Comparison of skin effects of immediate treatment modalities in experimentally induced hydrofluoric acid skin burns

Abstract

Hydrofluoric acid (HF) burns cause immediate damage and painful long‐term sequellae. Traditionally, chelating agents have been used as the initial treatment for such burns. We have introduced epidermal growth factor (EGF) into an HF model to compare EGF with Ca²⁺ and Mg²⁺ treatments; 40 Sprague Dawley rats were divided into five groups. Each rat suffered a 6 × 4 cm² burn induced by 40% HF. Group 1 had no treatment, group 2 had saline injected beneath the burn, group 3 received magnesium sulphate injections, group 4 received calcium gluconate and group 5 received EGF. Specimens were evaluated via planimetry and biopsy at intervals of 4, 8, 24 and 72 hours. Fluid losses were significantly less in the Mg²⁺ and EGF groups. The EGF group had the smallest burn area, least oedema, least polymorphonuclear granulocyte (PMN) infiltration, most angiogenesis and highest fibroblast proliferation of any group (P < 0·005). EGF limited HF damage morphologically and histologically more effectively than Ca²⁺ or Mg²⁺. This finding indicates that HF treatment via growth factors may be an improvement over chelation therapy.

Introduction

Hydrofluoric acid (HF) is among the strongest inorganic acids. HF is used in petrochemical processes, glass etching and mineral processing 1, 2, 3. It is a frequent ingredient in household rust removers 3, 4. The physiopathology of HF burns differs from that of other acid burns. Initially, like other inorganic acids, it has a corrosive effect on superficial tissues. Uniquely, the high permeability of fluoride ions leads to liquefaction necrosis, along with the decalcification and erosion of underlying bone 1, 5. The literature on the treatment of HF skin burns describes the topical application of 2·05% calcium gluconate 6, subcutaneous injections of calcium gluconate and magnesium sulphate 7, 8, intra‐arterial calcium gluconate infusion 9, 10, and intravenous magnesium sulphate infusion 5. The role of epidermal growth factor (EGF) in wound healing is well‐known 11, 12. Currently, research on the effect of EGF on burn wounds is very extensive 13. Although there have been many experimental articles about the effects of EGF on thermal burn injuries, there have not been any articles about the effects of EGF on HF burn injuries.

Current EGF applications related to wounds have led us to investigate its role in a HF burn model. In this study, the treatment of experimental chemical burn wounds created with 40% HF solutions with subcutaneous injections of EGF was compared to treatment with conventional calcium gluconate and magnesium sulphate injections.

Material and methods

Forty male Sprague Dawley rats weighing between 350 and 400 g were used in this study. The National Research Council's guidelines for the care and use of laboratory animals were followed. The Ethical Committee of the Training and Research Hospital approved the experimental protocol. Animals were anaesthetised with an intraperitoneal injection of a combination of ketamine (100 mg/ml) and xylazine (20 mg/ml) at a total dose of 0·02 ml/100 g of body weight.

Experimental protocols

A 6 × 4 cm² area of dorsal hair was removed with an electric shaver, and all procedures were performed under sterile conditions 14. A volume of 60 µl of 40% HF (VWR, Paris, France) was applied to each spot with a calibrated pipette and immediately spread over the delimited surface using a polypropylene stick. After 2 minutes of contact with HF, the skin was carefully blotted dry for 1 minute with a surgical towel 6. The rats were divided into five groups. For the first group (group 1, control, n = 8), no treatment was administered. For the second group (group 2, n = 8), 0·05 ml/cm² saline was injected subcutaneously. For the third group (group 3, n = 8), 0·05 ml/cm² calcium gluconate (Adeka Inc., Istanbul, Turkey) was injected subcutaneously 15. For the fourth group (group 4, n = 8), 0·05 ml/cm² 15% magnesium sulphate solution (Biofarma Inc., Istanbul, Turkey) was injected subcutaneously 8. For the fifth group (group 5, n = 8), 100 µ/kg recombinant human epidermal growth factor (rhEGF, Peprotech, London, UK) diluted with saline was injected 16. All applications were performed by blinded examiners. After the treatment, in all groups, the wound areas were covered with sterile, dry gauze containing no topical agents. Specially designed burn outfits were placed on the rats and stabilised to ensure that there would be no pressure on the wounds. For analgesia, 30 mg codeine phosphate and 500 mg paracetamol were added to every 500 ml of drinking water.

