Clinical Significance of Post-Operative Hyperglycemia in Nondiabetic Patients Undergoing Definitive Surgery for Gastrointestinal Fistula

Abstract

Background:

To evaluate the association between peri-operative hyperglycemia and adverse events in patients with gastrointestinal (GI) fistulas without a pre-operative diagnosis of diabetes mellitus who were undergoing definitive surgery.

Methods:

Pre-operative and all post-operative blood glucose concentrations (BG) were retrieved for 363 consecutive patients undergoing GI reconstruction from September 2012 to December 2015. Normoglycemic (BG <125 mL/dL), mild hyperglycemia (125–199 mL/dL), and severe hyperglycemia (≥200 mL/dL) were defined using the highest BG found within the first 48 h post-operatively. Outcomes of interest included 30-d mortality rate and re-operation, time of enteral nutrition resumption, and infectious and non-infectious complications.

Results:

More than half of the nondiabetic patients (61.4%) experienced hyperglycemia post-operatively. The degree of hyperglycemia correlated with patient age, American Society of Anesthesiologists class, and surgical interventions. Hyperglycemia was associated with re-operation and post-operative complications, the frequency of these complications increasing in parallel with the degree of hyperglycemia. Additionally, post-operative hyperglycemia was associated independently with surgical site infections (p = 0.014), anastomotic leak (p = 0.010), delayed resumption of enteral nutrition (p < 0.001), and longer hospital stay (p < 0.001).

Conclusion:

Elevated post-operative BG was frequent after surgery in patients with GI fistulas. Post-operative hyperglycemia is significantly associated with unfavorable outcomes, and this risk is related to the degree of BG elevation. Our findings suggest that vigilant post-operative BG monitoring and early appropriate glycemic control are critical for patients, even nondiabetic patients, undergoing definitive surgery for GI fistula.

Management of gastrointestinal (GI) fistula is a complex and challenging problem commonly encountered by the general and specialized surgeon. These patients have intra-abdominal and systemic infection; water, electrolytes, and acid–base disturbances, and severe malnutrition that can result in sepsis or even multiple organ dysfunction [1]. Surgical intervention offers a definitive therapeutic effect but is appropriate only for some patients, in whom the fistula may not or does not close after the conservative treatment [2]. Gastrointestinal fistula is potentially life-threatening. Therefore, physicians should pay sufficient attention to strengthening peri-operative management and avoiding post-operative complications.

Numerous studies have demonstrated that uncontrolled peri-operative hyperglycemia can be associated with more infectious complications in both diabetic and nondiabetic patients undergoing general and colorectal surgical [3 –6]. This effect generally is attributed to an impaired immune response, in combination with excessive release of stress hormones and inflammatory cytokines resulting in diminished insulin action [7]. Bacterial contamination commonly occurs in GI fistula and leads to post-operative infections that typically exceed those observed with other abdominal operations. Furthermore, considering that GI fistula surgery involves patients and conditions of various complexities and operations of differing severities and durations that could provoke a significant stress response, this group provides an ideal cohort for the evaluation of the influence of hyperglycemia on outcomes after surgery.

To our knowledge, the true risk and significance of peri-operative hyperglycemia in patients with GI fistulas without diabetes mellitus undergoing elective enterectomies remains largely unclear. The main purpose of this study was to determine whether post-operative hyperglycemia is associated with adverse outcomes, the time of enteral nutrition (EN) resumption, and hospital length of stay (LOS) in a uniform cohort of nondiabetic patients having surgery for GI fistulas.

Patients and Materials

Study design and patient population

Data of 363 patients undergoing bowel resection at a teaching hospital between September 1, 2012 and December 30, 2015, were collected retrospectively from the medical recordssystem of our hospital. Diabetic patients were identified on admission as outlined later and separated from previously nondiabetic patients. On the basis of the highest blood glucose value (BG) within the first 48 h post-operatively, patients were grouped into “normoglycemia,” “mild hyperglycemia,” and “severe hyperglycemia” as defined below. This study was approved by the Institutional Review Board of Jinling Hospital, Najing, and informed consent was obtained from all the participants.

