- Email: [email protected]
Bacteremia in transplant recipients: A prospective study of demographics, etiologic agents, risk factors, and outcomes
B a c t e r e m i a in t r a.n. .s. .p. l a n t . r e c ,"p ,"e n t s . • A prospective study of demographics) etiologic agents, risk factors, and outcomes Marilyn M. Wagener, MPH Victor L. Yu, MD Pittsburgh, Pennsylvania
Background: Bacteremic infections are a major cause of death among organ transplant
recipients. We sought to identify the risk factors associated with death and examine the timing of the bacteremic episode after operation to recognize patients who may benefit from perioperafive prophylactic antibiotic therapy. Methods: A total of 125 episodes of bacteremia or fungemia in 16 heart, 26 kidney, and 70 liver recipients were monitored prospectively in 1 year. Results: The urinary tract was the most frequent portal for kidney recipients, the gastrointestinal and biliary tracts were frequent for liver recipients, and the lung was frequent in heart recipients. Heart and liver recipients were more severely ill at the time of bacteremia and had bacteremia sooner after operation. Death at 14 days after onset of bacteremia was 33% in heart recipients, 24% in liver recipients, and 11% in kidney recipients. Risk of death was associated with the severity of the underlying condition of the transplant recipient, the source of the bacteremia, and the microbial agent. Pseudomonas aeruginosa and Enterobacter species had fatality rates of 47% and 63%, respectively. P. aeruginosa and Enterobacter were also most commonly associated with failures of perioperative antibiotic prophylaxis. Conclusions: There are distinct clinical patterns of bacteremia in transplant recipients. The emergence of P. aeruginosa and Enterobacter species in the immediate postoperative period appeared to be a significant cause of morbidity and death among transplant recipients. (AJIC AM J INFECTCONTROL1992;20:239-47)
I n f e c t i o n r e m a i n s a m a j o r c a u s e of m o r b i d i t y a n d d e a t h a m o n g o r g a n t r a n s p l a n t r e c i p i e n t s . 1-a Bacteremia has been associated with an especially h i g h m o r t a l i t y r a t e in this p a t i e n t p o p u l a t i o n . 4 W e p r o s p e c t i v e l y t r a c k e d all b a c t e r e m i a s in o u r p o p u l a t i o n of t r a n s p l a n t r e c i p i e n t s to define this p a t i e n t p o p u l a t i o n a n d t h e i r r i s k f a c t o r s f o r death. We f u r t h e r a n a l y z e d the m i c r o b i a l o r i g i n s o f the b a c t e r e m i c e p i s o d e s w i t h r e s p e c t to the t i m e of o n s e t of b a c t e r e m i a a n d the p o s s i b i l i t y t h a t the bacteremia might have been prevented by postoperative prophylactic antibiotic therapy) From the Division of Infectious Disease, University of Pittsburgh School of Medicine, Pittsburgh, Pa. Reprint requests: Victor L. Yu, MD, University of Pittsburgh School of Medicine, Scaife 968, Pittsburgh, PA 15261.
17/46/36079
METHODS Study population All t r a n s p l a n t r e c i p i e n t s w h o w e r e h o s p i t a l i z e d with blood cultures positive for pathogens during a 1 - y e a r p e r i o d (19 8 7) w e r e p r o s p e c t i v e l y t r a c k e d until d i s c h a r g e o r death. The s i g n i f i c a n c e o f a p o s i t i v e b l o o d c u l t u r e w a s b a s e d o n the clinical j u d g m e n t of the i n f e c t i o u s d i s e a s e a n d s u r g i c a l t r a n s p l a n t services. All p a tients w h o r e c e i v e d specific a n t i b i o t i c t h e r a p y f o r the b a c t e r e m i a w e r e i n c l u d e d in the study. Additional cultures positive for the same organism w i t h i n a 3 - d a y p e r i o d w e r e c o n s i d e r e d to r e p r e sent the same episode.
Study definitions and procedures The s t a n d a r d p e r i o p e r a t i v e p r o p h y l a c t i c a n t i b i otic r e g i m e n s w e r e as follows: k i d n e y r e c i p i e n t s , 239
A,.c
Wagener and Yu
240 Table
October 1992
1. Demographics of bacteremic transplant recipients Parameter
Heart recipients
Incidence during study 11% (11/98) Age (yr) Mean 48,7 Range 16-63 Male 67% (12/18) White 100% (18/18) WBC (1000 cells/ram 3) Mean 15.0 Range 0.5-38 <3 11% (2/18) Hernatocrit Mean 34.1 Median 33.7 Range 24-49 Creatinine level (rng/dl) Median 1.7 Mean 3.01 Range 0.5-15 Liver function tests--median (range) Bilirubin (mg/di) 2.1 (0.6-36) SGOT (IU/L) 38.8 (11-6465) SGPT (IU/L) 19.5 (5-3894) Receiving CyA 100% (18/18) CyA level (l~g/ml) Median 460 Range 99-1072 Receiving steroids 89% (16/18) Receiving azathioprine 50% (9/18) Receiving OKT3 6% (1/18) In ICU 44% (8/18) ReslSirator 33% (6/18)
Kidney recipients
Liver recipients
6% (11/90)
24% (69/284)
42.2 21-67 44% (12/27) 63% (17/27)
45 18-67 55% (44/80) 88% (70/80)
12.8 2.6-31 4% (1/27)
12.7 0.1-33 11% (9/80)
/.
p Value*
"
NS NS .01 NS NS
29.6 26.9 19-47
27.8 27.9 17-40
3.4 3,89 1.2-13.6
2.3 3.0 0.4-41
0.02 NS
1.0 (0.4-39) 31 (7-521) 16.0 (5-217) 81% (22/27)
5.3 (0.6-44) 258 (8-5747) 45 (2-4110) 97% (79/80)
<0.01" NS <0.01" 0.01
583 43-1447 96% (26/27) 18% (5/27) 4% (1/27) 15% (4/27) 11% (3/27)
762 115-2275 95% (76/80) 5% (4/80) 32% (26/80) 50% (40/80) 30% (24/80)
0.011-
0.06* NS <0.01 <0.01 <0.01 NS Cont'd
NS, Not Significant (p > 0.05); WBC,whiteblood cells; SGOT,aspartateaminotransferase;GTP,guanosinetriphosphate; SGPT,alanineaminotransferase;CyA, cyclosp0rin A; ICU, intensivecare unit; IV, intravenous. *By X~ test except as indicated. iKruskal-Wallis test.
cefazolin; heart recipients, cefamandole; and liver -recipients, cefotaxime plus ampicillin. The first dose w a s given before operation and therapy was continued for 2 to 5 days at the discretion of the surgeon. A bacteremic episode w a s considered nosocomial if the onset was longer than 48 hours after hospitalization or within 1 w e e k of discharge. All other episodes were considered community acquired. A bacteremic episode was considered to have an early onset if the positive blood culture o c c u r r e d within 14 days after the organ transplantation operation. Failure of antibiotic prophylaxis was defined as a blood culture positive for pathogens within 5 days after organ transplantation operation or an additional operation for which the same prophylactic regimen was used. Mortality rates w e r e defined by deaths within 14 days of positive blood culture.
A severity of illness index at the time of bacteremia was b a s e d on the assignment of points as follows: fever ( > 3 8 ° C = 1 point, _>40 ° C = 2 points), mental status (disorientation = 1 point, stupor -- 2 points, c o m a = 4 points), hypotension (2 points), respirator support (2 points), and cardiac arrest (4 points). Patients were considered to be seriously ill if the point total was greater than or equal to 4. This grading system has been shown in other studies of b a c t e r e m i a to be highly predictive of outcome. 5' 6 Statistical
analysis
Statistical analysis w a s done with the P R O P H E T system (National Institutes of Health, Division of R e s e a r c h Resources). Univariate analysis of contingency data was done with ×2 or Fisher exact tests. The t test was used for compar-
Volume 20 Number 5
Table
Posttransplant bacteremia 2 4 1
1. Continued
Parameter . . . . .
