- Email: [email protected]
Closure or nonclosure of the visceral peritoneum at cesarean delivery
Closure or nonclosure of the visceral peritoneum at cesarean delivery Fritz Nagele, MD, a Hermann Karas, MD, a Dietmar Spitzer, M D , b Alfons Staudach, M D , b Sonja Karasegh, MD, c Adolf Beck, MD, c and Peter Husslein, MD ~
Vienna, Austria OBJECTIVE: Our purpose was to determine whether nonclosure of the visceral peritoneum at low transverse cesarean delivery has advantages over suture peritonization with regard to postoperative morbidity. STUDY DESIGN: A prospective randomized trial of 549 women undergoing cesarean section was carried out; 262 were randomized to nonclosure and 287 to closure of the visceral peritoneum. Perioperative, intraoperative, and postoperative management decisions were made without reference to treatment groups. Statistical analysis compared intraoperative and postoperative outcome between the two groups. RESULTS: Operating and anesthesia times were significantly shorter in patients receiving nonclosure. The incidence of febrile morbidity and cystitis and the need for antibiotics and narcotics were all significantly greater when the peritoneum was closed. Hospital stay was significantly shorter after nonclosure. CONCLUSION: Nonclosure of the visceral peritoneum is associated with lower febrile and infectious morbidity. Routine closure of the visceral peritoneum should be abandoned at cesarean delivery. (AM J OBSTETGYNECOL1996;174:1366-70.) Key words: Cesarean delivery, peritoneal closure, postoperative morbidity
Cesarean delivery is the most common intraperitoneal surgical procedure in obstetrics and gynecology.1 Reapproximation of the visceral peritoneum after closure of the low transverse uterine incision has been widely performed on a routine basis, the theory being that suture peritonization is an important measure to separate the bowel from the wound cavity, thereby preventing adhesion formation. However, the advantages of this technique have not yet been proved by prospective randomized trials. In fact, prior animal experiments and general surgery reports have shown that suture peritonization tends to cause tissue ischemia, necrosis, inflammation, and foreign body reactions to the suture material. These factors may slow down the healing process and are considered important precursors for adhesion formation. By contrast, clean excision of peritoneal surfaces without suturing the cut edges provides for more rapid peritoneal repair and does not lead to tissue ischemia and infection, decreasing the risk for development of adhesion formation. 2-5
From the Departments of Obstetrics and Gynecology, UniversityHospital of Vienna/' General Hospital of Salzburg,b and General Hospital of Vienna-Hanusch. Receivedfor publication February 9, 1995; revised September11, 1995; accepted September28, 1995. Reprint requests: Fritz Nagele, MD, University Hospital of Vienna, Department of Obstetrics and Gynecology, Spitalgasse 23, A-1090 Vienna, Auslria. Copyright 9 1996 by Mosby-YearBook, Inc. 0002-9378/96 $5.00 + 0 6/1/69635 1366
On the basis of the observation that postoperative recovery after abdominal hysterectomy without visceral suture peritonization was conspicuously uneventful,6 we conducted this prospective randomized study in women undergoing cesarean delivery. Our goal was to evaluate whether nonclosure of the visceral peritoneum has benefits over routine closure with regard to the intraoperative and postoperative clinical course; anesthesia and operating times were measured and postoperative complications and febrile morbidity were analyzed. Material and methods
Between January 1993 and June 1994, 549 patients who underwent cesarean delivery in three obstetrics departments were included in this study (University Hospital of Vienna, n = 247; General Hospital of Salzburg, n = 213; General Hospital of Vienna-Hanusch, n = 89). All patients who underwent cesarean section within this period were randomized to one of two groups irrespective of the surgical preference of the delivering obstetrician. After patients gave informed consent, randomization occurred on the day of cesarean delivery with group assignment based on calendar days. On odd-numbered days the visceral peritoneum was left unsutured, and on even-numbered days the peritoneum was closed with a continuous absorbable 2-0 polyglactin suture. Also, closure of the parietal peritoneum was performed in a uniform manner with continuous polyglactin sutures No. 0 (all suture material by Ethicon, Peterborough, Ontario, Canada). Sub-
Volume 174, Number 4 AmJ ObstetGynecol
jects with low vertical hysterotomy (n = 2) or additional surgical procedures (ovarian cysts n = 4; cesarean hysterectomy, n = 2 ; myoma enucleation, n = 1) and women having received antibiotic treatment within 1 week before admission (n = 7) were excluded. A total of 287 women were randomized to the closed group and 262 to the open group. The type of anesthesia was chosen by the anesthesiologist without reference to treatment group. All surgical procedures were low transverse uterine cesareans and were done by the resident staff. When indicated, sterilization was performed by tubal coagulation and transection, the Pomeroy procedure, or Filshie clips. The decision regarding prophylactic antibiotics, defined as administration on the day of surgery or the first postoperative day, was up to the policy of the individual departments and i n d e p e n d e n t of treatment group. Therapeutic antibiotic treatment was defined as administration of up to 2 days postoperatively. Also, postoperative treatment decisions were made without reference to treatment group and included a single analgesic dose within 24 hours postoperatively (15 mg of piritramide subcutaneously or 100 mg of tramadol hydrochloride intramuscularly or a 500 ml infusion containing 4 gm of metamizole plus orphenadrine citrate) and a single dose of bowel stimulants on the second day (16 ml of sodium picosulfate orally or four doses of 12 mg of sennoside A plus B orally plus, in some cases, saline solution enema on the same day). Moreover, bowel stimulants were used in any patient with postoperative gas pains and constipation and no radiologic signs of ileus. During hospitalization tile following parameters were evaluated: duration of general anesthesia (from the first intravenous narcotic administration to extubation), total operating time (from incision to closure of the skin), n u m b e r of patients receiving prophylactic and therapeutic antibiotics, n u m b e r of patients requiring additional narcotic or bowel stimulant doses postoperatively (oral and parenteral), and febrile morbidity (defined both by n u m b e r of patients with temperature >38 ~ C for >2 days postoperatively and by daily average temperature values during the first postoperative week per group). Other variables recorded were chorioamnionitis (defined by maternal fever >37.5 ~ C and one or more of the following criteria: maternal tachycardia, fetal tachycardia, purulent vaginal discharge, foul-smelling amniotic fluid, maternal leukocytosis, uterine tenderness, and absence of other sources of infection), urinary tract infections and cystitis diagnosed by bacterial growth on a three-agar medium (Uricult plus, Orion Diagnostica, Espoo, Finland; results of >107 bacteria/ml were regarded as positive), endometritis (indicated by purulent vaginal discharge or uterine tenderness and fever), and wound problems (indicated by serous or purulent drainage from the skin incision or erythema or induration, with and without
Nagele et al. 1367
fever). Hospitalization was defined as the period from the day of cesarean section to the day of discharge from the hospital, All patients were evaluated each day they were hospitalized; the nursing staff was not aware of the patients' treatment groups. At discharge all patients were instructed to immediately contact the department at which they underwent operation in case they had any complaints or late complications. Statistical analysis was done with the Statistical Package for Social Sciences (SPSS/PC+, SPSS, Inc., Chicago). Analytic comparisons used the g2 test, Mantel-Haenszel test, Mann-Whitney U test, Fisher's exact test, and binomial test, with p < 0.05 considered significant.