Body weight assessment

To assess the bodyweight loss of the rats, their wet weights at 0 hours (before burn) and 72 hours (post‐burn) were measured with an electronic weighing machine (Electronic weighing system, Naugra Export, Ambala, India). Post‐burn biopsies were weighed and added to the wet weight at 72 hours. Before all the rats were euthanised, their final weights were realised, and the related data were obtained.

Analysis of burned area (cm²)

The wound area of the rats was photographed planimetrically from a 10 cm distance at 4, 8, 24 and 72 hours with a camera (Sony Cybershot DSC‐T100, London, UK). The photographs obtained of the burn wound area from every rat were analysed with the Image‐Pro Express program (Media Cybernetics Inc., Philadelphia,; Figure 1).

Figure 1.

The evaluation of burn areas (A) at 0 hours and (B) at 8 hours with Image‐Pro Express program.

Histopathological assessment

In all groups, 0·5 × 1 cm² rectangular biopsies were obtained at 4 hours from the upper left corner, at 8 hours from the lower left corner, at 24 hours from the lower right corner and at 72 hours from the center of the burned area. These specimens were taken for histopathological examination at 4, 8, 24 and 72 hours. After the fixation of the skin tissue with 4% paraformaldehyde, 4‐μ sections were sliced using a microtome. The sections were painted with haematoxylin and examined under a high‐performance light microscope (×10, ×20 and ×100 magnifications) by two blinded examiners. According to modified Verhofstad scoring 17, the samples were examined based on the following seven parameters: neutrophil [polymorphonuclear granulocytes (PMN)], lymphocyte density, vascularity, fibroblast proliferation, the amount of oedema, collagen density and necrosis (Table 1).

Table 1.

Histopathologic evaluation Verhofstad modified scoring table

Score	Necrosis	PMN	Lymphocyte	Oedema	Collagen density	Fibroblast proliferation	Vascularity
0	None	Normal	Normal	None	None	None	None
1	Superficial	Light	Light	Light	Light	Light	Light
2	Pronounced	Pronounced	Pronounced	Pronounced	Pronounced	Pronounced	Pronounced
3	Massive	Massive	Massive	Dense	Dense	Dense	Dense

PMN, polymorphonuclear granulocytes.

Statistical assessment

The statistical analysis of the obtained data was performed using the Statistical Package for Social Sciences (SPSS) 14·00 (SPSS, Inc., Chicago, IL) program. Data were analysed using the Kruskal–Wallis, Friedman, Wilcoxon and Mann–Whitney U‐tests. P values <0·005 were regarded as statistically significant.

Results

Results of weight assessment

The 0 and 72‐hour mean data values in group 1 were 369·08 and 346·08 gm, those in group 2 were 376·08 and 355·03 gm, those in group 3 were 376·05 and 3 62·06 gm, those in group 4 were 370·00 and 358·08 gm and those in group 5 were 376·06 and 369·06 gm. According to these data, there was a statistically significant difference in bodyweight loss between groups 1 and 2 (P < 0·005). The bodyweight losses of groups 1 and 2 were statistically significantly higher than those of the other groups (P < 0·005). However, there were no statistically significant differences between groups 3 and 4 (P > 0·005). The bodyweight losses of groups 3 and 4 were statistically significantly higher than that in group 5 (P < 0·005, Figure 2).

Figure 2.

The bodyweight loss of the groups.

