Eighty-six new CAPD-associated staphylococcal peritonitis episodes occurred in 63 patients between 1996 and 2000. Forty-three of the patients were females, 39 were Caucasians, and 28 had diabetes. The distribution of patients according to age was as follows: birth to 20 years (n = 4), 21 to 40 years (n = 12), 41 to 59 years (n = 22), and 60 years or older (n = 22). Treatment time on CAPD was less than one year in 28 patients and longer than one year in 35. Forty-four patients used a double bag system and 19 used a standard bag exchange system.
Microbiological investigation
Thirty-five episodes were caused by
S. aureus and 51 by CoNS. Among CoNS,
S. epidermidis was the most frequent species (24 cases), followed by
S. haemolyticus (11 cases) and other species (16 cases) (Table
1).
Table 1
Causative agent of the 86 new episodes of peritonitis caused by S. aureus and coagulase-negative staphylococci.
S. aureus
| 35 |
CoNS | 51 (100) |
S. epidermidis
| 24 (47) |
S. haemolyticus
| 11 (21.5) |
S. warneri
| 5 (9.8) |
S. hominis
| 5 (9.8) |
S. xylosus
| 2 (3.9) |
S. cohnii
| 2 (3.9) |
S. simulans
| 1 (1.9) |
S. lugdunensis
| 1 (1.9) |
Oxacillin susceptibility was observed in 30 (85.7%) of the 35 episodes due to S. aureus, in 10 (41.7%) of the 24 episodes due to S. epidermidis, and in 12 (44.4%) of the 27 episodes due to other CoNS (p = 0.0002). Five cases of intermediate susceptibility were detected among CoNS and one among S. aureus strains.
Positive slime production was observed in three cases due to S. aureus (8.6%) and in nine (17.6%) due to CoNS (p = 0.345), including S. epidermidis in four, S. haemolyticus in two, S. warneri in two, and S. lugdunensis in one.
With respect to toxigenic profile, 25 (71.4%) of the 35
S. aureus strains were toxin producers, whereas only seven (13.7%) of the 51 CoNS strains produced some type of toxin (
p < 0.00001). The rate of enzyme production was higher in
S. aureus strains than in
S. epidermidis or other CoNS, except for α-hemolysin whose production was similar in all strains. The rates of toxin and enzyme production by
S. aureus,
S. epidermidis and other CoNS species are shown in Table
2.
Table 2
Rates of toxin and enzyme production by S. aureus, S. epidermidis and other coagulase-negative staphylococci strains isolated from 86 new episodes of peritonitis.
Toxins | | | | |
SEA | 4 (11.4) | 1 (4.1) | - (-) | 0.15 |
SEB | 12 (34.3) | - (-) | 2 (7.4) | 0.0006 |
SEC | 8 (22.8) | - (-) | 6 (22.2) | 0.04* |
SED | - (-) | - (-) | - (-) | - |
TSST-1 | 12 (34.3) | 1 (4.1) | - (-) | 0.0008 |
Enzymes
| | | | |
α-Hemolysin | 17 (48.6) | 8 (33.3) | 8 (29.8) | 0.26 |
β-Hemolysin | 29 (82.3) | 6 (25) | 7 (25.9) | < 0.00001 |
Lipase | 34 (97.1) | 4 (16.7) | 5 (18.5) | < 0.00001 |
Lecithinase | 34 (97.1) | 2 (8.3) | 6 (22.2) | < 0.00001 |
DNAse | 34 (97.1) | - (-) | 4 (14.8) | < 0.00001 |
TNAse | 34 (97.1) | - (-) | 4 (14.8) | < 0.00001 |
Clinical outcome
Overall, 57 (66.3%) episodes were resolved, 14 (16.3%) relapsed, 12 (13.9%) required removal of the catheter, and three (3.5%) progressed to death. Among the 35 S. aureus cases, 17 (48.6%) were resolved, eight (22.8%) relapsed, seven (20%) required catheter removal, and three (8.6%) progressed to death. Regarding CoNS episodes, 40 (78.4%) were resolved, six (11.8%) relapsed, and five (9.8%) required catheter removal. Among these 11 non-resolved episodes, seven were due to S. epidermidis, two to S. haemolyticus, and two to other CoNS. There were significantly more CoNS cases resolved than S. aureus episodes (p < 0.001).
There were 52 episodes involving oxacillin-susceptible strains, 32 (61.5%) of them were resolved, 10 (19.2%) relapsed, and 10 (19.2%) required catheter removal. Thirty-four infections were caused by oxacillin-resistant strains, 20 (58.8%) of them were resolved, nine (26.5%) relapsed, and five (14.7%) required catheter removal. The resolution rate was similar for oxacillin-susceptible and -resistant strains (p = 0.9713).
Vancomycin was used in 56 episodes, 18 caused by S. aureus and 38 by CoNS. This antibiotic was prescribed because of bacterial resistance in 28 cases, lack of improvement in 20, and other undefined causes in eight. The time between diagnosis and the first vancomycin dose was 3.7 ± 1.7 days for S. aureus episodes and 3.9 ± 1.2 days for CoNS episodes (p = 0.793).
Two regression models were constructed. In Model 1 slime production was the only pathogenic factor, whereas in Model 2 the production of toxins and enzymes was included. Only lecithinase, α-hemolysin and TNAse were included since associations were observed between lecithinase and lipase, α-hemolysin and β-hemolysin, and TNAse and DNAse (p = 0.001). All toxins were included, except for toxin D which was not produced by any of the strains. Since an interaction effect was observed between S. aureus species and oxacillin susceptibility (p = 0.01), the influence of oxacillin susceptibility was evaluated at each vancomycin treatment co-variable level in both models.