Results Statistical analysis compared the characteristics and variables of 262 patients in whom the visceral peritoneum was left open with those of 287 patients with classical suture peritonization. Table I lists patient characteristics, anesthetic data, type of cesarean delivery, and further surgical details such as the method of uterine incision closure and the n u m b e r of patients undergoing tubal ligation. The indications for cesarean delivery are shown in Table II. Significant differences between the groups were noted only with regard to the type of anesthesia: for no apparent reason a significantly greater proportion of patients in the closure group received general anesthesia, whereas a significantly greater percentage of patients in the nonclosure group were given spinal anesthesia. Administration of antibiotics, postoperative narcotics, and bowel stimulants and perioperative and postoperative complications are outlined in Table III. Although there was no significant difference in the n u m b e r of patients receiving prophylactic antibiotics, therapeutic antibiotic requirements were significantly higher in patients in the closure group. Statistical analysis showed that the greater use of general anesthesia in the closed group did not account for the difference in antibiotic usage or fever between the open and closed groups. Patients in the closure group also required significantly more postoperative narcotics (oral or parenteral). Among the patients who required additional postoperative analgesia and who had previously received either general or spinal anesthesia, no differences were noted between the open and closed groups (general anesthesia: open group, n = 17, vs closed group, n = 58; spinal anesthesia: open group, n = 0, vs closed group, n = 4; ~2 = 1.16, Fisher's exact test: p=0.57). We therefore believe that the significantly higher pain medication requirements in the closed group are in no way related to the difference regarding the method of anesthesia used in the two groups. No significant difference was noted between the groups in the use of bowel stimulants, but cystitis was seen significantly more frequently with closure than without. Both temperature >38 ~ C for >2 postoperative days and the daily
1368
Nagete et al,
:\Wi[ 19!~6 Am.l OI)stet Gvnccol
Table I. Patient characteristics and procedure statistics Open group (n = 262)
Maternal age (yr) Range Maternal weight (kg) Range Parity Range Gestational age (wk) Range Anesthesia General Spinal Epidural Cesarean delivery Primary Acute Repeat Combined Uterine incision closure One-layer Two-layers Tubal sterilization
Closed group ( ~ - 287)
28.6 + 6.1 17-45 75.9 _+14.1 46-162 1.8+ 1.1 1-6 38.7 _+3.3 26-43
28.3 _+5.9 15-47 75.5 + 13.8 44-137 1.9_+ 1.1 1-8 38.8 _+3.4 27-43
173 (66.0%) 68 (25.9%) 21 (8.0%)
233 (81.2%) 42 (14.6%) 12 (4.2%)
84 142 12 24
75 163 13 36
(32.1%) (54.2%) (4.6%) (9.1%)
183 (69.8%) 79 (30,2%) 19 (7.3%)
Sig'n!ficance
NS NS NS NS p=0,003 p=0,01 NS
(26.1%) (56.8%) (4.5%) (12.5%)
NS NS NS NS
190 (66.2%) 97 (33.8%) 17 (5.9%)
NS NS
NS, Not significant.
Table II, Indications for cesarean delivery Open group (n = 262)
Fetal distress Breech presentation Arrest of labor Abnormal position/presentation Preterm (uncontrollable) labor Cranial-pelvicdisproporfion Placental factors Elective cesarean delivery Gestosis Multiple pregnancy Other
63 48 38 22 23 15 12 9 8 8 16
(24.0%) (18.3%) (14.5%) (8.4%) (8.8%) (5.7%) (4.6%) (3.4%) (3.1%) (3.1%) (6.l%)
Closed group (n = 287)
69 43 52 29 19 18 11 11 10 9 16
(24.0%) (14.9%) (18.1%) (10.1%) (6.6%) (6.3%) (3.8%) (3.8%) (3.5%) (3.2%) (5.6%)
Significance
NS NS NS NS NS NS NS NS NS NS NS
NS, Not significant.