Heart recipients
Kidney recil~ients
Liver recipients
p Vaiue*
Shock Mental status decrease Fever (> 37° C) Retransplant Severity of illness z 4 Polymicrobial Diabetes Hemodialysis Community-acquired Portal Urinary tract Pneumonia Wound/skin Abdominal/biliary IV site Unknown Other Timing (days after transplant) Mean Median Range < 14 days < 30 days < 60 days > 1 year Organism also isolated Sputum Lung Urine Wound Vascular catheter 2-week mortality rate
17% (3/18) 50% (9/18) 89% (16/18) 11% (2/18) 39% (7/18) 6% (1/18) 11% (2/18) 1% (1/18) 22% (5/18)
15% (4/27) 14% (4/27) 93% (25/27) 30% (8/27) 11% (3/27) 7% (2/27) 33% (9/27) 22% (6/27) 44% (12/27)
10% (8/80) 45% (36/80) 92% (74/80) 29% (23/80) 29% (23/80) 11% (9/80) 6% (5/80) 11% (9/80) 6% (5/80)
NS 0.02 NS NS 0.02 NS 0.03 NS < 0.01
11% (2/18) 22% (4/18) 17% (3/18) 0% (0/18) 6% (1/! 8) 39% (7/18) 6% (1/18)
41% (11/27) 4% (1/27) 26% (7/27) 0% (0/27) 7% (2/27) 22% (6/27) 0% (0/27)
7% (6/80) 10% (8/80) 6% (5/80) 27% (22/80) 6% (5/80) 41% (33/80) 1% (1/80)
< 0.01 0.14 0.03 < 0.01 NS NS NS
295 60 2-1074 22% (4/18) 44% (8/18) 50% (9/18) 39% (7/18)
573 392 17-2856 0% (0/27) 11% (3/27) 30% (8/27) 56% (15/2i)
88 26 1-630 35% (28/80) 59% (47/80) 75% (60/80) 9% (8/80)
44% (8/18) 11% (2/18) 33% (6/18) 18% (3/18) 1! % (2/18) 33% (6/18)
4% (1/27) 0% (0/27) 41% (11/27) 18% (5/27) 7% (2/27) 11% (3/27)
41% (33/80) 7% (6/80) 14% (11/80) 15% (12/80) 10% (8/80) 24% (19/80)
ing the means of continuous variables that approximate a normal distribution, as determined by the Wilk-Shapiro test for normality (age, hematocrit, serum creatinine level, and. white blood cell count). The Mann-Whitney test was used for comparing all other continuous variables (liver function tests, duration of hospitalization, and cyclosporin levels). The Kruskal-Wallis test (an extension of the nonparametric Mann-Whitney test) was used to compare continuous variables between the three transplant groups (Table 1). RESULTS
• One h u n d r e d sixty-one episodes of bacteremia were found in 146 transplant recipients within the 1-year study period. Thirteen patients ;had multiple episodes of bacteremia, with two transplant recipients having three episodes each. Thirty-three of the episodes were considered clinically insignificant, including infections with coagulase-negative staphylococci (20 patients), Bacillus species (four patients), Corynebacterium
< 0.01 * < 0.01 < 0.01 < 0.01 < 0.01 <0.01 NS <0.01 NS NS NS
(diphtheroid) species (three patients), Propionibacterium species (two patients), Flavobacterium species (one patient), Lactobacillus species (one patient), Clostridium species (one patient), and an unidentified gram-positive coccus (one patient). Antibiotic therapy was not administered for these episodes and these 33 patients were therefore excluded from further analysis. Three patients with multiorgan transplants (one heart-lung, two kidney-liver),, who each had one episode of bacteremia, were also excluded. The 125 remaining episodes occurred in 111 patients. Eighteen bacteremias occurred in 16 heart recipients, 27 occurred in 26 kidney recipients, and 80 occurred in 69 liver recipients. Thirteen patients had more than one episode of bacteremla. These episodes were separated by 11 to 335 days (mean, 62.4 days, median, 32 days). Of the episodes, 62% involved different organisms and 38% occurred during distinct hospital admissions. The five cases in which the same organism recurred were treated as independent events
AJIC October 1992
Wagener and Yu
242
0.9 0.8 .~ 0.7~ .=- 0.6 ..Q
'° t
.,......-. ..:,:.!"
:"
:
! 0
heart
I
0.5- : ' ~ E a. 0.4- •" 0.30.2- ~1 ,~I" 0.1- I Hf 0
liver
,..- ....
r kidney r ........
I
I
J" .............
I
100 200 300 Days post transplantation
a
I
380
Fig, 1. Kaplan-Meier curves comparing the number of days between transplantation and the onset of bacteremia or fungemia. Almost 80% of bacteremias in liver recipients occurred within the first 90 days. The x axis represents the number of days after transplantation. The y axis represents the cumulative probability, based on the total number of patients that had bacteremia during the study, of a transplant recipient having bacteremia.
because more than 28 days had passed between the two positive blood cultures and clinical evidence of relaps e was not obvious. The incidence of bacteremia during the study period was 11% (11/98) for heart recipients, 6% (11/190) for kidney recipients, and 24% (69/284) for liver recipients. The remaining bacteremias occurred in patients who underwent transplantation before the study year. T y p e of o r g a n t r a n s p l a n t
Differences between major transplant sites are summarized in Table 1. The ages and sex distributions of the heart, liver, and kidney recipients who had bacteremia were comparable. Kidney recipients with bacteremia were in general less severely ill than were liver and heart transplant recipients (Table 1); only 15 % of kidney recipients were in the intensive care unit, received mechanical ventilation, or had decreased mental status or vital sign abnormality; this compares with 44% and 50% heart and liver recipients, respectively. Kidney recipients were also less likely to have the bacteremic organism isolated from their sputum and more likely to have it in the urine. Pneumonia was most frequent in heart transplant recipients and infections originating in abdominal/biliary tract were seen exclusively in liver recipients. Forty-four percent of the bacteremias in kidney recipients were considered community acquired,
compared with 22% among the heart recipients and 6% among the liver recipients? Immunosuppressive chemotherapy regimens differed among the transplant sites, with the kidney recipients less likely to receive cyclosporin A. Liver recipients had higher m e a n cyclosporin levels and were more likely to have been given OKT3 to combat rejection. Bacteremias occurred later in the kidney recipient, with about 50% occurring more than 1 year after transplantation. About 75% of the bacteremias in liver recipients occurred within the first 2 months after transplantation. The Kaplan-Meier curves depicting the number of days between transplantation and the onset of bacteremia were significantly different for the groups (p < 0.01, Gehan-Breslow test; Fig. 1). Portal
The source of the bacteremia or fungemia varied with the organ transplanted (Table 1). The most common portal identified was the abdominal/biliary site, which was involved in 18% (22/125) of the bacteremic episodes; this portal was seen exclusively in liver recipients. The urinary tract was the source in 15% of the bacteremias; 58% (11/19) of these occurred in the kidney recipient. Wound or skin infections were implicated in 12% of the bacteremias, with kidney recipients having the highest percentage of wound infections: 26% (7/27), compared with 17% (3/18) in heart recipients and 6% (5/80) in liver recipients. Pneumonia was the origin identified in 10% (13/125) of the bacteremias: 25% (4/18) of heart recipients, 10% (8/80) of liver recipients, and 4% (1/27) of kidney recipients. Six percent of the bacteremias were attributed to intravascular catheters. One case was related to a gastrointestinal infection and one case to meningitis. Thirty-eight percent of all bacteremias (46/125) were from an unidentified portal of entry. Microbial agent
Aerobic gram-negative bacilli constituted 48% of the pathogens found by blood culture in kidney recipients, 49% in the liver recipients, and 39% in the heart recipients (Fig. 2). Escherichia coli, Klebsiella, and other Enterobacteriaceae were t h e predominant gram-negative organisms seen in kidney recipients. Pseudomonas aeruginosa and Enterobacter species accounted for most of the gram-negative bacteria seen in liver recipients. P. aeruginosa and E. coli were the gram-negative bacilli seen most often in heart recipients.
Volume 20 Number 5
Posttransplant bacteremia 2 4 3 S.aureus s.epider midis Enterobacter sp P.aeruginosa Gram neg enterics E.coli Gram positive cocci Fungi Enterococcus sp Anaerobes Miscellaneous A.hydrophilia S.pneumoniae
~!i~!i~!i~i~!!i~;i~iJ!i!}~!~!i~!~!i~!;!i~:~!i~!~.1i:~!~i~i~i~i~i!~i~!z~!i!~!i!!;!~iFiJi!~i!i;!!i~i~i~i~:i~!~!:!~!~
i 0
5
I
Heart
I
I
I
I
I
10
15
20
25
30
~
Liver
~
Kidney
Fig. 2. The major Jnfectious agents seen in 125 blood cultures from organ transplant recipients. Gram-negative enteric bacteria: Citrobacter freundii (four), Klebsiella pneumoniae (four), Serratia marcescens (two), and Proteus mirabilis (two). Gram-positive cocci: Streptococcus viridans group (four), Staphylococcus intermedius grou p (three), Streptococcus pyogenes (one), and Streptococcus agalactiae (one). Fungi: Candida albicans (two), Candida tropicalis (two), Candida stelloidea (two), and Candida glabrata (one). Anaerobic bacteria: Bacillus fragilis (three) and Clostridiurn species (one), Miscellaneous: Corynebacterium group JK (two), Listeria (one).
