Diabetes is considered one of the major risks of surgical site infections (SSIs).1 Moreover, in patients undergoing cardiac surgery, diabetes is identified as an independent risk factor associated with SSIs.1,2,3,4,5,6,7,8 Previous studies suggested a link between poor glycaemic control and adverse outcomes in patients with diabetes undergoing coronary artery bypass grafting (CABG) surgery, including a higher rate of SSIs.9,10,11 However, the estimated prevalence of diabetes among patients undergoing cardiac surgery was 20.1–48.0%.12,13,14,15
Several studies have indicated the effect of hyperglycaemia during the preoperative, intraoperative or postoperative phase, with a higher incidence of postoperative infections, longer intensive care unit (ICU) and hospital stays, and higher morbidity and mortality after cardiac surgery.16,17,18,19,20,21
Glycaemic control for patients undergoing cardiac surgery is the main focus to improve clinical outcomes, particularly SSIs. A variety of studies demonstrated useful results with the use of intravenous continuous insulin infusion (CII) as a standard of care to prevent hyperglycaemia without causing hypoglycaemia.22,23
The vast majority of examined literature urged the national medical societies to modify their guidelines to use more conservative glycaemic control. The American Association of Clinical Endocrinologists (AACE) and the American Diabetes Association (ADA) recommended the use of intravenous CII to maintain tight glycaemic control in patients with diabetes. The recommendation is to start the intravenous insulin infusions at the blood glucose threshold ≤180mg/dl to target blood glucose level between 140–180mg/dl (7.8–10.0mmol/l) for the treatment of persistent hyperglycaemia in those patients.24,25 Also, the 2017 ADA practice guidelines showed that more stringent goals of blood glucose levels from 110–140mg/dl (6.1–7.8mmol/l) may be appropriate for selected patients with diabetes, as long as this is achieved without significant risk of hypoglycaemia.24 In 2011, the American College of Physicians (ACP) recommended against intensive insulin control and suggested a target blood glucose level of 140–200mg/dl (7.8–11.1mmol/l) for hospitalised patients in the ICU.26 Specifically, for patients undergoing cardiac surgery, the Society of Thoracic Surgeons (STS) published guidelines regarding glycaemic control that recommended using CII in the operating room to maintain blood glucose levels <180mg/dl.27
In contrast, in a prospective, non-randomised, interventional study of 5510 patients with diabetes undergoing cardiac surgery (Portland Project) that used an intravenous CII protocol to achieve a target blood glucose level <150mg/dl, safer results were revealed with mortality reduced by 60% and deep sternal wound infection (DSWI) by 77%.28 This project outcome conforms to the 2016 recommendations of the World Health Organization (WHO) and the Critical Care Society for SSIs Prevention that suggested using targeted glycaemic control protocols to achieve a blood glucose level of ≤150mg/dl (8.3mmol/l) during the perioperative period, which is linked with better health outcomes, including low SSI prevalence.29,30
Due to the discrepancies in the literature and the lack of a consistent evidence base that determines the optimal range of blood glucose levels to improve clinical outcomes in patients with diabetes undergoing cardiac surgery, the purpose of this systematic review was to identify the best available evidence on the optimal range of tight glycaemic control protocols using CII on reducing the incidence of SSIs in these patients.
Methods
We thoroughly reviewed studies that investigated the different glycaemic control conditions using CII with the incidence of SSIs in patients with diabetes undergoing cardiac surgery. The review focused on the influence of diabetes and hyperglycaemia on SSIs, and current studies related to glycaemic control in the intraoperative phase.
Search strategy
A systematic search was conducted across the literature in three steps:
- An initial limited search of Cumulative Index to Nursing and Allied Health Literature (CINAHL), PubMed and Google Scholar, using the following keywords: ‘cardiac surgery’, ‘Surgical Site Infections’, ‘glucose’ and ‘glycaemic control’. Then, the authors analysed the text words in the title, abstract, and the index terms used, to extract a comprehensive list of keywords
- A thorough and extensive search of databases, including CINAHL, Medline, PubMed, Cochrane Database and Google Scholar
- Only human adult studies were chosen, specifically looking at blood sugar levels controlled intraoperatively and surgical wound infections in patients with diabetes undergoing cardiac surgery.