Inclusion and exclusion criteria

The inclusion criteria were at least 18 y of age; having an abdominal computed tomography (CT) scan or fistulography that confirmed the diagnosis of GI fistula, and being scheduled for a definitive intestinal reconstruction. Etiologies were grouped into inflammatory bowel disease, trauma, malignant disease, pancreatitis, and others. Patients treated with enterostomy without bowel resection were excluded from this study. Patients who had a history of abdominal surgery within the previous 3 mos also were excluded. All anastomoses were created using a linear stapler.

Pre-operative management

The fistula was examined fully for number, location, and type using contrast radiography and an abdominal CT scan. When patients were admitted to our units, we corrected any imbalance of electrolytes, controlled the infection using surgical drainage, and reinforced their nutritional status using parental or enteral nutrition. Parenteral nutrition was used initially and switched to enteral nutrition at specific times on the basis of the clinical judgment of the physician. A selective operation was considered to repair the fistula when the nutritional status of the patients had improved.

Clinical data and outcome measurements

All the data were reviewed and collected by the same physician. Demographic features included gender, age, body mass index (BMI), history of smoking and alcohol consumption, etiology and type of GI fistula, interval from diagnosis of GI fistula to operation (mos), American Society of Anesthesiologists (ASA) status, and pre-operative serum hemoglobin and albumin concentrations. Peri-operative data included the operative time, estimated blood loss, intra-operative transfusion requirements, numbers of anastomoses, and type of surgery. Post-operative indices of attention were anastomotic leaks, intensive care unit length of stay (ICU LOS), LOS, re-admission, 30-d mortality rate, re-operation, time of EN resumption, and hospitalization expenses. The post-operative infectious complications recorded were superficial, deep, and organ/space surgical site infections (SSIs); pneumonia; and urinary tract infections. When bowel function had recovered, as defined by physicians after surgery, EN was given to the patients.

Definition of hyperglycemia and diabetes mellitus

Patients received BG assessment using both finger checks and laboratory tests every 8 h for 48 h after the surgical procedure, and any episodes of hypoglycemia were recorded. According to the highest BG post-operatively, patients were categorized into three groups: normoglycemic (<125 mg/dL); mild hyperglycemia (125–199 mg/dL); and severe hyperglycemia (≥200 mg/dL). Hyperglycemia occurring later than 48 h post-operatively was not included in this analysis [8]. Admission random or fasting BG values and the presence of a known diagnosis of diabetes mellitus helped to identify patients with pre-operative diabetes mellitus.

Statistical analysis

Descriptive statistics were used to evaluate the demographic data. The Student t-test (normal distribution) or Mann-Whitney U test (abnormal distribution) was used for group-wise comparisons. For comparisons of the three BG groups, analysis of variance (normal distribution) or Kruskal-Wallis test (nonparametric distributions of data) was applied. The χ² test or the Fisher exact test was used to contrast nominal data. A multiple logistic regression model was employed to detect independent predictors of SSIs and anastomotic leak. The adjusted influence of hyperglycemia on the time of EN resumption and the LOS were calculated with multiple linear regression. All analyses were considered statistically significant if p was <0.05.

Results

Patient characteristics

A total of 363 patients receiving definitive surgery for GI fistula were included in the analysis. Forty-nine patients (13.5%) had diabetes mellitus on admission. Despite the absence of diabetes mellitus, 193 (61.5%) of the other patients suffered post-operative hyperglycemia of >125 mg/dL. Hyperglycemia was observed more frequently among patient with a fistula of the ileocolon anastomosis who underwent GI reconstruction. Older patients were vulnerable to post-operative hyperglycemia, and patients who developed hyperglycemia were prone to be in a greater ASA class than those with normoglycemia. No significant difference was observed in pre-operative hemoglobin or albumin concentrations among the three groups (Table 1).

Table 1.

