- Email: [email protected]
Loss of collagen from experimental intestinal anastomoses: Early events
SXPERIMENTAL
AND
MOLECULAR
PATHOLOGY
42, 411-418 (1985)
Loss of Collagen from Experimental Intestinal Anastomoses: Early Events T. H. L. B. VEREECKEN, W. L. E. M. HESP, P. H. M. SCHILLINGS,~ AND H. H. M. DE BOER
TH. HENDRIKS,
Department of General Surgery, ‘Department of Pathology, St. Radboud University Hospital. Nijmegen. The Netherlands Received July I I, 1984 Collagen lysis, which always occurs to some extent in the wound area, is thought to be the underlying cause for breakdown of intestinal anastomoses. Therefore, we have studied the loss of collagen around ileal and colonic anastomoses in New Zealand White rabbits during the first 48 hr after operation. In the ileum, significant lysis of collagen in the anastomotic area, as represented by a decreased level of hydroxyproline. occurs from 12 hr postoperatively onward. Maximal loss of hydroxyprohne. as compared to preoperative values, is 27% measured 24 hr after operation. In the colon, significant lysis of collagen occurs after 3 hr. The lowest level of hydroxyproline measured during the experimental period is found 48 hr after operation, where the concentration is decreased by 38%. Changes in ileum are restricted to the anastomotic area, while in the colon the decrease in hydroxyproline extends along the intestinal wall, particularly in a proximal direction. The fact that total protein concentrations do not vary significantly indicates that the lowered hydroxyproline levels are specific. Microscopic examination of the wound area shows that the cellular response during the first 24 hr after wounding is restricted to granulocytes. It is suggested that granulocyte collagenase is mainly responsible for the observed lysis of collagen after intestinal anastomosis. (0 1982 Academic Prear. Inc.
INTRODUCTION Anastomotic leakage in the intestine seems to occur almost exclusively in the colon. While the high incidence of leakage of colonic anastomoses is amply documented (e.g., Schrock et af., 1973), we have found no data in the literature concerning the leakage rate of ileal anastomoses. From this, one might infer that dehiscence of ileal anastomoses is not considered to be a frequent complication of surgery of the small bowel. Therefore, inventarization and comparison of the phenomena occurring after construction of an anastomosis in both parts of the gut might yield valuable information concerning the processes responsible for disturbance of anastomotic integrity. The connective tissue macromolecule collagen plays a central role in wound healing (Shoshan, 1981). The underlying cause for the occurrence of anastomotic leakage is thought to be a change in collagen metabolism, resulting in a massive loss of collagen from the anastomotic area during the first days after construction of the anastomosis. Earlier work has shown that such a process occurs around experimental colonic anastomoses (Irvin and Hunt, 1974), the first measurements being performed 2 days after operation (Jiborn et al., 1980). We have shown recently that after ileal anastomosis a similar loss of collagen occurs, despite the aforementioned fact that clinical signs of leakage of small bowel anastomoses apparently are not seen, or are seen infrequently (Hesp et al., 1984). Accepting the hypothesis that lysis of collagen around the anastomosis may ultimately lead to such a weakening of the integrity of the intestinal wall that 411 0014-4800/85 $3.00 Copyright 0 1985 by Academx Press. Inc. All rights of reproduction in any form reserved.
412
HENDRIKS
ET AL.
dehiscence may result, it seems logical to investigate whether it is possible to decrease the anastomotic leakage rate by diminishing collagen breakdown. As a first step, characterization of the collagenolysis and identification of factors which influence this process appear necessary, The present report documents the occurrence of substantial &olIagenolysis, reflected by decreased ~on~ent~tions of hydro~yproline, within the first 24 hr after const~ction of both ileal and colonic anastomoses in the rabbit. MATERIALS AND METHUDS For this study, 73 male New Zealand White rabbits were used with an average weight of 2380 g (SD, 244; range, 1800-3000 g); 41 rabbits received an ileal anast~mosis, 32 a colonic anastomosis. Surgical procedures were performed under sterile conditions. After an overnight fast the rabbits were intubated and anesthetized with an oxygen-nitrous oxide-fluothane mixture. The abdomen was opened through a midline incision of approximately 4 cm. Ileal a~asto~os~s. The distal ileum was severed 10 cm proximally to the appendix tip. A 2-cm-long segment (located proximally to the lesion) was removed for determination of hydroxyproline content. Continuity was restored by an endto-end anastomosis, using one layer of interrupted inverting sutures with Proiene 5 x 0. About 20 cm in a proximal direction, a second anastomosis was similarly constructed. Colonic anastomosis. A 2-cm-long segment was removed from the descending colon, for measurement of hydroxyproline, and continuity was restored as described above. This anastomosis was placed 3-5 cm above the Douglas pouch. Subsequently a second anastomosis was constructed 3 cm under the splenic flexure. The peritoneum and fascia were closed in one layer with a &ontinuous suture, using Vicryl 3 x 0, and the skin was closed with silk 2 x 0. ~stope~tiveiy the animals were fed ad l~b~f~~~. The animals with an ileal anastomosis were sacrificed, by means of an overdose of pentothal, after 3 hr (n = 51, 6 hr fn = 5)* 12 hr (n = 61, 15 hr (n = 61, 18 hr (n = 61, 24 hr fn = 71, and 48 hr in = 6). Rabbits with a coionic anastomosis were sacrificed after 3 hr (n = 6), 6 hr (n = 71, 12 hr (n = 61, 24 hr (n = 7), and 48 hr (n = 6). The intestinal segment containing the first, more distai, anastomosis was removed and three I-cm samples were coliected for hydroxyproiine assay: one containing the anastomosis and two representing the adjacent proximal and distal parts, respectively. The second, more proximal, anastomosis was used for histology. Ail samples for biochemical assayswere frozen immediately in liquid nitrogen. Subsequently, samples were pulverized in a Braun miGrodismembrator, Iyophiiized, and stored at -30°C. Hydroxyproline was measured (in duplicate) essentially according to Prockop and Udenfriend (1960), using a hydroxyproline standard from Caibiochem. Protein was assayed using bovine serum albumin (Type V, Sigma) as a standard (Lowry et al., 1951). Anastomotic segments for histology were removed and collected in 4% (w/v) formaldehyde. Samples were dehydrated successively with acetone, methyl benzoate, and totuene, and fixed in paraf%In.Thereafter they were sliced into 6-p.m sections, which were stained with hematoxyiin-eosine and screened for occurrence of various cell types.