Moreover, the following additional core-term definitions were considered in the selection strategy of studies:
- The superficial SSIs occurred within 30 days of the surgery; the incision only involved skin and subcutaneous tissue; the patient experienced at least one of the following—a purulent drainage from the incision, positive culture of tissue or fluid, an incision deliberately opened by the surgeon; the patient had at least one of the following signs or symptoms of infection—pain, redness, heat, tenderness or localised swelling; or there was a physician's diagnosis of superficial SSIs31
- The deep SSIs occurred within 30 days of surgery if no implant was left in the incision, or within one year if an implant was left in the incision, and the infection appeared to be related to the surgery, and the incision involved the deep soft tissue (muscle layer and fascia; the patient experienced at least one of the following—a purulent discharge from the incision, an incision was dehisced spontaneously, or was deliberately opened by the surgeon; the patient had at least one of the following signs or symptoms—localised pain, fever (>38°C) or tenderness; an abscess or other evidence of deep infection found by direct examination, histopathologic or radiologic examination; or physician's diagnosis of deep incisional SSIs.31
The search was further restricted to studies published in the English language with a date of publication of 2001 or later. Table 1 displays search terms for each keyword.
Table 1. Terms of search strategy
| Keywords | Cardiac surgery | Infection | Glucose | Glycaemic control |
|---|---|---|---|---|
| Search terms | ‘open heart surgery’ | ‘surgical site infections (SSIs)’ | ‘blood glucose level’ | ‘insulin intervention’ |
| ‘cardiac surgery’ | ‘sternal wound infection (SWI)’ | ‘diabetes mellitus (DM)’ | ‘insulin infusion’ | |
| ‘coronary artery bypass grafting (CABG) surgery’ | ‘superficial sternal wound infection (SSWI)’ | ‘hyperglycaemia’ | ‘insulin protocols’ | |
| ‘cardiac valve surgery’ | ‘deep sternal wound infection (DSWI)’ | ‘glycaemic control’ | ‘insulin drip’ | |
| ‘cardiovascular surgery’ | ‘harvest site infection (HSI)’ | ‘hypoglycaemia’ | ‘glucose-insulin-potassium drip’ |
Study selection
The guideline of Preferred Reporting Items for Meta-Analysis (PRISMA) was used in this review, including the PRISMA checklist and flowchart. As shown in the PRISMA flowchart (Fig 1), the comprehensive systematic search from the five different databases yielded 1763 relevant articles and 14 articles were identified by a manual search of the references of studies due to their relation to the selected keywords and aforementioned selection strategies. We removed 16 duplicate articles and excluded 1718 additional articles after reviewing titles, abstract and keywords. A total of 43 full-text articles remained for further review. After reviewing the full-text articles for eligibility, we excluded 34 articles that did not meet the inclusion criteria, leaving nine articles for inclusion as shown in Fig 1.
Methodological quality assessment
Each study selected for retrieval for the quality of the methodology was independently assessed by two researchers before inclusion in the review. The methodological quality assessment of the studies that were included was carried out using standardised critical appraisal tools from the Joanna Briggs Institute Meta Analysis of Statistics Assessment and Review Instrument (JBI–MAStARI).32
Data extraction and synthesis
The extracted data included specific details about the studies' samples, methods, different interventions and results of significance to the review's specific objectives. When there was no consistency in the assessment or data extraction process, a discussion between the researchers was conducted to resolve any discrepancies.
Results
In this review, nine studies were included: seven randomised controlled trials (RCTs),27,33,34,35,36,37,38 one retrospective cohort study,39 and one 17-year prospective cohort study at different time points.22Table 2 displays the characteristics and results of the studies.
