- Email: [email protected]
Peritoneal dialysis catheter sepsis: A medical and surgical dilemma
Peritoneal Dialysis Catheter Sepsis: A Medical and Surgical Dilemma
John L. Smith, MD, FACS and Michael J. Flanigan, MD, Iowa City, Iowa
Peritoneal dialysis in various forms, including longterm intermittent peritoneal dialysis, continuous ambulatory peritoneal dialysis, and continuous cyclic peritoneal dialysis, has established itself as an attractive alternative to hemodialysis in selected patients with end-stage renal disease. This is specifically true for infants and children in whom creation and maintenance of vascular access may be tenuous, in selected diabetics in whom the addition of insulin to the dialysate may assist blood glucose control, and in situations in which either medical center or home hemodialysis is not feasible [I]. Other advantages include improved patient mobility with continuous ambulatory peritoneal dialysis and continuous cyclic peritoneal dialysis, less need for protein and potassium restrictions, higher hemoglobin and hematocrit concentrations, better control of blood pressure, and maintenance of blood measurements at more steady-state values. Also, anticoagulation with heparin or citrate is not required and no routine obligatory blood loss is encountered [2]. In addition, peritoneal dialysis allows greater direct patient involvement than does medical center hemodialysis, and as a consequence, it may improve the patient’s psychosocial environment and selfimage. The first known description of the entity now known as peritoneal dialysis was published in the Muchener Medizinische Wochenschrift by G. Ganter in 1923 [3]. This article described the intermittent infusion and removal of saline solution into and out of the peritoneal cavity of a uremic guinea pig, which resulted in apparent clinical improvement [4]. In 1946 and 1948, Seligman et al [5], Fine Prom the Departments of Surgery and Medicine, University of Iowa College of Medicine, Iowa City, Iowa. Requests for reprints should be addressed to John L. Smith, MD, Department Of Surgery, 1532 JCP. University of Iowa Hospitals and Clinics, Iowa City, Iowa 52242. Presented at the 39th Annual Meeting of the Southwestern Surgical Congress, Coronado, California, April 26-29, 1987.
602
et al [6], and Frank et al [ 71published a technique of continuous peritoneal irrigation. This procedure was extrapolated from laboratory animals to 18 patients who were believed to have reversible renal failure. Four patients survived and five were found at postmortem examination to have irreversible renal lesions. Positive findings on peritoneal culture were found in 12 of the 18 patients during the treatment period [7]. Peritonitis occurred in this early clinical application and has remained a persistent problem and limiting factor in early widespread acceptance of peritoneal dialysis as a universally applicable clinical modality. In 1964, the introduction of long-term intermittent peritoneal dialysis by Palmer et al [8] made peritoneal dialysis a clinically significant form of therapy for end-stage renal disease. Further refinements of the technique with the introduction of continuous ambulatory peritoneal dialysis by Popovich et al [9] in 1978 and continuous cyclic peritoneal dialysis by Diaz-Buxo et al [IO] in 1981 likewise contributed to advancing the application of peritoneal dialysis. In addition, various modifications of the indwelling catheter (Tenckhoff’s modification of Palmer’s original catheter) and of the delivery systems (the switch from glass to plastic delivery containers and the introduction of automatic cyclers) rendered peritoneal dialysis less complicated for the patient and decreased the risk of infection. In a collective review by Vas [II], these refinements reduced the incidence of peritonitis in patients on long-term intermittent peritoneal dialysis from 5.2 to 7.5 episodes per patient per year of exposure to 0.23 to 1.2 episodes per patient per year of exposure. Additionally, in patients on continuous ambulatory peritoneal dialysis, the incidence decreased to 1.2 from 6.3 episodes per patient per year of exposure compared with the original 5.2 from 7.5 incidences per year. Although peritonitis, tunnel-cuff infection, and
The American Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
exit-site infection remain persistent problems, most infections can be treated with conservative measures such as exteriorization of the catheter tunnel or intraperitoneal antibiotics administered in the dialysate by the patient while at home [12]. However, a significant proportion of patients require hospitalization for more intensive medical management using parenteral antibiotics or interruption of peritoneal dialysis. Furthermore, a number of patients in this group will require surgical intervention for catheter removal or exploratory laparotomy if removal does not resolve the peritonitis. Material and Methods From October 2, 1978 to June 30, 1986, 207 patients seen by the University of Iowa Nephrology Service were placed on long-term intermittent peritoneal dialysis, continuous ambulatory peritoneal dialysis, or continuous cyclic peritoneal dialysis. The patients entered on long-term intermittent peritoneal dialysis received approximately 10 hours/day of medical center or home peritoneal dialysis 3 to 4 days/week. Those on continuous ambulatory peritoneal dialysis maintained continuous 2 liter volume fluid passes in the abdomen that were exchanged into plastic bags by the patient 3 to 5 times per day at approximately equal time intervals. Those patients on continuous cyclic peritoneal dialysis likewise received continuously indwelling fluid. Three to four 2l/z to 3 hour passes were performed by the automatic cycler at night, and a 12 to 14 hour indwelling pass remained during the day. Two catheter types were used: the double-cuffed nondisced Tenckhoff type and the double-cuffed double-disced Toronto type. In general, the nondisced Tenckhoff catheters were placed under local anesthesia by the nephrology service and the disced Toronto catheters were placed in the operating room under local or general anesthesia. The patients were observed through September 1,1986, totaling 3,552 months at risk for peritonitis. Peritonitis was diagnosed by abdominal pain associated with fever and leukocytosis, turbidity of the peritoneal effluent associated with leukocytosis (greater than 100 white blood cells/mm? and greater than 50 percent polymorphonuclear leukocytes), or by a positive finding on microbial culture from filtered peritoneal fluid. Treatment was initially randomized by the nephrology service in an ongoing study to primary single-drug intraperitoneal cefazolin or vancomycin. Results
Of the 207 patients started on peritoneal dialysis I), 45 were started on long-term intermittent peritoneal dialysis, with the first patient beginning October 2,1978 and the last beginning September 28,1986. One hundred twenty-six patients were entered on continuous ambulatory peritoneal dialysis beginning February 2,1979, and 32 were begun on continuous cyclic peritoneal dialysis beginning August 4,1982. Four patients were dialyzed in an unspecified manner between November 1983 and De(Table
Vofume 154, December 1967
TABLE I
Cause of Renal Failure
Diabetes mellitus Glomerulonephritis’ Atherosclerosis Polycystic disease Systemic lupus Obstruction Unknown Other+
LIPD
CAPD
CCPD
22 9
45 25 10
a
. 3
. 1 1
a
a a
6 3 2 2
5 4 14
1 7
Other
Total
1
76 40 14 13 11 7 6 29
. 1
. .