Characteristics of 314 Patients

	Blood glucose concentration (mg/dL)
	Normoglycemia (<125)	Mild HG (125–199)	Severe HG (≥200)	p
No. patients (%)	121 (38.5)	148 (47.1)	45 (14.3)
Male (%)	83 (68.6)	102 (68.9)	31 (68.9)	0.998
Mean BMI ± standard deviation	21.0 ± 3.0	20.8 ± 3.3	21.1 ± 3.9	0.141
Age (y)	42.1 ± 15.6	47.0 ± 14.2	50.0 ± 12.3	<0.001
Smoking status (%)				0.105
Smoker and ex-smoker	95 (76.0)	102 (68.9)	29 (64.4)
Nonsmoker	26 (24.0)	46 (31.1)	16 (35.6)
Alcohol consumption (%)				0.347
Yes	95 (76.0)	107 (72.3)	31 (64.6)
No	26 (24.0)	41 (27.7)	14 (35.4)
ASA class (%)				<0.001
<3	140 (96.6)	125 (84.5)	30 (66.7)
≥3	5 (4.1)	23 (15.5)	15 (33.3)
Pre-operative Hb (g/L)	127.8 ± 17.8	126.5 ± 13.4	124.3 ± 19.1	0.452
Pre-operative albumin (g/L)	38.2 ± 4.5	39.1 ± 6.3	37.5 ± 6.7	0.193
Etiology (%)				0.165
Inflammatory bowel disease	45 (37.2)	43 (29.1)	8 (17.8)
Trauma	21 (17.4)	33 (22.3)	11 (24.4)
Malignancy	13 (10.7)	28 (18.9)	10 (22.2)
Pancreatitis	2 (1.7)	2 (1.4)	1 (2.2)
Other	40 (33.0)	42 (28.3)	15 (33.2)
Type of fistula (%)				0.737
Upper gastrointestinal	2 (1.7)	3 (2.0)	1 (2.2)
Small intestine	41 (33.9)	43 (29.1)	11 (24.4)
Colon	40 (33.1)	40 (27.0)	12 (26.8)
Ileocolon anastomosis	36 (29.7)	60 (40.5)	20 (44.4)
Rectum	2 (1.6)	2 (1.4)	1 (2.2)
Interval from fistula development to operation (mos)	18.7 ± 21.5	15.1 ± 30.4	17.4 ± 28.7	0.548

Statistically significant p values are shown in boldface type.

ASA = American Society of Anesthesiologists; BMI = body mass index; Hb = hemoglobin; HG = hyperglycemia.

Association of hyperglycemia with outcomes in nondiabetic patients

On the basis of the degree of post-operative glycemia in nondiabetic patients, an analysis of the common outcome variables was provided (Table 2). Using univariable analysis, statistical significance was found between post-operative hyperglycemia and transfusion requirement, longer surgery time, multiple anastomoses, and type of surgery. Hyperglycemia (mild or severe) was associated with a greater rate of infectious complications (overall SSIs, organ/space SSIs, and anastomotic leak), re-operation, a longer time to EN resumption, and a greater ICU LOS and LOS. Additionally, the risk of complications rose with an increasing concentration of BG. The probability of SSIs was 22.1% in patients with normoglycemia and 44.4% in patients with severe hyperglycemia during the first 48 h post-operatively. Anastomotic leak, similarly, was nine times as frequent in patients with severe hyperglycemia compared with the normoglycemic group (15.6% vs. 1.7%; p = 0.004).

Table 2.

