The Effects of Hyperglycemia on Skin Graft Survival in the Burn Patient “The authors elected to determine the relative effects of hypergly- cemia andlor elevated wound Gram-positive bacterial counts on success of skin graft survival in 74 bur patients. Results of serum glucose and quantitative wound biopsies on the day of ‘admission and on postoperative day 4 were charted. Cases were separated into the following groups for analysis: normoglyeemia plus normal bacterial counts, elevated bacterial counts only, hyperglycemia only, and hyperglycemia plus elevated bacterial counts, Successful graft “take” was defined as survival of 80% to 100% of the grafted area as astessed on postoperative day 4. Significant results included decreased incidence of graft take for groups with hyperglycemia only (62.5%), elevated bacterial counts only (63.3%), a5 well as hyperglycemia plus elevated bacterial counts (54.5%) when compared with the group with ‘ormogiycemia plus normal bacterial counts (92.8%: p = 0.020, ‘p= 0.042, p = 0.012 respectively) for physiological parameters ‘measured on postoperative day 4 only. Additionally, incidence of raft take was reassessed and found to be decreased significantly in groups with hyperglycemia (60.0%) vs. groups with normogly- emia (84.6%), regardless of Gram-positive bacterial counts (p 0.034) Mowiav , Andrews K, Milner tal. The effects of hyperglycemia on skin raft survival in the burn patient. nn Plast Surg 2000;45:629-632 From *Seutharn Ilinois University, Institute for Plastic and Reconstructive Surgery, Springfield, IL: and the tDepartrent of Surgery and fStriners Buns Hospital, Galeston, Recoved May 5, 2000, publication Jul 7, 2000, Address corespondence and reprint requests {0 Or Mules, Insitute for Paste and Reconstructive Surgery, PO Box 19653, Springtild IL 62794, i revised form Jul 7, 2000. Accepted for Hyperglycemia, as an acute phase response, has been a consistently observed phenomenon in burn patients. The increase in glucose levels has been attributed by some clinicians to increased rates of glucose production resulting from i creased rates of gluconeogenesis as well as to insulin resistance. Interestingly, the degree of gluconeogenesis has not been correlated with the severity of body burn or the number of days after Arian Mowlavi, MD* Kris Andrews, MD* Stephen Milner, MB, BS, BDS, FRCS (Ed)* D. N. Herndon, MDT J. P. Heggers, PhD, CWS, FAAM#t bum injury.' To provide precursors for gluconeo- genesis, substantial protein catabolism occurs,’ re- sulting in potential inhibition of wound healing. In addition, hyperglycemia and its predisposition for ized that \creased serum ver, others have hypoth: insulin resistance, which induces glucose levels, should provide for increased avail- ability of glucose at the wound, presumably allow- ing for enhancement of local healing and immunity.’ A supporting study has shown that administration of a combination of exogenous in- sulin and glu tively, results in shorter graft donor healing time.* To establish the effects of hyperglycemia on skin graft survival in the burn patient, we elected to perform a retrospective study in which serum glucose levels and local Gram-positive bacterial counts would be assessed as two independent physiological parameters on the day of admission and postoperative day 4. We would limit our study of hyperglycemia to the acute phase, ex- se, when administered postopera cluding from our study patients known to have diabetes mellitus. Bacterial counts would be as- sessed to determine the effects of local bacterial colonization on graft survival and, more impor- tant, to identify any contributing effects. The outcome measure of graft “take” on postoperative day 4 would be assessed with respect to the previously mentioned physiological parameters. Patients and Methods Cases of consecutive patients admitted to the Shriners Burn Institute who underwent serum glucose-level testing as well as biopsy of the bun for quantitative cultures on the day of admission and of the wound beds on postoperative day 4 Copyright © 2000 by Lippincott Williams & Wilkins, Inc. 629Annals of Plastic Surgery were reviewed. Candidates for the study in- cluded patients with 10% or more total body surface area (TBSA) burn. Exclusion criteria for our study included (1) wound cultures pos for Gram-negative bacteria, (2) history of diabetes mellitus type 1 or type 2, (3) insulin administra- tion during hospital stay, (4) inhalation injury, and (5) hematoma or seroma formation after sur- gery for skin grafting. Seventy-four acute burn cases were included in our data analysis. Patients were separated into the following exposure groups: normoglycemia plus normal bacterial counts (NG/NB), elevated bacterial counts only (NG/EB), hyperglycemia only (HG/NB), and hy- perglycemia plus elevated bacterial count (HG/ EB). Hyperglycemia was defined as serum glucose = 126 mg per deciliter.