Table 2. Summary of reviewed glycaemic control studies for cardiac surgery patients
| Study authors | Mode and time of intervention | Type of patients | Diabetic/non-diabetic | Glycaemic control protocols (mg/dl) | Main findings |
|---|---|---|---|---|---|
| Furnary et al., 200422 | SQI every 4 hours during hospitalisation versus CII during surgery until third postoperative day (perioperative period) | Cardiac surgery | 4864 diabetic patients | SQI: <200 from 1987-1991CII: 150-200 from 1991-1998CII: 125-175 from 1999-2000CII: 100-150 in 2001 | CII had reduced morbidity and mortality compared with SQITargeting BG ≤150mg/dl reduced DSWI 66%, mortality rate 57%, and LOS |
| Gandhi et al., 200736 (RCT) | CII during surgery (intraoperative phase) | Cardiac surgery | 371 diabetic and non-diabetic patients | Intensive: 80-100Conventional: <200 | Intensive therapy had increased the rate of strokes by 4%, incidence of deaths by 2%, and heart block requiring pacemaker by 3% compared with conventional therapy. Morbidity, ICU/hospital LOS, and DSWI same for intensive and conventional |
| Chan et al. 200933 (RCT) | CII during surgery and 36 hours after surgery (perioperative period) | Cardiac surgery | 109 diabetic and non-diabetic patients | Intensive: 80-130Conventional: 160-200 | No significant differences between the two treatment groups in the incidence of postoperative infections, ICU LOS, hypoglycaemic events, duration of MV, number of blood transfusions, neurological dysfunction or 30-day mortality rates (p>0.05) |
| Finfer et al., 200935 (RCT) | CII after surgery for 3 days (postoperative phase) | Adult medical and surgical ICUs patients | 6104 diabetic and non-diabetic patients | Intensive: 81-108Conventional: ≥180 | Mortality rate in intensive group was 27.5% compared with 24.9% in conventional group (p=0.02). Severe hypoglycaemia was 6.8% in intensive group compared with 0.5% in conventional group (p<0.001). No significant difference between the two treatment groups in the LOS in the ICU or hospital, postoperative infections, median days of MV, or renal-replacement therapy |
| Bhamidipati et al., 201139 | CII during surgery and 3 days after the surgery (perioperative period) | CABG surgery | 4658 diabetic and non-diabetic patients | Intensive: ≤126Moderate: 127-179Liberal: ≥180 | Moderate group had the lowest mortality rate by 2% compared with intensive 2.9% and liberal 3.4% (p=0.02). Moderate group had the lowest major complications by 11.1% compared with intensive 19.4% and liberal 14.2% (p<0.001). The incidences of postoperative DSWI, stroke, and ICU LOS were similar across the study groups (p>0.05) |
| Lazar et al., 201127 (RCT) | CII during surgery and 18 hours after surgery | CABG surgery | 82 diabetic patients | Intensive: 90-120Conventional: 120-180 | No significant differences between treatment groups in the incidence of major adverse events included DSWI, time on the MV and LOS in the ICU/hospital. Increased hypoglycaemia with intensive control (p<0.0001) |
| Desai et al., 201234 (RCT) | CII after surgery for a minimum of 72 hours (postoperative phase) | CABG surgery | 189 diabetic and non-diabetic patients | Liberal group: 121-180Strict group: 90-120 | The liberal group was not inferior to the strict group for perioperative complications. The liberal group was superior to the strict group for glucose control and target range management. No significant differences between the two treatment groups in the incidence of DSWI, AF, renal failure, pneumonia, prolonged ventilation, LOS and mortality |
| Umpierrez et al., 201537 (RCT) | CII after surgery, until the patient was able to eat and/or transferred out the ICU (postoperative phase) | CABG surgery | 305 diabetic and non-diabetic patients | Intensive: 100-140Conventional: 141-180 | No significant differences between the two treatment groups in the incidence of SSIs, adverse complications, hypoglycaemia, ICU/hospital LOS or readmission rate. Non-diabetic patients in intensive group had a lower complication rate of 34% compared with 55% in conventional group (p=0.008) |
| Gadallah et al., 201638 (RCT) | CII during surgery until the patient was extubated in the ICU (perioperative period) | CABG surgery | 135 diabetic patients | Intensive: 110-149Conventional: 150-180 | Intensive therapy had reduced the rate of AF from 36.76% to 19.4% (p=0.025), SWI from 39.71% to 20.9% (p=0.012), leg wound infection from 35.29% to 19.4% (p=0.038), need for inotropic support from 66.18% to 41.79% (p=0.004) and time spent on MV from 27.94% to 13.43% (p=0.038) compared with the conventional therapy |
| CABG–coronary artery bypass graft; RCT–randomised controlled trial; SSIs–surgical site infections; SWI–sternal wound infection; DSWI–deep sternal wound infection; CII–continuous insulin infusion; SQI–subcutaneous insulin; ICU–intensive care unit; LOS–length of stay; MV–mechanical ventilation; AF–atrial fibrillation; BG–blood glucose |
All studies investigated tight glycaemic control using a CII on the incidence of SSIs in adult patients with or without diabetes undergoing cardiac surgery. All studies also investigated tight glycaemic control in the intraoperative phase, postoperative phase and perioperative period to the conventional protocol. The conventional protocol included the administration of CII for maintaining blood glucose levels higher than the tight protocol.