COLLAGEN
LOSS AFTER INTESTINAL
ANASTOMOSIS
413
Methods used for statistical evaluation of the data are given together with results in the appropriate section, RESULTS The average hydroxyprol~ne concentration in unwounded intestine is 8.01 (SD, 1.42; range, 5.69-12.25; n = 41) pg/mg dry wt for ileum and 13.04 (SD, 2.14; range, 9.48- 16.58; n = 32) E.tg/mgdry wt for colon. The variations in hydroxyproline levels measured in the anastomotic area during the first 48 hr after construction of an anastomosis are represented as both absolute (Fig. 1) and relative (Fig. 2) changes. Statistical comparison by means of a paired-sample two-sided Student f test, of the values found at operation and at sacrifice shows similar results with both approaches. In the ilea segment,‘ hydroxyproline levels in the anastomotic area are signi~cantiy lowered from 12 hr after operation. In the colon, this decrease is significant 3 hr after operation. The course of the hydroxyproline loss in both intestinal segments was analyzed further by means of one-way analysis of variance. In the ileum, it appears that the groups of measurements obtained at the seven time points, which are represented by their mean values in Figs. 1 and 2, cannot be considered to be random samples from the same population (P = 0.0164), which means that hydroxyproline levels change significantly over this time period. Further comparison of the groups at the various time points by means of the contrast test of Tukey (1949) shows the loss of hydroxyproline at I8 and 24 hr after operation to be significantly Change by~rc$ 1,ugfmg dry weight)
i noursatter
operatson
FIG. f I Absolute changes in hydroxy~~l~ne content of aoastomuti~ segments during the first 48 hr after operation. Results are expressed as average change f&mg dry wt) with standard error of the mean; the time scale is logarithmic. Lewis of significance for the differences between anastomosis and unwounded tissue are calculated by means of a paired-sample two-sided Student t test and are represented by * 0.01 < P s 0.05; ** 0.001< P C 0.01;and *** P s 0.001.
414
HENDRIKS Perceniuat [hwo]
change
ET AL.
----.
T
o-
-10-
+
t **
i l
I
t
* -2o*
r * Ileum
-3o-
~ t
I
Colon
t * \ T -LO-
I
3
I
6
I,,
12 15 18
/
24 Hours after
L : I
I.8 operation
FIG. 2. Relative changes in hydroxyproline content of anastomotic segments during the first 48 hr after operation. Results are expressed as mean percentage change, calculated with respect to the preoperative value, with standard error of the mean; the time scale is logarithmic. Levels of signiticance for the differences between anastomosis and unwounded tissue are calculated by means of a paired-sample two-sided Student t test and represented by * 0.01 < P c 0.05; ** 0.001 < P < 0.01; and *** P =s 0.001.
greater than the loss 3 hr postoperatively. Maximal decrease of hydroxyproline levels occurs 24 hr after operation and amounts to 27% of the preoperative value. The same statistical procedures have been applied to the five groups of data, obtained for colonic anastomosis at various times after operation. Analysis of variance shows a very significant (P < 0.001) change in hydroxyproline levels with time during the period of investigation. Loss of hydroxyproline at 24 and 48 hr is signi~cantly greater than that 3, 6, or 12 hr after operation. Here, maximal lowering of hydroxyproline, amounting to 38% of preoperative values, was found 48 hr postoperatively. It is also possible to compare the course of hydroxyproline changes around ileal and colonic anastomoses by means of two-way analysis of variance, which was performed using the data obtained in both ileum and colon 3, 6, 12, 24, and 48 hr after operation. The loss of hydroxyproline appears to be systematically greater around colon anastomoses, if both absolute changes (P < 0.001) and relative changes (P = 0.03) are used for the comparison. Also, the shape of the time curve for the ileal anastomoses is signi~cantly (P = 0.005) different from the shape of the curve for colonic anastomoses, if absolute changes are compared (cf Fig. 1). In the ileum, changes in hydroxyproline levels are almost completely restricted to the anastomotic area. The only possible exception is the proximal segment 24 hr after operation, where the average loss of hydroxyproline is almost (P = 0.06) significant (Table I). In the colon, the decrease in hydroxyproline measured in
COLLAGEN
LOSS AFTER INTESTINAL
415
ANASTOM~SIS
TABLE I Postoperative Changes in Hydroxyproline Levels in Segments Proximal and Distal to Intestinal Anastomoses
Ileum
Level of significance (0
Proximal
-0.84 r 0.36
0.06
Proximal Proximal Distal Proximal Distal Proximal Distal
- 1.43 -1.16 -0.85 -2.49 - 1.68 -3.61 -2.16
(hr)
24 Colon (hr) 3 6 12 24 48
Segment
Concentration change (whz dry wt 2 SEM)
k 2 ” + 2 ‘2
0.62 0.50 0.41 0.39 0.67 0.73 0.72
0.07 0.06 0.09 <0.001 0.04 0.004 0.03
Note. Only those experimental groups are represented which show a signi~cant or almost signi~cant effect in one or both segments, as calculated by means of a paired-sample two-sided Student f test.