average temperature values during the first postoperative week were significantly higher with closure of the periton e u m (p< 0.001). Also, endometritis and wound infections tended to be m o r e c o m m o n in the d o s e d group; the differences between groups did not achieve significance. During hospitalization serious complications occurred in two patients. O n e 38-year-old patient (case 486, o p e n group) u n d e r w e n t repeat laparotomy on the first postoperative day because of a secondary hemorrhage. A hematoma in the abdominal wall was removed, and a small bleeding vessel at the uterine incision was ligated. The second case was a 15-year-old patient (case 526, closed group) in w h o m a large h e m a t o m a in the abdominal wall necessitated reoperation on the third postoperative day. Finally, there was a single case of readmission (a 30-yearold woman, case 75, open group) 17 days after the patient's discharge. H e r complaints were abdominal pain and diarrhea, and enteritis was diagnosed by means of stool cultures. She was treated in the usual m a n n e r and
discharged 5 days later. No patients in either group had p e M c abscess or peritonitis, and n o n e required reoperation because of ileus. The mean operating time (_+SD) was significantly greater in the closed group (56.9 _+ 17.9 minutes, range 13 to 115 minutes) than in the o p e n group (50.6 + t6.8 minutes, range 14 to 105 minutes) (p < 0.001). Comparison of the operating times in patients in whom cesarean section and tubal sterilization were p e r f o r m e d revealed no signit: icant difference between the closed group (62.6-+ 16.3 minutes, range 30 to 98 minutes) and the open group (61.0 + 29.1 minutes, range 22 to 115 minutes). General anesthesia time was significantly longer in the closed group (67.4 + 17.5 minutes, range 20 to 125 minutes) than in the open group (62.6 + 17.6 minutes, range 17 to 110 minutes) (p= 0.02). Also, postoperative hospitalization was significandy longer in the closed g r o u p (7.9 4- 1.8 days, range 4 to 23 days) than in the open group (7.2 + 1.6 days, range 2 to 14 days) (p < 0.001).
Volume 174, Number 4 AmJ ObstetGynecol
Nagele et al. 1369
Table III. Administration of antibiotics, narcotics, and bowel stimulants and perioperative and postoperative complications Open group (n = 262)
Antibiotics Prophylaxis Therapy Narcotics Bowel stimulants Chorioamnionifis Fever Cystitis Endometrifis Wound infection
177 61 20 18 32 22 8 8 5
(67.6%) (23.3%) (7.6%) (6.9%) (12.2%) (8.4%) (3.1%) (3.1%) (1.9%)
Closed group (n = 287)
180 97 64 18 28 45 22 17 14
(62.7%) (33.8%) (22.3%) (6.3%) (9.8%) (15.7%) (7.7%) (5.9%) (4.9%)
Significance
NS p= 0.006 p< 0.001 NS NS p=0.009 p=0.01 NS NS
NS, Not significant.
Analysis by institutions showed that significant differences between the open and closed groups were reflected in each center.
Comment This study examines the question of the closure or nonclosure of the visceral peritoneum at cesarean delivery from a critical-clinical point of view and compares the perioperative, intraoperative, and postoperative course in the two treatment groups. The most important aspects reviewed were operative and anesthesia times; antibiotic, bowel stimulant, and narcotic requirements; febrile morbidity; and postoperative complications. Our findings indicate that cesarean section without suture reapproximation of the peritoneal cut edges provides a n u m b e r of significant advantages. One is the possible decrease in operative time, which in turn is associated with shorter anesthesia exposure and diminished intraoperative anesthetic requirements. Of interest, pain medication requirements were lower in the open group, suggesting that nonclosure may be associated with less postoperative abdominal pain, possibly because no tension is placed on the peritoneal wound edges. Antibiotic requirements were also lower in the nonclosure group, which is most likely related to the lower febrile and infectious morbidity in this group. In theory, the higher rate of clinically asymptomatic febrile cases in the closed group might be due to the formation of subperitoneal pockets resulting from the suture; these pockets could fill with blood and wound secretions from the uterine incision and serve as a nutrient media for bacteria. Urinary tract infection and cystitis are common complications in obstetric and gynecologic surgery. Although the relationship between suture peritonization and the high incidence of cystitis in the closed group remains unclear, one hypothesis is that surgical manipulation on the peritoneal part of the bladder is more likely to lead to cystitis and urinary tract infection. One argument against nonclosure of the visceral peritoneum has been that it increases the risk of adhesion
formation. To reduce this risk, a variety of chemical additives 79 and, more recently, intraperitoneal barrier therapylO, 11have been used. However, the most important factor in adhesion prevention is impeccable surgical technique, including minimal tissue traumatization and avoidance of ischemia and inflammation by eliminating crushing forceps pressure, stitch tension, and knot pressure.l~ These demands are best met by leaving the visceral peritoneum open. Hubbard et al.~3 examined the process of autologous reperitonization in animal experiments. They observed that regeneration of peritoneal defects is completed in 5 to 6 days. After showing that the entire wound surface endothelializes simultaneously and not gradually from the cut edges as in skin wounds, they concluded that large peritoneal defects heal as fast as small ones. According to Holtz ~4and Elkins et al., 4 adhesion formation after peritoneal closure is primarily the result of foreign body reactions to the suture material, interruption of vascular supply, or ischemia, and tissue inflammation. They found less tissue necrosis and inflammatory reactions in peritoneal cut edges that were not closed primarily than in those that were, even when minimally reactive suture material was used. On the basis of these and similar findings, we doubt that there is any connection between nonclosure of the visceral peritoneum and postoperative adhesion formation. 15-17 Conversely, sutured peritoneal surfaces might cause adhesion formation) s To our knowledge, only two previous articles report on the effects of peritoneal closure or nonclosure in women undergoing cesarean delivery. In the study of Pietrantoni et al. 19 only the parietal peritoneum was left open, whereas Hull and Varner 2~left both the visceral and parietal peritoneum unsutured. In agreement with our findings, both studies reported a significant decrease in operating time in the open group. Hull and Varner found that pain medication requirements were higher in closure subjects, which is also consistent with our data. However, the closed group in their study required more bowel stimulants, a finding contrasting with our investigation,
1370
Nageleet al.