Description of selected studies
The nine studies included samples that ranged from 82–6104 patients. The average age among patients from eight studies was 55.6–64.1 years; however, the average age was not reported by one study.22 The percentage of males from eight studies ranged from 28.2–84.5%, while female representation ranged from 15.5–71.8%; however, the percentage of males and/or females was not reported by one study.22
Of the included studies, six were conducted in the US,7,22,33,36,37,39 one in Egypt,38 one in Brazil33 and one was an international study.36 SSIs were identified as sternal wound infections (SWIs) after the cardiac surgery in six studies.22,27,33,36,37,39 A study specified the site of SSIs as SWIs and leg wound infections.38 The results of postoperative infection that included SSIs in combination with other sources of infection, such as urinary tract infection, pneumonia and sepsis, were reported in two studies.33,35 While this was not the outcome identified at the beginning of the review, we decided to include these studies to obtain a more comprehensive understanding of the effects of tight glycaemic control on postoperative infections.
Time of intervention versus target range for tight glycaemic control
Of the nine studies included in this systematic review, five studies were conducted during the perioperative period,22,27,33,38,39 three studies in the postoperative phase34,35,37 and one study in the intraoperative phase.36
The target range of tight glycaemic control in this review varied as defined in each study; six studies targeted the lower value of blood glucose level between 80–120mg/dl,22,33,34,35,36,39 and the other three studies targeted the upper value of blood glucose level between 100–150mg/dl.22,37,38
Tight glycaemic control to reduce the incidence of surgical site infections
Of the included studies, seven RCTs and two cohort studies were conducted to explore tight glycaemic control on the outcome of SSI rates for adult patients with diabetes undergoing cardiac surgery. An RCT and one cohort study showed significant evidence with the use of tight glycaemic control that reduced infections, morbidity and mortality.22,38 In contrast, two RCTs and one cohort study indicated that there was no improvement in clinical outcomes, or possibly that it led to worse outcomes with the use of tight glycaemic control.35,36,39 The other four RCTs showed similar results with the use of tight glycaemic control without any improvement in the patients' clinical outcomes.27,33,34,37
The results of the included studies according to the time of intervention show a difference in the benefit of tight glycaemic control. Of the five perioperative studies, two showed a benefit;22,38 the other two showed a similarity without any differences;27,33 and the last study showed adverse results with tight glycaemic control.39 In addition, two of the three postoperative studies showed a similarity without any differences,34,37 and one study showed adverse results with tight glycaemic control.35 The study conducted in the intraoperative phase showed adverse results with tight glycaemic control.36
Moreover, out of six studies that targeted low blood glucose levels between 80–120mg/dl, three studies did not show any significant differences,27,33,34 and another three studies did not produce any improvement in patient outcomes with tight glycaemic control.35,36,39 On the other hand, two of the three studies that targeted higher blood glucose levels between 100–150mg/dl showed an improvement in patient outcomes,22,38 and one study did not show any significant differences in patient outcomes with tight glycaemic control.37
Discussion
For the nine studies conducted to explore glycaemic control for patients with diabetes undergoing cardiac surgery, there is equivocal evidence indicating the benefits of tight glycaemic control. Some studies showed significant evidence for the use of tight glycaemic control in these patients that led to an improvement in patients' clinical outcomes, and reduced infections, morbidity and mortality.22,38 However, this evidence was challenged by the results of other studies which indicated that there was no improvement in clinical outcomes, or possibly that it led to worse outcomes.35,36,39 Other studies showed similar results with tight glycaemic control without any improvement in the patients' clinical outcomes. Moreover, the conventional protocol showed a superiority in the control of blood sugar and targeting the range management with a low incidence of hypoglycaemia.27,33,34,37 Glycaemic control during the intraoperative period remains inadequately addressed in the selected nine articles. However, studies conducted to explore glycaemic control for patients undergoing cardiac surgery showed controversial