1 1
Membranoproliferative, membranous, immunoglobulin A nephritis. chronic glomerulonephritis, and focal segmental sclerosis. + Interstitial nephritis, analgesics, hemolytic uremic syndrome, vasculitis, lithiasis, pyelonephritis, Alport’s syndrome, Goodpasture’s syndrome, and multiple myeloma. CAPD = continuous ambulatory peritoneal dialysis; CCPD = continuous cyclic peritoneal dialysis; LIPD = long-term intermittent peritoneal dialysis. l
cember 1984. These four patients began dialysis while in an acutely moribund state in an intensive care unit. They all survived and were converted to long-term peritoneal dialysis. There were 100 male and 107 female patients with an age range of 2 to 86 years. The mean age for the entire group was 45.1 years, and the median age was 46 years with no statistically significant age differences among the three groups. There was a wide variety of diseases represented, but 76 of 207 patients (36.7 percent) were either type I (50 patients) or type II (26 patients) diabetics, with their renal failure secondary to diabetic glomerulopathy (Table I). There were 3,552 months of total patient exposure to the risk of peritonitis. There were 310 episodes of peritonitis documented during this period of exposure (Table II). Comparable with the peritonitis rates reported in the literature, this translates into 1.05 infections per patient per year of risk among the total group [11]. When observed from a different perspective, 113 of the patients followed experienced no episode of peritonitis, whereas 94 of 207 patients (45.4 percent) experienced all 310 episodes. The bacteriologic characteristics of the peritonitis varied considerably and comprised a wide variety of gram-positive, gram-negative, and fungal organisms (Table III). As expected, Staphylococcus epidermidis accounted for the most common isolate (90 of 310 cases, 29 percent). This, however, may be a somewhat smaller proportion of isolates than has been previously reported [13]. Gram-positive organisms accounted for 174 of the 310 cases (56.1 percent), gram-negative organisms accounted for 60 cases (19.4 percent), and fungal organisms for 6 cases (1.9 percent). Culture findings demonstrated no growth in 70 of the 310 cases (22.6 percent) that otherwise presented with abdominal pain and peritoneal fluid leukocytosis consistent with peritonitis.
603
Smith and Flanigan
TABLE II
Risk of Peritonitis
Patients Months at risk Episodes Risk/patient/year CAP0 = continuous peritoneal dialysis.
TABLE III
LIP0
CAP0
CCPD
Unspecified
Total
45 940.8 46 0.59
126 2,049.a 218 1.28
32 543.5 44 0.97
4 17.9 2 1.34
207 3,552 310 1.05
ambulatory
peritoneal
dialysis: CCPD = continuous
Bacteriologic and Mycologic Data on 310 Episodes of Peritonitis
Gram-Positive (n = 174) Staphylococcus epidermidis Staphylococcus aureus Streptococcus species Other gram-positive Gram-Negative
90 48 27 9 (n = 60) 12 12 9 a 6 13
Escherichia coli Other Enterobacteriaceae’ Acinetobacter Mixed gram-negative+ Pseudomonas Other gram-negative Fungal and Culture-Negative
(n = 76) 6 70
Fungal No orowth Proteus, Klebsiella, Enterobacter, + includes Bacteroides species. l
and Serratia.
Antibiotics were used in all 310 documented cases of peritonitis. Cefazolin, vancomycin, and tobramytin were the most commonly used, either as single agents or in combination. Occasionally, rifampin in combination with vancomycin, was used when appropriate culture and sensitivity studies indicated its necessity. Amphotericin was used when fungal organisms were present. Cefazolin was used as a primary single agent on 141 occasions. Eleven relapses (7.8 percent) and two Candida superinfections (1.4 percent) occurred. Vancomycin was used 112 times as the single primary agent. Only two relapses (1.8 percent) and no superinfections occurred (chi-square analysis, p <0.03) in the vancomycin-treated patients. With this single-agent regimen, 130 of 310 peritonitis episodes (41.9 percent) were successfully treated without hospital admission. Conversely, 180 of 310 of the episodes (58.1 percent) were severe enough, with symptoms of nausea and emesis or with signs of systemic sepsis, relapse, or superinfection, to require hospital admission. Of these 180 admissions, 128 cases were successfully treated with intensive intraperitoneal 604
cyclic
peritoneal
dialysis; LIP0 = long-term
intermittent
antibiotics and intravenous antibiotics, and in accordance with presently accepted etiologic theories of this type of peritonitis, with interruption of peritoneal dialysis (see “Discussion” section). Maintenance hemodialysis supported the patients during this period of time. Fifty-two of 310 of the episodes (16.8 percent) could not be managed with conservative treatment. These 52 septic episodes occurred in 47 patients. Forty patients underwent one operation, with 28 catheter removals and 12 laparotomies, and 6 underwent two operations, wherein 4 had two catheter removals and 2 had a catheter removed at one time and later, during a totally separate episode of peritonitis, underwent laparotomy. One patient died before either catheter removal or laparotomy could be performed. This patient was a 43 year old extremely debilitated type I diabetic who presented with Pseudomonas peritonitis and rapidly progressed to generalized sepsis and multiorgan system failure. The patient died without an attempt at surgical intervention due to patient and family decision. Of the 28 patients who had single procedures other than laparotomy, 13 had removal of the peritoneal dialysis catheter and were converted to hemodialysis. Fifteen, however, were returned to peritoneal dialysis. Ten of these patients had no further episodes of peritonitis, and 5 experienced one to nine episodes of sepsis which responded to conservative measures. Twelve patients underwent exploratory laparotomy. Of the six patients who underwent two procedures separated by time and, frequently, other episodes of peritonitis, three were returned to peritoneal dialysis. Two of these patients had no further infection, one had further episodes but was treated conservatively, and the fourth was converted to hemodialysis. The last two patients proceeded to laparotomy. All patients who underwent exploratory laparotomy were maintained on hemodialysis until death or were converted permanently to hemodialysis if they survived or had a functional renal transplant (one patient). Defined in another light, 15 of these 47 patients experienced major morbidity greater than simple peritoneal catheter removal. Twelve of 15 underwent 16 exploratory laparotomies. Six died and six survived. Three patients did not undergo exploratory laparotomy. The first patient had removal of the The American Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
TABLE IV
Epbdea of Perlionitb, lhwapy, and Outcome
Organism
Episodes
Culture negative Gram-positive Gram-negative Fungal l
70 174 60 6
(22.6) (56.1) (19.4) (1.9)
Conservative Therapy 64170 152/174 42160 O/6
(91.4) (67.3) (70) (0)
Operative Therapy 6170 22/174 16160 6/6
(6.6) (12.7) (30) (100)
Deaths 0 0 4116 516
(0) (0) (22.2) (83.3)
Values in parentheses are percentages.