the anastomotic area extends further along the intestinal wall, particularly in proximal direction. These effects are most pronounced from 24 hr postoperatively onward (Table I). We have also measured total protein concentrations in control segments, obtained at operation, and anastomotic segments. The protein concentration averages 685 (SD: 96, range 498-842, n = 41) pglmg dry weight in unwounded ileum and 703 (SD: 86, range: 561-966, pt = 32) pg/mg dry weight in colon. From this, it can be calculated that collagen comprises approximately 8 and 13% of total protein in the wall of ileum and colon, respectively. Table II shows that no significant changes in total protein occur, with the possible exception of a slight increase in ileum 3 hr and in colon 6 hr after operation. The occurrence of various types of cells in the wound area has been examined microscopically in the second set of more proximally constructed anastomoses. The results are summa~zed in Table III. A similar picture arises for both ileum and colon. Granulocytes are present immediately after wounding. Their number increases until 12-24 hr postoperatively. An occasional monocyte may be observed after 24 hr, but their number starts to grow significantly only between 24 and 48 hr. The same goes for fibroblasts, although a number of these cells is already present 24 hr after operation. DISCUSSION A lowered hydroxyproline concentration around experimental colonic anastomoses, as measured from 2 days postoperatively onward, has been documented previously (e.g., Jiborn et al., 1980, and references therein). Stromberg and Klein (1982) have argued that a change in hydroxyproline concentration does not necessarily mean that the actual amount of hydroxyproline changes: a decreased concentration might also be caused by an increase in wet weight or dry weight.
416
HENDRIKS
ET AL.
TABLE II A Comparison of Protein Concentrations in Unwounded Tissue and Anastomotic Segments in the Various Experimental Groups Protein concentration
n ileum (hr) 3 6 12 15 18 24 48
Unwounded tissue
Anastomosis
Difference
5 5 6 6 6 7 6
710 603 686 730 679 676 700
f f zt 2 e + t
30 61 27 33 46 48 16
744 681 671 733 673 680 731
2 f ? f e + 2
36 80 17 25 69 50 22
34 78 -9 3 -6 4 31
16 6 7 6
172 685 706 692 665
5rt 2 2 t
49 28 36 30 13
712 826 715 662 630
‘-+ + -c i
33 39 33 47 33
54 27 9 -30 -35
+ zt ? 2 + k +
Level of significance P
14 59 36 50 31 30 26
0.07 n.s.* ns. ns. n.s.
2e 37 13 57 33 + 60 2 40
Ki n.s. us. ns.
ll.S.
ns.
Colon (hr) 63 12 24 48
Nom. Data are represented as mean value with standard error of the mean. Statistical evaluation of the difference between unwounded and anastomotic tissue is performed with a paired-sample twosided Student t test. * Not significant, 0.1 < P.
In our study total protein concentrations do not change signi~cantly in any of the experimental groups (Table II). This supports the conclusion that the hydroxyproline effect is rather specific and indicative of true lysis of collagen. The present data represent the first description of changes in collagen levels during the initial 48 hr after construction of intestinal anastomoses. Figure I shows that the lysis of collagen in the anastomotic area starts immediately after wounding, at least in the colon. So far, changes in collagen concentration during wound healing after anastomosis of the small bowel have not been described in the literature. Our results clearly indicate that a similar process occurs in the ileum, although it takes 12 hr before hydroxyproline levels have decreased significantly. The absolute size of the effect appears greater in the colon than the ileum. Since the concentration of hydroxyproline in the ileum is lower than in the colon, the relative size of the effect, at least from 6 to 24 hr after operation, appears about equal in both intestinal segments (Fig. 2). Still, statistical comparison of the data obtained over the entire experimental period show the loss of hydroxyproline to be systematically greater around colonic anastomoses. Apart from the apparent slower onset of the collagenolysis, the duration of a state of decreased hydroxyproline levels seem shorter in ileum. After 24 hr the concentration starts to rise again, while in the colon the lowest point is at 48 hr after operation. Moreover, preoperative levels are restored after 3 days in the ileum and only after 7 days in the colon (Hesp et al., 1984). These findings, together with the fact that hydroxyproline levels in segments proximal and distal to the anastomosis do not change significantly in the ileum. clearly demonstrate that,
COLLAGEN
LOSS AFTER
INTESTINAL
417
ANASTOMOSIS
TABLE III Occurrence of Different Types of Cells in Sections of Intestinal Anastomoses at Various Times after Operation Monocytesi macrophages
n
Granulocytes
Fibroblasts
Ileum (hr) 3 6 12 15 18 24 48
5 5 6 3 4 3 3
+ + ++ +l++ 4-l” +
f +
+ ++
Colon (hr) 3 6 12 24 48
4 7 2 3 3
+ 4” ++ + f
I!? f
+ ++
-
Note. - , absent; 2, occasional cell present; + , present; + + , abundantly present.