results regarding the benefits of tight glycaemic control.22,27,33,34,35,36,37,38,39 Despite the agreement to avoid hyperglycaemia in patients undergoing cardiac surgery, irrespective of the surgical period, it is still unclear how the best glycaemic control should be implemented. The target blood glucose levels in the tight glycaemic control group varied between studies as defined in each study; some studies used the value of 80–120mg/dl27,33,35,36,39 and other studies used 100–150mg/dl.22,37,38 Of the studies targeting a low level of tight glycaemic control with values 80–120mg/dl, only one postoperative study showed an improvement in patient outcomes for reduced infections, morbidity and mortality,34 while a large postoperative international RCT showed worse patient outcomes with increased mortality and hypoglycaemia events.35 The intraoperative phase or perioperative period studies targeting blood glucose levels ranging between 80–120mg/dl did not display any improvement in patient outcomes and led to worse results by increased morbidity and mortality.27,33,36,39 On the other hand, studies conducted in the intraoperative phase or perioperative period, that indicated an improvement in patient outcomes with a low incidence of infections, morbidity and mortality, were targeted by a higher level of tight glycaemic control with values 100–150mg/dl.22,38 In addition, another postoperative study using higher blood glucose levels ranging between 100–140mg/dl did not show any significant differences in the incidence of adverse outcomes among patients.37
To summarise, the literature reports controversial findings in relation to the type of glycaemic control protocol used. However, the reviewed literature tends to support the implementation of tight glycaemic control protocols, particularly in the postoperative period, which revealed fewer potential complications associated with cardiac surgery in patients with diabetes. On the other hand, the literature also supported the application of moderate glycaemic control protocol in the intraoperative phase to obtain better glycaemic stability with fewer, less severe potential complications among patients undergoing cardiac surgery who had diabetes.
Implications
It is important to note that the presence of an effective glycaemic control protocol assists the healthcare team working with patients with diabetes undergoing cardiac surgery to anticipate and properly intervene on a regular basis using tight glycaemic control (80–120mg/dl) to prevent hyperglycaemia and other potential complications. Cardiac surgeons will be more confident to use the tight glycaemic control protocol while performing cardiac surgery for vulnerable patients with diabetes and, at the same time, they will be enabled to avoid hyperglycaemia complications postoperatively, including SSIs. On the other hand, the health status of patients with diabetes undergoing cardiac surgery tends to be safer with better health outcomes.
Limitations
Although the present paper presents a rigorous research review of the main variables in the multiple key databases of peer-review articles, there are some limitations in the selected articles. There was a lack of representation of low-and-middle income countries, such as Jordan, for addressing both tight and conventional glycaemic control protocols among patients with diabetes undergoing CABG surgery. Extracting inclusion/exclusion criteria of tight and conventional glycaemic control protocols was challenging in most of the articles. In addition, the selection of participants and their sample sizes was hard to understand, in which the G power calculation was often not clarified. The value ranges of glycaemic levels for both implemented protocols were varied in some publications. Furthermore, certain patients' health characteristics, as with the psychosocial responses, were not addressed in all articles, which could have an impact on blood glycaemic levels among patients intraoperatively.
Conclusion
The findings of this systematic review would seem to indicate that tight glycaemic control (80–120mg/dl) may be most effective in the postoperative phase. On the other hand, moderate glycaemic control (100–150mg/dl) should be maintained in the intraoperative phase. The results emphasise the importance of time-based protocol implementation to ensure fewer adverse health outcomes and provide better quality of care for patients with diabetes undergoing cardiac surgery.
Reflective questions
- What are the potential adverse health outcomes among patients with diabetes post-coronary artery graft bypass (CABG) surgery?
- How can glycemic control for patients undergoing cardiac surgery improve their health outcomes?
- What are the major modified guidelines of various medical societies for better glycemic control?