Journal of Antimicrobial Chemotherapy (1985) 15, 751-758

Vancomycin and netilmicin asfirstline treatment of peritonitis in CAPD patients

Lars Braunerf, Gunnar Kahlmeter*, Tore Lindhohn and Ole Simonsen

Departments of Nephrology and * Medical Microbiology, The University Hospital, Lund, Sweden

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

The first line treatment of peritonitis in patients on continuous ambulatory peritoneal
dialysis (CAPD) in our hospital was recently altered from a combination of genta-
micin and clindamycin, given as continuous peritoneal lavage, to one of vancomycin
and netilmicin given in peritonea] dialysis fluid with prolonged dwell time (4-6 h).
The change was prompted by the emergence of multiply resistant Staphylococcus
epidermidis among CAPD patients and nursing staff. In 9 of 19 episodes of peritonitis
treated with gentamicin/clindamycin, the infecting organism could still be isolated
from peritoneal fluid 5-15 days after commencement of therapy. All of 35 culture
verified episodes treated with vancomycin/netilmicin were cleared bactenologically
within 3 days (/"< 0-0005). The vancomycin and netilmicin serum levels achieved were
6-5-37-0 mg/1 and 1-0-8-1 mg/1, respectively. Apart from an asthmatic reaction, poss-
ibly triggered by vancomycin, no side effects were seen. However, audiometry was not
performed regularly and the possible effect of netilmicin on the residual renal function
was not systematically investigated.

Introduction

Continuous ambulatory peritoneal dialysis (CAPD) was first described by Popovich et

al. (1976). Oreopoulos et al. (1978) improved on the technique and now the method has
gained world wide acceptance. Peritonitis remains the major complication. Among the
pathogens isolated Staphylococcus epidermidis accounts for approximately 40%, Staph.
aureus for 12% and Streptococcus spp. for 10% (Nolph & Sorkin, 1981; Oreopoulos,
Vas & Khanna, 1981; Oreopoulos et al., 1981; Gokal et al., 1982). Various antibiotic
regimens, often including parenteral as well as lavage administration, and often
including a cephalosporin or an aminoglycoside antibiotic have been used as first line
therapy (Nolph & Sorkin, 1981; Oreopoulos, Vas & Khanna, 1981; Gokal et al., 1982).
At our hospital CAPD has been in use since April 1980. Initially a combination of
clindamycin and gentamicin was used for treatment of presumed bacterial peritonitis.
The emergence of several strains of multiply resistant Staph. epidermidis as the cause
of peritonitis in the CAPD patients and the subsequent demonstration of intra-
departmental dissemination of these strains, prompted a change in antibiotic regimen.

fDr Brauner died on 26 November 1984

Requests for reprints to Dr G Kahlmeter, Dept of Clinical Microbiology, The Central Hospital, 35185,
VAXJO, Sweden
751
0305-7453/85/060751 + 08 $02 00/0 © 1985 The British Society for Antimicrobial Chemotherapy
752 L. Braoner et al.

Table I. Dose regimens of antibiotics used as first line treatment of

peritonitis in CAPD patients

Period 1
Loading: clindamycin 300 mg tid iv
genlamicin 1 mg/kg iv
Maintenance: clindamycin 300 mg tid iv+10 mg/1 dialysate(L)
gentamienn 4 mg/1 dialysate(L)
Period 2
Loading: vancomycin 500 mg iv
netilmicin 1 -7 mg/kg/2 1 of dialysate (4 h)
Maintenance: vancomycin 15 mg/1 dialysate (4-6 h)
netilmicin 7-5 mg/1 dialysate (4 h) (day 1)
4 0 mg/1 dialysate (6 h) (day 2 - )

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

(L), Continuous lavagc; (4 h) = 4 h dwell time; (6 h) = 6 h dwell time.

The choice of vancomycin and netilmicin seemed logical, since the isolated multiply
resistant Staph. epidermidis strains exhibited low minimum inhibitory concentration
(MIC) values for vancomycin and netilmicin. Furthermore, vancomycin and netilmicin
weuld also be effective against Staph. aureus and netilmicin against Enterobacteriaceae.
In addition both drugs had been shown to be stable in peritoneal dialysis fluid (Glew &
Pavuk, 1981).
Material and methods
Protocol and inclusion criteria. Consecutive patients in need of dialysis entered the study.
The most common selection criteria for CAPD were diabetes mellitus (22%), cardio-
vascular disease and inability to continue haemodialysis. CAPD was also used in two
paediatric patients. A few patients entered CAPD because of a strong wish for the
independence offered by CAPD as compared to haemodialysis.

Period 1. From April 1980 to February 1982 a combination of gentamicin (Schering

Corporation, U.S.A.) and clindamycin (The Upjohn Company, U.S.A.) was given as
first line therapy (Table I). Continuous lavage was used for peritoneal administration. In
this period, 14 patients (10 male) aged 31-73 years (mean 58 years) underwent CAPD for
a total of 96 months (Table II).

Period 2. From February 1982 to September 1983, a combination of vancomycin (Eli

Lilly & Company, U.S.A.) and netilmicin (Schering Corporation, U.S.A.) was used as
first line therapy (Table I). Prolonged dwell times (4-6 h) were used for peritoneal
administration (Oreopoulos et al., 1981). From the second day onwards only four
daily exchanges were performed. Following the first negative culture of peritoneal
fluid, therapy was continued for seven days. In this period, 31 patients (20 male) aged
5-75 years (mean 52 years), have been treated with CAPD for a total of 284 months
(Table II).