April 1996 Am.] ()bstt't (A nccol
which showed t h a t r e t u r n to bowel flmction was similar in b o t h groups. C o n t r a r y to o u r findings, Hull a n d V a r n e r f o u n d n o difference b e t w e e n the g r o u p s in the i n c i d e n c e of febrile morbidity. O n e e x p l a n a t i o n for this discrepancy may be f o u n d in the small series of patients used in t h e i r study (closed group, n = 59, vs o p e n g r o u p , n = 54). Furt h e r m o r e , febrile morbidity was d e f i n e d by j u s t o n e par a m e t e r ( t e m p e r a t u r e _>38~ C at least o n c e d u r i n g postoperative hospitalization), whereas we also i n c l u d e d the daily average t e m p e r a t u r e values o f the first postoperative week. In summary, the results of o u r study indicate that nonclosure of the visceral p e r i t o n e u m m i g h t n o t only r e d u c e o p e r a t i n g time a n d anesthesia e x p o s u r e b u t also offer o t h e r significant benefits with r e g a r d to the postoperative course, i n c l u d i n g a r e d u c t i o n in postoperative narcotics and, most i m p o r t a n t , a lower febrile a n d infectious morbidity a n d t h e r e f o r e d e c r e a s e d antibiotic r e q u i r e m e n t s . Finally, these factors also explain why hospitalization was s h o r t e r in the o p e n group. We c o n c l u d e t h a t r o u t i n e closure o f the visceral p e r i t o n e u m b e a b a n d o n e d in w o m e n u n d e r g o i n g cesarean delivery.
6. 7.
8. 9. 10.
11. 12. 13. 14. 15.
We t h a n k H e r b e r t Poltnig, o u r biostatistician, for his assistance in p e r f o r m i n g the statistical analyses of this study.
16.
REFERENCES
1. Placek PJ, Taffel SM. Recent patterns in cesarean delivery ill the United States. Obstet Gynecol Clin North Am 1987;15: 607-27. 2. Ellis H. The aetiology of postoperative abdominal adhesions. BrJ Surg 1962;50:10-6. 3. Ellis H. Internal overhealing: the problem ofintraperitoneal adhesions. WorldJ Surg 1980;4:303-6. 4. Elkins TE, Stovall TG, Warren J, Ling FW, Meyer NL. A histologic evaluation of peritoneal injury and repair: implications for adhesion formation. Obstet Gynecol 1987;70: 225-8. 5. McDonald MN, Elkins TE, Wortham GE Stovall TG, Ling
17. 18. 19, 20.
I'~ ~,McNeeley S(;. Adhesion tk)rmation and prevemion after peritoneal injury and repair ill the rabbit. J Reprod Med 1988;33:43(5-9. Nagele F, Husslem P. ls closure ot" visceral peritoneum in abdominal hysterectomy necessary: a retrospective pilot smdv. Gelmrtshilfe Frauenheilkd 1991:51:925-8. Golan A, Wexler S, I,otan G, Abramov L, Langer R, Favid MP. Calcium antagonist: eft(~ct on adhesion tormation. Acta Obstet G~llecol Scand 1989;68:529-32. Caballero J, Tulandi T. Effects of Ringer's lactate and fibrin glue on postsurgical adhesions. J Reprod Med 1992;37:141-3. Adhesion Study Group. Reduction of postoperative adhesions with intraperitoneal 32% dextran 70: a prospective, randomized trial_ Fertil Steril 1983;4(/:612-9. Sekiba K, Obstetrics and Gynecology Adhesion Pre~enlion Committee. Use of Interceed (TC7) absorbable adhesion barrier to reduce postoperative adhesion reformation in intertility and endometriosis surgery. Obstet Gwlecol 1992; 79:518-22. Surgical Membrane Study Group. Prophylaxis of pelvic sidewall adhesions with Gore-Tex surgical membrane: a multicenter clinical investigation. Fertil Steril 1992;57:921-3. Stone K. Adhesions in gynecologic surgery. Curr Opin Obstet Gynecol 1993;5:322-7. Hubbard BT, Khan MZ, Carag RV, Albites EV, Hricko (;M. The pathology of peritoneal repair: its relation to the formation of adhesions. Ann Surg 1967;165:908-16. Holtz G. Adhesion induction by suture of varying tissue reactivity and caliber. IntJ Fertil 1982;27:134-5. Tulandi T, Hum HS, Gelfand MM. Closure of laparotomy incisions with or ~ithout peritoneal suturing and se(ondlook laparoscopy. AsL] OBs'rEI"C,~x~(:ol. 1988;158:536-7. Thompson N], Paterson-Brown S, Harbourne T, Whawell SA, Kalodiki E, Dudley HAF. Reduced hnman peritoneal plasminogen activating activity.: possible mechanism of adhesion formation. BrJ Surg 1989;76:382-4. Vipond MN, Whawell SA, Thompson.IN, Dudlev HAF. Peritoneal fibrinolytic activity and intraabdominal adhesions. Lancet 1990;335:1120-2. Conolly WB, Stephens FO. Factors influencing tile incidence of intraperitoneal adhesions: an experimental study. Surgery 1968;63:9769. Pietrantoni M, Parsons MT, O'Brien ~A,T,Collins E, I~mppel RA, Spellacy WN. Peritoneal closure or non-closure at cesarean. Ohstet Gynecol 1991;77:293-6. Hull DB, Varner MW. A randomized study of closm'e of the peritoneum at cesarean delivery. Obstet GynecoI 1991;77: 818-21.
