Int J Colorectal Dis (2004) 19:250–257DOI 10.1007/s00384-003-0533-9

Accepted: 21 July 2003Published online: 24 September 2003© Springer-Verlag 2003

Abstract Background and aims:Anastomotic dehiscence followingcolorectal surgery is a significantcause of morbidity and mortality.Phenytoin has wound-healing pro-moting and collagenase inhibitoryeffects. This study assessed these ef-fects on healing of experimental co-lonic anastomoses in a rat model.Materials and methods: Ninety Wis-tar rats weighing 240–290 g were di-vided into six groups: 3rd-day con-trol group (n=15), 3rd-day oral ad-ministration of phenytoin (n=15),3rd-day rectal administration of phe-nytoin (n=15), 7th-day control group(n=15), 7th-day oral administrationof phenytoin (n=15), and 7th-dayrectal administration of phenytoin(n=15). In oral phenytoin groups theagent was given at 10 mg/kg dailyper orogastric route by 4-F fine feed-ing catheter; in rectal phenytoin RAPgroups the agent was administered at10 mg/0.5 cc daily to the anastomos-es transrectally via a fine anal cathe-ter. Results: There were significantlyhigher anastomosis bursting pressurevalues and hydroxyproline contentsin phenytoin groups than in controls.

In histopathological examination itwas seen that phenytoin treatmentcaused greater collagen deposition,fibroblast, and blood vessel ingrowththan in controls. Immunohistochemi-cal analysis showed the stimulatoryeffect of phenytoin in expression ofvascular endothelial growth factorand basic fibroblast growth factor.Anastomosis bursting pressure, his-topathological analysis, hydroxypro-line content, and immunohistochemi-cal results were better in the groupswith rectal administration than inthose with oral administration. Con-clusion: These results had showed usthat phenytoin administration result-ed in enhanced stability of colonicanastomoses during the first postop-erative week and rectal administra-tion showed better results than oraladministration.

Keywords Colon anastomosis ·Phenytoin · Vascular endothelialgrowth factor · Basic fibroblastgrowth factor

O R I G I N A L A RT I C L E

Mustafa TuranSerpil Unver SaraydinEmel CanbayKursat KaradayiEray BulutÖge CetinkayaSahande ElagözMetin Sen

Positive effects of phenytoin on experimental colonic anastomoses

Introduction

Anastomotic dehiscence following colorectal surgery is asignificant cause of morbidity and mortality [1, 2]. Therate of clinically apparent anastomotic leakage rangesfrom 3.4% to 8% [1, 3]. At least one-third of deaths fol-lowing colorectal surgery are attributed to anastomotic

leakage [1, 3]. Despite the progress made to this dayanastomotic dehiscence following colonic anastomosis isa serious clinical complication. It has been reported thatapproximately 40–50% of patients treated by phenytoin,an anticonvulsant, develop esthetically disfiguring en-largement of gingivae [4, 5]. Phenytoin may also signifi-cantly promote wound healing following topical applica-

M. Turan (✉) · E. Canbay · K. KaradayiM. SenDepartment of General Surgery, Faculty of Medicine,Cumhuriyet University,58140 Sivas, Turkeye-mail: mturan_1999@yahoo.comTel.: +90-346-2250434Fax: +90-346-2262162

M. TuranGokcebostan Mah. 2. Sok. Hurriyet Apt. Daire No 2,58050 Sivas, Turkey

S. Unver Saraydin · E. BulutDepartment of Histology, Faculty of Medicine,Cumhuriyet University,58140 Sivas, Turkey

Ö. CetinkayaDepartment of Biochemistry, Faculty of Medicine,Cumhuriyet University,58140 Sivas, Turkey

S. ElagözDepartment of Pathology, Faculty of Medicine,Cumhuriyet University,58140 Sivas, Turkey

tion [6, 7]. Houck et al. [8] observed that doubling timeof diploid human fibroblasts decreased by 38% in thepresence of phenytoin 5 mg/l. Subsequently phenytoinwas found to promote the healing of dental extractionsockets, to increase the tensile strength of experimentalskin and corneal wounds [6, 7, 8]. To date the effects ofphenytoin on healing of colonic anastomoses have notbeen evaluated. This study investigated the effects ofphenytoin on the healing of experimental colon anasto-moses.

