Insulin and hypertonic glucose in the management of aseptic fat liquefaction of post‐surgical incision: a meta‐analysis and systematic review

Abstract

A meta‐analysis and systematic review assessing randomised controlled trials (RCTs) was sought to determine whether subcutaneous injection of insulin with hypertonic glucose promotes healing in postoperative incisions with aseptic fat liquefaction. We searched the Cochrane library, Pubmed, EMBASE, National Science Digital Library (NSDL) and China Biological Medicine Database (CBMdisc) for literature published from 1 January 1990 to 30 September 2011. RCTs that evaluated subcutaneous injection of insulin with hypertonic glucose as a treatment for postoperative wound with fat liquefaction were sought. Wound healing was the primary endpoint. Jadad score and Cochrane Collaboration's tool were used for assessing quality of studies and risk of bias. We abstracted data regarding time to wound healing, cost and adverse effects. The random‐effects inverse variance model was used for all analyses using weighted mean difference and 95% confidence interval. Eight trials (414 participants) were identified that met the inclusion criteria. Subcutaneous injection of insulin with hypertonic glucose significantly reduces time to healing by 6·33 days compared with conventional drainage, with less cost. There was no report concerning adverse effects. Subcutaneous injection of insulin with hypertonic glucose may improve the healing process in postoperative wounds with aseptic fat liquefaction.

INTRODUCTION

Aseptic fat liquefaction, as a main cause of prolonged healing of aseptic post‐surgical incision, is the necrosis of adipose tissue without infection and exhibits an incidence of 0·52–1·11% in all postoperative wounds 1, 2, 3. It is more common in older or overweight patients complicated with diabetes or malnutrition 4, 5, 6, which, besides the disease and operation, increases stress and economic burdens on patients and their families. Many preclinical and clinical interventions have shown their efficacy in promoting wound healing, including topical application of honey, hormones, insulin‐zinc, negative pressure therapy, low‐level laser energy, antibiofilm, microRNA and statins 7, 8, 9, 10, 11, 12, 13, 14. There has been anecdotal and sporadic evidence of insulin in promoting problematic wound healing over decades (15). However, data from clinical studies in English literature are still limited.

In this meta‐analysis, we investigated into the therapeutic efficacy of the subcutaneous application of insulin with hypertonic glucose in post‐surgical wound healing with aseptic fat liquefaction, and reviewed related literature.

METHODS

Sources

We searched Medline, EMBASE, Cochrane Library, National Science Digital Library (NSDL), and China Biological Medicine Database (CBMdisc) for relevant randomised controlled trials (RCTs) aiming to promote the healing of postoperative wounds with fat liquefaction by subcutaneous application of insulin. Queries included articles published from 1 January 1990 to 30 September 2011 in English and Chinese peer‐reviewed publications (including abstracts). Keywords used were ‘subcutaneous’ (or ‘regional’), ‘insulin’, ‘glucose’, ‘operation’ (or ‘surgery’ or ‘surgical’ or ‘postoperative’ or ‘post‐surgical’) and ‘wound’ (or ‘incision’) and ‘fat liquefaction’ (or ‘necrosis’). We also hand‐searched bibliographies of original studies, reviews (including meta‐analyses) and relevant conference abstracts, and contacted some investigators directly. The date last searched was 15 October 2011.

Study selection

Two authors independently selected relevant studies, extracted data and assessed trial quality by means of modified Jadad score (Table 1). The modified Jadad used is an 8‐item scale designed to assess randomisation, blinding, withdrawals/dropouts, inclusion/exclusion criteria, adverse effects and statistical analysis. The score for each article can range from 0 (lowest quality) to 8 (highest quality). Scores of 4 to 8 represent high quality 16, 17. The Cochrane Collaboration's tool for assessing risk of bias (18) was also referred to address potential bias (Table 2). When necessary, supplementary study information was obtained by contacting the RCT authors. Questionable RCTs were confirmed by discussion with a third author. Inclusion criteria: the diagnosis of fat liquefaction of post‐surgical wounds was clear; patients were without wound infection, fever or inflammation (indicated by blood leukocyte and postoperative morbidity; also refer to Figure 1). We abstracted data about study design and methods, inclusion and exclusion criteria, patient characteristics, treatment methods and comparative dosage regimens, patient outcomes, cost and adverse events.