peritoneal catheter but died secondary to a myocardial infarction before a planned laparotomy could be performed. The second patient progressed to generalized sepsis from a combination of decubitus ulcers, osteomyelitis, and peritonitis. This patient refused laparotomy and died from septic hypotensive cardiac arrest. The third patient died without any operative intervention at her own request. The age range of these 15 patients was 14 to 79 years (mean 48.3 years, median 43 years). Ten of these patients were female and 5 were male. Dialysis type included long-term intermittent peritoneal dialysis in 1 patient, continuous ambulatory peritoneal dialysis in 11 patients, and continuous cyclic peritoneal dialysis in 3 patients. The disease entities included polycystic disease in five patients, diabetes mellitus in four patients, hypertension in two patients, systemic lupus in one patient, immunoglobulin A nephropathy in one patient, Goodpasture’s syndrome in one patient, and atherosclerosis in one patient. It is important to note that 10 of 13 patients (76.9 percent) with polycystic kidney disease experienced 1 to 10 episodes of peritonitis. Two of the 13 patients required dialysis catheter removal. One had successful return to peritoneal dialysis and the second was converted to hemodialysis. Five of 13 patients (38.5 percent) progressed to exploratory laparotomy. Four of the five died. Three patients had 3 to 10 episodes of peritonitis but were successfully treated with medical management. Examining the results from a bacteriologic standpoint produced some interesting findings (Table IV). In general, the peritonitis produced by gramnegative and fungal organisms demonstrated significantly higher morbidity and mortality rates than those episodes produced by gram-positive organisms or episodes in which culture revealed no growth (chi-square analysis, p <0.005). In all but 3 of the 15 cases involving laparotomy or death, multiple episodes of peritonitis preceded the final event.
comments Peritonitis has traditionally been and continues to be a serious and life-threatening complication involving all three types of peritoneal dialysis. Two hundred fifty-eight of 310 episodes (83.2 percent) Volume 154, December 1987
responded to home or inpatient conservative management. In the remaining 52 of 310 episodes (16.8 percent), however, potentially life-threatening peritonitis occurred. Several theories as to the source of infection have been recorded in the literature. Skin contamination of the catheter or cannula is one of the most common [12,13]. In addition, exit-site infection and tunnel infections can lead to peritonitis [12]. These infections are usually caused by gram-positive organisms and, in our series, were the most likely to respond appropriately to conservative methods of treatment such as intraperitoneal antibiotics or exteriorization of the catheter tract. Intraabdominal pathologic abnormalities, including diverticulosis, appendicitis, perforated ulcer, gynecologic infection, and cholecystitis, have all been implicated in isolated instances [12-151. Socioeconomic status has been statistically identified as a risk factor [14]. Opsonic deficiency of peritoneal macrophages caused by the continuously indwelling dialysate, membrane croasing, and depletion of immunoglobulins have all been demonstrated to contribute to the inability of the peritoneal membrane to clear normally removable bacteria and fungi. Compliment fixation capability in the peritoneal cavity has likewise been demonstrated to correlate with the ability of the patient to clear peritoneal sepsis. In the 195Os, Schweinberg et al [16] demonstrated transmural migration of Escherichia coli through the bowel wall to be theoretically possible [14]. Although the significance of this finding remains unclear, it is an interesting point which might help to explain the peritonitis that some investigators have attributed to diverticulosis
[17l. Our results are remarkable in regard to the difficulties encountered during the management of problem peritonitis cases. Contrary to previous reports, no specific trend from gram-positive to gramnegative organism could be discerned in the present study [18]. Likewise, the occurrence of prior grampositive or gram-negative sepsis did not specifically identify the population at risk for life-threatening peritonitis. The presence of fungi in the culture specimen did point to a poor outcome. In this subgroup, fungal infection occurred after multiple episodes of peritonitis in five of six patients. This occurred as a superinfection in three of six patients or 605
Smith and Flanigan
as mixed flora with gram-negative organisms in three of six patients. The timing of aggressive surgical intervention appears to play a major role in determining the outcome in difficult peritonitis cases. In the present study, if there was no dramatic response to conservative therapy within the first few days, the outcome tended to be far less satisfactory. Those patients with peritonitis who were treated for more than 2 weeks without a major response died. Although early, aggressive surgical therapy did not guarantee patient survival, survivors came only from this group. It also appears that gram-positive and no-growth peritonitis may be somewhat more benign in its course and may be more safely treated in an aggressively managed nonoperative fashion than can those cases of peritonitis that involve gram-negative and fungal organisms (Table IV). Fungal sepsis carried the worst prognosis. In fact, the single survivor required two separate exploratory laparotomies and two hospitalizations before the disease process finally resolved.