although qualitatively similar processes occur in ileum and colon, the loss of collagen is less extensive and more quickly undone in the ileum. It should be emphasized that Figs. I and 2 do not necessarily reflect the true course of collagen breakdown. The concentrations are the net result of lysis and synthesis. Jiborn et al. (1980) have shown that considerable collagen synthesis occurs shortly (48 hr) after construction of colonic anastomosis in the rat. Thus, actual collagen breakdown may be greater than depicted in Fig. 1. Hunt et al. (1980) have argued that the collagenous equilibrium is most important to the process of colon repair: anastomotic integrity is essentially a race between collagen synthesis and lysis. If synthesis does not overtake lysis, anastomotic leakage occurs. Consequently, control of lysis could be a crucial step in preventing anastomotic dehiscence. The first step in the breakdown of collagen molecules, cleavage of the triple helix, is performed by the enzyme collagenase. The cellular response after wounding, a sequential appearance of polymorphonuclear neutrophilic leukocytes (granulocytes), macrophages, and fibroblasts has been well characterized (Shoshan, 1981). Microscopic examination of the anastomotic area shows that in our experiments the first monocytes and fibroblasts only appear 24 hr after wounding, when collagen lysis is well under way. Granulocytes are already present after a few hours and these cells are capable of producing a true collagenase (Horwitz et al., 1977). While the existence of collagen-bound collagenase has also been demonstrated in animal tissues (Perez-Tamayo, 1978) our results exclude macrophages and fibroblasts as sources of collagenase during the first 24 hr after wounding, a period when massive collagen lysis already occurs. Thus, seeking to control collagenolysis after intestinal anastomosis, further research into regulation, and particularly inhibition, of granulocyte and collagen-bound collagenase, if present, appears necessary.
418
HENDRIKS
ET AL.
ACKNOWLEDGMENTS The authors are grateful to Mr. A. A. Klompmakers and Mr. P. Thissen for expert technical assistance, and to Dr. Ph. van Elteren (Department of Statistical Support, University of Nijmegen) for statistical analysis of the experimental data.
REFERENCES HESP, W. L. E. M., HENDRIKS. T., LUBBERS, E. J. C., and DE BOER, H. H. M. (1984). Wound healing in the intestinal wall: A comparison between experimental heal and colonic anastomosis. Dis. Colon Rectum 27, 99-104. HORWITZ, A. L., HANCE, A. J., and CRYSTAL, R. G. (1977). Granulocyte collagenase: Selective digestion of type I relative to type III collagen. Proc. Nat/. Acad. Sci. USA 74, 897-901. HUNT, T. K., HAWLEY, P. R., HALE, J., GOODSON, W., and THAKRAL. K. K. (1980). Colon repair: The collagenous equilibrium. In “Wound Healing and Wound Infection. Theory and surgical practice” (T. K. Hunt, ed.), pp. 153-159. Appleton-Century-Crofts, New York. IRVIN, T. T., and HUNT, T. K. (1974). Reappraisal of the healing process of anastomosis of the colon. Surg.
Gynecol.
Obstet.
138, 741-746.
JIBORN, H., AHONEN, J., and ZEDERFELDT, B. (1980). Healing of experimental colonic anastomoses. IV. Effect of suture technique on collagen metabolism in the colonic wall. Amer. J. Surg. 139,406413.
LOWRY, 0. H., ROSEBROUGH,N. J., FARR, A. L., and RANDALL, R. J. (1951). Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265-275. PEREZ-TAMAYO, R. (1978). Pathology of collagen degradation. Amer. J. Pathol. 92, 509-566. PROCKOP,D. J., and UDENFRIEND, S. (196Ct). A specific method for the analysis of hydroxyproline in tissues and urine. Anal. Biochem. 1, 228-239. SCHROCK, T. R., DEVENEY, C. W., and DUNPHY, J. E. (1973). Factors contributing to leakage of colonic anastomoses. Ann. Surg. 177, 513-518. SHOSHAN, S. (1981). Wound healing. In “International Review of Connective Tissue Research” (D. A. Hall and D. S. Jackson, eds.), Vol. 9, pp. l-26. Academic Press, New York. STROMBERG, B. V., and KLEIN, L. R. (1982). Collagen formation during the healing of colonic anastomoses. Dis. Co/on Rectum 25, 301-304. TUKEY, J. W. (1949). Comparing individual means in the analysis of variance. Biometrics 5, 99- 114.