Results of burn area analysis

The average 4‐hour data values were 23·03 cm² in group 1, 23·02 cm² in group 2, 17·04 cm² in group 3, 21·06 cm² in group 4 and 15·04 cm² in group 5. According to the burned area comparison, the data values in groups 1 and 2 were statistically significantly higher than those in the other groups (P < 0·005). The average value of group 4 was significantly higher than that of groups 3 and 5 (P < 0·005). The average value of group 3 was also statistically significantly higher than that of group 5 (P < 0·005).

The average 8‐hour data values were 22·06 cm² in group 1, 22·04 cm² in group 2, 16·05 cm² in group 3, 19·07 cm² in group 4 and 12·09 cm² in group 5. According to the burned area comparison, the average data values in groups 1 and 2 were statistically significantly higher than those in the other groups (P < 0·005). The average value of group 4 was significantly higher than those in groups 3 and 5 (P < 0·005). The average value of group 3 was also statistically significantly higher than that in group 5 (P < 0·005).

The average 24‐hour data values were 21·09 cm² in group 1, 21·09 cm² in group 2, 15·07 cm² in group 3, 17·08 cm² in group 4 and 10·08 cm² in group 5. According to the burned area comparison, the average data values in groups 1 and 2 were statistically significantly higher than those in the other groups (P < 0·005). There was no statistically significant difference between groups 3 and 4 (P > 0·005). The average values of groups 3 and 4 were also statistically significantly higher than that of group 5 (P < 0·005).

The average 72‐hour data values were 21·02 cm² in group 1, 21·02 cm² in group 2, 14·04 cm² in group 3, 16·04 cm² in group 4 and 9·04 cm² in group 5. According to the burned area comparison, the average values of groups 1 and 2 were statistically significantly higher than those in the other groups (P < 0·005). There was no statistically significant difference between groups 3 and 4 (P > 0·005). The average values of groups 3 and 4 were also statistically significantly higher than that of group 5 (P < 0·005, Figure 3).

Figure 3.

The distribution of burned areas by group based on time.

Results of pathological assessment

All the groups were scored histopathologically according to the modified Verhofstad scoring system.

According to the histopathological data at 4 hours, the results for PMN, lymphocytes, oedema, collagen density, fibroblast proliferation and vascularity showed that there were no statistically significant differences between groups (P > 0·005). Also, the necrosis results were not statistically significantly different among groups 1, 2, 3 and 4 (P > 0·005). However, the necrosis results of groups 1, 2, 3 and 4 were statistically significantly higher than that of group 5 (P < 0·005, Table 2).

Table 2.

The comparison of histopathological data among groups at 4 hours^{*, †}

Group	Necrosis	PMN	Lymphocyte	Oedema	Collagen density	Fibroblast proliferation	Vascularity
Control group	2·87 ± 0·35	0	0·12 ± 0·35	1·37 ± 0·51	0	0	0
Sham group	2·87 ± 0·35	0·12 ± 0·35	0	1·5 ± 0·53	0	0	0
Calcium group	2·5 ± 0·53	0·12 ± 0·35	0	2·12 ± 0·6	0	0	0
Magnesium group	2·75 ± 0·46	0·25 ± 0·46	0	2·12 ± 0·6	0	0	0
EGF group	1·25 ± 0·46	0	0	1·37 ± 0·74	0	0·37 ± 0·51	0

EGF, epidermal growth factor; PMN, polymorphonuclear granulocytes.

^*Neutrophil (PMN).

^†±Standard deviation (SD).

According to the histopathological data at 8 hours, the results for lymphocytes, oedema, collagen density, fibroblast proliferation and vascularity showed that there were no statistically significant differences between groups (P > 0·005). Also, the necrosis results were not statistically significantly different among groups 1, 2, 3 and 4 (P > 0·005). However, the necrosis results of groups 1, 2, 3 and 4 were statistically significantly higher than those of group 5 (P < 0·005). The PMN results of groups 1 and 2 were statistically significantly higher than those of the other groups (P < 0·005). The fibroblast proliferation results of group 5 were statistically significantly higher than those of the other groups (P < 0·005, Table 3).