Summary Peritoneal dialysis remains a viable and valuable alternative to hemodialysis in selected patients; however, the development of intraperitoneal sepsis should raise serious questions as to whether a particular patient should remain with this particular mode of dialysis. Six conclusions can be drawn from this retrospective review. (1) Vancomycin appears to be the first single drug of choice, especially in cases of gram-positive peritonitis. (2) In our experience, the dialysis catheter should be removed in patients who do not demonstrate major resolution of their peritoneal sepsis by 3 to 4 days. (3) If removal of the dialysis catheter does not resolve the issue within 2 to 3 days, exploratory laparotomy should be seriously considered. (4) If fungal organisms are present, exploration and debridement of the peritoneal cavity should be carried out and the patient should be aggressively treated with systemic amphotericin. This should be undertaken early in the course of the peritonitis. (5) Patients with polycystic kidney disease may be better served by hemodialysis. (6) Patients who experience multiple septic episodes should be, when feasible, electively converted to hemodialysis or should undergo transplantation. References 1. Baum M. Powell D, Calvin S, et al. Continuous ambulatory peritoneal dialysis in children: comparison with hemodialysis. N Engl J Med 1982; 307: 1537-42. 2. Olcott C IV, Feldman CA, Coplon NS, Oppenheimer ML, Mehigan JT. Continuous ambulatory peritoneal dialysis: technique of catheter insertion and management of associatedsurgical complications. Am J Surg 1983; 148: 98-100. 3. Ganter G. Uber die Beseitigung giftiger Stoffe aus dem Blute durch Dialyse. Munchener Medizinische Wochenschrift 1923: 70: 1478-80.
606
4. Boen ST. History of peritoneal dialysis. In: Nolph KD, ed. Peritoneal dialysis. Boston: Martinus Nijhoff, 1985: l-22. 5. Seligman AM, Frank HA, Fine J. Treatment of experimental uremia by means of peritoneal irrigation. J Clin Invest 1946; 25: 21 l-9. 8. Fine J, Frank HA, Seligman AM. The treatment of acute renal failure by peritoneal irrigation. Ann Surg 1946; 124: 85778. 7. Frank HA, Seligman AM, Fine J. Further experiences with peritoneal irrigation for acute renal failure. Ann Surg 1948; 128: 561-608. 8. Palmer RA, Quinton WE, Gray JE. Prolonged peritoneal dialysis for chronic renal failure. Lancet, 1964; 1: 700-2. 9. Popovich RP, Moncrief JW, Nolph KD, Ghods AJ, Twardowski ZJ, Pyle WK. Continuous ambulatory peritoneal dialysis. Ann Intern Med 1978; 88: 449-58. 10. Diaz-Buxo JA, Farmer CD, Walker PJ, Chandler JT, Halt JL. Continuous cyclic peritoneal dialysis: a preliminary report. Artif Organs 1981; 5: 157-61. 11. Vas SI. Peritonitis. In: Nolph KD, ed. Peritoneal dialysis. Boston: Martinus Nijhoff, 1985: 403-39. 12. Piraino B, Bernardini J, Sorkin M. The influence of peritoneal catheter exit-site infections on peritonitis, tunnel infections, and catheter loss in patients on continuous ambulatory peritoneal dialysis. Am J Kidney Dis 1988; 8: 435-40. 13. Gokal R. Peritonitis in continuous ambulatory peritoneal dialysis. J Antimicrob Chemother 1982; 9: 417-22. 14. Rubin J, Ray R, Barnes T, et al. Peritonitis in continuous ambulatory peritoneal dialysis patients. Am J Kidney Dis 1983; 6: 602-9. 15. Keane WF, Comty CM, Verbrugh HA, Peterson PK. Opsonic deficiency of peritoneal dialysis effluent in continuous ambulatory peritoneal dialysis. Kidney Int 1984; 25: 539-43. 16. Schweinburg FB, Seligman AM, Fine J. Transmural migration of intestinal bacteria: a study based on the use of radioactive escherichia coli. N Engl J Med 1950; 242: 747-51. 17. Schulak JA, Flanigan MJ, Nghiem DD, Corry RJ. Ambulatory peritoneal dialysis. Arch Surg 1984; 119: 1400-4. 18. Rubin J, Rogers WA, Taylor HM, et al. Peritonitis during continuous ambulatory peritoneal dialysis. Ann Intern Med 1980; 92: 7-13.
DiscUSSiOll Russell E. Postier (Oklahoma City, OK): Dr. Smith, I think that it is not surprising that cefazolin is not a good agent to use intraperitoneally because it is an antibiotic that has a significant binding to protein sites and thus probably becomes relatively biounavailable fairly early after infusion. You recommend laparotomy if there is no early resolution of the peritonitis with removal of the peritoneal dialysis catheter; however, you did not detail what the findings were at laparotomy. I would assume that peritoneal irrigation can be just as successful through the peritoneal dialysis catheter as through open laparotomy. Some believe that the presence of polymicrobial infections indicates something more than simple peritonitis and may well indicate that the catheter eroded into an area where it shouldn’t be. How many of your patients had polymicrobial infections? Did that have any effect on overall outcome? Were any episodes of peritonitis related to exit site infections, and if so, did therapy of the exit site infection significantly alter the outcome?
John L. Smith (closing): Dr. Postier, of our patients with polymicrobial infections, only one demonstrated documented evidence of diverticular disease, and in a The Amerlcan Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
second patient, the catheter was inserted incorrectly as was proved at laparotomy. In regard to exit site infections, there was no definite correlation in our particular group of patients with exit site infection and secondary peritonitis. I realize there is at least one published study that suggests such a correlation. In our group, the exit site infections were all due to skin contamination. A relationship between peritonitis and infection was suspected in 1 of the 12 patients with cuff infection who underwent laparotomy, however, that relationship was never proved. In regard to the operative findings, the most common was
Volume 154, December 1987
multiple small-interloop abscesses. There were two larger subdiagphragmatic abscesses. One patient was known preoperatively to have a large pelvic abscess that was fungal in origin. In regard to antibiotics, our nephrology service is still involved with an ongoing randomized study of cefazolin and vancomycin. This is usually used as primary intraperitoneal home therapy. I am pretty sure from what we found in the present study that they are going to begin using vancomycin. It is far better, especially against the most common organism, which is Staph. epidermidis.