Materials and methods

Ninety Wistar Albino rats with a median weight of 265 (range240–290) g were maintained accordance with the recommenda-tions of the Guide for the Care and Use of Laboratory Animalsand the experiments conformed the Statement on Animal Experi-mentation by the NH&MRC and approved by the CumhuriyetUniversity Faculty of Medicine Ethics Committee. All animalswere weighed daily and had free access to a standard laboratorydiet and water. There were six groups, depending on whether theanimals were killed on the 3rd or the 7th day, and whether they re-ceived oral (OAP) or rectal (RAP) administration of phenytoin:

– Group A: 3rd-day control (n=15)– Group B: 3rd-day OAP (n=15)– Group C: 3rd-day RAP (n=15)– Group D: 7th-day control (n=15)– Group E: 7th-day OAP (n=15)– Group F: 7th-day RAP (n=15)

Operative procedures

Mechanical and antibacterial bowel preparation was not per-formed. After 12 h fasting all rats were anesthetized with 5 mg/kgxylasine (Rhompun, Abdi Ibrahim, Istanbul, Turkey) and40 mg/kg ketamine hydrochloride (Ketalar, Eczacibasi, Istanbul,Turkey). To prevent postoperative dehydration 5 ml Ringer’s lac-tate solution was injected subcutaneously. In all rats the abdomenwas opened by a midline incision; complete colonic transectionwas performed transversely approximately 3 cm above the perito-neal reflection. Colonic feaces were swabbed by manual milkingmethod. Anastomoses were performed with interrupted invertingsutures (6-0 monoflament polypropylene; Prolene, Ethicon). Theabdominal muscle layers and skin incision were closed separatelywith running sutures (3-0 silk, Ethicon). All procedures were per-formed by a single surgeon. Five rats which died during the opera-tion were replaced immediately with the new ones. In OAP groupsphenytoin was given at 10 mg/kg daily per orogastric route by 4-Ffine feeding catheter and in RAP groups at 10 mg/0.5 cc daily tothe anastomoses transrectally via a fine anal catheter. Groups A,B, and C were anesthetized on day 3 and groups D, E, and F onday 7 after operation for in vivo analytical procedures. The ani-mals were then killed with an excess dose of ether for in vitro ana-lytic procedures.

Analyses

The strength of each anastomosis was assessed by measuring itsbursting pressure using a fluid infusion pump (Abbott, USA), op-erating at 3 ml/min with a pressure transducer (Abbott MonitoringKit “Transpac IV,” Abbott). The bursting pressure was measured

in situ without disturbing any adhesions formed around the anasto-mosis. A fine catheter was passed via the rectum, to 2 cm belowthe level of anastomosis. The colon was ligated around the cathe-ter below and also 2 cm proximal to the anastomosis. Saline withmethylene blue was transfused through the catheter. The pressurewas recorded in mmHg using monitor (Petas KMA 460 R, Anka-ra, Turkey). The pressure was observed and the site of leakage wasvisualized with a magnification lens or identified by a sudden lossof pressure. The anastomotic segment was isolated from the sur-rounding tissue and a part of it was collected for hydroxyprolineassay as a marker of collagen content [9].

Histopathological analyses

A segment of each anastomotic ring was cut off for histologicalexamination and fixed in 10% formaldehyde. The samples for his-tology were dehydrated and embedded in paraffin. From all paraf-fin blocks 5-µm sections were cut, and staining was performedwith hematoxylin and eosin. In addition, Verchoof’s Van Giesonand Masson’s trichrome staining for assessment of the tissue reac-tion was performed. Anastomoses were graded histologically in ablind fashion using a modified 0–4 Ehrlich and Hunt [10] numeri-cal scale: 0, no evidence; 1+, occasional evidence; 2+, light scat-tering; 3+, abundant evidence; 4+, confluent cells or fibers. Evalu-ated parameters were inflammatory cell infiltrate (white bloodcount), collagen deposition, edema, fibroblast and blood vessel in-growth.