AND
MOLECULAR
PATHOLOGY
42, 411-418 (1985)
Loss of Collagen from Experimental Intestinal Anastomoses: Early Events T. H. L. B. VEREECKEN, W. L. E. M. HESP, P. H. M. SCHILLINGS,~ AND H. H. M. DE BOER
TH. HENDRIKS,
Department of General Surgery, ‘Department of Pathology, St. Radboud University Hospital. Nijmegen. The Netherlands Received July I I, 1984 Collagen lysis, which always occurs to some extent in the wound area, is thought to be the underlying cause for breakdown of intestinal anastomoses. Therefore, we have studied the loss of collagen around ileal and colonic anastomoses in New Zealand White rabbits during the first 48 hr after operation. In the ileum, significant lysis of collagen in the anastomotic area, as represented by a decreased level of hydroxyproline. occurs from 12 hr postoperatively onward. Maximal loss of hydroxyprohne. as compared to preoperative values, is 27% measured 24 hr after operation. In the colon, significant lysis of collagen occurs after 3 hr. The lowest level of hydroxyproline measured during the experimental period is found 48 hr after operation, where the concentration is decreased by 38%. Changes in ileum are restricted to the anastomotic area, while in the colon the decrease in hydroxyproline extends along the intestinal wall, particularly in a proximal direction. The fact that total protein concentrations do not vary significantly indicates that the lowered hydroxyproline levels are specific. Microscopic examination of the wound area shows that the cellular response during the first 24 hr after wounding is restricted to granulocytes. It is suggested that granulocyte collagenase is mainly responsible for the observed lysis of collagen after intestinal anastomosis. (0 1982 Academic Prear. Inc.
INTRODUCTION Anastomotic leakage in the intestine seems to occur almost exclusively in the colon. While the high incidence of leakage of colonic anastomoses is amply documented (e.g., Schrock et af., 1973), we have found no data in the literature concerning the leakage rate of ileal anastomoses. From this, one might infer that dehiscence of ileal anastomoses is not considered to be a frequent complication of surgery of the small bowel. Therefore, inventarization and comparison of the phenomena occurring after construction of an anastomosis in both parts of the gut might yield valuable information concerning the processes responsible for disturbance of anastomotic integrity. The connective tissue macromolecule collagen plays a central role in wound healing (Shoshan, 1981). The underlying cause for the occurrence of anastomotic leakage is thought to be a change in collagen metabolism, resulting in a massive loss of collagen from the anastomotic area during the first days after construction of the anastomosis. Earlier work has shown that such a process occurs around experimental colonic anastomoses (Irvin and Hunt, 1974), the first measurements being performed 2 days after operation (Jiborn et al., 1980). We have shown recently that after ileal anastomosis a similar loss of collagen occurs, despite the aforementioned fact that clinical signs of leakage of small bowel anastomoses apparently are not seen, or are seen infrequently (Hesp et al., 1984). Accepting the hypothesis that lysis of collagen around the anastomosis may ultimately lead to such a weakening of the integrity of the intestinal wall that 411 0014-4800/85 $3.00 Copyright 0 1985 by Academx Press. Inc. All rights of reproduction in any form reserved.
412
HENDRIKS
ET AL.
dehiscence may result, it seems logical to investigate whether it is possible to decrease the anastomotic leakage rate by diminishing collagen breakdown. As a first step, characterization of the collagenolysis and identification of factors which influence this process appear necessary, The present report documents the occurrence of substantial &olIagenolysis, reflected by decreased ~on~ent~tions of hydro~yproline, within the first 24 hr after const~ction of both ileal and colonic anastomoses in the rabbit. MATERIALS AND METHUDS For this study, 73 male New Zealand White rabbits were used with an average weight of 2380 g (SD, 244; range, 1800-3000 g); 41 rabbits received an ileal anast~mosis, 32 a colonic anastomosis. Surgical procedures were performed under sterile conditions. After an overnight fast the rabbits were intubated and anesthetized with an oxygen-nitrous oxide-fluothane mixture. The abdomen was opened through a midline incision of approximately 4 cm. Ileal a~asto~os~s. The distal ileum was severed 10 cm proximally to the appendix tip. A 2-cm-long segment (located proximally to the lesion) was removed for determination of hydroxyproline content. Continuity was restored by an endto-end anastomosis, using one layer of interrupted inverting sutures with Proiene 5 x 0. About 20 cm in a proximal direction, a second anastomosis was similarly constructed. Colonic anastomosis. A 2-cm-long segment was removed from the descending colon, for measurement of hydroxyproline, and continuity was restored as described above. This anastomosis was placed 3-5 cm above the Douglas pouch. Subsequently a second anastomosis was constructed 3 cm under the splenic flexure. The peritoneum and fascia were closed in one layer with a &ontinuous suture, using Vicryl 3 x 0, and the skin was closed with silk 2 x 0. ~stope~tiveiy the animals were fed ad l~b~f~~~. The animals with an ileal anastomosis were sacrificed, by means of an overdose of pentothal, after 3 hr (n = 51, 6 hr fn = 5)* 12 hr (n = 61, 15 hr (n = 61, 18 hr (n = 61, 24 hr fn = 71, and 48 hr in = 6). Rabbits with a coionic anastomosis were sacrificed after 3 hr (n = 6), 6 hr (n = 71, 12 hr (n = 61, 24 hr (n = 7), and 48 hr (n = 6). The intestinal segment containing the first, more distai, anastomosis was removed and three I-cm samples were coliected for hydroxyproiine assay: one containing the anastomosis and two representing the adjacent proximal and distal parts, respectively. The second, more proximal, anastomosis was used for histology. Ail samples for biochemical assayswere frozen immediately in liquid nitrogen. Subsequently, samples were pulverized in a Braun miGrodismembrator, Iyophiiized, and stored at -30°C. Hydroxyproline was measured (in duplicate) essentially according to Prockop and Udenfriend (1960), using a hydroxyproline standard from Caibiochem. Protein was assayed using bovine serum albumin (Type V, Sigma) as a standard (Lowry et al., 1951). Anastomotic segments for histology were removed and collected in 4% (w/v) formaldehyde. Samples were dehydrated successively with acetone, methyl benzoate, and totuene, and fixed in paraf%In.Thereafter they were sliced into 6-p.m sections, which were stained with hematoxyiin-eosine and screened for occurrence of various cell types.