Table 3.

The comparison of histopathological data among groups at 8 hours^{*, †}

Group	Necrosis	PMN	Lymphocyte	Oedema	Collagen density	Fibroblast proliferation	Vascularity
Control group	2·75 ± 0·46	1·12 ± 0·35	0·12 ± 0·35	1·37 ± 0·51	0	0	0
Sham group	2·87 ± 0·35	1·12 ± 0·35	0·12 ± 0·35	1·5 ± 0·53	0	0	0
Calcium group	2·5 ± 0·53	0·12 ± 0·35	0·12 ± 0·35	2·12 ± 0·6	0	0·12 ± 0·35	0
Magnesium group	2·75 ± 0·46	0·25 ± 0·46	0·25 ± 0·46	2·12 ± 0·6	0	0·12 ± 0·35	0
EGF group	1·12 ± 0·35	0	0	1·37 ± 0·74	0·37 ± 0·51	1	0

EGF, epidermal growth factor; PMN, polymorphonuclear granulocytes.

^*Neutrophil (PMN).

^†±Standard deviation (SD).

According to the histopathological data at 24 hours, the results for vascularity showed that there were no statistically significant differences between groups (P > 0·005). The necrosis results for groups 1, 2 and 5 were statistically significantly higher than those of groups 3 and 5 (P < 0·005). Also, the necrosis results for group 3 were statistically significantly higher than those of group 5 (P < 0·005). The PMN results for groups 1 and 2 were statistically significantly higher than those of the other groups (P < 0·005). The PMN results for group 4 were statistically significantly higher than those of groups 3 and 5 (P < 0·005). The lymphocyte results for groups 1, 2, 3 and 4 were statistically significantly higher than those of group 5 (P < 0·005). Also, the oedema results for groups 3 and 4 were statistically significantly higher than those of group 5 (P < 0·005). The collagen density results for group 5 were statistically significantly higher than those of the other groups (P < 0·005). The fibroblast proliferation results for group 5 were statistically significantly higher than those of the other groups (P < 0·005). The fibroblast proliferation results of groups 3 and 4 were statistically significantly higher than those of groups 1 and 2 (P < 0·005, Table 4).

Table 4.

The comparison of histopathological data among groups at 24 hours^{*, †}

Group	Necrosis	PMN	Lymphocyte	Oedema	Collagen density	Fibroblast proliferation	Vascularity
Control group	2·37 ± 0·51	1·37 ± 0·51	1·25 ± 0·46	2·5 ± 0·53	0·25 ± 0·46	0·37 ± 0·51	0
Sham group	2·5 ± 0·53	1·37 ± 0·51	1·25 ± 0·46	2·37 ± 0·74	0·25 ± 0·46	0·37 ± 0·51	0
Calcium group	1·87 ± 0·64	0·12 ± 0·35	1 ± 0·75	1·62 ± 0·51	0·62 ± 0·51	1·25 ± 1·03	0
Magnesium group	2·25 ± 0·46	0·75 ± 0·88	1·25 ± 0·46	1·62 ± 0·51	0·37 ± 0·51	1·25 ± 1·03	0
EGF group	1·12 ± 0·35	0·12 ± 0·35	0·25 ± 0·46	1 ± 0·53	1·5 ± 0·53	2·37 ± 0·74	0·37 ± 0·51

EGF, epidermal growth factor; PMN, polymorphonuclear granulocytes.

^*Neutrophil (PMN).

^†±Standard deviation (SD).

At 72 hours, the EGF group showed a statistically significantly lower rate of necrosis, neutrophils, lymphocytes and oedema (P < 0·005), while collagen density, fibroblast proliferation and vascularity were increased significantly as compared with the other groups (P < 0·005). In the calcium group, at 72 hours, the necrosis rate, neutrophils, lymphocytes and oedema were not statistically significantly different from those in the magnesium group (P > 0·005). The neutrophils, lymphocytes and oedema results of the control and sham groups were statistically significantly different from those of the calcium and magnesium groups (P < 0·005). However, the neutrophils, lymphocytes and oedema results of the calcium and magnesium groups were statistically significantly different from those of the EGF group (P < 0·005). The collagen density, fibroblast proliferation and vascularity results for the calcium and magnesium groups were statistically significantly different from those of the control and sham groups (P < 0·005). However, the collagen density, fibroblast proliferation and vascularity results of the EGF group were statistically significantly different from those of the other groups (P < 0·005, Table 5).