Vienna, Austria OBJECTIVE: Our purpose was to determine whether nonclosure of the visceral peritoneum at low transverse cesarean delivery has advantages over suture peritonization with regard to postoperative morbidity. STUDY DESIGN: A prospective randomized trial of 549 women undergoing cesarean section was carried out; 262 were randomized to nonclosure and 287 to closure of the visceral peritoneum. Perioperative, intraoperative, and postoperative management decisions were made without reference to treatment groups. Statistical analysis compared intraoperative and postoperative outcome between the two groups. RESULTS: Operating and anesthesia times were significantly shorter in patients receiving nonclosure. The incidence of febrile morbidity and cystitis and the need for antibiotics and narcotics were all significantly greater when the peritoneum was closed. Hospital stay was significantly shorter after nonclosure. CONCLUSION: Nonclosure of the visceral peritoneum is associated with lower febrile and infectious morbidity. Routine closure of the visceral peritoneum should be abandoned at cesarean delivery. (AM J OBSTETGYNECOL1996;174:1366-70.) Key words: Cesarean delivery, peritoneal closure, postoperative morbidity
Cesarean delivery is the most common intraperitoneal surgical procedure in obstetrics and gynecology.1 Reapproximation of the visceral peritoneum after closure of the low transverse uterine incision has been widely performed on a routine basis, the theory being that suture peritonization is an important measure to separate the bowel from the wound cavity, thereby preventing adhesion formation. However, the advantages of this technique have not yet been proved by prospective randomized trials. In fact, prior animal experiments and general surgery reports have shown that suture peritonization tends to cause tissue ischemia, necrosis, inflammation, and foreign body reactions to the suture material. These factors may slow down the healing process and are considered important precursors for adhesion formation. By contrast, clean excision of peritoneal surfaces without suturing the cut edges provides for more rapid peritoneal repair and does not lead to tissue ischemia and infection, decreasing the risk for development of adhesion formation. 2-5
From the Departments of Obstetrics and Gynecology, UniversityHospital of Vienna/' General Hospital of Salzburg,b and General Hospital of Vienna-Hanusch. Receivedfor publication February 9, 1995; revised September11, 1995; accepted September28, 1995. Reprint requests: Fritz Nagele, MD, University Hospital of Vienna, Department of Obstetrics and Gynecology, Spitalgasse 23, A-1090 Vienna, Auslria. Copyright 9 1996 by Mosby-YearBook, Inc. 0002-9378/96 $5.00 + 0 6/1/69635 1366
On the basis of the observation that postoperative recovery after abdominal hysterectomy without visceral suture peritonization was conspicuously uneventful,6 we conducted this prospective randomized study in women undergoing cesarean delivery. Our goal was to evaluate whether nonclosure of the visceral peritoneum has benefits over routine closure with regard to the intraoperative and postoperative clinical course; anesthesia and operating times were measured and postoperative complications and febrile morbidity were analyzed. Material and methods
Between January 1993 and June 1994, 549 patients who underwent cesarean delivery in three obstetrics departments were included in this study (University Hospital of Vienna, n = 247; General Hospital of Salzburg, n = 213; General Hospital of Vienna-Hanusch, n = 89). All patients who underwent cesarean section within this period were randomized to one of two groups irrespective of the surgical preference of the delivering obstetrician. After patients gave informed consent, randomization occurred on the day of cesarean delivery with group assignment based on calendar days. On odd-numbered days the visceral peritoneum was left unsutured, and on even-numbered days the peritoneum was closed with a continuous absorbable 2-0 polyglactin suture. Also, closure of the parietal peritoneum was performed in a uniform manner with continuous polyglactin sutures No. 0 (all suture material by Ethicon, Peterborough, Ontario, Canada). Sub-
Volume 174, Number 4 AmJ ObstetGynecol
jects with low vertical hysterotomy (n = 2) or additional surgical procedures (ovarian cysts n = 4; cesarean hysterectomy, n = 2 ; myoma enucleation, n = 1) and women having received antibiotic treatment within 1 week before admission (n = 7) were excluded. A total of 287 women were randomized to the closed group and 262 to the open group. The type of anesthesia was chosen by the anesthesiologist without reference to treatment group. All surgical procedures were low transverse uterine cesareans and were done by the resident staff. When indicated, sterilization was performed by tubal coagulation and transection, the Pomeroy procedure, or Filshie clips. The decision regarding prophylactic antibiotics, defined as administration on the day of surgery or the first postoperative day, was up to the policy of the individual departments and i n d e p e n d e n t of treatment group. Therapeutic antibiotic treatment was defined as administration of up to 2 days postoperatively. Also, postoperative treatment decisions were made without reference to treatment group and included a single analgesic dose within 24 hours postoperatively (15 mg of piritramide subcutaneously or 100 mg of tramadol hydrochloride intramuscularly or a 500 ml infusion containing 4 gm of metamizole plus orphenadrine citrate) and a single dose of bowel stimulants on the second day (16 ml of sodium picosulfate orally or four doses of 12 mg of sennoside A plus B orally plus, in some cases, saline solution enema on the same day). Moreover, bowel stimulants were used in any patient with postoperative gas pains and constipation and no radiologic signs of ileus. During hospitalization tile following parameters were evaluated: duration of general anesthesia (from the first intravenous narcotic administration to extubation), total operating time (from incision to closure of the skin), n u m b e r of patients receiving prophylactic and therapeutic antibiotics, n u m b e r of patients requiring additional narcotic or bowel stimulant doses postoperatively (oral and parenteral), and febrile morbidity (defined both by n u m b e r of patients with temperature >38 ~ C for >2 days postoperatively and by daily average temperature values during the first postoperative week per group). Other variables recorded were chorioamnionitis (defined by maternal fever >37.5 ~ C and one or more of the following criteria: maternal tachycardia, fetal tachycardia, purulent vaginal discharge, foul-smelling amniotic fluid, maternal leukocytosis, uterine tenderness, and absence of other sources of infection), urinary tract infections and cystitis diagnosed by bacterial growth on a three-agar medium (Uricult plus, Orion Diagnostica, Espoo, Finland; results of >107 bacteria/ml were regarded as positive), endometritis (indicated by purulent vaginal discharge or uterine tenderness and fever), and wound problems (indicated by serous or purulent drainage from the skin incision or erythema or induration, with and without
Nagele et al. 1367
fever). Hospitalization was defined as the period from the day of cesarean section to the day of discharge from the hospital, All patients were evaluated each day they were hospitalized; the nursing staff was not aware of the patients' treatment groups. At discharge all patients were instructed to immediately contact the department at which they underwent operation in case they had any complaints or late complications. Statistical analysis was done with the Statistical Package for Social Sciences (SPSS/PC+, SPSS, Inc., Chicago). Analytic comparisons used the g2 test, Mantel-Haenszel test, Mann-Whitney U test, Fisher's exact test, and binomial test, with p < 0.05 considered significant.