Outcome Variables for Patients Operated on for Gastrointestinal Fistula Not Having Diabetes Mellitus

	Blood glucose concentration (mg/dL)
	Normoglycemia (<125)	Mild HG (125–199)	Severe HG (≥200)	p
Intra-operative (%)	121 (38.5)	148 (47.1)	45 (14.3)	0.141
EBL >300 mL	8 (6.6)	12 (8.1)	5 (11.1)	0.633
Transfusion	25 (20.6)	53 (35.8)	23 (51.1)	<0.001^a
Length of surgery (>180 min)	24 (19.8)	35 (22.6)	17 (37.8)	0.026^b
Numbers of anastomoses				0.001^a
Single	115 (95.0)	124 (83.8)	34 (75.6)
Multiple	6 (5.0)	24 (16.2)	11 (24.4)
Type of surgery				0.005^a
Laparoscopic	41 (33.9)	35 (23.6)	5 (10.4)
Open laparotomy	80 (66.1)	113 (76.4)	43 (89.6)
Post-operative (%)
Pneumonia	2 (1.7)	0	1 (4.5)	0.245
Anastomotic leak	2 (1.7)	16 (10.8)	7 (15.6)	0.004^a
UTI	0	2 (1.4)	0	0.317
Time of EN resumption	4.5 ± 2.1	7.8 ± 2.9	8.8 ± 2.7	<0.001^a
Overall SSIs	26 (21.5%)	53 (35.8%)	20 (44.4%)	0.006^a
Superficial	20 (16.5%)	27 (18.2)	8 (17.8%)	0.925
Deep	4 (3.3%)	8 (5.4%)	4 (8.9%)	0.227
Organ/space	2 (1.7%)	18 (12.2%)	8 (17.8%)	0.001^a
Length of hospital stay (d)	26.9 ± 9.5	30.0 ± 14.4	36.8 ± 15.3	<0.001^a
ICU length of stay (d)	3.1 ± 2.5	4.2 ± 2.7	4.9 ± 2.8	<0.001^a
Re-operation (%)	3 (2.5)	17 (11.5)	8 (16.7)	0.004^a
Re-admission (%)	8 (6.6)	12 (8.1)	7 (14.6)	0.239
Deaths (%)	1 (0.83)	1 (0.68)	2 (4.2)	0.145
Total hospitalization costs (yuan)	123,901 ± 41,784	160,856 ± 60,747	153,765 ± 65,064	<0.001

Significant difference between normoglycemia and mild or severe hyperglycemia by analysis of variance.

Significant difference between normoglycemia and severe hyperglycemia by analysis of variance.

EBL = estimated blood loss; EN = enteral nutrition; HG = hyperglycemia; ICU = intensive care unit; SSIs = surgical site infection; UTI = urinary tract infection.

Finally, we evaluated the economic burden of the three groups. The total hospitalization costs in the normoglycemia group were significantly less than in the other two groups. Moreover, in 193 nondiabetic patients with post-operative hyperglycemia, 47 (24.4%) had more than two episodes of hypoglycemia within 48 h postoperatively, and no significance was found in adverse outcomes compared with those of 146 patients (75.6%) who had one isolated elevated value (data not shown).

Independent predictors of specific outcomes in nondiabetic patients

Significant associations in univariable analysis (Table 2) were included in the multivariable regression analysis to explore independent predictors of such specific complications as SSIs, anastomotic leakage, and re-operation (Tables 3 and 4). The SSIs and re-operation were associated with post-operative hyperglycemia (p = 0.014 and p = 0.035) by multivariable analysis (Tables 3 and 4). Anastomotic leak was related independently to an ASA score of 3 or greater (p = 0.04) and post-operative hyperglycemia (p = 0.010). Multiple linear regression analysis confirmed that EN resumption was initiated 2 d and 3 d later in patients with mild hyperglycemia and severe hyperglycemia, respectively, compared with normoglycemic patients (Table 5). The LOS was extended to 10 d only in severe hyperglycemia (Table 5).

Table 3.