Gram-positive organisms were seen in about 50% of the bacteremias among liver, heart, and kidney recipients. Staphylococcus aureus was the predominant gram-positive organism seen in both heart and liver recipients and was second only to Staphylococcus epidermidis in kidney recipients. Fungi were present in 6% (5/80) of the positive blood cultures from liver recipients and in 11% (2/18) from heart recipients. None of the kidney transplant recipients had fungemia. The fungi isolated were as follows: Candida albicans (two), Candida tropicalis (two), Candida stellatoidea (two), and Candida glabrata (one). Anaerobic bacteria were present in 3% (5/125) of the posttransplant bacteremias-Bacteroides fragilis (three), Clostridium species (two). Ten percent (12/126) of the positive blood cultures were polymicrobial. Multiple organisms were isolated in 11% (9/80) of the bacteremias in liver recipients, 6% (1/18) of thos6 in heart recipients, and 7% (2/27) of those in kidney recipients. Community versus hospital acquired
Survival rate was somewhat higher for patients with• c0mmunity-acquired bacteremia (90%;
19/21) than for those with nosocomial bacteremia (75%; 78/104), although statistical significance was not attained ( p - - 0 . 1 5 , xa). Patients with community-acquired bacteremia had significantly lower severity of illness index values t h a n did those with nosocomial bacteremia (p < 0.01, X2); only 9% (2/21) of the former group were classified as severely ill. In general the baeteremic infection was the reason for admission for patients with community-acquired bacteremia. Communityacquired organisms were predominantly E. coli, S. aureus, and other Enterobacteriaceae (Fig. 3). There were no P. aeruginosa infections and only one Enterobacter among the patients with community-acquired bacteremia. Timing of b a c t e r e m l a onset Failures of antibiotic prophylaxis. Fifteen episodes of bacteremia occurred within 5 days of
operation and had a mortality rate of 67%. The predomifient organisms were Enterobacter species (40% [6/15]) and Pseudomonas aeruginosa (33% [5/15]; Table 2). The surgical procedures included eight liver transplants, two heart transplants, and five posttransplant exploration and repair procedures (four liver, one kidney). These
AJIC 244
October 1992
Wagener and Yu S.aureus S.epidermidis P.aeruginosa Enterobacter sp Gram neg enterics
~//////~///////~///~////~////////////////~/~//~//////////////~
i
Gram pos cocci Fungi Enterococcus sp Anaerobes A.hydrophilia JK diptheroids Listeria S.pneumoniae 0
I
I
I
I
5
10
15
20
~
nosocomial
community acquired
25
Fig. 3. Sixty-two percent of the organisms involved in community-acquired bacteremias were gram-negative enteric bacteria. Gram-negative enteric bacteria: Escherichia coli (11), Citrobacter freundii (four), Klebsiella pneumoniae (four), Serratia marcescens (two), and Proteus mirabilis (two). Gram-positive cocci: Streptococcus viridans group (four), Staphylococcus intermedius group (three), Streptococcus pyogenes (one), and Streptococcus agalactiae (one). Fungi: Candida albicans (two), Candida tropicalis (two), Candida stelloidea (two), and Candida glabrata (one). Anaerobic bacteria: Bacillus fragilis (three) and Clostridium species (one).
five surgical procedures occurred 7 to 60 days after transplantation. Early versus late onset. Twenty-six percent (32/125) ofl the bacteremias occurred within 14 days of transplantation and were considered to be early onset. This included bacteremias in 22% (4/18) of the heart and 35% (28/80) of the liver transplant recipients. None of the bacteremias in kidney recipients were classified as early onset. _The predominant organisms seen in these early bacteremias were Enterobacter species (31% [10/32]), P. aeruginosa (28% [9/32]), and Staphylococcus aureus (25% [8/32]). Twenty-two percent (7/32) of the early-onset bacteremias had a pulmonary source, as opposed to only 6% (6/93) of the bacteremias with later onset. Seventy-four percent (93/125) of the bacteremias occurred longer than 14 days after transplantation and were considered late onset. The predominant organisms seen after 2 weeks were Staphylococcus epidermidis (18% [17/93]), Staphylococcus aureus (17% [16/93]), E. coli (12% [11/93]), and P. aeruginosa (11% [10/93]). The late-onset bacteremias were more likely to have a urinary tract portal (19% [18/93]) than were early bacteremias (3% [1/32]).
Mortality rates
The 2-week mortality rate after bacteremia for all patients was 23%. The mortality rate was 33% (6/18) among the heart recipients, 24% (19/80) among the liver recipients, and 11% (3/27) among the kidney recipient s . Within 60 days 37% of the bacteremic patients had died, with most of the subsequent deaths occurring in the liver recipients. Associated with increased mortality rate were retransplantation (p < 0.01, X2), polymicrobial infection (p < 0.05, X2), a n d p n e u m o n i a (p < 0.01, ×2). Patients who died also had higher severity of illness index scores (p < 0.01, X2), higher serum creatinine levels (p - 0.01, t test) and higher liver function test abnormalities, even when controlled for site of transplantation (p < 0.01, Mann-Whitney). Death occurred in 47% (15/32) of the patients with early-onset bacteremia and 15% (14/93) of the patients with late-onset bacteremia (p < 0.01, X2). Patients with gram-negative bacteremia and fungemia had a significantly higher mortality rate (35% [22/62]) than did those with gram-positive bacteremia (8% [4/53]; p < 0.01, X2). P. aerugi- " nosa and Enterobacter species were associated
Volume 20 Number 5
with most of the fatal gram-negative bacteremias (Table 3). DISCUSSION
Infection has been recognized as a major source of morbidity and, death since the early days of organ transplantation. Improvements in surgical techniques and immunosuppressive protocols have been followed by a reduction in severe bacterial infections. The incidence of bacteremia among transplant recipients, however, remains relatively high. In the 1970s liver transplantation was complicated by an incidence of bacteremia greater than 70%, 7 generally reflecting contami-nation by gastrointestinal organisms. This incidence was subsequently reduced to 25% to 35%.1, 4, a Heart recipients had a 30% incidence of bacteremia in the early 1970s; a 1987 study2 still reported a 20% rate of posttransplant bacteremia. In kidney recipients, the rate of bacteremia has decreased from 26% to 60% in the late 1960s 9"10to about 12% in the 1980s. ~ At one of the largest organ transplant centers in the world, we had the opportunity to compare the demographic and risk factors for bacteremia in three major groups of transplant recipients. We compared the microbial agent, the timing after transplantation, the portal of infection, and the subsequent mortality rate for bacteremic episodes in kidney, heart, and liver recipients. As expected, the major sources of the bacteremic organism differed among the three groups. Urinary tract infection was the primary source of bacteremia among kidney recipients, abdominal and biliary infections were the major sources among liver recipients, and pneumonia was the primary source among heart recipients. Forty-one percent of the bacteremias in kidney recipients originated in the urinary tract; an infected dialysis access site was the source in 21%. This is a smaller percentage of urinary tract infections than that previously reported (50%) for kidney transplant recipients, a' 9 Most kidney recipients in this study had bacteremia more than 1 year after transplantation, rather than during the immediate postoperative period. Return to dialysis appears to be related to an increase in the number of gram-positive skin infections and may be at least partially responsible for a lower percentage of gram-negative organisms than previously reported (50% vs 60% to 70%). a, 10-12 Staphylococcus aureus was found in about 20% of the bacteremias for each of the three groups. Enterobacter species were seen almost exclusively
Posttransplant bacteremia 2 4 5 T a b l e 2. Microbial agent and morality rate of
bacteremia for failures of prophylaxis (bacteremia occurring within 5 days of transplantation or other operation*) Bacteremia rate
Liver recipients Enterobacter species 5% (4/80) Pseudomonas aeruginosa 3% (2/80) Staphylococcus aureus 4% (3/80) Fungi 3% (2/80) Polymicrobiall1% (1/80) TOTAL 15% (12/80) Heart recipients. Pseudomonas aeruginosa 6% (1/18) Serratia marcescens 6% (1/18) TOTAL " 11% (2/18) Kidney Polymicrobialt4% (1/27) TOTAL 12% (15/125)
Mortality rate
100% (4/4) 50% (1/2) 0% (0/3) 50% (1/2) 100% (1/1) 58% (7/12) 100% (1/1) 100% (1/1) 100% (2/2) 100% (1/1) 67% (10/15)
*Other operations accounted for five episodes of bacteremia after surgical exploration/repairafter transplantation. These patients received the same antibiotic regimen as used during organ transplantation and operations occurred 7 to 60 days after transplantation. l-Polymicrobialbacteremia included two episodes (one liver recipient, one kidney recipient) in which Pseudomonas aeruginosa and Enterobacter species were simultaneouslyisolated.
