic coronary artery bypass graft patients improves perioperative outcomes and decrease recurrent ischemic events. Circulation 2004;109:1497–502. Lazar HL, McDonnell MM, Chipkin S, Fitzgerlad C, Bliss C, Cabral H. Effects of aggressive versus moderate glycemic control on clinical outcomes in diabetic coronary artery bypass graft patients. Ann Surg 2011;254:458–64. Ouattara A, Lecomte P, Le Manach Y, Landi M, Jacqueminet S, Platonov I, et al. Poor intraoperative blood glucose control is associated with a worsened hospital outcome after cardiac surgery in diabetic patients. Anesthesiology 2005;103:687–94. Smith DK, Bowen J, Bucher LHT, Jurkovitz C, Reed III J, et al. A study of perioperative hyperglycemia in patients with diabetes having colon, spine, and joint surgery. J Perianesth Nurs 2009;24:362–9. Standards of medical care in diabetes – 2011. Diabetes Care 2011;34(Suppl 1):S11–61. D’Alessandro C, Leprince P, Golmard JL, Ouattara A, Aubert S, Pavie A, et al. Strict glycemic control reduces EuroSCORE expected mortality in diabetic patients undergoing myocardial revascularization. J Thorac Cardiovasc Surg 2007;134:29–37. Furnary AP, Gao G, Grunkemeier GL, Wu Y, Zerr KJ, Bookin SO, et al. Continuous insulin infusion reduces mortality in patients with diabetes undergoing coronary artery bypass grafting. J Thorac Cardiovasc Surg 2003;125:1007–21. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A. Continuous intravenous insulin infusion reduces the incidence of deep sternal wound infection in diabetic patients after cardiac surgical procedures. Ann Thorac Surg 1999;67:352–60. discussion 360-352. Lazar HL, McDonnell M, Chipkin SR, Furnary AP, Engelman RM, Sadhu AR, et al. The society of thoracic surgeons practice guidelines series: blood glucose management during adult cardiac surgery. Ann Thorac Surg 2009;87: 663–9. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177–88. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088–101.
diabetes research and clinical practice 102 (2013) 8–15
[22] American Diabetes Association. Standard of medical care in diabetes – 2012. Diabetes Care 2012;35. [23] Haga KK, McClymont KL, Clarke S, Grounds RS, Ng KY, Glyde DW, et al. The effect of tight glycaemic control, during and after cardiac surgery, on patient mortality and morbidity: a systematic review and meta-analysis. J Cardiothorac Surg 2011;6:3.
15
[24] Hua J, Chen G, Li H, Fu S, Zhang L -M, Scott M, et al. Intensive intraoperative insulin therapy versus conventional insulin therapy during cardiac surgery: a meta-analysis. J Cardiothorac Vasc Anesth 2012. [25] Marik PE, Preiser JC. Toward understanding tight glycemic control in the ICU: a systematic review and metaanalysis. Chest 2010;137:544–51.