Immunohistochemical analysis

Histological specimens were taken from each anastomotic seg-ments at 3 and 7 days and fixed in a 10% neutral formalin solutionfor 48 h then embedded in paraffin. Immunohistochemical stainsfor basic fibroblast growth factor (bFGF) and vascular endothelialgrowth factor (VEGF) were performed on each section. All sam-ples were washed three times with phosphate-buffered solution(Sigma Chemical, St. Louis, Mo., USA). Hydrogen peroxide blocksolution was used as the blocking solution (Lab vision, Calif.,USA). The samples were then incubated for 60 min with the pri-mary antibody (bFGF and VEGF; Quartett, Berlin, Germany).This was followed by three washes in phosphate-buffered solutionand a 60-min incubation with secondary antibody. The slides werestained with AEC chromogen (Lab vision, Calif., USA). After de-velopment the slides were rinsed with water, counterstained withhematoxylin, and mounted.

Statistical analysis

For statistical analysis of the results the mean, the standard devia-tion, and standard error of the mean were calculated. Results werecompared using the Kruskall-Wallis and Mann-Whitney U tests,with Windows SPSS 10.0 program.

Results

The effects of phenytoin on abdominal wound healingand anastomotic complications are shown in Table 1. Nosignificant difference was observed. Although there wasa 15–20 g weight reduction in rats on the 3rd day, theyreached the preoperative weight by the 7th day.

251

252

Results of anastomotic bursting pressure analyses

For all 3-day-old anastomosis bursting occurred at thesuture line in all groups. In the 7-day-old colonic seg-ments rupture was noted outside the anastomoses gener-ally in the proximal segment and in the longitudinal line.There were statistically significant differences in anasto-motic bursting pressure values between the 3rd-day con-trol group (A, 42.00±2.19) and 3rd-day OAP (B,54.27±3.24; P<0.05) and between the 7th-day controlgroup (D, 168.70±5.42) and 7th-day OAP (E,184.77±4.54; P<0.05; Fig. 1). There were also statisti-cally significant differences in anastomotic burstingpressure values between group A (42.00±2.19) and 3rd-day RAP (C, 64.20±3.19; P<0.05) and between group D(168.70±5.42) and 7th-day RAP (F, 196.70±4.21;P<0.05; Fig. 1). When the anastomotic bursting pressurevalues of phenytoin-treated groups were compared (B vs.C, E vs. F), values were found to be significantly higherin RAP than in OAP groups (P<0.05, Kruskall-Wallisand Mann-Whitney U tests).

Results of histopathological analyses

Histological data from the colonic anastomoses areshown in Table 2 and Fig. 2. Histological examination

showed that by the 3rd-day both OAP and RAP causedgreater collagen deposition, fibroblast ingrowth, and vas-cularization than was observed in controls (P<0.05).There was no significant difference in anastomotic whiteblood cell count or edema amount between the 3rd-daygroups (P>0.05). By the 7th-day positive effects of phe-nytoin on collagen deposition, fibroblast ingrowth, andvascularization were again observed (P<0.05); differ-

Table 1 Complications in abdominal wound and anastomoses (differences nonsignificant, Kruskall-Wallis and Mann-Whitney U tests)

3rd-day 7th-day

Control OAP RAP Control OAP RAP (A,n=15) (B, n=15) (C, n=15) (D, n=15) (E, n=15) (F, n=15)

Wound infection 1 1 1 1 1 1Wound dehiscence 1 − − − − −Intestinal obstruction − − − 1 − 1Anastomotic stenosis − − − 1 − −Anastomotic dehiscence 1 − − 1a − −

a Closed perforation without any signs of peritonitis

Table 2 Histological evaluation results of the 3th day and 7th day colonic anastomoses