COLLAGEN
LOSS AFTER INTESTINAL
ANASTOMOSIS
413
Methods used for statistical evaluation of the data are given together with results in the appropriate section, RESULTS The average hydroxyprol~ne concentration in unwounded intestine is 8.01 (SD, 1.42; range, 5.69-12.25; n = 41) pg/mg dry wt for ileum and 13.04 (SD, 2.14; range, 9.48- 16.58; n = 32) E.tg/mgdry wt for colon. The variations in hydroxyproline levels measured in the anastomotic area during the first 48 hr after construction of an anastomosis are represented as both absolute (Fig. 1) and relative (Fig. 2) changes. Statistical comparison by means of a paired-sample two-sided Student f test, of the values found at operation and at sacrifice shows similar results with both approaches. In the ilea segment,‘ hydroxyproline levels in the anastomotic area are signi~cantiy lowered from 12 hr after operation. In the colon, this decrease is significant 3 hr after operation. The course of the hydroxyproline loss in both intestinal segments was analyzed further by means of one-way analysis of variance. In the ileum, it appears that the groups of measurements obtained at the seven time points, which are represented by their mean values in Figs. 1 and 2, cannot be considered to be random samples from the same population (P = 0.0164), which means that hydroxyproline levels change significantly over this time period. Further comparison of the groups at the various time points by means of the contrast test of Tukey (1949) shows the loss of hydroxyproline at I8 and 24 hr after operation to be significantly Change by~rc$ 1,ugfmg dry weight)
i noursatter
operatson
FIG. f I Absolute changes in hydroxy~~l~ne content of aoastomuti~ segments during the first 48 hr after operation. Results are expressed as average change f&mg dry wt) with standard error of the mean; the time scale is logarithmic. Lewis of significance for the differences between anastomosis and unwounded tissue are calculated by means of a paired-sample two-sided Student t test and are represented by * 0.01 < P s 0.05; ** 0.001< P C 0.01;and *** P s 0.001.
414
HENDRIKS Perceniuat [hwo]
change
ET AL.
----.
T
o-
-10-
+
t **
i l
I
t
* -2o*
r * Ileum
-3o-
~ t
I
Colon
t * \ T -LO-
I
3
I
6
I,,
12 15 18
/
24 Hours after
L : I
I.8 operation
FIG. 2. Relative changes in hydroxyproline content of anastomotic segments during the first 48 hr after operation. Results are expressed as mean percentage change, calculated with respect to the preoperative value, with standard error of the mean; the time scale is logarithmic. Levels of signiticance for the differences between anastomosis and unwounded tissue are calculated by means of a paired-sample two-sided Student t test and represented by * 0.01 < P c 0.05; ** 0.001 < P < 0.01; and *** P =s 0.001.
greater than the loss 3 hr postoperatively. Maximal decrease of hydroxyproline levels occurs 24 hr after operation and amounts to 27% of the preoperative value. The same statistical procedures have been applied to the five groups of data, obtained for colonic anastomosis at various times after operation. Analysis of variance shows a very significant (P < 0.001) change in hydroxyproline levels with time during the period of investigation. Loss of hydroxyproline at 24 and 48 hr is signi~cantly greater than that 3, 6, or 12 hr after operation. Here, maximal lowering of hydroxyproline, amounting to 38% of preoperative values, was found 48 hr postoperatively. It is also possible to compare the course of hydroxyproline changes around ileal and colonic anastomoses by means of two-way analysis of variance, which was performed using the data obtained in both ileum and colon 3, 6, 12, 24, and 48 hr after operation. The loss of hydroxyproline appears to be systematically greater around colon anastomoses, if both absolute changes (P < 0.001) and relative changes (P = 0.03) are used for the comparison. Also, the shape of the time curve for the ileal anastomoses is signi~cantly (P = 0.005) different from the shape of the curve for colonic anastomoses, if absolute changes are compared (cf Fig. 1). In the ileum, changes in hydroxyproline levels are almost completely restricted to the anastomotic area. The only possible exception is the proximal segment 24 hr after operation, where the average loss of hydroxyproline is almost (P = 0.06) significant (Table I). In the colon, the decrease in hydroxyproline measured in
COLLAGEN
LOSS AFTER INTESTINAL
415
ANASTOM~SIS
TABLE I Postoperative Changes in Hydroxyproline Levels in Segments Proximal and Distal to Intestinal Anastomoses
Ileum
Level of significance (0
Proximal
-0.84 r 0.36
0.06
Proximal Proximal Distal Proximal Distal Proximal Distal
- 1.43 -1.16 -0.85 -2.49 - 1.68 -3.61 -2.16
(hr)
24 Colon (hr) 3 6 12 24 48
Segment
Concentration change (whz dry wt 2 SEM)
k 2 ” + 2 ‘2
0.62 0.50 0.41 0.39 0.67 0.73 0.72
0.07 0.06 0.09 <0.001 0.04 0.004 0.03
Note. Only those experimental groups are represented which show a signi~cant or almost signi~cant effect in one or both segments, as calculated by means of a paired-sample two-sided Student f test.