in liver recipients. P. aeruginosa was seen in 20% of the bacteremias in liver recipients and in 11% • of heart recipients. Staphylococcus aureus and Staphylococcus epidermidis were the organisms most frequently seen in kidney recipients, rather than enteric gram-negative bacteria as reported in earlier studies.3' 10. 11Pseudomonas and Enterobacter accounted for about 40% of bacteremias in liver recipients and most of the deaths (65%). These two organisms were also more likely to be present in patients considered to have early onset (infection within 2 weeks of transplantation). Twelve bacteremia episodes occurring within 5 postoperative days were considered to be failures of antibiotic prophylaxis. Of these bacteremias, 58% (7/12) were caused by P. aeruginosa, Enterobacter species, or both (Table 2). The current prophylaxis regimen of cefotaxime/ampicillin used in the liver transplant recipients appeare d to be associated with superinfection with P. aeruginosa and Enterobacter species. P. aeruginosa is resistant to cefotaxime and ampicillin, the regimen used at the time of liver transplantation. In other studies, previous use of third-generation cephalosporins has been linked to the emergence of Enterobacter resistant to those agents.13 The overall mortality rate after bacteremic episodes was 23%, deaths occurred primarily
246
.&JI@ October 1992
Wagener and Yu
T a b l e 3. Mortality rate among bacteremic transplant recipients subclassified by microbial agent Single organism
Polymicroblal*
Total
Enterobacter species Pseudomonas aeruginosa Fungi Other gram-positive cocci
60% 38% 45% 40%
(9/15) (5/13) (4/7) (2/5)
75% (3/4) 67% (4/6) 0% (0/5)
63% 47% 43% 20%
(12/19) (9/19) (4/7) (2/10)
Enterococcus species Staphylococcus epidermidis Staphylococcus aureus Gram-negative enteric bacteria Anaerobes Miscellaneous Other gram-negative bacilli
0% (0/5) 7% (1/15) 4% (1/23) 5% (1/22) 0% (0/4) 0% (0/2) 50% (1/2)
50% (1/2) 20% (1/5) 0% (0/1) 0% (0/1) 0% (0/1) -
14% (1/7) 10% (2/20) 4% (1/24) 4% (1/23) 0% (0/4) 0% (0/3) 50% (1/2)
Fungi: Candida albicans (two), Candida tropicalis (two), Candida stellatiodea (two), and Candida glabrata (one). Other gram-positive cocci: Streptococcus viridans group (four), Streptococcus intermedius group (three), Streptococcus pyogenes (one), Streptococcus agalactiae (one) and Streptococcus pneumoniae (one). Gram-negative enteric bacteria: Escherichia coil (11),. Citrobacter freundii (four), Klebsiella pneumonia (four), Serratia marcescens (two), and Proteus rnirabilis (two). Other gram-negative bacilli: Aeremonas hydrophila (two). Miscellaneous: Corynebacterium group JK (two) and Listeria (one). Anaerobes: Bacteroides fragilis (three) and Clostridium species (one). *There were 12 polymicrobial bacteremias with 25 organisms. Pseudomonas aeruginosa and/or Enterobacter species were involved in all fatal cases (42% [5/12]) of polymicrobial bacteremias.
among the heart and liver transplant recipients. Mortality rate was directly related to the underlying condition of the patient. These patients had significantly higher serum creatinine levels, more abnormal results of liver function tests (even when controlled for site of transplant), and higher severity of illness scores. Patients who were bacteremic within 14 days of transplantation had a 47% mortality rate. Retransplantation was also significantly associated with a higher mortality rate; 39% (13/53) of bacteremic patients with a second transplanted organ died. Bacteremic pneumonia carried a mortality rate of 60%, compared with 5% among those with a urinary tract source and 20% among patients with an abdominal or biliary portal. Aerobic gram-negative bacilli and yeasts Were more likely to be associated with death. Transplant recipients with P. aeruginosa bacteremia had twice the mortality rate and those with Enterobacter bacteremia had four times the mortality rate of patients with other agents of bacteremia. Patients with P. aeruginosa or Enterobacter were significantly more likely to have a high severity of illness index score and to have recently received antimicrobial agents. Polymicrobial bacteremias had a 42% (5/12) mortality rate. All five fatal cases involved P. aeruginosa, Enterobacter species, or both. Mortality rate was significantly higher for earlyonset bacteremia (47% [15/32]) than for bacteremia occurring later (14% [13/93]). Enterobacter species bacteremia occurring within the first 2 weeks after transplantation had a 70% (7/10)
mortality rate, P. aeruginosa bacteremia had a 67% (6/9) mortality rate, and fungemia had a 67% (2/3) mortality rate; none of the Staphylococcus aureus bacteremias were fatal. Enterobacter species bacteremia occurring after 14 days still had a mortality rate of 55% (5/9); late-onset P. aeruginosa bacteremia had a 30% (3/10) mortality rate. Thirty-four percent (10/29) of the deaths that occurred in bacteremic or fungemic transplant recipients were associated with episodes occurring within 5 days of operation (Table 2). The overwhelming majority of these deaths (8/10) occurred in patients with Enterobacter, P. aeruginose, or both. In our institution the current prophylactic regimen for liver transplantation should be reevaluated; increased emphasis on the prevention of P. aeruginosa and Enterobacter species is warranted. The emergence of these two organisms during the immediate postoperative period appeared to be a significant cause of morbidity and death among transplant recipients. In summary, the demographics, clinical characteristics, microbiology, and prognoses of bacteremia are notably different for the three major groups of organ recipients. The infection control practitioner must be aware that the transplant recipient represents a susceptible host for many opportunistic infections. In addition to opportunistic infections most commonly associated with immunosuppression (cytomegalovirus, Listeria, Legionella, etc.), there are distinct clinical patterns in transplant recipients with bacteremia. Knowledge of these patterns may allow the prac-
Volume 20 Number 5
Posttransplant bacteremia
titioner to tailor specific interventional approaches to the therapy and prevention of these bacteremias. We thank Sara Vaccarello for her assistance with data collection, Joseph Chow for his critical review of the manuscript, and Shirley Brinker for manuscript preparation.
References 1. Kusne S, Dummer JS, Singh N, et al_ Infections after liver transplantation: an analysis of 101 consecutive cases. Medicine 1988;67:132-43. 2. Hofflin JM, Potasman I, Baldwin JC, et al. Infectious complications in heart transplant recipients receiving cyclosporine and corticosteroids. Ann Intern Med 1987; 106:209-16. 3. Peterson PK, Anderson RC. Infection in renal transplant recipients: current approaches to diagnosis, therapy, and prevention. Am J Med 1986;81:2-10. 4. Paya CV, Hermans PE, Washington JA, et al. Incidence, distribution, and outcome of episodes of infection in 100 orthotopic liver transplantations. Mayo Clin Proc 1989;64: 555-64. 5. Korvick JA, Marsh JW, Starzl TE, Yu VL. Pseudomonas aeruginosa bacteremia in patients undergoing liver transplantation: an emerging problem. Surgery 1991; 109: 62-8.
247
6. Hilf M, Yu VL, Sharp J, et al. Antibiotic therapy for Pseudomonas aeruginosa bacteremia: outcome correlations in a prospective study of 200 patients. Am J Med 1989;87:540-6. 7. Schroter GP, Hoelscher M, Putnam CW, et al. Infections complicating orthotopic liver transplantation: a study emphasizing graft-related septicemia. Arch Surg 1976; 111:1337-47. 8. Colonna JO, Winston D J, Brill JE, et al. Infectious complications in liver transplantation_ Arch Surg 1988; 123:360-4. 9. Myerowitz RL, Medeiros AA, O'Brien TF. Bacterial infection in renal homotransplant recipients: a study of fiftythree bacteremic episodes. Am J Med 1972;53:306-14. 10. Anderson RJ, Schafer LA, Olin DB, Eickhoff TC. Septicemia in renal transplant recipients. Arch Surg 1973;106: 692-4. 11. Rubin RH, Wolfson JS, Cosimi AB, Tolkoff-Rubin NE. Infection in the renal transplant recipient. Am J Med 1981;70:405-11. 12. Panjwani DD, Kalowi M, Kumar MS, et al. Septicemia in renal :transplant recipients: epidemiology and prognosis. Transplant Proc 1989;21~2112-3. 13. Chow JW, Fine MJ, Shlaes DM, et al. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med 1991; 115:58590.
) Bound volumes available to subscribers Bound volumes of AJIC: AMERICANJOURNALOF INFECTION CONTROLare available tO subscribers (only) for the 1992 issues from the Publisher, at a cost of $24.00 for domestic, $32.68 for Canadian, and $31.00 for international subscribers for Vol. 20 (FebruaryDecember). Shipping charges are included. Each bound volume contains a subject and an author index. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Drive, St. Louis, Missouri 63146-3318, USA; phone 800-325-4177, ext. 4351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular journal subscription.