Table 5.

The comparison of histopathological data among groups at 72 hours^{*, †}

Group	Necrosis	PMN	Lymphocyte	Oedema	Collagen density	Fibroblast proliferation	Vascularity
Control group	2·37 ± 0·51	2·25 ± 0·46	1·87 ± 0·64	1·75 ± 0·46	0·5 ± 0·53	0·75 ± 0·7	0·37 ± 0·51
Sham group	2·5 ± 0·53	2·25 ± 0·46	1·87 ± 0·64	2 ± 0·53	0·37 ± 0·51	0·62 ± 0·74	0·37 ± 0·51
Calcium group	1·62 ± 0·51	0·62 ± 0·74	1·12 ± 0·64	1·25 ± 0·7	1·62 ± 0·91	1·62 ± 1·06	1·62 ± 0·74
Magnesium group	1·75 ± 0·7	0·75 ± 0·88	1·5 ± 0·53	1·25 ± 0·7	1·25 ± 1·03	1·37 ± 1·06	1·45 ± 0·7
EGF group	1	0·12 ± 0·35	0·25 ± 0·46	1 ± 0·53	2·75 ± 0·46	2·62 ± 0·74	2·62 ± 0·51

EGF, epidermal growth factor; PMN, polymorphonuclear granulocytes.

^*Neutrophil (PMN).

^†±Standard deviation (SD).

Discussion

Many growth factors are involved in the processes of normal wound healing. Growth factors have therefore been proposed as therapeutic agents to promote wound repair 18, 19, 20, 21, 22. EGF is a polypeptide composed of 53 amino acids. It was first isolated from the mouse submaxillary gland by Stanley Cohen in 1962 as part of his Nobel Prize winning work with growth factors 19. The investigation of the role of exogenous EGF in acute wound healing began as early as 1973 with in vivo models of corneal epithelial hyperplasia after injury in response to topical EGF 23. Enhanced wound healing has also been noted in dermal wounds treated with topical or subcutaneous EGF 24. Franklin and Lynch recorded faster epithelial regeneration and less scar contracture in EGF‐treated wounds in rabbits 25. With the introduction of recombinant human EGF in the 1980s, the range of studies increased to include burn wounds. EGF is secreted by platelets, macrophages, monocytes and fibroblasts and acts in an autocrine and paracrine manner on epidermal cells, smooth muscle cells and fibroblasts. Specifically, EGF interacts with its receptor across the entire epidermis, particularly in the basal layer, promoting epithelial growth through the activation of several pathways. The binding of EGF to its receptor, results in the prompt dimerization and autophosphorylation of the latter. This process activates the mitogen‐activated protein kinase pathway, ultimately affecting the phosphorylation of numerous transcription factors and calcium release via activated protein kinase C. EGF also promotes epidermal regeneration and corneal epithelialisation via a number of actions. Such actions include enhancing epithelial cell proliferation and migration to the wound, stimulating the production of proteins such as fibronectin, and increasing the number of fibroblasts in the wound 26.

Studies with partial‐thickness human burn wounds and topical EGF confirmed a decrease in wound healing time 27. In thermal wound models in rats, it has been shown that EGF is effective, but the healing effect on chemical burns is not known 25, 27. Wang SL et al. investigated the acceleration of burn wound healing with the use of pomades containing rhEGF and found a significant acceleration of burn wound healing 28. Alemderoğlu et al. studied the effect of kitosan gel combined with EGF on experimental second‐degree burn wounds on rats. In the group in which the kitosan gel and the EGF were combined, faster and better epithelisation was observed 29. In our study, we also found that the epithelisation of burn areas was faster and better with EGF applications.