the anastomotic area extends further along the intestinal wall, particularly in proximal direction. These effects are most pronounced from 24 hr postoperatively onward (Table I). We have also measured total protein concentrations in control segments, obtained at operation, and anastomotic segments. The protein concentration averages 685 (SD: 96, range 498-842, n = 41) pglmg dry weight in unwounded ileum and 703 (SD: 86, range: 561-966, pt = 32) pg/mg dry weight in colon. From this, it can be calculated that collagen comprises approximately 8 and 13% of total protein in the wall of ileum and colon, respectively. Table II shows that no significant changes in total protein occur, with the possible exception of a slight increase in ileum 3 hr and in colon 6 hr after operation. The occurrence of various types of cells in the wound area has been examined microscopically in the second set of more proximally constructed anastomoses. The results are summa~zed in Table III. A similar picture arises for both ileum and colon. Granulocytes are present immediately after wounding. Their number increases until 12-24 hr postoperatively. An occasional monocyte may be observed after 24 hr, but their number starts to grow significantly only between 24 and 48 hr. The same goes for fibroblasts, although a number of these cells is already present 24 hr after operation. DISCUSSION A lowered hydroxyproline concentration around experimental colonic anastomoses, as measured from 2 days postoperatively onward, has been documented previously (e.g., Jiborn et al., 1980, and references therein). Stromberg and Klein (1982) have argued that a change in hydroxyproline concentration does not necessarily mean that the actual amount of hydroxyproline changes: a decreased concentration might also be caused by an increase in wet weight or dry weight.
416
HENDRIKS
ET AL.
TABLE II A Comparison of Protein Concentrations in Unwounded Tissue and Anastomotic Segments in the Various Experimental Groups Protein concentration
n ileum (hr) 3 6 12 15 18 24 48
Unwounded tissue
Anastomosis
Difference
5 5 6 6 6 7 6
710 603 686 730 679 676 700
f f zt 2 e + t
30 61 27 33 46 48 16
744 681 671 733 673 680 731
2 f ? f e + 2
36 80 17 25 69 50 22
34 78 -9 3 -6 4 31
16 6 7 6
172 685 706 692 665
5rt 2 2 t
49 28 36 30 13
712 826 715 662 630
‘-+ + -c i
33 39 33 47 33
54 27 9 -30 -35
+ zt ? 2 + k +
Level of significance P
14 59 36 50 31 30 26
0.07 n.s.* ns. ns. n.s.
2e 37 13 57 33 + 60 2 40
Ki n.s. us. ns.
ll.S.
ns.
Colon (hr) 63 12 24 48
Nom. Data are represented as mean value with standard error of the mean. Statistical evaluation of the difference between unwounded and anastomotic tissue is performed with a paired-sample twosided Student t test. * Not significant, 0.1 < P.
In our study total protein concentrations do not change signi~cantly in any of the experimental groups (Table II). This supports the conclusion that the hydroxyproline effect is rather specific and indicative of true lysis of collagen. The present data represent the first description of changes in collagen levels during the initial 48 hr after construction of intestinal anastomoses. Figure I shows that the lysis of collagen in the anastomotic area starts immediately after wounding, at least in the colon. So far, changes in collagen concentration during wound healing after anastomosis of the small bowel have not been described in the literature. Our results clearly indicate that a similar process occurs in the ileum, although it takes 12 hr before hydroxyproline levels have decreased significantly. The absolute size of the effect appears greater in the colon than the ileum. Since the concentration of hydroxyproline in the ileum is lower than in the colon, the relative size of the effect, at least from 6 to 24 hr after operation, appears about equal in both intestinal segments (Fig. 2). Still, statistical comparison of the data obtained over the entire experimental period show the loss of hydroxyproline to be systematically greater around colonic anastomoses. Apart from the apparent slower onset of the collagenolysis, the duration of a state of decreased hydroxyproline levels seem shorter in ileum. After 24 hr the concentration starts to rise again, while in the colon the lowest point is at 48 hr after operation. Moreover, preoperative levels are restored after 3 days in the ileum and only after 7 days in the colon (Hesp et al., 1984). These findings, together with the fact that hydroxyproline levels in segments proximal and distal to the anastomosis do not change significantly in the ileum. clearly demonstrate that,
COLLAGEN
LOSS AFTER
INTESTINAL
417
ANASTOMOSIS
TABLE III Occurrence of Different Types of Cells in Sections of Intestinal Anastomoses at Various Times after Operation Monocytesi macrophages
n
Granulocytes
Fibroblasts
Ileum (hr) 3 6 12 15 18 24 48
5 5 6 3 4 3 3
+ + ++ +l++ 4-l” +
f +
+ ++
Colon (hr) 3 6 12 24 48
4 7 2 3 3
+ 4” ++ + f
I!? f
+ ++
-
Note. - , absent; 2, occasional cell present; + , present; + + , abundantly present.