607
John L. Smith, MD, FACS and Michael J. Flanigan, MD, Iowa City, Iowa
Peritoneal dialysis in various forms, including longterm intermittent peritoneal dialysis, continuous ambulatory peritoneal dialysis, and continuous cyclic peritoneal dialysis, has established itself as an attractive alternative to hemodialysis in selected patients with end-stage renal disease. This is specifically true for infants and children in whom creation and maintenance of vascular access may be tenuous, in selected diabetics in whom the addition of insulin to the dialysate may assist blood glucose control, and in situations in which either medical center or home hemodialysis is not feasible [I]. Other advantages include improved patient mobility with continuous ambulatory peritoneal dialysis and continuous cyclic peritoneal dialysis, less need for protein and potassium restrictions, higher hemoglobin and hematocrit concentrations, better control of blood pressure, and maintenance of blood measurements at more steady-state values. Also, anticoagulation with heparin or citrate is not required and no routine obligatory blood loss is encountered [2]. In addition, peritoneal dialysis allows greater direct patient involvement than does medical center hemodialysis, and as a consequence, it may improve the patient’s psychosocial environment and selfimage. The first known description of the entity now known as peritoneal dialysis was published in the Muchener Medizinische Wochenschrift by G. Ganter in 1923 [3]. This article described the intermittent infusion and removal of saline solution into and out of the peritoneal cavity of a uremic guinea pig, which resulted in apparent clinical improvement [4]. In 1946 and 1948, Seligman et al [5], Fine Prom the Departments of Surgery and Medicine, University of Iowa College of Medicine, Iowa City, Iowa. Requests for reprints should be addressed to John L. Smith, MD, Department Of Surgery, 1532 JCP. University of Iowa Hospitals and Clinics, Iowa City, Iowa 52242. Presented at the 39th Annual Meeting of the Southwestern Surgical Congress, Coronado, California, April 26-29, 1987.
602
et al [6], and Frank et al [ 71published a technique of continuous peritoneal irrigation. This procedure was extrapolated from laboratory animals to 18 patients who were believed to have reversible renal failure. Four patients survived and five were found at postmortem examination to have irreversible renal lesions. Positive findings on peritoneal culture were found in 12 of the 18 patients during the treatment period [7]. Peritonitis occurred in this early clinical application and has remained a persistent problem and limiting factor in early widespread acceptance of peritoneal dialysis as a universally applicable clinical modality. In 1964, the introduction of long-term intermittent peritoneal dialysis by Palmer et al [8] made peritoneal dialysis a clinically significant form of therapy for end-stage renal disease. Further refinements of the technique with the introduction of continuous ambulatory peritoneal dialysis by Popovich et al [9] in 1978 and continuous cyclic peritoneal dialysis by Diaz-Buxo et al [IO] in 1981 likewise contributed to advancing the application of peritoneal dialysis. In addition, various modifications of the indwelling catheter (Tenckhoff’s modification of Palmer’s original catheter) and of the delivery systems (the switch from glass to plastic delivery containers and the introduction of automatic cyclers) rendered peritoneal dialysis less complicated for the patient and decreased the risk of infection. In a collective review by Vas [II], these refinements reduced the incidence of peritonitis in patients on long-term intermittent peritoneal dialysis from 5.2 to 7.5 episodes per patient per year of exposure to 0.23 to 1.2 episodes per patient per year of exposure. Additionally, in patients on continuous ambulatory peritoneal dialysis, the incidence decreased to 1.2 from 6.3 episodes per patient per year of exposure compared with the original 5.2 from 7.5 incidences per year. Although peritonitis, tunnel-cuff infection, and
The American Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
exit-site infection remain persistent problems, most infections can be treated with conservative measures such as exteriorization of the catheter tunnel or intraperitoneal antibiotics administered in the dialysate by the patient while at home [12]. However, a significant proportion of patients require hospitalization for more intensive medical management using parenteral antibiotics or interruption of peritoneal dialysis. Furthermore, a number of patients in this group will require surgical intervention for catheter removal or exploratory laparotomy if removal does not resolve the peritonitis. Material and Methods From October 2, 1978 to June 30, 1986, 207 patients seen by the University of Iowa Nephrology Service were placed on long-term intermittent peritoneal dialysis, continuous ambulatory peritoneal dialysis, or continuous cyclic peritoneal dialysis. The patients entered on long-term intermittent peritoneal dialysis received approximately 10 hours/day of medical center or home peritoneal dialysis 3 to 4 days/week. Those on continuous ambulatory peritoneal dialysis maintained continuous 2 liter volume fluid passes in the abdomen that were exchanged into plastic bags by the patient 3 to 5 times per day at approximately equal time intervals. Those patients on continuous cyclic peritoneal dialysis likewise received continuously indwelling fluid. Three to four 2l/z to 3 hour passes were performed by the automatic cycler at night, and a 12 to 14 hour indwelling pass remained during the day. Two catheter types were used: the double-cuffed nondisced Tenckhoff type and the double-cuffed double-disced Toronto type. In general, the nondisced Tenckhoff catheters were placed under local anesthesia by the nephrology service and the disced Toronto catheters were placed in the operating room under local or general anesthesia. The patients were observed through September 1,1986, totaling 3,552 months at risk for peritonitis. Peritonitis was diagnosed by abdominal pain associated with fever and leukocytosis, turbidity of the peritoneal effluent associated with leukocytosis (greater than 100 white blood cells/mm? and greater than 50 percent polymorphonuclear leukocytes), or by a positive finding on microbial culture from filtered peritoneal fluid. Treatment was initially randomized by the nephrology service in an ongoing study to primary single-drug intraperitoneal cefazolin or vancomycin. Results
Of the 207 patients started on peritoneal dialysis I), 45 were started on long-term intermittent peritoneal dialysis, with the first patient beginning October 2,1978 and the last beginning September 28,1986. One hundred twenty-six patients were entered on continuous ambulatory peritoneal dialysis beginning February 2,1979, and 32 were begun on continuous cyclic peritoneal dialysis beginning August 4,1982. Four patients were dialyzed in an unspecified manner between November 1983 and De(Table
Vofume 154, December 1967
TABLE I
Cause of Renal Failure
Diabetes mellitus Glomerulonephritis’ Atherosclerosis Polycystic disease Systemic lupus Obstruction Unknown Other+
LIPD
CAPD
CCPD
22 9
45 25 10
a
. 3
. 1 1
a
a a
6 3 2 2
5 4 14
1 7
Other
Total
1
76 40 14 13 11 7 6 29
. 1
. .