Table 1.

Clinical comparison of patients in insulin + hypertonic glucose and control groups

RCT	Treatment		Operation performed	IG † (n)	C ‡ (n)	Time/cost to healing		Clinical indices		Jadad Score §
	IG †	C ‡				IG (d)	C (d)	IG	C
Liu 2001 (22)	50% glucose 20 ml + 8 IU insulin	Saline	Obstetrical and gynecological	13	12	9·46 ± 4·96	17·50 ± 7·51	−	−	3/6
Ding 2006 (23)	50% glucose 20 ml + 3 IU insulin infrared physiotherapy	Saline + infrared physiotherapy	C‐section	25	20	3·51 ± 0·08	10·33 ± 2·42	−	−	3/6
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive
Hu 2008 (24)	50% glucose 20 ml + 4 IU insulin	Saline + secondary suture	General surgery	20	18	13·30 ± 2·80	18·40 ± 8·00	−	−	3/6
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive
Yang 2008 (25)	50% glucose 20 ml + 8 IU insulin	Saline + secondary suture	Gynecological surgery	37	23	10·40 ± 1·20	17·40 ± 2·20	4·1 ± 2·0 ‖;	3·8 ± 2·2 ‖;	3/6
	Butterfly‐shaped adhesive
						108·3 ± 27·1 ¶	607 ± 57·4 ¶	37·8 ± 3·4 ††	38·2 ± 3·0 ††
Chen 2008 (26)	50% glucose 20 ml + 5 IU insulin	Saline	C‐section	33	20	3·36 ± 0·13	9·58 ± 1·42	−	−	3/6
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive
Zh 2009 (27)	50% glucose 20 ml + 2 IU insulin	Saline	C‐section	28	20	9·80 ± 2·30	13·90 ± 5·80	−	−	2/5
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive
Wu 2010 (28)	50% glucose 20 ml + 20 U insulin infrared physiotherapy	Saline + infrared physiotherapy	Obstetrical and gynecological	20	20	9·2 ± 1·7	15·3 ± 3·9	−	−	3/6
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive
Lei 2011 (29)	50% glucose 20 ml + 5 U insulin	Saline	Obstetrical and gynecological	76	29	3·4 ± 0·1	9·6 ± 1·4	−	−	2/5
	Butterfly‐shaped adhesive	Butterfly‐shaped adhesive

There were no differences in incidence of diabetes, severe anaemia and cough after operation between the insulin + hypertonic glucose group and the control group of enrolled studies.

^†Insulin + hypertonic glucose.

^‡Control.

^§Jadad Score (3 items/6 items).

^¶Cost, unit in RMB, shown by mean ± SD.

^‖;Thickness of abdominal fat (cm).

^††Age.

^* P < 0·05 for comparisons of ¶, P > 0·05 for comparison of ‖; and ††.

Table 2.

General information of enrolled RCTs (according to the Cochrane Collaboration's tool for assessing risk of bias)

	Sequence generation	Allocation concealment	Blinding of participants, personnel and assessors	Incomplete outcome data	Selective outcome reporting	Other sources of bias
Liu 2001 (22)	Random table	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Ding 2006 (23)	Random table	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Hu 2008 (24)	Computer random number generator	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Yang 2008 (25)	Coin tossing	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Chen 2008 (26)	Computer random number generator	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Zhou 2009 (27)	Unclear †	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Wu 2010 (28)	Computer random number generator	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No
Lei 2011 (29)	Unclear †	Unclear	Measurement not influenced by lack of blinding	No missing outcome data	All pre‐specified outcomes reported	No

^†High risk.

Figure 1.

Flow chart of study recruitment and data selection. *Only studies from peer‐reviewed journals or conferences with a modified Jadad score of above 4 (6‐item scale) were included; studies that did not meet the criteria set in the search section were excluded; studies without enough information about patients' clinical conditions were excluded. Authors were further contacted if their names appeared more than once in the included studies to rule out duplicate data. ^#6‐item Jadad score.