Results Statistical analysis compared the characteristics and variables of 262 patients in whom the visceral peritoneum was left open with those of 287 patients with classical suture peritonization. Table I lists patient characteristics, anesthetic data, type of cesarean delivery, and further surgical details such as the method of uterine incision closure and the n u m b e r of patients undergoing tubal ligation. The indications for cesarean delivery are shown in Table II. Significant differences between the groups were noted only with regard to the type of anesthesia: for no apparent reason a significantly greater proportion of patients in the closure group received general anesthesia, whereas a significantly greater percentage of patients in the nonclosure group were given spinal anesthesia. Administration of antibiotics, postoperative narcotics, and bowel stimulants and perioperative and postoperative complications are outlined in Table III. Although there was no significant difference in the n u m b e r of patients receiving prophylactic antibiotics, therapeutic antibiotic requirements were significantly higher in patients in the closure group. Statistical analysis showed that the greater use of general anesthesia in the closed group did not account for the difference in antibiotic usage or fever between the open and closed groups. Patients in the closure group also required significantly more postoperative narcotics (oral or parenteral). Among the patients who required additional postoperative analgesia and who had previously received either general or spinal anesthesia, no differences were noted between the open and closed groups (general anesthesia: open group, n = 17, vs closed group, n = 58; spinal anesthesia: open group, n = 0, vs closed group, n = 4; ~2 = 1.16, Fisher's exact test: p=0.57). We therefore believe that the significantly higher pain medication requirements in the closed group are in no way related to the difference regarding the method of anesthesia used in the two groups. No significant difference was noted between the groups in the use of bowel stimulants, but cystitis was seen significantly more frequently with closure than without. Both temperature >38 ~ C for >2 postoperative days and the daily
1368
Nagete et al,
:\Wi[ 19!~6 Am.l OI)stet Gvnccol
Table I. Patient characteristics and procedure statistics Open group (n = 262)
Maternal age (yr) Range Maternal weight (kg) Range Parity Range Gestational age (wk) Range Anesthesia General Spinal Epidural Cesarean delivery Primary Acute Repeat Combined Uterine incision closure One-layer Two-layers Tubal sterilization
Closed group ( ~ - 287)
28.6 + 6.1 17-45 75.9 _+14.1 46-162 1.8+ 1.1 1-6 38.7 _+3.3 26-43
28.3 _+5.9 15-47 75.5 + 13.8 44-137 1.9_+ 1.1 1-8 38.8 _+3.4 27-43
173 (66.0%) 68 (25.9%) 21 (8.0%)
233 (81.2%) 42 (14.6%) 12 (4.2%)
84 142 12 24
75 163 13 36
(32.1%) (54.2%) (4.6%) (9.1%)
183 (69.8%) 79 (30,2%) 19 (7.3%)
Sig'n!ficance
NS NS NS NS p=0,003 p=0,01 NS
(26.1%) (56.8%) (4.5%) (12.5%)
NS NS NS NS
190 (66.2%) 97 (33.8%) 17 (5.9%)
NS NS
NS, Not significant.
Table II, Indications for cesarean delivery Open group (n = 262)
Fetal distress Breech presentation Arrest of labor Abnormal position/presentation Preterm (uncontrollable) labor Cranial-pelvicdisproporfion Placental factors Elective cesarean delivery Gestosis Multiple pregnancy Other
63 48 38 22 23 15 12 9 8 8 16
(24.0%) (18.3%) (14.5%) (8.4%) (8.8%) (5.7%) (4.6%) (3.4%) (3.1%) (3.1%) (6.l%)
Closed group (n = 287)
69 43 52 29 19 18 11 11 10 9 16
(24.0%) (14.9%) (18.1%) (10.1%) (6.6%) (6.3%) (3.8%) (3.8%) (3.5%) (3.2%) (5.6%)
Significance
NS NS NS NS NS NS NS NS NS NS NS
NS, Not significant.