Independent Risk Factors Associated with SSIs and Anastomotic Leak in Nondiabetic Patients

Characteristic	OR	95% CI	p
Overall SSIs
Age <50 y	1.05	0.62–1.80	0.847
ASA score ≥3	1.13	0.55–2.31	0.746
Smoking	1.27	0.74–2.19	0.387
Type of surgery (laparotomy)	0.61	0.34–1.09	0.095
Surgery time ≥180 min	1.14	0.79–2.52	0.246
Multiple anastomoses	0.85	0.41–1.79	0.675
Hyperglycemia group			Overall 0.014
Normoglycemia (reference)	1.0
Mild hyperglycemia	1.99	1.12–3.54	0.019
Severe hyperglycemia	3.02	1.36–6.70	0.007
Anastomotic leak
Age <50 y	1.12	0.46–2.75	0.797
ASA score ≥3	4.10	1.35–12.3	0.04
Smoking	1.18	0.47–2.96	0.729
Type of surgery (laparotomy)	0.88	0.31–2.48	0.811
Surgery time ≥180 min	1.19	0.45–3.17	0.727
Multiple anastomoses	1.58	0.57–4.40	0.383
Hyperglycemia group			Overall 0.010
Normoglycemia (reference)	1.0
Mild hyperglycemia	7.59	1.68–34.34	0.009
Severe hyperglycemia	13.4	2.50–71.65	0.002

Statistically significant p values are shown in boldface type.

ASA = American Society of Anesthesiologists; CI = confidence interval; OR = odds ratio; SSI = surgical site infection.

Table 4.

Independent Risk Factors Associated with Reoperation in Nondiabetic Patients

Characteristic	OR	95% CI	p
Age <50 y	0.89	0.39–2.03	0.775
ASA score ≥3	1.69	0.64–4.47	0.289
Smoking	1.42	0.61–3.29	0.415
Types of surgery (laparotomy)	1.18	0.40–3.44	0.766
Surgery time ≥180 min	1.24	0.50–3.06	0.640
Multiple anastomoses	0.759	0.24–2.43	0.642
Hyperglycemia group			Overall 0.035
Normoglycemia (reference)	1.0
Mild hyperglycemia	4.50	1.25–16.13	0.021
Severe hyperglycemia	6.52	1.52–27.98	0.012

Statistically significant p values are shown in boldface type.

ASA = American Society of Anesthesiologists; CI = confidence interval; OR = odds ratio.

Table 5.

Multiple Linear Regression Analysis of Time of Enteral Nutrition Resumption and Length of Hospital Stay

	Time of EN resumption (d)		Length of hospital stay (d)
	Estimated coefficient (95% CI)	p	Estimated coefficient (95%)	p
Age (y)		0.502		0.021
<50 (reference)	0		0
≥50	0 (−4–1)		4 (1–7)
ASA score		0.052		0.623
1–2 (reference)	0		0
≥3	1 (0–2)		−1 (−6–4)
Smoking		0.023		0.052
No (reference)	0		0
Yes	1 (0–2)		3 (−1–7)
Types of surgery		0.360		0.140
Laparoscopic (reference)	0		0
Laparotomy	0 (−4–1)		3 (−9–6)
Surgery time (min)		0.001		0.531
<180 (reference)	0		0
≥180	1 (0–2)		−1 (−5–2)
Number of anastomoses		0.001		0.350
Single (reference)	0		0
Multiple	2 (1–3)		−2 (−6–2)
Hyperglycemia group		<0.001		<0.001
Normoglycemia (reference)	0		0
Mild hyperglycemia	2 (1–3)	<0.001	2 (−8–6)	<0.134
Severe hyperglycemia	3 (1–4)	<0.001	10 (4–11)	<0.001

Figures in boldface type are statistically significant.

ASA = American Society of Anesthesiologists; CI = confidence interval; EN = enteral nutrition.

Effect of post-operative hyperglycemia on nondiabetic compared with diabetic patients

We performed the adverse outcomes analysis in nondiabetic vs. diabetic patients with post-operative hyperglycemia (Table 6). The ICU LOS and the hospitalization costs showed differences between these two groups. However, none of other adverse outcomes differed in diabetic vs. nondiabetic patients.

Table 6.