For primary outcome, we estimated the time from initial treatment to wound healing. Secondary outcomes included thickness of abdominal fat, cost of treatment, patient characteristics and adverse effects. In the included RCTs, the diagnosis of fat liquefaction of a postoperative wound should meet all of the following four criteria (19):

• 1Developed 3–7 days after operation. No subjective symptoms except yellow exudate;
• 2Open or pseudo closure of surgical incision, with fat droplets in exudate;
• 3No redness or tenderness presented, and no signs of necrosis along the incision and subcutaneous tissue;
• 4Large amount of fat droplets from exudates observed under microscope, and no growth of bacteria after three consecutive bacterial cultures.

Statistical analyses were conducted using Review Manager version 5·0 software (Cochrane Collaboration). Pooled weighted mean difference (WMD) and 95% confidence intervals (CIs) were determined by choice between the fixed‐effects or random‐effects model of inverse variance method, whichever was most conservative (20) (showing less efficacy, with a higher P value, and the random‐effects model qualified for this purpose) and shown in forest plot. The latter is a graphical display designed to illustrate the relative strength of treatment effects in multiple quantitative scientific studies addressing the same question (21). Statistical between‐study heterogeneity was assessed by I ² test (22) and χ ² test. Publication bias was assessed by funnel plot 23, 24. An asymmetric funnel indicates a relationship between treatment effect and study size, suggesting the possibility of either publication bias or a systematic difference between smaller and larger studies (‘small study effects'), or the use of an inappropriate effect measure. For all the tests performed, statistical significance was achieved if the P value was < 0·05 (for overall effect of intervention) or < 0·10 (for heterogeneity test, due to the small number of RCTs) (20).

RESULTS

Eight RCTs were included and analysed quantitatively 25, 26, 27, 28, 29, 30, 31, 32 (characteristics in Table 2, selection process in Figure 1), including 414 cases. All wounds were closed by the time of healing. According to diagnosis criteria, efficacy of insulin with hypertonic glucose was evaluated according to days to healing for comparison purpose. Debridement and wash with disinfectant were performed before intervention. Solution of insulin and hypertonic glucose mixture was given by subcutaneous injection along both sides of incision. Solution in control groups was given by flushing. All studies determined wound healing by visualisation of skin regeneration and closure of the incision. Funnel plot indicating selection bias was shown in figure of the corresponding comparison.

Primary outcomes

According to healing criterion, the efficacy of subcutaneous injection of insulin and hypertonic glucose indicated by WMD [95% CI] was −6·33 [−6·66, −5·99] (P < 0·01 for overall effect), with no heterogeneity (P = 0·48, I ² = 0%, Figure 2).

Figure 2.

Comparison of subcutaneous injection of insulin with hypertonic glucose and conventional drainage in postoperative patients with incision fat liquefaction. (A) Forest plot. (B) Funnel plot of enrolled studies for A. Vertical line indicates no difference between compared treatments. Horizontal lines show 95% CIs. Squares indicate point estimates, and the size of the squares indicates the weight of each study in the meta‐analysis. IV Random, random‐effects inverse variance model; CI, confidence interval.

Secondary outcomes

Although in individual studies it was mentioned that there was no significant difference between the two groups in clinical characteristics of patients, such as age and degree of obesity, only one study (28) reported the above items quantitatively (Table 1). There was no report of adverse effects in any of the included studies.

DISCUSSION

Factors contributing to fat liquefaction in postoperative wounds include:

• 1Obesity, with thick abdominal subcutaneous fat (33). 43–75% of wound fat liquefaction occurs in overweight patients (34).
• 2Unnecessary and overuse of electrotome in surgery (35). Heat generated by the electrotome may cause thermal injury in superficial subcutaneous adipose tissue, leading to its partial degeneration. Meanwhile, because of thermal coagulation and thrombosis in adipose tissue capillaries, blood supply diminishes, causing aseptic necrosis of fat tissue and more exudation. Therefore, for general surgeries, its application should be as gentle and quick as possible.
• 3Prolonged exposure of the incision, along with mechanical irritation, such as compression and forceps clipping, is prone to cause more oxidation, decomposition and aseptic inflammation of fatty tissue, and finally fat liquefaction (36).
• 4Old age and chronic diseases, such as anaemia, hypoproteinaemia and diabetes.