Background: Bacteremic infections are a major cause of death among organ transplant
recipients. We sought to identify the risk factors associated with death and examine the timing of the bacteremic episode after operation to recognize patients who may benefit from perioperafive prophylactic antibiotic therapy. Methods: A total of 125 episodes of bacteremia or fungemia in 16 heart, 26 kidney, and 70 liver recipients were monitored prospectively in 1 year. Results: The urinary tract was the most frequent portal for kidney recipients, the gastrointestinal and biliary tracts were frequent for liver recipients, and the lung was frequent in heart recipients. Heart and liver recipients were more severely ill at the time of bacteremia and had bacteremia sooner after operation. Death at 14 days after onset of bacteremia was 33% in heart recipients, 24% in liver recipients, and 11% in kidney recipients. Risk of death was associated with the severity of the underlying condition of the transplant recipient, the source of the bacteremia, and the microbial agent. Pseudomonas aeruginosa and Enterobacter species had fatality rates of 47% and 63%, respectively. P. aeruginosa and Enterobacter were also most commonly associated with failures of perioperative antibiotic prophylaxis. Conclusions: There are distinct clinical patterns of bacteremia in transplant recipients. The emergence of P. aeruginosa and Enterobacter species in the immediate postoperative period appeared to be a significant cause of morbidity and death among transplant recipients. (AJIC AM J INFECTCONTROL1992;20:239-47)
I n f e c t i o n r e m a i n s a m a j o r c a u s e of m o r b i d i t y a n d d e a t h a m o n g o r g a n t r a n s p l a n t r e c i p i e n t s . 1-a Bacteremia has been associated with an especially h i g h m o r t a l i t y r a t e in this p a t i e n t p o p u l a t i o n . 4 W e p r o s p e c t i v e l y t r a c k e d all b a c t e r e m i a s in o u r p o p u l a t i o n of t r a n s p l a n t r e c i p i e n t s to define this p a t i e n t p o p u l a t i o n a n d t h e i r r i s k f a c t o r s f o r death. We f u r t h e r a n a l y z e d the m i c r o b i a l o r i g i n s o f the b a c t e r e m i c e p i s o d e s w i t h r e s p e c t to the t i m e of o n s e t of b a c t e r e m i a a n d the p o s s i b i l i t y t h a t the bacteremia might have been prevented by postoperative prophylactic antibiotic therapy) From the Division of Infectious Disease, University of Pittsburgh School of Medicine, Pittsburgh, Pa. Reprint requests: Victor L. Yu, MD, University of Pittsburgh School of Medicine, Scaife 968, Pittsburgh, PA 15261.
17/46/36079
METHODS Study population All t r a n s p l a n t r e c i p i e n t s w h o w e r e h o s p i t a l i z e d with blood cultures positive for pathogens during a 1 - y e a r p e r i o d (19 8 7) w e r e p r o s p e c t i v e l y t r a c k e d until d i s c h a r g e o r death. The s i g n i f i c a n c e o f a p o s i t i v e b l o o d c u l t u r e w a s b a s e d o n the clinical j u d g m e n t of the i n f e c t i o u s d i s e a s e a n d s u r g i c a l t r a n s p l a n t services. All p a tients w h o r e c e i v e d specific a n t i b i o t i c t h e r a p y f o r the b a c t e r e m i a w e r e i n c l u d e d in the study. Additional cultures positive for the same organism w i t h i n a 3 - d a y p e r i o d w e r e c o n s i d e r e d to r e p r e sent the same episode.
Study definitions and procedures The s t a n d a r d p e r i o p e r a t i v e p r o p h y l a c t i c a n t i b i otic r e g i m e n s w e r e as follows: k i d n e y r e c i p i e n t s , 239
A,.c
Wagener and Yu
240 Table
October 1992
1. Demographics of bacteremic transplant recipients Parameter
Heart recipients
Incidence during study 11% (11/98) Age (yr) Mean 48,7 Range 16-63 Male 67% (12/18) White 100% (18/18) WBC (1000 cells/ram 3) Mean 15.0 Range 0.5-38 <3 11% (2/18) Hernatocrit Mean 34.1 Median 33.7 Range 24-49 Creatinine level (rng/dl) Median 1.7 Mean 3.01 Range 0.5-15 Liver function tests--median (range) Bilirubin (mg/di) 2.1 (0.6-36) SGOT (IU/L) 38.8 (11-6465) SGPT (IU/L) 19.5 (5-3894) Receiving CyA 100% (18/18) CyA level (l~g/ml) Median 460 Range 99-1072 Receiving steroids 89% (16/18) Receiving azathioprine 50% (9/18) Receiving OKT3 6% (1/18) In ICU 44% (8/18) ReslSirator 33% (6/18)
Kidney recipients
Liver recipients
6% (11/90)
24% (69/284)
42.2 21-67 44% (12/27) 63% (17/27)
45 18-67 55% (44/80) 88% (70/80)
12.8 2.6-31 4% (1/27)
12.7 0.1-33 11% (9/80)
/.
p Value*
"
NS NS .01 NS NS
29.6 26.9 19-47
27.8 27.9 17-40
3.4 3,89 1.2-13.6
2.3 3.0 0.4-41
0.02 NS
1.0 (0.4-39) 31 (7-521) 16.0 (5-217) 81% (22/27)
5.3 (0.6-44) 258 (8-5747) 45 (2-4110) 97% (79/80)
<0.01" NS <0.01" 0.01
583 43-1447 96% (26/27) 18% (5/27) 4% (1/27) 15% (4/27) 11% (3/27)
762 115-2275 95% (76/80) 5% (4/80) 32% (26/80) 50% (40/80) 30% (24/80)
0.011-
0.06* NS <0.01 <0.01 <0.01 NS Cont'd
NS, Not Significant (p > 0.05); WBC,whiteblood cells; SGOT,aspartateaminotransferase;GTP,guanosinetriphosphate; SGPT,alanineaminotransferase;CyA, cyclosp0rin A; ICU, intensivecare unit; IV, intravenous. *By X~ test except as indicated. iKruskal-Wallis test.
cefazolin; heart recipients, cefamandole; and liver -recipients, cefotaxime plus ampicillin. The first dose w a s given before operation and therapy was continued for 2 to 5 days at the discretion of the surgeon. A bacteremic episode w a s considered nosocomial if the onset was longer than 48 hours after hospitalization or within 1 w e e k of discharge. All other episodes were considered community acquired. A bacteremic episode was considered to have an early onset if the positive blood culture o c c u r r e d within 14 days after the organ transplantation operation. Failure of antibiotic prophylaxis was defined as a blood culture positive for pathogens within 5 days after organ transplantation operation or an additional operation for which the same prophylactic regimen was used. Mortality rates w e r e defined by deaths within 14 days of positive blood culture.
A severity of illness index at the time of bacteremia was b a s e d on the assignment of points as follows: fever ( > 3 8 ° C = 1 point, _>40 ° C = 2 points), mental status (disorientation = 1 point, stupor -- 2 points, c o m a = 4 points), hypotension (2 points), respirator support (2 points), and cardiac arrest (4 points). Patients were considered to be seriously ill if the point total was greater than or equal to 4. This grading system has been shown in other studies of b a c t e r e m i a to be highly predictive of outcome. 5' 6 Statistical
analysis
Statistical analysis w a s done with the P R O P H E T system (National Institutes of Health, Division of R e s e a r c h Resources). Univariate analysis of contingency data was done with ×2 or Fisher exact tests. The t test was used for compar-
Volume 20 Number 5
Table
Posttransplant bacteremia 2 4 1
1. Continued
Parameter . . . . .