average temperature values during the first postoperative week were significantly higher with closure of the periton e u m (p< 0.001). Also, endometritis and wound infections tended to be m o r e c o m m o n in the d o s e d group; the differences between groups did not achieve significance. During hospitalization serious complications occurred in two patients. O n e 38-year-old patient (case 486, o p e n group) u n d e r w e n t repeat laparotomy on the first postoperative day because of a secondary hemorrhage. A hematoma in the abdominal wall was removed, and a small bleeding vessel at the uterine incision was ligated. The second case was a 15-year-old patient (case 526, closed group) in w h o m a large h e m a t o m a in the abdominal wall necessitated reoperation on the third postoperative day. Finally, there was a single case of readmission (a 30-yearold woman, case 75, open group) 17 days after the patient's discharge. H e r complaints were abdominal pain and diarrhea, and enteritis was diagnosed by means of stool cultures. She was treated in the usual m a n n e r and
discharged 5 days later. No patients in either group had p e M c abscess or peritonitis, and n o n e required reoperation because of ileus. The mean operating time (_+SD) was significantly greater in the closed group (56.9 _+ 17.9 minutes, range 13 to 115 minutes) than in the o p e n group (50.6 + t6.8 minutes, range 14 to 105 minutes) (p < 0.001). Comparison of the operating times in patients in whom cesarean section and tubal sterilization were p e r f o r m e d revealed no signit: icant difference between the closed group (62.6-+ 16.3 minutes, range 30 to 98 minutes) and the open group (61.0 + 29.1 minutes, range 22 to 115 minutes). General anesthesia time was significantly longer in the closed group (67.4 + 17.5 minutes, range 20 to 125 minutes) than in the open group (62.6 + 17.6 minutes, range 17 to 110 minutes) (p= 0.02). Also, postoperative hospitalization was significandy longer in the closed g r o u p (7.9 4- 1.8 days, range 4 to 23 days) than in the open group (7.2 + 1.6 days, range 2 to 14 days) (p < 0.001).
Volume 174, Number 4 AmJ ObstetGynecol
Nagele et al. 1369
Table III. Administration of antibiotics, narcotics, and bowel stimulants and perioperative and postoperative complications Open group (n = 262)
Antibiotics Prophylaxis Therapy Narcotics Bowel stimulants Chorioamnionifis Fever Cystitis Endometrifis Wound infection
177 61 20 18 32 22 8 8 5
(67.6%) (23.3%) (7.6%) (6.9%) (12.2%) (8.4%) (3.1%) (3.1%) (1.9%)
Closed group (n = 287)
180 97 64 18 28 45 22 17 14
(62.7%) (33.8%) (22.3%) (6.3%) (9.8%) (15.7%) (7.7%) (5.9%) (4.9%)
Significance
NS p= 0.006 p< 0.001 NS NS p=0.009 p=0.01 NS NS
NS, Not significant.
Analysis by institutions showed that significant differences between the open and closed groups were reflected in each center.
Comment This study examines the question of the closure or nonclosure of the visceral peritoneum at cesarean delivery from a critical-clinical point of view and compares the perioperative, intraoperative, and postoperative course in the two treatment groups. The most important aspects reviewed were operative and anesthesia times; antibiotic, bowel stimulant, and narcotic requirements; febrile morbidity; and postoperative complications. Our findings indicate that cesarean section without suture reapproximation of the peritoneal cut edges provides a n u m b e r of significant advantages. One is the possible decrease in operative time, which in turn is associated with shorter anesthesia exposure and diminished intraoperative anesthetic requirements. Of interest, pain medication requirements were lower in the open group, suggesting that nonclosure may be associated with less postoperative abdominal pain, possibly because no tension is placed on the peritoneal wound edges. Antibiotic requirements were also lower in the nonclosure group, which is most likely related to the lower febrile and infectious morbidity in this group. In theory, the higher rate of clinically asymptomatic febrile cases in the closed group might be due to the formation of subperitoneal pockets resulting from the suture; these pockets could fill with blood and wound secretions from the uterine incision and serve as a nutrient media for bacteria. Urinary tract infection and cystitis are common complications in obstetric and gynecologic surgery. Although the relationship between suture peritonization and the high incidence of cystitis in the closed group remains unclear, one hypothesis is that surgical manipulation on the peritoneal part of the bladder is more likely to lead to cystitis and urinary tract infection. One argument against nonclosure of the visceral peritoneum has been that it increases the risk of adhesion
formation. To reduce this risk, a variety of chemical additives 79 and, more recently, intraperitoneal barrier therapylO, 11have been used. However, the most important factor in adhesion prevention is impeccable surgical technique, including minimal tissue traumatization and avoidance of ischemia and inflammation by eliminating crushing forceps pressure, stitch tension, and knot pressure.l~ These demands are best met by leaving the visceral peritoneum open. Hubbard et al.~3 examined the process of autologous reperitonization in animal experiments. They observed that regeneration of peritoneal defects is completed in 5 to 6 days. After showing that the entire wound surface endothelializes simultaneously and not gradually from the cut edges as in skin wounds, they concluded that large peritoneal defects heal as fast as small ones. According to Holtz ~4and Elkins et al., 4 adhesion formation after peritoneal closure is primarily the result of foreign body reactions to the suture material, interruption of vascular supply, or ischemia, and tissue inflammation. They found less tissue necrosis and inflammatory reactions in peritoneal cut edges that were not closed primarily than in those that were, even when minimally reactive suture material was used. On the basis of these and similar findings, we doubt that there is any connection between nonclosure of the visceral peritoneum and postoperative adhesion formation. 15-17 Conversely, sutured peritoneal surfaces might cause adhesion formation) s To our knowledge, only two previous articles report on the effects of peritoneal closure or nonclosure in women undergoing cesarean delivery. In the study of Pietrantoni et al. 19 only the parietal peritoneum was left open, whereas Hull and Varner 2~left both the visceral and parietal peritoneum unsutured. In agreement with our findings, both studies reported a significant decrease in operating time in the open group. Hull and Varner found that pain medication requirements were higher in closure subjects, which is also consistent with our data. However, the closed group in their study required more bowel stimulants, a finding contrasting with our investigation,
1370
Nageleet al.