3rd day 7th day

Control (A) OAP (B) RAP (C) Pa Control (D) OAP (E) RAP (F) Pa

Inflammatory 3.0±0.08 2.80±0.06 2.71±0.09 n.s. 1.82±0.09 1.63±0.06 1.41±0.09 n.s.cell infiltrateFibroblast 1.15±0.09 1.85±0.06 2.50±0.06 <0.05 2.22±0.05 2.79±0.08 3.4±0.03 <0.05Collagen 1.28±0.08 1.80±0.04 2.20±0.06 <0.05 2.13±0.04 2.89±0.06 3.66±0.06 <0.05depositionVascularization 1.10±0.06 1.94±0.05 2.54±0.08 <0.05 2.0±0.1 2.84±0.05 3.62±0.05 <0.05Edema 3.06±0.07 2.98±0.04 2.78±0.08 n.s. 2.02±0.02 1.70±0.02 1.32±0.04 n.s.

a Kruskall-Wallis test

Fig. 1 In situ anastomotic bursting pressure (mmHg; mean±SEM). P<0.05 (Kruskall-Wallis and Mann-Whitney U tests)

253

ences in anastomotic white blood cell count and edemaamount between groups D and E and groups D and Fwere not statistically significant (P>0.05). When the his-tological data of OAP and RAP were compared (groupsB and C and groups E and F), collagen deposition, fibro-blast ingrowth, and vascularization was found to be bet-ter in the RAP than the OAP groups (P<0.05, (Kruskall-Wallis and Mann-Whitney U tests; Table 2 and Fig. 2).

Results of anastomotic hydroxyproline analyses

There was significantly greater anastomotic hydroxypro-line content (P<0.05; Fig. 3) in group B (10.19±0.34)than in group A (8.39±0.49), in group E (11.51±0.39)than in group D (10.37±0.41), in group C (11.89±0.41)than in group A (8.39±0.49), and in group F(12.61±0.38) than in group D (10.37±0.41). Comparisonof groups B and C and of groups E and F showed signifi-cantly better results in the RAP than in the OAP groups(P<0.05, Kruskall-Wallis and Mann-Whitney U tests;Fig. 3).

Results of immunohistochemical analysis

While there was no VEGF immunolocalization in theconnective tissue of anastomotic lines in 3rd-day controlanimals, columnar epithelium showed a weak VEGF im-munolocalization (Fig. 4a, b). In immunohistochemicalstaining of sections in anastomotic lines of 3rd-day phe-nytoin-treated groups, there was a weak VEGF immuno-localization in the colonic epithelium whereas the sub-mucosal region contained a relatively dense staining(Fig. 4c, d). In immunohistochemical staining of anasto-

Fig. 2 a Third-day control; H&E, ×10. b Third-day OAP; Mas-son’s trichome, ×25. c Third-day RAP; Verchoof’s Van Gieson,×10. d Seventh-day control; Masson’s trichome, ×10. e Seventh-day OAP; Masson’s trichome, ×25. f Seventh-day RAP; Masson’strichome, ×25. Histolopathological examination of anastomoticsegments of 3rd-day phenytoin treatment and control groupsshows intestinal wall showed influx of white blood cells (a–c).Collagen deposition (arrows, e, f), fibroblast ingrowth, and vascu-larization were more prominent in treatment groups

Fig. 3 Hydroxyproline level of each group (mean ±SEM, µg/mgdry weight). P<0.05, Kruskall-Wallis and Mann-Whitney U tests

254

motic lines of 7th-day phenytoin-treated animals, therewas markedly dense VEGF localization (Fig. 5c, d), ac-cording to 7th-day control animals (Fig. 5a, b). VEGFlocalization in the submucosal region was distributedboth in the epithelium and lamina propria. Comparisonof the phenytoin-treated groups (B vs. C, E vs. F)showed denser VEGF localization in RAP than in OAPgroups.

In immunohistochemical analysis of anastomotic linesof 3rd-day control animals bFGF localization was evi-dent especially in the connective tissue of the mucosaland submucosal regions (Fig. 6a). Although bFGF stain-ing of 3rd-day phenytoin group animals (Fig. 6b, c) was

similar to that in 3rd-day control animals there wasmarkedly denser bFGF immunolocalization in the 7th-day phenytoin groups (Fig. 6e, f) than in the 7th-daycontrol group (Fig. 6d) especially in the submucosa.When immunohistochemical analysis results of phenyto-in-treated groups were compared (B vs. C, E vs. F), 7th-day RAP animals (group F) showed denser bFGF local-ization than OAP animals (group E).