1 1
Membranoproliferative, membranous, immunoglobulin A nephritis. chronic glomerulonephritis, and focal segmental sclerosis. + Interstitial nephritis, analgesics, hemolytic uremic syndrome, vasculitis, lithiasis, pyelonephritis, Alport’s syndrome, Goodpasture’s syndrome, and multiple myeloma. CAPD = continuous ambulatory peritoneal dialysis; CCPD = continuous cyclic peritoneal dialysis; LIPD = long-term intermittent peritoneal dialysis. l
cember 1984. These four patients began dialysis while in an acutely moribund state in an intensive care unit. They all survived and were converted to long-term peritoneal dialysis. There were 100 male and 107 female patients with an age range of 2 to 86 years. The mean age for the entire group was 45.1 years, and the median age was 46 years with no statistically significant age differences among the three groups. There was a wide variety of diseases represented, but 76 of 207 patients (36.7 percent) were either type I (50 patients) or type II (26 patients) diabetics, with their renal failure secondary to diabetic glomerulopathy (Table I). There were 3,552 months of total patient exposure to the risk of peritonitis. There were 310 episodes of peritonitis documented during this period of exposure (Table II). Comparable with the peritonitis rates reported in the literature, this translates into 1.05 infections per patient per year of risk among the total group [11]. When observed from a different perspective, 113 of the patients followed experienced no episode of peritonitis, whereas 94 of 207 patients (45.4 percent) experienced all 310 episodes. The bacteriologic characteristics of the peritonitis varied considerably and comprised a wide variety of gram-positive, gram-negative, and fungal organisms (Table III). As expected, Staphylococcus epidermidis accounted for the most common isolate (90 of 310 cases, 29 percent). This, however, may be a somewhat smaller proportion of isolates than has been previously reported [13]. Gram-positive organisms accounted for 174 of the 310 cases (56.1 percent), gram-negative organisms accounted for 60 cases (19.4 percent), and fungal organisms for 6 cases (1.9 percent). Culture findings demonstrated no growth in 70 of the 310 cases (22.6 percent) that otherwise presented with abdominal pain and peritoneal fluid leukocytosis consistent with peritonitis.
603
Smith and Flanigan
TABLE II
Risk of Peritonitis
Patients Months at risk Episodes Risk/patient/year CAP0 = continuous peritoneal dialysis.
TABLE III
LIP0
CAP0
CCPD
Unspecified
Total
45 940.8 46 0.59
126 2,049.a 218 1.28
32 543.5 44 0.97
4 17.9 2 1.34
207 3,552 310 1.05
ambulatory
peritoneal
dialysis: CCPD = continuous
Bacteriologic and Mycologic Data on 310 Episodes of Peritonitis
Gram-Positive (n = 174) Staphylococcus epidermidis Staphylococcus aureus Streptococcus species Other gram-positive Gram-Negative
90 48 27 9 (n = 60) 12 12 9 a 6 13
Escherichia coli Other Enterobacteriaceae’ Acinetobacter Mixed gram-negative+ Pseudomonas Other gram-negative Fungal and Culture-Negative
(n = 76) 6 70
Fungal No orowth Proteus, Klebsiella, Enterobacter, + includes Bacteroides species. l
and Serratia.
Antibiotics were used in all 310 documented cases of peritonitis. Cefazolin, vancomycin, and tobramytin were the most commonly used, either as single agents or in combination. Occasionally, rifampin in combination with vancomycin, was used when appropriate culture and sensitivity studies indicated its necessity. Amphotericin was used when fungal organisms were present. Cefazolin was used as a primary single agent on 141 occasions. Eleven relapses (7.8 percent) and two Candida superinfections (1.4 percent) occurred. Vancomycin was used 112 times as the single primary agent. Only two relapses (1.8 percent) and no superinfections occurred (chi-square analysis, p <0.03) in the vancomycin-treated patients. With this single-agent regimen, 130 of 310 peritonitis episodes (41.9 percent) were successfully treated without hospital admission. Conversely, 180 of 310 of the episodes (58.1 percent) were severe enough, with symptoms of nausea and emesis or with signs of systemic sepsis, relapse, or superinfection, to require hospital admission. Of these 180 admissions, 128 cases were successfully treated with intensive intraperitoneal 604
cyclic
peritoneal
dialysis; LIP0 = long-term
intermittent
antibiotics and intravenous antibiotics, and in accordance with presently accepted etiologic theories of this type of peritonitis, with interruption of peritoneal dialysis (see “Discussion” section). Maintenance hemodialysis supported the patients during this period of time. Fifty-two of 310 of the episodes (16.8 percent) could not be managed with conservative treatment. These 52 septic episodes occurred in 47 patients. Forty patients underwent one operation, with 28 catheter removals and 12 laparotomies, and 6 underwent two operations, wherein 4 had two catheter removals and 2 had a catheter removed at one time and later, during a totally separate episode of peritonitis, underwent laparotomy. One patient died before either catheter removal or laparotomy could be performed. This patient was a 43 year old extremely debilitated type I diabetic who presented with Pseudomonas peritonitis and rapidly progressed to generalized sepsis and multiorgan system failure. The patient died without an attempt at surgical intervention due to patient and family decision. Of the 28 patients who had single procedures other than laparotomy, 13 had removal of the peritoneal dialysis catheter and were converted to hemodialysis. Fifteen, however, were returned to peritoneal dialysis. Ten of these patients had no further episodes of peritonitis, and 5 experienced one to nine episodes of sepsis which responded to conservative measures. Twelve patients underwent exploratory laparotomy. Of the six patients who underwent two procedures separated by time and, frequently, other episodes of peritonitis, three were returned to peritoneal dialysis. Two of these patients had no further infection, one had further episodes but was treated conservatively, and the fourth was converted to hemodialysis. The last two patients proceeded to laparotomy. All patients who underwent exploratory laparotomy were maintained on hemodialysis until death or were converted permanently to hemodialysis if they survived or had a functional renal transplant (one patient). Defined in another light, 15 of these 47 patients experienced major morbidity greater than simple peritoneal catheter removal. Twelve of 15 underwent 16 exploratory laparotomies. Six died and six survived. Three patients did not undergo exploratory laparotomy. The first patient had removal of the The American Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
TABLE IV
Epbdea of Perlionitb, lhwapy, and Outcome
Organism
Episodes
Culture negative Gram-positive Gram-negative Fungal l
70 174 60 6
(22.6) (56.1) (19.4) (1.9)
Conservative Therapy 64170 152/174 42160 O/6
(91.4) (67.3) (70) (0)
Operative Therapy 6170 22/174 16160 6/6
(6.6) (12.7) (30) (100)
Deaths 0 0 4116 516
(0) (0) (22.2) (83.3)
Values in parentheses are percentages.