April 1996 Am.] ()bstt't (A nccol
which showed t h a t r e t u r n to bowel flmction was similar in b o t h groups. C o n t r a r y to o u r findings, Hull a n d V a r n e r f o u n d n o difference b e t w e e n the g r o u p s in the i n c i d e n c e of febrile morbidity. O n e e x p l a n a t i o n for this discrepancy may be f o u n d in the small series of patients used in t h e i r study (closed group, n = 59, vs o p e n g r o u p , n = 54). Furt h e r m o r e , febrile morbidity was d e f i n e d by j u s t o n e par a m e t e r ( t e m p e r a t u r e _>38~ C at least o n c e d u r i n g postoperative hospitalization), whereas we also i n c l u d e d the daily average t e m p e r a t u r e values o f the first postoperative week. In summary, the results of o u r study indicate that nonclosure of the visceral p e r i t o n e u m m i g h t n o t only r e d u c e o p e r a t i n g time a n d anesthesia e x p o s u r e b u t also offer o t h e r significant benefits with r e g a r d to the postoperative course, i n c l u d i n g a r e d u c t i o n in postoperative narcotics and, most i m p o r t a n t , a lower febrile a n d infectious morbidity a n d t h e r e f o r e d e c r e a s e d antibiotic r e q u i r e m e n t s . Finally, these factors also explain why hospitalization was s h o r t e r in the o p e n group. We c o n c l u d e t h a t r o u t i n e closure o f the visceral p e r i t o n e u m b e a b a n d o n e d in w o m e n u n d e r g o i n g cesarean delivery.
6. 7.
8. 9. 10.
11. 12. 13. 14. 15.
We t h a n k H e r b e r t Poltnig, o u r biostatistician, for his assistance in p e r f o r m i n g the statistical analyses of this study.
16.
REFERENCES
1. Placek PJ, Taffel SM. Recent patterns in cesarean delivery ill the United States. Obstet Gynecol Clin North Am 1987;15: 607-27. 2. Ellis H. The aetiology of postoperative abdominal adhesions. BrJ Surg 1962;50:10-6. 3. Ellis H. Internal overhealing: the problem ofintraperitoneal adhesions. WorldJ Surg 1980;4:303-6. 4. Elkins TE, Stovall TG, Warren J, Ling FW, Meyer NL. A histologic evaluation of peritoneal injury and repair: implications for adhesion formation. Obstet Gynecol 1987;70: 225-8. 5. McDonald MN, Elkins TE, Wortham GE Stovall TG, Ling
17. 18. 19, 20.
I'~ ~,McNeeley S(;. Adhesion tk)rmation and prevemion after peritoneal injury and repair ill the rabbit. J Reprod Med 1988;33:43(5-9. Nagele F, Husslem P. ls closure ot" visceral peritoneum in abdominal hysterectomy necessary: a retrospective pilot smdv. Gelmrtshilfe Frauenheilkd 1991:51:925-8. Golan A, Wexler S, I,otan G, Abramov L, Langer R, Favid MP. Calcium antagonist: eft(~ct on adhesion tormation. Acta Obstet G~llecol Scand 1989;68:529-32. Caballero J, Tulandi T. Effects of Ringer's lactate and fibrin glue on postsurgical adhesions. J Reprod Med 1992;37:141-3. Adhesion Study Group. Reduction of postoperative adhesions with intraperitoneal 32% dextran 70: a prospective, randomized trial_ Fertil Steril 1983;4(/:612-9. Sekiba K, Obstetrics and Gynecology Adhesion Pre~enlion Committee. Use of Interceed (TC7) absorbable adhesion barrier to reduce postoperative adhesion reformation in intertility and endometriosis surgery. Obstet Gwlecol 1992; 79:518-22. Surgical Membrane Study Group. Prophylaxis of pelvic sidewall adhesions with Gore-Tex surgical membrane: a multicenter clinical investigation. Fertil Steril 1992;57:921-3. Stone K. Adhesions in gynecologic surgery. Curr Opin Obstet Gynecol 1993;5:322-7. Hubbard BT, Khan MZ, Carag RV, Albites EV, Hricko (;M. The pathology of peritoneal repair: its relation to the formation of adhesions. Ann Surg 1967;165:908-16. Holtz G. Adhesion induction by suture of varying tissue reactivity and caliber. IntJ Fertil 1982;27:134-5. Tulandi T, Hum HS, Gelfand MM. Closure of laparotomy incisions with or ~ithout peritoneal suturing and se(ondlook laparoscopy. AsL] OBs'rEI"C,~x~(:ol. 1988;158:536-7. Thompson N], Paterson-Brown S, Harbourne T, Whawell SA, Kalodiki E, Dudley HAF. Reduced hnman peritoneal plasminogen activating activity.: possible mechanism of adhesion formation. BrJ Surg 1989;76:382-4. Vipond MN, Whawell SA, Thompson.IN, Dudlev HAF. Peritoneal fibrinolytic activity and intraabdominal adhesions. Lancet 1990;335:1120-2. Conolly WB, Stephens FO. Factors influencing tile incidence of intraperitoneal adhesions: an experimental study. Surgery 1968;63:9769. Pietrantoni M, Parsons MT, O'Brien ~A,T,Collins E, I~mppel RA, Spellacy WN. Peritoneal closure or non-closure at cesarean. Ohstet Gynecol 1991;77:293-6. Hull DB, Varner MW. A randomized study of closm'e of the peritoneum at cesarean delivery. Obstet GynecoI 1991;77: 818-21.