* Elevated bacte- rial counts were defined as burn or wound levels = 1 X 10° Gram-positive organisms per gram of wound tissue,” whereas normal bacterial counts were defined as levels less than this value. Fi- nally, successful graft take was defined as sur- vival of 80% to 100% of the grafted area. All patients underwent surgical intervention within 48 hours of sustaining burn injury. We applied the results of quantitative wound biop- sies on the day of admission to establish baseline wound contamination indices. It should be noted that we did not treat wounds contaminated with levels = 1 x 10° Gram-positive organisms per gram of wound tissue with definitive autologous skin grafting in patients whose biopsy results were available before surgery. An independent staff statistician assigned pa- tients to four physiological condition groups and analyzed the data with Sigma Stat (version 20, 1995, Jandel Scientific). Chi-squared analysis was performed to evaluate the significance of both datasets with respect to incidence of graft take. If a dataset was found to be significant, then Fish- er’s exact test was used for specific comparisons between any two groups within that dataset with respect to incidence of graft take. Results Initial mean percent TBSA burn was not found to be significantly different between the NG/NB (30 630 Volume 45 / Number 6 / December 2000 Table 1. Incidence of Graft Take Relative to Postoperative Day 4 Physiological Conditions Elevated Incidence Bacterial Sample of Graft Hyperglycemia’ Count” Size ‘% NGINB — = 28 (92.8 NG/EB — + u HGINB + - 24 HGIEB + + u 545 raed analysis, dataset dotermined to be (specific group significant and acceptable for Fisher's exact t comparison). p ~ 0.025. “Average of serum ghicose =126 mg per deciliter. “Perioperative quantitative culture from areas of graft loss with mos 10° organisms per gram of tissue obtained on postoperative day 4 “Percent of sample with 80% to 100% graft survival on postoperative day 4 NG/NB = normal scrum glucose plus normal bacterial count: NG/EB = normal sorum glucose plus elevated bac HGINB ~ hyperglycemia plus normal bacterial co hyperglycemia plus elevated bacterial count: ~ present absent: + + 18%), NG/EB (31 + 16%), HG/NB (32 + 21%), and HG/EB (33 + 14%) groups. The value of serum glucose levels ranged from 82 to 225 mg per deciliter. The bacterial counts ranged from 1 x 10' to 1 X 10° Gram-positive organisms per gram of tissue. Isolated organisms included meth- icillin sensitive Staphylococcus aureus and S. epidermidis. ‘The results for postoperative day 4 data re- vealed graft takes for the groups to be: NG/NB, 92.8%; NG/EB, 63.6%; HG/NB, 62.5%; and HG/ EB, 54.5% (Table 1). These results are significant based on chi-squared analysis of the dataset (p 0.025). With respect to group comparisons, sig- nificant results, as determined by Fisher's exact tost, include decreased incidence of graft take for groups with hyperglycemia only (62.5%), ele- vated bacterial counts only (63.6%), and hyper- glycomia plus elevated bacterial counts (54.5%: when compared with the group with normogly- cemia plus normal bacterial counts (92.8%; p = 0.020, p = 0.042, p = 0.012 respectively). No significant difference was noted between groups with hyperglycemia only, elevated bacterial counts only, and hyperglycemia plus elevated bacterial counts (Table 2). Therefore, the inc dence of graft take was reassessed and found to be significantly decreased for all patients with hy- perglycemia (60.0%) vs. all patients with normo- glycemia (84.6%), regardless of bacterial counts,Mowiavi et al: Hyperglycemia and Skin Graft Survival Table 2. Specific Group Comparisons of Graft Take Within Postoperative Day 4 Dataset Sample Incidence of Group Size Graft Take, % pValue’ NGINB vs. HGINB 28 vs. 24 vs. 0.020 NG/NB vs. NG/EB 28 vs. 11 92.8 vs. 63.6 0.032 NGINB vs. HG/EB 28 vs. 11 92.8 vs. 54.5 o.012 HGINB vs. HG/EB 24s. 11 625 vs. 54.5 0.721 NGIEB vs. HG/EB tvs. 63.6 vs. 54.5 1.000. NGIEB vs. HG/NB. ys. 24 63.6 vs. 62.5 1.000. NG/NB + NG/EB vs. HG/NB + HG/EB 39 vs. 35, 84.6 vs. 60.0 0.034 “Based on Fisher's exact tots, NGINB = ‘serum glucose plus nocmal bacterial count: NG/EB = normal serum glucose plus elevated bacterial count; HGINB = hyperglycemia plus normal bacterial cou for physiological parameters measured on post- operative day 4 (p = 0.034; see Table 2). The results for the day of admission revealed a trend for decreased incidence of skin graft sur- vival for patients with hyperglycemia only (69.2%), elevated bacterial counts only (80.0%), and hyperglycemia plus elevated bacterial counts (60.0%) when compared with the group with normoglycemia plus normal bacterial counts (90.0%). However, chi-squared analysis deter- mined this dataset to not be significant (p = 0.069). Discussion Several investigators have implied that the hy- perglycemic state can predispose to bacterial in- fection. Specifically, it has been reported that hyperglycemia may have enhanced