Discussion

Colonic anastomotic healing has been extensively inves-tigated in many clinical and experimental studies overthe past two centuries. Considerable improvements intechnique have been made, and the adverse effects ofmultiple local and systemic factors have been document-ed. Despite the progress that has been made anastomoticdehiscence following colonic anastomosis is a seriousclinical complication. Various factors are known to influ-ence the healing of colonic anastomosis. While technicalfailure may occur, there are many factors which may in-

Fig. 4 a Third-day control; ×20. b Third-day control; ×100. cThird-day OAP; ×40. d Third-day RAP; ×20. a, b Immunohisto-chemical staining of connective tissue shows no immunolocaliza-tion and a weak VEGF immunolocalization in columnar epitheli-um (Ep). c, d Sections of 3rd-day phenytoin-treated animals showrelatively dense staining in submucosa (Sub; arrows, d) and weakVEGF localization in epithelial cells (Ep; arrows). VEGF localiza-tion of RAP (d) was denser than that in the OAP group (c)

255

fluence disruption of an anastomosis, including malnutri-tion, protein depletion, trauma, blood loss, sepsis, fecalloading, foreign bodies, and possibly the proteolytic en-zyme collagenase [11, 12].

Collagen plays a central role in the healing sequenceand newly formed collagen fibers are responsible for thedevelopment of wound strength. A study by Jiborn et al.[13] found that after left colonic resection and anastomo-sis there was a decrease in collagen content of the colon,and that this decrease was more prominent in the proxi-mal than the distal segment of the anastomosis. It hasalso been shown that in the first 4 days after an operationcollagen lysis is prominent, and that collagen synthesisincreases after 7th day; therefore definitive anastomotichealing occurs through the careful equilibrium betweencollagen synthesis and breakdown [14]. In vivo and clin-ical studies suggest that topical phenytoin accelerates theinflammatory process, granulation tissue formation, andreepithelization [7, 15, 16]. Bosi et al. [4] also observedincreased accumulation of sulfated glycoeaminoglycansin phenytoin treated fibroblasts and determined that thiswas caused by increased synthesis.

In this study significantly higher in bursting pressurewas found of phenytoin-treated groups than in con-trols. Histological examination showed greater collagen-ization, epithelization, and capillary proliferation inphenytoin groups than in controls. Tissue hydroxypro-line level studies also supported this increase in colla-genization.

Most of the events in wound healing are generallycontrolled by growth factors such as VEGF and bFGF.Therefore in this study effects of phenytoin treatmenton VEGF and bFGF in colonic anastomotic healingwere also examined. VEGF is an endothelial-specificgrowth factor that is strongly angiogenic in vivo [17].Similarly, there exists much evidence of a role for bFGFin tumor angiogenesis [18]. bFGF is an effective mito-gen for fibroblasts, vascular endothelial cells, and intes-tinal epithelial cells [19, 20]. bFGF has also been re-ported to enhance the healing of gastrointestinal injuriesin an animal model [21]. VEGF and bFGF are suggest-ed to stimulate the nitric oxide pathway, enzyme activi-ty, and the levels of cyclic GMP during angiogenesis. Ithas been shown that the levels of cyclic GMP in the cellare doubled by VEGF but have no effect on cyclicAMP. In contrast, bFGF has no effect on cyclic GMPaccumulation but slightly increases cyclic AMP produc-tion [22]. These findings suggest that VEGF and bFGFare highly involved in the angiogenic pathways. Thewound healing process also includes angiogenesis viacell migration, proliferation, and differentiation. Thepresent study found that while there was no positive

Fig. 5 a Seventh-day control; ×40. b Seventh-day OAP; ×100. cSeventh-day RAP; ×20. Immunohistochemical staining of anasto-motic lines of 7th-day phenytoin-treated animals showed diffuseVEGF localization (arrows) compared to 7th-day control animals(a) in colonic epithelium (Ep) and lamina propria (Lp). VEGF lo-calization of the OAP group (b) was less dense than that of theRAP group (c)

256

VEGF immunolocalization in the colonic region in con-trols, the colonic anastomotic region densely showedVEGF localization in the submucosal connective tissueafter 3 days of phenytoin treatment. It was even denserin the 7 day phenytoin treatment group. On the otherhand, there was no difference between the control and3rd-day phenytoin-treatment group when bFGF immu-nolocalization was compared between the groups. How-ever, bFGF staining of colonic anastomotic region wasdenser in the 7-day phenytoin-treatment than in the con-trol group. These results demonstrate that phenytoin ad-ministration increases VEGF and bFGF expression ofVEGF and bFGF.