In Model 1 (Table
3), controlling for co-variables, the odds of resolution was not influenced by host factors such as age, gender, diabetes, exchange system, or CAPD treatment time. The odds of resolution were 9.5 times higher for
S. epidermidis than for
S. aureus episodes (
p = 0.0263), whereas similar resolution odds were observed for
S. epidermidis and the other CoNS (
p = 0.085). Among strains isolated from infections treated with vancomycin, no significant difference was observed between oxacillin-susceptible and -resistant strains (p = 0.89). In contrast, among strains isolated from infections not treated with vancomycin, there was a significant difference between strains susceptible and resistant to oxacillin (p = 0.0113). In this case, the chance of cure of infections caused by oxacillin-susceptible strains was 137 times higher than that of infections caused by resistant strains. With respect to slime production, the chance of cure of infections caused by non-producers was estimated to be up to 68 times higher than that of infections caused by slime producers (p = 0.0015).
Table 3
Odds comparison of peritonitis resolution by logistic regression analysis (Model 1).
Age (birth to 20 years/>60 years) | -0.0510 | 0.96 | |
Age (21 to 40 years/>60 years) | 1.7097 | 0.09 | |
Age (41 to 59 years/>60 years) | 0.7511 | 0.84 | |
Gender (male/female) | 1.2523 | 0.19 | |
Race (Caucasian/non-Caucasian) | -0.0818 | 0.92 | |
Diabetes mellitus (no/yes) | -0.6709 | 0.12 | |
System (standard/double bag) | 1.2585 | 0.22 | |
Treatment time on CAPD | | | |
<1 year | 0.2074 | 0.89 | |
>1 year | 0.9547 | 0.21 | |
Treatment with vancomycin | | | |
Oxacillin susceptible/oxacillin resistant | -0.1070 | 0.89 | |
No treatment with vancomycin | | | |
Oxacillin susceptible/oxacillin resistant | 1.9440 | 0.01 | 137 (3.0; 6,202.4) |
Etiological agent | | | |
S. epidermidis/S. aureus
| 2.2548 | 0.02 | 9.5 (1.3; 69.6) |
Other CoNS/S. epidermidis
| -1.6113 | 0.08 | |
Slime production (no/yes) | 4.2139 | 0.001 | 68 (5.0; 914.8) |
As observed for the first model, in Model 2 (Table
4) resolution odds were not influenced by host factors. Among strains isolated from infections treated with vancomycin, no significant difference was observed between those susceptible and resistant to oxacillin (p = 0.1523), whereas there was a significant difference among strains not treated with vancomycin (p = 0.0039). With respect to slime production, the chance of cure of infections caused by non-producers was estimated to be 184 times higher than that of infections caused by producers (p = 0.0012). In contrast to Model 1,
S. aureus did not differ from
S. epidermidis in terms of the probability of peritonitis resolution (p = 0.7014), whereas the chance of cure of infections caused by other CoNS species was estimated to be 46 times higher than that of infections caused by
S. epidermidis (p = 0.0175). Alpha-hemolysin production was an independent predictor of resolution odds, with episodes caused by non-producers presenting an 8.2 times higher chance of resolution than those caused by producers (p = 0.0423). No significant effects on the probability of peritonitis resolution were observed for the remaining enzymes and toxins.
Table 4
Odds comparison of peritonitis resolution by logistic regression analysis (Model 2).
Age (birth to 20 years/>60 years) | 0.0602 | 0.97 | |
Age (21 to 40 years/>60 years) | 4.2565 | 0.06 | |
Age (41 to 59 years/>60 years) | 1.9474 | 0.09 | |
Gender (male/female) | 1.2461 | 0.37 | |
Race (Caucasian/non-Caucasian) | -1.3322 | 0.21 | |
Diabetes mellitus (no/yes) | -0.6711 | 0.47 | |
System (standard/double bags) | 2.3397 | 0.12 | |
Treatment time on CAPD | | | |
<1 year | 0.0531 | 0.98 | |
>1 year | -0.3045 | 0.77 | |
Treatment with vancomycin | | | |
Oxacillin susceptible/oxacillin resistant | 1.6316 | 0.15 | |
No treatment with vancomycin | | | |
Oxacillin susceptible/oxacillin resistant | 10.0189 | 0.004 | 23,906 (25.4; exp{16.9274}) |
Etiological agent | | | |
S. epidermidis S. aureus
| -1.6481 | 0.70 | |
Other CoNS/S. epidermidis
| 3.8238 | 0.017 | 46 (2.0; 1,069.4) |
Slime production (no/yes) | 5.2149 | 0.001 | 184 (7.8; 4,354.2) |
Enzyme production (no/yes) | | | |
α-Hemolysin | 2.1092 | 0.04 | 8.2 (1.1; 63.1) |
TNAse | 2.0500 | 0.28 | |
Lecithinase | 3.5545 | 0.06 | |
Toxin production (no/yes) | | | |
Enterotoxin A | 2.3914 | 0.26 | |
Enterotoxin B | 0.5669 | 0.13 | |
Enterotoxin C | 4.0501 | 0.06 | |
TSST-1 | 1.3694 | 0.48 | |