Comparisons of Post-Operative Adverse Outcomes in Nondiabetic vs. Diabetic Patients with Any Degree of Post-Operative Hyperglycemia (≥125 mg/dL)

Adverse outcome	Nondiabetic (N = 193)	Diabetic (N = 32)	p
Pneumonia	1	0	0.999
Anastomotic leak	23	2	0.544
UTI	2	1	0.370
Time of EN resumption	7.9 ± 2.9	7.6 ± 4.1	0.634
Overall SSI	73	8	0.162
Superficial	35	4	0.615
Deep	12	3	0.454
Organ/space	26	1	0.140
Length of hospital stay (d)	35.8 ± 15.1	37.9 ± 14.2	0.245
ICU length of stay (d)	4.4 ± 2.8	5.5 ± 2.3	0.038
Re-operation	25	1	0.139
Re-admission	19	4	0.752
Deaths	3	0	1.00
Total hospitalization costs (yuan)	170,633 ± 68,754	198,893 ± 79,787	0.037

Figures in boldface are statistically significant.

EN = enteral nutrition; ICU = intensive care unit; SSI = surgical site infection; UTI = urinary tract infection.

Discussion

Post-operative hyperglycemia has been confirmed to be associated with adverse outcomes in subgroups of surgical patients. A greater incidence of sepsis and pneumonia, a greater rate of re-operation, and increased mortality rates have been reported in colorectal surgery patients with peri-operative hyperglycemia [8]. However, the risk of uncontrolled or poorly controlled hyperglycemia in nondiabetic patients undergoing definitive intestinal surgery, especially for GI fistula, remains poorly understood. The purpose of this study was to explore the association of post-operative hyperglycemia with adverse events in patients with GI fistula undergoing definitive reconstruction. Because greater attention would be attracted to the monitoring and control of glucose in diabetics, patients with a history of diabetes mellitus, or a measured HbA1C concentration indicating diabetes were excluded. In this study, we provided evidence that peri-operative hyperglycemia was common after surgery, and our results also suggested that post-operative hyperglycemia was an independent predictor of SSIs, anastomotic leakage, re-operation, long time to EN resumption, and longer LOS.

In our study, only a small proportion (38.5%) of nondiabetic patients did not have an elevated glucose concentration; that is, they maintained normoglycemic during the first 48 h post-operatively, which mirrored the occurrence of hyperglycemia at a single center in colorectal surgery patients found by Kiran et al. [8]. When examining this population, we noticed a significant difference between the patients with elevated and normal post-operative BG concentrations. Specifically, subjects with hyperglycemia were older and had more serious ASA scores. Additionally, they were more likely to have had multiple anastomoses, intra-operative packed red blood cell transfusion, longer surgery times, and open laparotomies. These factors probably make it difficult to keep normoglycemic after surgery and contribute to adverse outcomes. Even so, after these confounding elements were adjusted for, post-operative hyperglycemia (≥125 or ≥200 mg/dL) was independently associated with adverse outcomes and high LOS. This suggests that appropriate peri-operative glucose control measures should be employed to prevent the unfavorable outcomes, even in patients who do not have a pre-operative diagnosis of diabetes mellitus.

The distribution of post-operative BG values between diabetic and nondiabetic groups was similar in our study. Although diabetic patients stayed longer in the ICU, the occurrence of adverse outcomes and LOS between two groups showed no statistical difference. This finding supports the need for monitoring of and attention to BG concentrations after definitive surgery even in nondiabetic patients. Moreover, with greater degrees of hyperglycemia, the frequency of several adverse outcomes increased, further stressing the necessity of such measures.