Collagen is very resistant to the action of proteolyticenzymes, but breakdown is initiated by the specific en-zyme collagenase. A significant increase in collagenasewas found in the gastrointestinal tract after anastomosisin the colon [23]. It was attempted to inhibit the action ofcollagenase in the rabbit colon by using aprotinin, and animprovement in bursting pressure in animals receivingthis agent was observed [24, 25]. Phenytoin inhibits the

release of lysosomal and cytoplasmic enzymes [26]. Incat palatal mucosa phenytoin administration decreasedthe extracellular breakdown of collagen and sulfatedproteoglycans. This effect appeared to be mediated bythe inhibition of collagenase release by phenytoin [3].The collagenase inhibitory property of phenytoin mayalso contribute its positive effects on colonic anastomo-sis healing.

Gingival hypertrophy and fibrosis in lungs seen insystemic use of phenytoin [5, 15]. Therefore in this studyits systemic effects on experimental colonic anastomoseswere also examined. Our findings regarding anastomosisbursting pressure, hydroxyproline contents, histologicalexamination, and immunohistochemical analysis showthat systemic use also had positive effects in healing ofcolonic anastomoses in the first postoperative week.However, comparison of OAP and RAP groups demon-strated the greater positive effects of the latter, a differ-ence which we attribute to the higher drug concentrationthat is achieved in the anastomotic area with RAP thanwith OAP.

In conclusion, the data from this study demonstratethat the administration of phenytoin to rats with colonicanastomoses significantly enhance the healing of anasto-moses, and that the effects of RAP are better than thosewith OAP.

Acknowledgements This study was carried out in the ResearchLaboratories of the Faculty of Medicine, Cumhuriyet University,in Sivas, Turkey, and was supported by the Cumhuriyet UniversityResearch Foundation.

Fig. 6 a Third-day control; ×20. b Third-day OAP; ×100. c Third-day RAP; ×20. d Seventh-day control; ×20. e Seventh-day OAP;×40. f Seventh-day RAP; ×20. In immunohistochemical analysisof anastomotic lines of 3rd-day control animals bFGF localizationwas determined in connective tissue of mucosa (Muc) and submu-cosa (Sub). b, c bFGF staining of 3rd-day phenytoin-treated ani-mals were similar with the 3rd-day control group (a). Sections of7th-day phenytoin-treated animals (e, f) showed markedly morediffuse bFGF staining (arrows) than 7th-day control animals (d)especially in submucosa (Sub). bFGF localization in the 7th-dayRAP group (f) was denser than that in the 7th-day OAP group (e)

257

References

1. Mann B, Kleinschmidt S, Stremmel W(1996) Prospective study of hand-sutured anastomosis after colorectal resection. Br J Surg 83:29–31

2. Pakkastie TE, Luukkonen PE, JarvinenHJ (1994) Anastomotic leakage afteranterior resection of the rectum. Eur JSurg 160:293–297

3. Schrock TR, Deveney CW, Dunphy JE(1973) Factor contributing to leakageof colonic anastomoses. Ann Surg177:513–518

4. Bosi G, Evangelisti R, Valeno V, Carinci F, Pezzetti F, Calastrini C(1998) Diphenylhydantoin affects gly-cosaminoglycans and collagen produc-tion by human fibroblasts from cleftpalate patients. J Dent Res77:1613–1621

5. Goebel RW (1972) Sodium diphenyl-hydantoin association with oral heal-ing. J Oral Surg 30:191–195

6. DaCosta M, Regan M, Sader M, LeaderM, Bouchier-Hayes D (1998) Diphe-nylhydantoin sodium promotes earlyand marked angiogenesis and results in increased collagen deposition andtensile strength in healing wounds.Surgery 123:287–293