Role of insulin and hypertonic glucose in wound healing

Wound healing is a complicated biological process involving chemotaxis, angiogenesis and neovascularisation that comprise synthesis of extracellular matrix proteins and remodelling of tissues. The physiological properties of insulin suggest its potential favourable role in wound healing because of its stimulation of individual cell growth as well as anabolism of the organism as a whole 37, 38, 39. Glucose combined with insulin may also promote protein synthesis, inhibit protein degradation, increase anti‐inflammation capacity of local tissue, promote wound healing and reduce skin scarring. As much as 50% glucose is hypertonic, which may inhibit bacterial growth, prevent oedema of granulation tissue and stimulate its growth. Meanwhile, a butterfly‐shaped adhesive may eliminate dead space by avoiding damage to the problematic wound caused by conventional suture stitches, and thus speed healing (28).

It has been shown pre‐clinically that subcutaneous injection of insulin in mice led to regional longer vessels with more branches, along with increased numbers of associated alpha‐smooth muscle actin‐expressing cells, suggesting the appropriate differentiation and maturation of the new vessels. Also found was that insulin stimulates human microvascular endothelial cell migration and tube formation (38). Insulin has also been reported to have beneficial effects on cell proliferation and protein metabolism in skin donor site wound (40). When topically applied to incision wounds, insulin accelerates re‐epithelialisation and stimulates ‘maturation’ of the healing tissue (41). One clinical study has shown the efficacy of topical application of insulin (indicated by healing rate, mm²/day) in the treatment of non infected acute and chronic extremity wounds regardless of baseline wound size (42). In another study, wounds treated with topical insulin (without hypertonic glucose) healed 2·4 ± 0·8 days faster than the wounds treated with saline (P < 0·001) (43).

Safety of subcutaneous application of insulin combined with hypertonic glucose

Although none of the included studies reported hypoglycaemia, insulin should be used with caution in patients with hypoglycaemia, acute hepatitis, liver cirrhosis, haemolytic jaundice, pancreatitis and nephritis. In our analysis, no secondary wound infection was reported in the insulin–glucose group, showing the antiinfectious efficacy of hypertonic glucose.

Influencing factors

In our analysis, the dose of insulin ranged from 2 to 20 IU. Ideally, we may analyse the efficacy of insulin by dosage so that the optimal dose of insulin can be determined. However, because of the limited number of studies and patients in each study, a stratified analysis would not be much more convincing. According to our data, there does not appear to be a clear dose‐dependant improvement of healing within this range. It has been reported in each enrolled study that there were no differences in the rate of diabetes, severe anaemia, and cough after operation between the insulin–glucose group and control group, which excluded their influence on the result of our analysis. All enrolled studies were carried out in China because there were no qualified studies available elsewhere. The result of incision healing was seldom influenced by patients' in‐hospital activity, and the treatment could not be blinded because of daily wound care by doctors and nurses; therefore blinding was not considered as a bias‐raising factor in our analysis. The concealment of allocation was unclear in all studies, because not much attention has been paid to the notion of allocation concealment in China, and even if the authors carried out the allocation concealment, they did not report. As an effective method to eliminate bias in randomisation process, the notion of concealment should be reinforced among clinical researchers. There might also be possible bias based on individual physician skills and hospital conditions. As were shown by funnel plots, the estimated WMD is likely biased in favour of the insulin and hypertonic glucose treatment due to publication bias. More RCTs of higher quality and larger size are needed for further investigations and more convincing results.

In conclusion, our meta‐analysis suggests that subcutaneous injection of insulin along with hypertonic glucose is effective in the management of aseptic postoperative wounds with fat liquefaction without significant adverse effects or complications, and therefore might be recommended, especially to obese and old patients undergoing operations. Prophylactic application of insulin and hypertonic glucose might be recommended to patients with risk factors contributing to fat liquefaction in postoperative wounds.

ACKNOWLEDGEMENTS

This systemic review was carried out using the recommendations and protocols of the Cochrane Collaboration. The authors would like to thank Dr. Allen Nicholson, Dr. James Bloxton (both from Temple University), Dr. Patricia Dillon (LaSalle University) and Dr. Joshua Breslau (Princeton University) for their critical review and generous support. No funding source was involved. The authors have neither conflict of interest nor competing interest concerning this paper.