Isabelle Robin et al. treated experimental burn areas created on rats with 40% HF solutions by using 2·05% calcium gluconate gel and observed successful wound healing in 17 days. Topical therapy with 2·05% calcium gluconate gel is considered to be the first‐aid treatment for accidental HF skin burns 6. Seyb et al. studied subcutaneous injections of magnesium and calcium salts in the treatment of HF burns and had better results than were seen with the conventional topical treatments 7. For this reason, magnesium sulphate, calcium gluconate and EGF were applied subcutaneously in our study. In our opinion, the subcutaneous application of EGF affects wounds both locally and systemically. The works described above have shown that angiogenesis begins and collateral veins start to enlarge after 24 hours and that 3–7 days are required for the smallest capillary enlargement 30. Therefore, histopathological assessment was performed in our study to understand the changes in vascularity, oedema, collagen density, fibroblast proliferation and polymorphic nuclear leucocyte inflammatory activity, which are very important in wound healing. After 72 hours, there were no statistically significant differences between the parameters of the calcium gluconate and magnesium sulphate groups. However, parameters such as collagen density, fibroblast proliferation and vascularity, which demonstrated wound healing, were statistically significantly higher in the EFG group than in the other groups (Figure 4). In addition to reduced oedema and necrosis, there are signs that fibroblast and collagen depositions are higher in EGF‐treated animals. This shows that wound healing is started early by EGF and that the harmful effects of HF are prevented by EGF.

Figure 4.

Histopathological images of all groups at 72 hours. The yellow arrow shows oedema. The blue arrow shows collagen density. The red arrow shows vessels. The epithelial damage and oedema have been increased in groups 1 and 2. Epithelial thickness and collagen density have been increased in groups 3, 4 and 5. Vascularity has been increased in group 5 [haematoxylin and eosin (H&E), ×20 magnification].

Williams et al. compared the effectiveness of intravenous magnesium and that of intradermal calcium gluconate treatment in HF burns in rats. For third‐degree burns, high‐dose intravenous magnesium is more effective than intradermal calcium gluconate. High‐dose IV magnesium sulphate reduces the severity of HF burns as compared with conventional intradermal calcium gluconate therapy 5. Wang LF et al. evaluated the effect of calcium‐ and magnesium‐containing biological dressings and isotonic saline dressings on experimental HF burns in rats planimetrically and histopathologically at 4, 8, 24 and 72 hours. As a result, they proposed the use of such biological dressings containing calcium and magnesium in the emergency room or after the treatment of HF burns 31. Our study has shown that the histopathological results of the magnesium and calcium groups were statistically significantly better than those of the control and sham groups. In light of these findings, the calcium and magnesium groups, which were using some of the cheapest burn injury treatment methods, would be beneficial when used on hydrofluoric burn injuries. However, these methods are not more effective than EGF.

The major reason for bodyweight loss after a burn is fluid loss 32. In this study, to evaluate the body weight of rats with the experimental burns caused by 40% HF, the animals were weighed at 0 and 72 hours, and the authors observed a significant prevention of bodyweight loss in the EGF‐treated group (P < 0·005). These results have shown us that EGF is beneficial in terms of preventing fluid loss. All rats reached for the water, but the rats that had lost a greater amount of body weight had very slow movements. If the experiments had lasted another week, the rats could have faced mortality.

Conclusion

Industrial development and the extensive use of chemicals containing HF, both in factories and household applications, have increased the number of patients with HF burns. In the literature on the treatment of burns caused by this potent acid, there are few alternatives. The effects of calcium gluconate and magnesium sulphate have been proven. The authors believe that employing EGF in the first stage of HF burn treatment will be useful. The disadvantage of EGF use is its high price and the difficulty involved in obtaining it. In the absence of EGF, the use of magnesium and calcium sulphate is the treatment of choice.