Heart recipients
Kidney recil~ients
Liver recipients
p Vaiue*
Shock Mental status decrease Fever (> 37° C) Retransplant Severity of illness z 4 Polymicrobial Diabetes Hemodialysis Community-acquired Portal Urinary tract Pneumonia Wound/skin Abdominal/biliary IV site Unknown Other Timing (days after transplant) Mean Median Range < 14 days < 30 days < 60 days > 1 year Organism also isolated Sputum Lung Urine Wound Vascular catheter 2-week mortality rate
17% (3/18) 50% (9/18) 89% (16/18) 11% (2/18) 39% (7/18) 6% (1/18) 11% (2/18) 1% (1/18) 22% (5/18)
15% (4/27) 14% (4/27) 93% (25/27) 30% (8/27) 11% (3/27) 7% (2/27) 33% (9/27) 22% (6/27) 44% (12/27)
10% (8/80) 45% (36/80) 92% (74/80) 29% (23/80) 29% (23/80) 11% (9/80) 6% (5/80) 11% (9/80) 6% (5/80)
NS 0.02 NS NS 0.02 NS 0.03 NS < 0.01
11% (2/18) 22% (4/18) 17% (3/18) 0% (0/18) 6% (1/! 8) 39% (7/18) 6% (1/18)
41% (11/27) 4% (1/27) 26% (7/27) 0% (0/27) 7% (2/27) 22% (6/27) 0% (0/27)
7% (6/80) 10% (8/80) 6% (5/80) 27% (22/80) 6% (5/80) 41% (33/80) 1% (1/80)
< 0.01 0.14 0.03 < 0.01 NS NS NS
295 60 2-1074 22% (4/18) 44% (8/18) 50% (9/18) 39% (7/18)
573 392 17-2856 0% (0/27) 11% (3/27) 30% (8/27) 56% (15/2i)
88 26 1-630 35% (28/80) 59% (47/80) 75% (60/80) 9% (8/80)
44% (8/18) 11% (2/18) 33% (6/18) 18% (3/18) 1! % (2/18) 33% (6/18)
4% (1/27) 0% (0/27) 41% (11/27) 18% (5/27) 7% (2/27) 11% (3/27)
41% (33/80) 7% (6/80) 14% (11/80) 15% (12/80) 10% (8/80) 24% (19/80)
ing the means of continuous variables that approximate a normal distribution, as determined by the Wilk-Shapiro test for normality (age, hematocrit, serum creatinine level, and. white blood cell count). The Mann-Whitney test was used for comparing all other continuous variables (liver function tests, duration of hospitalization, and cyclosporin levels). The Kruskal-Wallis test (an extension of the nonparametric Mann-Whitney test) was used to compare continuous variables between the three transplant groups (Table 1). RESULTS
• One h u n d r e d sixty-one episodes of bacteremia were found in 146 transplant recipients within the 1-year study period. Thirteen patients ;had multiple episodes of bacteremia, with two transplant recipients having three episodes each. Thirty-three of the episodes were considered clinically insignificant, including infections with coagulase-negative staphylococci (20 patients), Bacillus species (four patients), Corynebacterium
< 0.01 * < 0.01 < 0.01 < 0.01 < 0.01 <0.01 NS <0.01 NS NS NS
(diphtheroid) species (three patients), Propionibacterium species (two patients), Flavobacterium species (one patient), Lactobacillus species (one patient), Clostridium species (one patient), and an unidentified gram-positive coccus (one patient). Antibiotic therapy was not administered for these episodes and these 33 patients were therefore excluded from further analysis. Three patients with multiorgan transplants (one heart-lung, two kidney-liver),, who each had one episode of bacteremia, were also excluded. The 125 remaining episodes occurred in 111 patients. Eighteen bacteremias occurred in 16 heart recipients, 27 occurred in 26 kidney recipients, and 80 occurred in 69 liver recipients. Thirteen patients had more than one episode of bacteremla. These episodes were separated by 11 to 335 days (mean, 62.4 days, median, 32 days). Of the episodes, 62% involved different organisms and 38% occurred during distinct hospital admissions. The five cases in which the same organism recurred were treated as independent events
AJIC October 1992
Wagener and Yu
242
0.9 0.8 .~ 0.7~ .=- 0.6 ..Q
'° t
.,......-. ..:,:.!"
:"
:
! 0
heart
I
0.5- : ' ~ E a. 0.4- •" 0.30.2- ~1 ,~I" 0.1- I Hf 0
liver
,..- ....
r kidney r ........
I
I
J" .............
I
100 200 300 Days post transplantation
a
I
380
Fig, 1. Kaplan-Meier curves comparing the number of days between transplantation and the onset of bacteremia or fungemia. Almost 80% of bacteremias in liver recipients occurred within the first 90 days. The x axis represents the number of days after transplantation. The y axis represents the cumulative probability, based on the total number of patients that had bacteremia during the study, of a transplant recipient having bacteremia.
because more than 28 days had passed between the two positive blood cultures and clinical evidence of relaps e was not obvious. The incidence of bacteremia during the study period was 11% (11/98) for heart recipients, 6% (11/190) for kidney recipients, and 24% (69/284) for liver recipients. The remaining bacteremias occurred in patients who underwent transplantation before the study year. T y p e of o r g a n t r a n s p l a n t
Differences between major transplant sites are summarized in Table 1. The ages and sex distributions of the heart, liver, and kidney recipients who had bacteremia were comparable. Kidney recipients with bacteremia were in general less severely ill than were liver and heart transplant recipients (Table 1); only 15 % of kidney recipients were in the intensive care unit, received mechanical ventilation, or had decreased mental status or vital sign abnormality; this compares with 44% and 50% heart and liver recipients, respectively. Kidney recipients were also less likely to have the bacteremic organism isolated from their sputum and more likely to have it in the urine. Pneumonia was most frequent in heart transplant recipients and infections originating in abdominal/biliary tract were seen exclusively in liver recipients. Forty-four percent of the bacteremias in kidney recipients were considered community acquired,
compared with 22% among the heart recipients and 6% among the liver recipients? Immunosuppressive chemotherapy regimens differed among the transplant sites, with the kidney recipients less likely to receive cyclosporin A. Liver recipients had higher m e a n cyclosporin levels and were more likely to have been given OKT3 to combat rejection. Bacteremias occurred later in the kidney recipient, with about 50% occurring more than 1 year after transplantation. About 75% of the bacteremias in liver recipients occurred within the first 2 months after transplantation. The Kaplan-Meier curves depicting the number of days between transplantation and the onset of bacteremia were significantly different for the groups (p < 0.01, Gehan-Breslow test; Fig. 1). Portal
The source of the bacteremia or fungemia varied with the organ transplanted (Table 1). The most common portal identified was the abdominal/biliary site, which was involved in 18% (22/125) of the bacteremic episodes; this portal was seen exclusively in liver recipients. The urinary tract was the source in 15% of the bacteremias; 58% (11/19) of these occurred in the kidney recipient. Wound or skin infections were implicated in 12% of the bacteremias, with kidney recipients having the highest percentage of wound infections: 26% (7/27), compared with 17% (3/18) in heart recipients and 6% (5/80) in liver recipients. Pneumonia was the origin identified in 10% (13/125) of the bacteremias: 25% (4/18) of heart recipients, 10% (8/80) of liver recipients, and 4% (1/27) of kidney recipients. Six percent of the bacteremias were attributed to intravascular catheters. One case was related to a gastrointestinal infection and one case to meningitis. Thirty-eight percent of all bacteremias (46/125) were from an unidentified portal of entry. Microbial agent
Aerobic gram-negative bacilli constituted 48% of the pathogens found by blood culture in kidney recipients, 49% in the liver recipients, and 39% in the heart recipients (Fig. 2). Escherichia coli, Klebsiella, and other Enterobacteriaceae were t h e predominant gram-negative organisms seen in kidney recipients. Pseudomonas aeruginosa and Enterobacter species accounted for most of the gram-negative bacteria seen in liver recipients. P. aeruginosa and E. coli were the gram-negative bacilli seen most often in heart recipients.
Volume 20 Number 5
Posttransplant bacteremia 2 4 3 S.aureus s.epider midis Enterobacter sp P.aeruginosa Gram neg enterics E.coli Gram positive cocci Fungi Enterococcus sp Anaerobes Miscellaneous A.hydrophilia S.pneumoniae
~!i~!i~!i~i~!!i~;i~iJ!i!}~!~!i~!~!i~!;!i~:~!i~!~.1i:~!~i~i~i~i~i!~i~!z~!i!~!i!!;!~iFiJi!~i!i;!!i~i~i~i~:i~!~!:!~!~
i 0
5
I
Heart
I
I
I
I
I
10
15
20
25
30
~
Liver
~
Kidney
Fig. 2. The major Jnfectious agents seen in 125 blood cultures from organ transplant recipients. Gram-negative enteric bacteria: Citrobacter freundii (four), Klebsiella pneumoniae (four), Serratia marcescens (two), and Proteus mirabilis (two). Gram-positive cocci: Streptococcus viridans group (four), Staphylococcus intermedius grou p (three), Streptococcus pyogenes (one), and Streptococcus agalactiae (one). Fungi: Candida albicans (two), Candida tropicalis (two), Candida stelloidea (two), and Candida glabrata (one). Anaerobic bacteria: Bacillus fragilis (three) and Clostridiurn species (one), Miscellaneous: Corynebacterium group JK (two), Listeria (one).