although qualitatively similar processes occur in ileum and colon, the loss of collagen is less extensive and more quickly undone in the ileum. It should be emphasized that Figs. I and 2 do not necessarily reflect the true course of collagen breakdown. The concentrations are the net result of lysis and synthesis. Jiborn et al. (1980) have shown that considerable collagen synthesis occurs shortly (48 hr) after construction of colonic anastomosis in the rat. Thus, actual collagen breakdown may be greater than depicted in Fig. 1. Hunt et al. (1980) have argued that the collagenous equilibrium is most important to the process of colon repair: anastomotic integrity is essentially a race between collagen synthesis and lysis. If synthesis does not overtake lysis, anastomotic leakage occurs. Consequently, control of lysis could be a crucial step in preventing anastomotic dehiscence. The first step in the breakdown of collagen molecules, cleavage of the triple helix, is performed by the enzyme collagenase. The cellular response after wounding, a sequential appearance of polymorphonuclear neutrophilic leukocytes (granulocytes), macrophages, and fibroblasts has been well characterized (Shoshan, 1981). Microscopic examination of the anastomotic area shows that in our experiments the first monocytes and fibroblasts only appear 24 hr after wounding, when collagen lysis is well under way. Granulocytes are already present after a few hours and these cells are capable of producing a true collagenase (Horwitz et al., 1977). While the existence of collagen-bound collagenase has also been demonstrated in animal tissues (Perez-Tamayo, 1978) our results exclude macrophages and fibroblasts as sources of collagenase during the first 24 hr after wounding, a period when massive collagen lysis already occurs. Thus, seeking to control collagenolysis after intestinal anastomosis, further research into regulation, and particularly inhibition, of granulocyte and collagen-bound collagenase, if present, appears necessary.
418
HENDRIKS
ET AL.
ACKNOWLEDGMENTS The authors are grateful to Mr. A. A. Klompmakers and Mr. P. Thissen for expert technical assistance, and to Dr. Ph. van Elteren (Department of Statistical Support, University of Nijmegen) for statistical analysis of the experimental data.
REFERENCES HESP, W. L. E. M., HENDRIKS. T., LUBBERS, E. J. C., and DE BOER, H. H. M. (1984). Wound healing in the intestinal wall: A comparison between experimental heal and colonic anastomosis. Dis. Colon Rectum 27, 99-104. HORWITZ, A. L., HANCE, A. J., and CRYSTAL, R. G. (1977). Granulocyte collagenase: Selective digestion of type I relative to type III collagen. Proc. Nat/. Acad. Sci. USA 74, 897-901. HUNT, T. K., HAWLEY, P. R., HALE, J., GOODSON, W., and THAKRAL. K. K. (1980). Colon repair: The collagenous equilibrium. In “Wound Healing and Wound Infection. Theory and surgical practice” (T. K. Hunt, ed.), pp. 153-159. Appleton-Century-Crofts, New York. IRVIN, T. T., and HUNT, T. K. (1974). Reappraisal of the healing process of anastomosis of the colon. Surg.
Gynecol.
Obstet.
138, 741-746.
JIBORN, H., AHONEN, J., and ZEDERFELDT, B. (1980). Healing of experimental colonic anastomoses. IV. Effect of suture technique on collagen metabolism in the colonic wall. Amer. J. Surg. 139,406413.
LOWRY, 0. H., ROSEBROUGH,N. J., FARR, A. L., and RANDALL, R. J. (1951). Protein measurement with the folin phenol reagent. J. Biol. Chem. 193, 265-275. PEREZ-TAMAYO, R. (1978). Pathology of collagen degradation. Amer. J. Pathol. 92, 509-566. PROCKOP,D. J., and UDENFRIEND, S. (196Ct). A specific method for the analysis of hydroxyproline in tissues and urine. Anal. Biochem. 1, 228-239. SCHROCK, T. R., DEVENEY, C. W., and DUNPHY, J. E. (1973). Factors contributing to leakage of colonic anastomoses. Ann. Surg. 177, 513-518. SHOSHAN, S. (1981). Wound healing. In “International Review of Connective Tissue Research” (D. A. Hall and D. S. Jackson, eds.), Vol. 9, pp. l-26. Academic Press, New York. STROMBERG, B. V., and KLEIN, L. R. (1982). Collagen formation during the healing of colonic anastomoses. Dis. Co/on Rectum 25, 301-304. TUKEY, J. W. (1949). Comparing individual means in the analysis of variance. Biometrics 5, 99- 114.