Surgical site infections are the third most common hospital-acquired infection, and elective gastrointestinal surgery has the highest incidence of SSIs, which contribute to greater LOS and resource utilization [9]. Post-operative hyperglycemia has been linked to SSIs in general surgery patients. Using multivariable analysis, Yarrow et al. revealed that hyperglycemia (≥200 mg/dL) at 48 h post-operatively was the strongest independent predictor of SSIs in 149 patients having colorectal surgery [10]. A comprehensive analysis explored the correlation between the post-operative BG concentration and SSIs, and the investigators reported that a single post-operative BG concentration above 200 mg/dL within 7 d post-operatively was independently associated with organ/space SSIs [11]. On the contrary, a large cohort study by Jeon et al., investigating peri-operative BG values of 13,800 hospitalized patients who underwent various surgical procedures, revealed that there was no association between post-operative hyperglycemia (≥180 mg/dL) and SSIs or mortality rate by multivariable analysis [12]. We observed an apparent association between post-operative hyperglycemia and organ/space SSIs but not of superficial and deep SSIs in the univariable analysis, and our multivariable results suggested that early post-operative hyperglycemia was independently associated with overall SSIs.

Compared with several previous studies, we provide a better estimate of the effect of peri-operative hyperglycemia on adverse outcomes in patients operated on for a GI fistula. First, a uniform cohort was selected that underwent definitive GI reconstruction, thereby reducing the potential bias caused by differences in the intensity of the operation and peri-operative management, which probably affected the post-operative BG concentrations and adverse outcomes. Second, patients with mild hyperglycemia (125 to 200 mg/dL) constituted a great proportion of our study cohort, and this category of nondiabetic patients was previously considered unproblematic. However, the result of our analysis revealed that post-operative mild hyperglycemia in patients deemed to be nondiabetic was significantly associated with adverse outcomes and increased LOS and total hospitalization costs.

Anastomotic leakage is a dreadful complication of GI surgery, leading to significant morbidity and mortality rate with long hospital stays and considerable healthcare costs [13]. Previous studies attempted to identify specific factors associated with anastomotic leakage in colorectal surgery but did not include post-operative hyperglycemia in their analysis. Brenner et al. showed that fistula recurrence after definitive surgery for GI fistula was the sole predictor of death and was influenced by four variables, namely, inflammatory bowel disease, location of the fistula, stapled anastomosis, and operative delay [14]. To our knowledge, our study is the first to demonstrate that post-operative hyperglycemia is independently associated with anastomotic leakage after repair of GI fistula, and the risk of leakage was greater in patients with severe hyperglycemia.

Acute hyperglycemia alters gastric motility and delays GI emptying in normal subjects and persons with diabetes mellitus, slowing the entry of nutrients into the circulation [15 –17]. Under conditions of hypoglycemia, gastric emptying typically is accelerated, which quickens the delivery of glucose into the circulation [18]. We found that the time to post-operative EN resumption was significantly longer in patients with hyperglycemia. Similarly, Won et al. in their retrospective observational study of 172 patients undergoing ventral hernia repair found that patients with hyperglycemia (140 mg/dL) within 48 h after surgery were more likely to have delays in the interval to the first solid meal [19]. However, fewer studies elucidate the association between peri-operative hyperglycemia and post-operative GI hypomotility, and additional trials are needed to prove this point.

There are several limitations to this study. First, it was retrospective, which could result in differences in baseline patient population characteristics and in medical practice and biases regarding treatment. Confirmation is required in larger prospectively recruited cohorts. Second, although several relevant covariates were adjusted for in our analysis, such as age, ASA score, types of surgery, etc., we were not able to adjust for other confounding factors such as co-morbidities and medication use prior to surgery. Additionally, there may be unknown confounding variables that might have affected our results. In addition, there are no standardized treatment plans or guidelines for treating nondiabetic patients with elevated BG concentrations post-operatively.

In conclusion, the prevalence of various complications in patients with GI fistulas undergoing definitive surgery is high, and our findings suggest that post-operative hyperglycemia contributes to these problems. Episodes of elevated BG concentrations were associated with adverse outcomes, and this risk related to the degree of hyperglycemia. These findings suggest that monitoring glucose concentrations and early consideration of glycemic management for patients undergoing surgical procedures for GI fistula are critical even in nondiabetic patients. Further prospective studies are needed to clarify the mechanism of the greater risk of hyperglycemia among nondiabetic patients after surgery.

Footnotes

Author Disclosure Statement

All authors declare no conflicts of interest.