bacterial pro- liferation of the staphylococci." In a study of 214 patients with septicemia, 80% of the patients with blood glucose levels more than 130 mg per deciliter showed Gram-positive organisms on blood cultures.” Recent observations have suggested a relation- ship between hyperglycemia and poor skin graft survival. However, we elected to determine the relative independent effects of hyperglycemia on success of skin graft survival in burn patients. The significant findings of our retrospective study are consistent with our clinical observa- tions and confirm our hypothesis that hyperglyce- mia in the early postburn phase is associated with graft loss, irrespective of Gram-positive wound bac- terial counts. Our findings indicate that hypergly- cemia and wound contamination with high counts HG/EB = hyporglycomia plus elevated bacterial count. of Gram-positive bacteria are relatively indepen- dent risk factors, associated with a comparable degree of graft loss in the burn patient. During the first week after burn injury, when insulin resistance is prevalent, local glucose up- take is low and we think that hyperglycemia actually impairs wound healing and induces graft loss. The specific mechanism by which hypergly- cemia induces graft loss has yet to be elucidated, but possibilities to consider include local tissue edema with inhibition of diffusion of wound- healing factors/agents at the wound site during the early postburn phase. Despite the body’s attempt to preserve available glucose for wound repair, the wounded cells, similar to other cells, may be unable to use this medium during the early postburn period. In- stead, elevated levels of glucose during the few days immediately after injury may induce addi- tional edema and may subsequently inhibit dif- fusion of other metabolites at the wound site. As insulin resistance resolves, however, over 1 to 2 weeks, the available glucose is used by the local tissue. One study of glucose use by skin burn wounds in rats best depicts this hypothesis. When the rate of local glucose uptake was mon- itored, the investigators observed no use at 0, 6, and 24 hours after injury. Instead, glucose uptake was only increased after 3 weeks." At this time, local glucose uptake increases and glucose is then used for wound repair. In an extensive literature search on the topic of hyperglycemia in association with skin grafts, few articles were identified. One of these re- ported a decrease in graft donor site healing time associated with exogenous insulin and glucose 631‘Annals of Plastic Surgery infusion postoperatively." However, this report failed to explore the possibility that exogenous insulin may not be affected by the resistance that is observed with endogenous insulin. In the cur- rent study, only charts for patients with serum glucose levels of 225 mg per dec reviewed to limit patients who received exoge- nous insulin therapy. Although we do not have an established protocol for treatment of hypergly- cemia, the trend is to treat patients with serum blood glucose levels higher than 200 mg per deciliter. Regardless of glucose level, we ex cluded all patients who received insulin therapy during their hospital stay to avoid confounding effects as suggested by a previous study. We think that a prospective study of the effects of glucose control therapy on skin graft survival the burn patient is warranted. Findings of the current study may direct treatment in a manner that will facilitate increased skin graft survival for burn patients. In addition, we think that elucidation of the mechanism by which hyper- glycemia hinders graft take is warranted. ‘er of less were 632 EE Volume 45 / Number 6 / December 2000 References 10 Wolfe RR. Burn injury and increased glucose production, J Trauma 1979;10(11 Suppl):898—890 Hoggers JP. Goodheart R, Carino E, ota. Is the limulus amebo- {te lysate the sole prodctor of septic episodes in major thermal injuries? J Bum Care Rehabil 1998;19(6:512-515 Carter EA. Insulin resistance in burns and trauma. Nutr Rev 1098:565(1):$170-S176 Pierro BJ, Barrow RE, Hawkins HK, etal. Effects of insuli fon wound healing. J Trauma 1998:44( 5 Peters AL, Schriger DL. The new diagnostic criteria for diabetes: the impact on management of diabetes and macro- vascular risk factors. Am J Med 1998;105(1A):15S-19S Phillips LG, Heggers JP. Robson MC, et al. The fect of lndogenous skin bacteria on burn wound infection. Ann Plast Surg 1989:23(1):35-38 Robson MC, Heggers JP. Variables in host resistance pe taining to sopticomia. I: blood glucose lovel. J Am Goriatr ‘Soc 1969:17(10):901-996 Robson MC, Heggers JP. Elfect of hyper vival of bacteria, Surg Kucan JO, Heggers JP, Robson MC. Blood glucose ja. Burns 19 an aid in the diagnosis of m1113 Carter EA. Tompkins RG. Babich JW, etal. Thermal injury in ats alters glucose utilization by skin. wound. and small Intestine, but not by skeletal muscle. Metabolism 1996; 45(0):1161-1167