Gram-positive organisms were seen in about 50% of the bacteremias among liver, heart, and kidney recipients. Staphylococcus aureus was the predominant gram-positive organism seen in both heart and liver recipients and was second only to Staphylococcus epidermidis in kidney recipients. Fungi were present in 6% (5/80) of the positive blood cultures from liver recipients and in 11% (2/18) from heart recipients. None of the kidney transplant recipients had fungemia. The fungi isolated were as follows: Candida albicans (two), Candida tropicalis (two), Candida stellatoidea (two), and Candida glabrata (one). Anaerobic bacteria were present in 3% (5/125) of the posttransplant bacteremias-Bacteroides fragilis (three), Clostridium species (two). Ten percent (12/126) of the positive blood cultures were polymicrobial. Multiple organisms were isolated in 11% (9/80) of the bacteremias in liver recipients, 6% (1/18) of thos6 in heart recipients, and 7% (2/27) of those in kidney recipients. Community versus hospital acquired
Survival rate was somewhat higher for patients with• c0mmunity-acquired bacteremia (90%;
19/21) than for those with nosocomial bacteremia (75%; 78/104), although statistical significance was not attained ( p - - 0 . 1 5 , xa). Patients with community-acquired bacteremia had significantly lower severity of illness index values t h a n did those with nosocomial bacteremia (p < 0.01, X2); only 9% (2/21) of the former group were classified as severely ill. In general the baeteremic infection was the reason for admission for patients with community-acquired bacteremia. Communityacquired organisms were predominantly E. coli, S. aureus, and other Enterobacteriaceae (Fig. 3). There were no P. aeruginosa infections and only one Enterobacter among the patients with community-acquired bacteremia. Timing of b a c t e r e m l a onset Failures of antibiotic prophylaxis. Fifteen episodes of bacteremia occurred within 5 days of
operation and had a mortality rate of 67%. The predomifient organisms were Enterobacter species (40% [6/15]) and Pseudomonas aeruginosa (33% [5/15]; Table 2). The surgical procedures included eight liver transplants, two heart transplants, and five posttransplant exploration and repair procedures (four liver, one kidney). These
AJIC 244
October 1992
Wagener and Yu S.aureus S.epidermidis P.aeruginosa Enterobacter sp Gram neg enterics
~//////~///////~///~////~////////////////~/~//~//////////////~
i
Gram pos cocci Fungi Enterococcus sp Anaerobes A.hydrophilia JK diptheroids Listeria S.pneumoniae 0
I
I
I
I
5
10
15
20
~
nosocomial
community acquired
25
Fig. 3. Sixty-two percent of the organisms involved in community-acquired bacteremias were gram-negative enteric bacteria. Gram-negative enteric bacteria: Escherichia coli (11), Citrobacter freundii (four), Klebsiella pneumoniae (four), Serratia marcescens (two), and Proteus mirabilis (two). Gram-positive cocci: Streptococcus viridans group (four), Staphylococcus intermedius group (three), Streptococcus pyogenes (one), and Streptococcus agalactiae (one). Fungi: Candida albicans (two), Candida tropicalis (two), Candida stelloidea (two), and Candida glabrata (one). Anaerobic bacteria: Bacillus fragilis (three) and Clostridium species (one).
five surgical procedures occurred 7 to 60 days after transplantation. Early versus late onset. Twenty-six percent (32/125) ofl the bacteremias occurred within 14 days of transplantation and were considered to be early onset. This included bacteremias in 22% (4/18) of the heart and 35% (28/80) of the liver transplant recipients. None of the bacteremias in kidney recipients were classified as early onset. _The predominant organisms seen in these early bacteremias were Enterobacter species (31% [10/32]), P. aeruginosa (28% [9/32]), and Staphylococcus aureus (25% [8/32]). Twenty-two percent (7/32) of the early-onset bacteremias had a pulmonary source, as opposed to only 6% (6/93) of the bacteremias with later onset. Seventy-four percent (93/125) of the bacteremias occurred longer than 14 days after transplantation and were considered late onset. The predominant organisms seen after 2 weeks were Staphylococcus epidermidis (18% [17/93]), Staphylococcus aureus (17% [16/93]), E. coli (12% [11/93]), and P. aeruginosa (11% [10/93]). The late-onset bacteremias were more likely to have a urinary tract portal (19% [18/93]) than were early bacteremias (3% [1/32]).
Mortality rates
The 2-week mortality rate after bacteremia for all patients was 23%. The mortality rate was 33% (6/18) among the heart recipients, 24% (19/80) among the liver recipients, and 11% (3/27) among the kidney recipient s . Within 60 days 37% of the bacteremic patients had died, with most of the subsequent deaths occurring in the liver recipients. Associated with increased mortality rate were retransplantation (p < 0.01, X2), polymicrobial infection (p < 0.05, X2), a n d p n e u m o n i a (p < 0.01, ×2). Patients who died also had higher severity of illness index scores (p < 0.01, X2), higher serum creatinine levels (p - 0.01, t test) and higher liver function test abnormalities, even when controlled for site of transplantation (p < 0.01, Mann-Whitney). Death occurred in 47% (15/32) of the patients with early-onset bacteremia and 15% (14/93) of the patients with late-onset bacteremia (p < 0.01, X2). Patients with gram-negative bacteremia and fungemia had a significantly higher mortality rate (35% [22/62]) than did those with gram-positive bacteremia (8% [4/53]; p < 0.01, X2). P. aerugi- " nosa and Enterobacter species were associated
Volume 20 Number 5
with most of the fatal gram-negative bacteremias (Table 3). DISCUSSION
Infection has been recognized as a major source of morbidity and, death since the early days of organ transplantation. Improvements in surgical techniques and immunosuppressive protocols have been followed by a reduction in severe bacterial infections. The incidence of bacteremia among transplant recipients, however, remains relatively high. In the 1970s liver transplantation was complicated by an incidence of bacteremia greater than 70%, 7 generally reflecting contami-nation by gastrointestinal organisms. This incidence was subsequently reduced to 25% to 35%.1, 4, a Heart recipients had a 30% incidence of bacteremia in the early 1970s; a 1987 study2 still reported a 20% rate of posttransplant bacteremia. In kidney recipients, the rate of bacteremia has decreased from 26% to 60% in the late 1960s 9"10to about 12% in the 1980s. ~ At one of the largest organ transplant centers in the world, we had the opportunity to compare the demographic and risk factors for bacteremia in three major groups of transplant recipients. We compared the microbial agent, the timing after transplantation, the portal of infection, and the subsequent mortality rate for bacteremic episodes in kidney, heart, and liver recipients. As expected, the major sources of the bacteremic organism differed among the three groups. Urinary tract infection was the primary source of bacteremia among kidney recipients, abdominal and biliary infections were the major sources among liver recipients, and pneumonia was the primary source among heart recipients. Forty-one percent of the bacteremias in kidney recipients originated in the urinary tract; an infected dialysis access site was the source in 21%. This is a smaller percentage of urinary tract infections than that previously reported (50%) for kidney transplant recipients, a' 9 Most kidney recipients in this study had bacteremia more than 1 year after transplantation, rather than during the immediate postoperative period. Return to dialysis appears to be related to an increase in the number of gram-positive skin infections and may be at least partially responsible for a lower percentage of gram-negative organisms than previously reported (50% vs 60% to 70%). a, 10-12 Staphylococcus aureus was found in about 20% of the bacteremias for each of the three groups. Enterobacter species were seen almost exclusively
Posttransplant bacteremia 2 4 5 T a b l e 2. Microbial agent and morality rate of
bacteremia for failures of prophylaxis (bacteremia occurring within 5 days of transplantation or other operation*) Bacteremia rate
Liver recipients Enterobacter species 5% (4/80) Pseudomonas aeruginosa 3% (2/80) Staphylococcus aureus 4% (3/80) Fungi 3% (2/80) Polymicrobiall1% (1/80) TOTAL 15% (12/80) Heart recipients. Pseudomonas aeruginosa 6% (1/18) Serratia marcescens 6% (1/18) TOTAL " 11% (2/18) Kidney Polymicrobialt4% (1/27) TOTAL 12% (15/125)
Mortality rate
100% (4/4) 50% (1/2) 0% (0/3) 50% (1/2) 100% (1/1) 58% (7/12) 100% (1/1) 100% (1/1) 100% (2/2) 100% (1/1) 67% (10/15)
*Other operations accounted for five episodes of bacteremia after surgical exploration/repairafter transplantation. These patients received the same antibiotic regimen as used during organ transplantation and operations occurred 7 to 60 days after transplantation. l-Polymicrobialbacteremia included two episodes (one liver recipient, one kidney recipient) in which Pseudomonas aeruginosa and Enterobacter species were simultaneouslyisolated.