This study was supported by a grant from the National Natural Science Foundation of China (No. 81270478).

References

Ren

, Liu

, Wang

, et al. Laparoscopy improves clinical outcome of gastrointestinal fistula caused by Crohn's disease. J Surg Res, 2016; 200:110–116.

Fan

, Ren

, Zhao

, et al. Fistula output microorganism-susceptible antimicrobial prophylaxis is associated with a lower risk of surgical site infection in gastrointestinal fistula patients undergoing one-stage definitive surgery. Surg Infect, 2014; 15:774–780.

Jamsen

, Nevalainen

, Eskelinen

, et al. Risk factors for perioperative hyperglycemia in primary hip and knee replacements. Acta Orthop, 2015; 86:175–182.

Kwon

, Thompson

, Dellinger

, et al. Importance of perioperative glycemic control in general surgery: A report from the Surgical Care and Outcomes Assessment Program. Ann Surg, 2013; 257:8–14.

Mohan

, Kaoutzanis

, Welch

, et al. Postoperative hyperglycemia and adverse outcomes in patients undergoing colorectal surgery: Results from the Michigan Surgical Quality Collaborative Database. Int J Colorectal Dis, 2015; 30:1515–1523.

Engoren

, Schwann

, Habib

. Hyperglycemia, hypoglycemia, and glycemic complexity are associated with worse outcomes after surgery. J Crit Care, 2014; 29:611–617.

Eshuis

, Hermanides

, van Dalen

, et al. Early postoperative hyperglycemia is associated with postoperative complications after pancreatoduodenectomy. Ann Surg, 2011; 253:739–744.

Kiran

, Turina

, Hammel

, et al. The clinical significance of an elevated postoperative glucose value in nondiabetic patients after colorectal surgery: Evidence for the need for tight glucose control?. Ann Surg, 2013; 258:599–604.

Uchino

, Ikeuchi

, Tsuchida

, et al. Surgical site infection following surgery for inflammatory bowel disease in patients with clean-contaminated wounds. World J Surg, 2009; 33:1042–1048.

10.

McConnell

, Johnson

, Porter

. Surgical site infections following colorectal surgery in patients with diabetes: Association with postoperative hyperglycemia. J Gastrointest Surg, 2009; 13:508–515.

11.

, Stein

, Trencheva

, et al. Differing risk factors for incisional and organ/space surgical site infections following abdominal colorectal surgery. Dis Colon Rectum, 2011; 54:818–825.

12.

Jeon

, Furuya

, Berman

, et al. The role of pre-operative and post-operative glucose control in surgical-site infections and mortality. PLoS One, 2012; 7:e45616.

13.

Singh

, Zeng

ISL

, Srinivasa

, et al. Systematic review and meta-analysis of use of serum C-reactive protein levels to predict anastomotic leak after colorectal surgery. Br J Surg, 2014; 101:339–346.

14.

Risk factors for recurrence after re Source Arch Surg SO. 2009 Jun 144 6 500 5 [PMIDM19528379].pdf

15.

Harder

, Hernando-Harder

, Franke

, et al. Effect of acute hyperglycemia on intestinal gas transit and tolerance in nondiabetic humans. Digestion, 2005; 71:179–186.

16.

Hyperglycemia inhibits mechanorecept Source Gastroenterology SO. 1995 Feb 108 2 350 9 [PMIDM7835576].pdf

17.

Boronikolos

, Menge

, Schenker

, et al. Upper gastrointestinal motility and symptoms in individuals with diabetes, prediabetes and normal glucose tolerance. Diabetologia, 2015; 58:1175–1182.

18.

Plummer

, Jones

, Annink

, et al. Glucagon-like peptide 1 attenuates the acceleration of gastric emptying induced by hypoglycemia in healthy subjects. Diabetes Care, 2014; 37:1509–1515.

19.

Won

, Lehman

, Geletzke

, et al. Association of postoperative hyperglycemia with outcomes among patients with complex ventral hernia repair. JAMA Surg, 2015; 150:433–440.