peritoneal catheter but died secondary to a myocardial infarction before a planned laparotomy could be performed. The second patient progressed to generalized sepsis from a combination of decubitus ulcers, osteomyelitis, and peritonitis. This patient refused laparotomy and died from septic hypotensive cardiac arrest. The third patient died without any operative intervention at her own request. The age range of these 15 patients was 14 to 79 years (mean 48.3 years, median 43 years). Ten of these patients were female and 5 were male. Dialysis type included long-term intermittent peritoneal dialysis in 1 patient, continuous ambulatory peritoneal dialysis in 11 patients, and continuous cyclic peritoneal dialysis in 3 patients. The disease entities included polycystic disease in five patients, diabetes mellitus in four patients, hypertension in two patients, systemic lupus in one patient, immunoglobulin A nephropathy in one patient, Goodpasture’s syndrome in one patient, and atherosclerosis in one patient. It is important to note that 10 of 13 patients (76.9 percent) with polycystic kidney disease experienced 1 to 10 episodes of peritonitis. Two of the 13 patients required dialysis catheter removal. One had successful return to peritoneal dialysis and the second was converted to hemodialysis. Five of 13 patients (38.5 percent) progressed to exploratory laparotomy. Four of the five died. Three patients had 3 to 10 episodes of peritonitis but were successfully treated with medical management. Examining the results from a bacteriologic standpoint produced some interesting findings (Table IV). In general, the peritonitis produced by gramnegative and fungal organisms demonstrated significantly higher morbidity and mortality rates than those episodes produced by gram-positive organisms or episodes in which culture revealed no growth (chi-square analysis, p <0.005). In all but 3 of the 15 cases involving laparotomy or death, multiple episodes of peritonitis preceded the final event.
comments Peritonitis has traditionally been and continues to be a serious and life-threatening complication involving all three types of peritoneal dialysis. Two hundred fifty-eight of 310 episodes (83.2 percent) Volume 154, December 1987
responded to home or inpatient conservative management. In the remaining 52 of 310 episodes (16.8 percent), however, potentially life-threatening peritonitis occurred. Several theories as to the source of infection have been recorded in the literature. Skin contamination of the catheter or cannula is one of the most common [12,13]. In addition, exit-site infection and tunnel infections can lead to peritonitis [12]. These infections are usually caused by gram-positive organisms and, in our series, were the most likely to respond appropriately to conservative methods of treatment such as intraperitoneal antibiotics or exteriorization of the catheter tract. Intraabdominal pathologic abnormalities, including diverticulosis, appendicitis, perforated ulcer, gynecologic infection, and cholecystitis, have all been implicated in isolated instances [12-151. Socioeconomic status has been statistically identified as a risk factor [14]. Opsonic deficiency of peritoneal macrophages caused by the continuously indwelling dialysate, membrane croasing, and depletion of immunoglobulins have all been demonstrated to contribute to the inability of the peritoneal membrane to clear normally removable bacteria and fungi. Compliment fixation capability in the peritoneal cavity has likewise been demonstrated to correlate with the ability of the patient to clear peritoneal sepsis. In the 195Os, Schweinberg et al [16] demonstrated transmural migration of Escherichia coli through the bowel wall to be theoretically possible [14]. Although the significance of this finding remains unclear, it is an interesting point which might help to explain the peritonitis that some investigators have attributed to diverticulosis
[17l. Our results are remarkable in regard to the difficulties encountered during the management of problem peritonitis cases. Contrary to previous reports, no specific trend from gram-positive to gramnegative organism could be discerned in the present study [18]. Likewise, the occurrence of prior grampositive or gram-negative sepsis did not specifically identify the population at risk for life-threatening peritonitis. The presence of fungi in the culture specimen did point to a poor outcome. In this subgroup, fungal infection occurred after multiple episodes of peritonitis in five of six patients. This occurred as a superinfection in three of six patients or 605
Smith and Flanigan
as mixed flora with gram-negative organisms in three of six patients. The timing of aggressive surgical intervention appears to play a major role in determining the outcome in difficult peritonitis cases. In the present study, if there was no dramatic response to conservative therapy within the first few days, the outcome tended to be far less satisfactory. Those patients with peritonitis who were treated for more than 2 weeks without a major response died. Although early, aggressive surgical therapy did not guarantee patient survival, survivors came only from this group. It also appears that gram-positive and no-growth peritonitis may be somewhat more benign in its course and may be more safely treated in an aggressively managed nonoperative fashion than can those cases of peritonitis that involve gram-negative and fungal organisms (Table IV). Fungal sepsis carried the worst prognosis. In fact, the single survivor required two separate exploratory laparotomies and two hospitalizations before the disease process finally resolved.