7. Shafer WG (1961) Effect of dilantinsodium analogues on cell proliphera-tion in tissue culture. Proc Soc ExpBiol Med 106:205–209

8. Kolbert GS (1968) Oral diphenylhy-dantoin in corneal wound healing inthe rabbit. Am J Ophthalmol66:736–738

9. Syk I, Agren MS, Adawi D, JeppssonB (2001) Inhibition of matrix metallo-proteinases enhances breaking strengthof colonic anastomoses in an experi-mental model. Br J Surg 88:228–234

10. Ehrlich HP, Tarver H, Hunt TK (1973)Effects of vitamin A and glucocortico-ids upon inflammation and collagensynthesis. Ann Surg 177:222–227

11. Yamamoto T, Keighley MR (1999) Stapled functional end-to-end anasto-mosis in Crohn’s disease. Surg Today29:679–681

12. Goligher JC, Graham NG, De DombalFT (1970) Anastomotic dehiscence after anterior resection of rectum andsigmoid. Br J Surg 57:109–118

13. Jiborn H, Ahonen J, Zederfeld B(1980) Healing of experimental colonicanastomoses collagen metabolism inthe colon after left colon resection. AmJ Surg 139:398–405

14. Narayanan AS, Meyers DF, Page RC(1988) Regulation of collagen produc-tion in fibroblasts cultured from normaland phenytoin-induced hyperplastichuman gingiva. J Periodontal Res23:118–121

15. Talas G, Brown RA, McGrouther DA(1999) Role of phenytoin in woundhealing-a wound pharmacology per-spective. Biochem Pharmacol57:1085–1094

16. Moy LS, Tan EM, Holness R, Uitto J(1985) Phenytoin modulates connec-tive tissue metabolism and cell prolif-eration in human skin fibroblast cul-tures. Arch Dermatol 121:79–83

17. Ferrara N, Henzel WJ (1989) Pituitaryfollicular cells secrete a novel heparin-binding growth factor specific for vas-cular endothelial cells. Biochem Bio-phys Res Commun 161:851–858

18. Kandel J, Bossy-Wetzel E, Radvanyi F,Klagsbrun M, Folkman J, Hanahan D(1991) Neovascularization is associat-ed with a switch to the export of bFGFin the multistep development of fibro-sarcoma. Cell 66:1095–1104

19. Grant M, Jerdan J, Merimee TJ (1987)Insulin-like growth factor-I modulatesendothelial cell chemotaxis. J Clin En-docrinol Metab 65:370–371

20. Dignass AU, Tsunekawa S, PodolskyDK (1994) Fibroblast growth factorsmodulate intestinal epithelial cellgrowth and migration. Gastroenterolo-gy 106:1254–1262

21. Folkman J, Szabo S, Stovroff M,McNeil P, Li W, Shing Y (1991) Duodenal ulcer. Discovery of a newmechanism and development of angio-genic therapy that accelerates healing.Ann Surg 214:414–425

22. Ziche M, Morbidelli L, Choudhuri R,Zhang HT, Donnini S, Granger HJ(1997) Nitric oxide synthase lies down-stream from vascular endothelialgrowth factor-induced but not basic fi-broblast growth factor-induced angio-genesis. J Clin Invest 99:2625–2634

23. Hawley PR, Faulk WP, Hunt TK, Dunphy JE (1970) Collagenase activity in

the gastro-intestinal tract. Br J Surg57:896–900

24. Delaney P, Lalor D (1976) Enzyme in-hibition in colorectal surgery. Br J Surg63:23–24

25. Young HL, Wheeler MH (1983) Effectof intravenous aprotinin (Trasylol) onthe healing of experimental colonicanastomoses in the rabbit. Eur SurgRes 15:18–23

26. Lerner U, Hanstrom L (1980) Influenceof diphenylhydantoin on lysosomal enzyme release during bone resorptionin vitro. Acta Pharmacol Toxicol(Copenh) 47:144–150