in liver recipients. P. aeruginosa was seen in 20% of the bacteremias in liver recipients and in 11% • of heart recipients. Staphylococcus aureus and Staphylococcus epidermidis were the organisms most frequently seen in kidney recipients, rather than enteric gram-negative bacteria as reported in earlier studies.3' 10. 11Pseudomonas and Enterobacter accounted for about 40% of bacteremias in liver recipients and most of the deaths (65%). These two organisms were also more likely to be present in patients considered to have early onset (infection within 2 weeks of transplantation). Twelve bacteremia episodes occurring within 5 postoperative days were considered to be failures of antibiotic prophylaxis. Of these bacteremias, 58% (7/12) were caused by P. aeruginosa, Enterobacter species, or both (Table 2). The current prophylaxis regimen of cefotaxime/ampicillin used in the liver transplant recipients appeare d to be associated with superinfection with P. aeruginosa and Enterobacter species. P. aeruginosa is resistant to cefotaxime and ampicillin, the regimen used at the time of liver transplantation. In other studies, previous use of third-generation cephalosporins has been linked to the emergence of Enterobacter resistant to those agents.13 The overall mortality rate after bacteremic episodes was 23%, deaths occurred primarily
246
.&JI@ October 1992
Wagener and Yu
T a b l e 3. Mortality rate among bacteremic transplant recipients subclassified by microbial agent Single organism
Polymicroblal*
Total
Enterobacter species Pseudomonas aeruginosa Fungi Other gram-positive cocci
60% 38% 45% 40%
(9/15) (5/13) (4/7) (2/5)
75% (3/4) 67% (4/6) 0% (0/5)
63% 47% 43% 20%
(12/19) (9/19) (4/7) (2/10)
Enterococcus species Staphylococcus epidermidis Staphylococcus aureus Gram-negative enteric bacteria Anaerobes Miscellaneous Other gram-negative bacilli
0% (0/5) 7% (1/15) 4% (1/23) 5% (1/22) 0% (0/4) 0% (0/2) 50% (1/2)
50% (1/2) 20% (1/5) 0% (0/1) 0% (0/1) 0% (0/1) -
14% (1/7) 10% (2/20) 4% (1/24) 4% (1/23) 0% (0/4) 0% (0/3) 50% (1/2)
Fungi: Candida albicans (two), Candida tropicalis (two), Candida stellatiodea (two), and Candida glabrata (one). Other gram-positive cocci: Streptococcus viridans group (four), Streptococcus intermedius group (three), Streptococcus pyogenes (one), Streptococcus agalactiae (one) and Streptococcus pneumoniae (one). Gram-negative enteric bacteria: Escherichia coil (11),. Citrobacter freundii (four), Klebsiella pneumonia (four), Serratia marcescens (two), and Proteus rnirabilis (two). Other gram-negative bacilli: Aeremonas hydrophila (two). Miscellaneous: Corynebacterium group JK (two) and Listeria (one). Anaerobes: Bacteroides fragilis (three) and Clostridium species (one). *There were 12 polymicrobial bacteremias with 25 organisms. Pseudomonas aeruginosa and/or Enterobacter species were involved in all fatal cases (42% [5/12]) of polymicrobial bacteremias.
among the heart and liver transplant recipients. Mortality rate was directly related to the underlying condition of the patient. These patients had significantly higher serum creatinine levels, more abnormal results of liver function tests (even when controlled for site of transplant), and higher severity of illness scores. Patients who were bacteremic within 14 days of transplantation had a 47% mortality rate. Retransplantation was also significantly associated with a higher mortality rate; 39% (13/53) of bacteremic patients with a second transplanted organ died. Bacteremic pneumonia carried a mortality rate of 60%, compared with 5% among those with a urinary tract source and 20% among patients with an abdominal or biliary portal. Aerobic gram-negative bacilli and yeasts Were more likely to be associated with death. Transplant recipients with P. aeruginosa bacteremia had twice the mortality rate and those with Enterobacter bacteremia had four times the mortality rate of patients with other agents of bacteremia. Patients with P. aeruginosa or Enterobacter were significantly more likely to have a high severity of illness index score and to have recently received antimicrobial agents. Polymicrobial bacteremias had a 42% (5/12) mortality rate. All five fatal cases involved P. aeruginosa, Enterobacter species, or both. Mortality rate was significantly higher for earlyonset bacteremia (47% [15/32]) than for bacteremia occurring later (14% [13/93]). Enterobacter species bacteremia occurring within the first 2 weeks after transplantation had a 70% (7/10)
mortality rate, P. aeruginosa bacteremia had a 67% (6/9) mortality rate, and fungemia had a 67% (2/3) mortality rate; none of the Staphylococcus aureus bacteremias were fatal. Enterobacter species bacteremia occurring after 14 days still had a mortality rate of 55% (5/9); late-onset P. aeruginosa bacteremia had a 30% (3/10) mortality rate. Thirty-four percent (10/29) of the deaths that occurred in bacteremic or fungemic transplant recipients were associated with episodes occurring within 5 days of operation (Table 2). The overwhelming majority of these deaths (8/10) occurred in patients with Enterobacter, P. aeruginose, or both. In our institution the current prophylactic regimen for liver transplantation should be reevaluated; increased emphasis on the prevention of P. aeruginosa and Enterobacter species is warranted. The emergence of these two organisms during the immediate postoperative period appeared to be a significant cause of morbidity and death among transplant recipients. In summary, the demographics, clinical characteristics, microbiology, and prognoses of bacteremia are notably different for the three major groups of organ recipients. The infection control practitioner must be aware that the transplant recipient represents a susceptible host for many opportunistic infections. In addition to opportunistic infections most commonly associated with immunosuppression (cytomegalovirus, Listeria, Legionella, etc.), there are distinct clinical patterns in transplant recipients with bacteremia. Knowledge of these patterns may allow the prac-
Volume 20 Number 5
Posttransplant bacteremia
titioner to tailor specific interventional approaches to the therapy and prevention of these bacteremias. We thank Sara Vaccarello for her assistance with data collection, Joseph Chow for his critical review of the manuscript, and Shirley Brinker for manuscript preparation.
References 1. Kusne S, Dummer JS, Singh N, et al_ Infections after liver transplantation: an analysis of 101 consecutive cases. Medicine 1988;67:132-43. 2. Hofflin JM, Potasman I, Baldwin JC, et al. Infectious complications in heart transplant recipients receiving cyclosporine and corticosteroids. Ann Intern Med 1987; 106:209-16. 3. Peterson PK, Anderson RC. Infection in renal transplant recipients: current approaches to diagnosis, therapy, and prevention. Am J Med 1986;81:2-10. 4. Paya CV, Hermans PE, Washington JA, et al. Incidence, distribution, and outcome of episodes of infection in 100 orthotopic liver transplantations. Mayo Clin Proc 1989;64: 555-64. 5. Korvick JA, Marsh JW, Starzl TE, Yu VL. Pseudomonas aeruginosa bacteremia in patients undergoing liver transplantation: an emerging problem. Surgery 1991; 109: 62-8.
247
6. Hilf M, Yu VL, Sharp J, et al. Antibiotic therapy for Pseudomonas aeruginosa bacteremia: outcome correlations in a prospective study of 200 patients. Am J Med 1989;87:540-6. 7. Schroter GP, Hoelscher M, Putnam CW, et al. Infections complicating orthotopic liver transplantation: a study emphasizing graft-related septicemia. Arch Surg 1976; 111:1337-47. 8. Colonna JO, Winston D J, Brill JE, et al. Infectious complications in liver transplantation_ Arch Surg 1988; 123:360-4. 9. Myerowitz RL, Medeiros AA, O'Brien TF. Bacterial infection in renal homotransplant recipients: a study of fiftythree bacteremic episodes. Am J Med 1972;53:306-14. 10. Anderson RJ, Schafer LA, Olin DB, Eickhoff TC. Septicemia in renal transplant recipients. Arch Surg 1973;106: 692-4. 11. Rubin RH, Wolfson JS, Cosimi AB, Tolkoff-Rubin NE. Infection in the renal transplant recipient. Am J Med 1981;70:405-11. 12. Panjwani DD, Kalowi M, Kumar MS, et al. Septicemia in renal :transplant recipients: epidemiology and prognosis. Transplant Proc 1989;21~2112-3. 13. Chow JW, Fine MJ, Shlaes DM, et al. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med 1991; 115:58590.
) Bound volumes available to subscribers Bound volumes of AJIC: AMERICANJOURNALOF INFECTION CONTROLare available tO subscribers (only) for the 1992 issues from the Publisher, at a cost of $24.00 for domestic, $32.68 for Canadian, and $31.00 for international subscribers for Vol. 20 (FebruaryDecember). Shipping charges are included. Each bound volume contains a subject and an author index. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Drive, St. Louis, Missouri 63146-3318, USA; phone 800-325-4177, ext. 4351, or 314-453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular journal subscription.