Summary Peritoneal dialysis remains a viable and valuable alternative to hemodialysis in selected patients; however, the development of intraperitoneal sepsis should raise serious questions as to whether a particular patient should remain with this particular mode of dialysis. Six conclusions can be drawn from this retrospective review. (1) Vancomycin appears to be the first single drug of choice, especially in cases of gram-positive peritonitis. (2) In our experience, the dialysis catheter should be removed in patients who do not demonstrate major resolution of their peritoneal sepsis by 3 to 4 days. (3) If removal of the dialysis catheter does not resolve the issue within 2 to 3 days, exploratory laparotomy should be seriously considered. (4) If fungal organisms are present, exploration and debridement of the peritoneal cavity should be carried out and the patient should be aggressively treated with systemic amphotericin. This should be undertaken early in the course of the peritonitis. (5) Patients with polycystic kidney disease may be better served by hemodialysis. (6) Patients who experience multiple septic episodes should be, when feasible, electively converted to hemodialysis or should undergo transplantation. References 1. Baum M. Powell D, Calvin S, et al. Continuous ambulatory peritoneal dialysis in children: comparison with hemodialysis. N Engl J Med 1982; 307: 1537-42. 2. Olcott C IV, Feldman CA, Coplon NS, Oppenheimer ML, Mehigan JT. Continuous ambulatory peritoneal dialysis: technique of catheter insertion and management of associatedsurgical complications. Am J Surg 1983; 148: 98-100. 3. Ganter G. Uber die Beseitigung giftiger Stoffe aus dem Blute durch Dialyse. Munchener Medizinische Wochenschrift 1923: 70: 1478-80.
606
4. Boen ST. History of peritoneal dialysis. In: Nolph KD, ed. Peritoneal dialysis. Boston: Martinus Nijhoff, 1985: l-22. 5. Seligman AM, Frank HA, Fine J. Treatment of experimental uremia by means of peritoneal irrigation. J Clin Invest 1946; 25: 21 l-9. 8. Fine J, Frank HA, Seligman AM. The treatment of acute renal failure by peritoneal irrigation. Ann Surg 1946; 124: 85778. 7. Frank HA, Seligman AM, Fine J. Further experiences with peritoneal irrigation for acute renal failure. Ann Surg 1948; 128: 561-608. 8. Palmer RA, Quinton WE, Gray JE. Prolonged peritoneal dialysis for chronic renal failure. Lancet, 1964; 1: 700-2. 9. Popovich RP, Moncrief JW, Nolph KD, Ghods AJ, Twardowski ZJ, Pyle WK. Continuous ambulatory peritoneal dialysis. Ann Intern Med 1978; 88: 449-58. 10. Diaz-Buxo JA, Farmer CD, Walker PJ, Chandler JT, Halt JL. Continuous cyclic peritoneal dialysis: a preliminary report. Artif Organs 1981; 5: 157-61. 11. Vas SI. Peritonitis. In: Nolph KD, ed. Peritoneal dialysis. Boston: Martinus Nijhoff, 1985: 403-39. 12. Piraino B, Bernardini J, Sorkin M. The influence of peritoneal catheter exit-site infections on peritonitis, tunnel infections, and catheter loss in patients on continuous ambulatory peritoneal dialysis. Am J Kidney Dis 1988; 8: 435-40. 13. Gokal R. Peritonitis in continuous ambulatory peritoneal dialysis. J Antimicrob Chemother 1982; 9: 417-22. 14. Rubin J, Ray R, Barnes T, et al. Peritonitis in continuous ambulatory peritoneal dialysis patients. Am J Kidney Dis 1983; 6: 602-9. 15. Keane WF, Comty CM, Verbrugh HA, Peterson PK. Opsonic deficiency of peritoneal dialysis effluent in continuous ambulatory peritoneal dialysis. Kidney Int 1984; 25: 539-43. 16. Schweinburg FB, Seligman AM, Fine J. Transmural migration of intestinal bacteria: a study based on the use of radioactive escherichia coli. N Engl J Med 1950; 242: 747-51. 17. Schulak JA, Flanigan MJ, Nghiem DD, Corry RJ. Ambulatory peritoneal dialysis. Arch Surg 1984; 119: 1400-4. 18. Rubin J, Rogers WA, Taylor HM, et al. Peritonitis during continuous ambulatory peritoneal dialysis. Ann Intern Med 1980; 92: 7-13.
DiscUSSiOll Russell E. Postier (Oklahoma City, OK): Dr. Smith, I think that it is not surprising that cefazolin is not a good agent to use intraperitoneally because it is an antibiotic that has a significant binding to protein sites and thus probably becomes relatively biounavailable fairly early after infusion. You recommend laparotomy if there is no early resolution of the peritonitis with removal of the peritoneal dialysis catheter; however, you did not detail what the findings were at laparotomy. I would assume that peritoneal irrigation can be just as successful through the peritoneal dialysis catheter as through open laparotomy. Some believe that the presence of polymicrobial infections indicates something more than simple peritonitis and may well indicate that the catheter eroded into an area where it shouldn’t be. How many of your patients had polymicrobial infections? Did that have any effect on overall outcome? Were any episodes of peritonitis related to exit site infections, and if so, did therapy of the exit site infection significantly alter the outcome?
John L. Smith (closing): Dr. Postier, of our patients with polymicrobial infections, only one demonstrated documented evidence of diverticular disease, and in a The Amerlcan Journal of Surgery
Peritoneal Dialysis Catheter Sepsis
second patient, the catheter was inserted incorrectly as was proved at laparotomy. In regard to exit site infections, there was no definite correlation in our particular group of patients with exit site infection and secondary peritonitis. I realize there is at least one published study that suggests such a correlation. In our group, the exit site infections were all due to skin contamination. A relationship between peritonitis and infection was suspected in 1 of the 12 patients with cuff infection who underwent laparotomy, however, that relationship was never proved. In regard to the operative findings, the most common was
Volume 154, December 1987
multiple small-interloop abscesses. There were two larger subdiagphragmatic abscesses. One patient was known preoperatively to have a large pelvic abscess that was fungal in origin. In regard to antibiotics, our nephrology service is still involved with an ongoing randomized study of cefazolin and vancomycin. This is usually used as primary intraperitoneal home therapy. I am pretty sure from what we found in the present study that they are going to begin using vancomycin. It is far better, especially against the most common organism, which is Staph. epidermidis.
607