Bacteriological cultures. Cultures of peritoneal fluid were performed immediately on the

suspicion of infection and then daily during hospitalization of the patient. Five millilitres
of peritoneal fluid was inoculated at the bedside into bottles containing standard blood
culture media.
 Vgocomycin and netilmirin in CAPD peritonitis 753

Table n . Patient material

Period 1 Period 2

No. of pts on CAPD: 14 31

Total no. of CAPD months: 96 284
CAPD months/pt: 6-9 9-2
Total no. of peritonitis episodes' 29 40
No. within training period; 5(17%) 8 (20%)
No. of episodes on CAPD: 24 32
No. of CAPD months/episode: 40 8-9
No. of peritonitis episodes with
adequate cultures:* 21 ((72%) 38 (95%)

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

no bacterial growth: 2 3
No. of peritonitis episodes available
for comparison of the efficacy of
antimicrobial regimens: 19 35

•In eight and two patients in period 1 and 2, respectively, adequate cultures were not obtained; three patients
died, one patient was transferred to hemodialysis and a few patients had commenced antibiotic treatment out-
side the university hospital.

Susceptibility determination. All isolated strains were tested for susceptibility to relevant
antibiotics with the disk diffusion method of Ericsson & Sherris (1971) with the modifi-
cations suggested by The Swedish Reference Group for Antibiotics (1981). 'Multiple
resistance' in Staph. epidermidis was defined as resistance (as judged by the disk diffusion
test) to four or more of phenoxymethylpenicillin, isoxazolylpenicillin, clindamycin,
erythromycin, tetracycline, fusidic acid, rifampicin and either or both of gentamicin
and tobramycin. MIC determinations were performed with the plate dilution technique
on PDM Sensitivity Test agar (AB Biodisk, Stockholm, Sweden) using a multipoint
inoculator and 105 bacteria/spot.

Determination of vancomycin and netilmicin serum levels. Vancomycin and netilmicin

serum concentrations were determined daily during therapy. Both determinations were
performed on the same sample which was obtained simultaneously with the changing of
antibiotic-containing dialysis bags. Netilmicin was determined with an agar plate dif-
fusion assay previously described (Kahlmeter, 1980) in which a vancomycin-resistant
Escherichia coli was used as test strain. After extraction of netilmicin with cellulose
phosphate powder (ICN Biochemicals, Cleveland, Ohio), as described by Stevens &
Young (1977), vancomycin was determined using a Staph. epidermidis as test strain. The
sensitivities of the netilmicin and vancomycin assays were 1 0 and 5 mg/1, respectively,
and the interassay variations (standard deviation/mean (%); n= 10 determinations) 5%
and 4%, respectively.

Statistical method. The Wilcoxon signed rank test was used to compare the bacterio-
logical clearance in period 1 and 2 (Table IV). Episodes of peritonitis within respective
training period were included (Table II). Five culture negative episodes and ten episodes
with inadequate bacteriological follow-up were excluded.
754 L. Brenner et al.

Table m. Micro-organisms isolated from peritoneal fluid of

CAPD patients with peritonitis

Period 1 Period 2
no. (%) no. (%)

Culture negative 2(5) 3(7)

Organism
Staph. epidermidis 15(34) 19(42)
Staph. aureus 2(5) 4(9)
Streptococcus spp. (non-A) 9(20) 4(9)
Diplococcus pneumoniae (type 5) 1(2) —
Cl. perfringens 1(2) —

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

E. coli 2(5) 1(2)
K. pneumoniae 4(9) 1(2)
Pseudomonas aeruginosa — 1(2)
Acinetobacter 1(2) 4(9)
Miscellaneous* 3(7) 7(16)
Can. albicansf 4(9) 1(2)

*Veillontllaparvula,B cereus, Staph atreus

tSuperinfections; one patient died, the rest were transferred to
hemodialysis and instituted on antifungal therapy, none were included in
the evaluation of the efficacy of the antibiotic regimens (Table IV)

Results
During period 1, the average time on CAPD was 6-9 months per patient (Table II), and
the incidence of peritonitis was one episode/40 CAPD months. During period 2, the
average time on CAPD was 9-2 months per patient and the incidence was one episode/
8-9 CAPD months. The average number of days in hospital was reduced from 17 in
period 1 to 12 in period 2.

Bacteriological results (Table III). Sixty to seventy per cent of the peritonitis episodes
were caused by Gram-positive micro-organisms, 15% by Gram-negatives and approxi-
mately 5% were culture negative. Staph. epidermidis was the most frequently isolated
pathogen in both periods, followed by Staph. aureus and Streptococcus spp. In period 1
35% of Staph. epidermidis were susceptible to both gentamicin and clindamycin, Staph.
aureus were susceptible to gentamicin but resistant to clindamycin, streptococci were
intermediately susceptible to gentamicin and susceptible to clindamycin while all Gram-
negative organisms were susceptible to gentamicin (cf. Table IV). In period 2 all bacteria
(except a Str. mitis) were susceptible to netilmicin and all Gram-positive bacteria (except
a Str. viridans) were susceptible to vancomycin. Organisms, isolated from the nine
peritonitis episodes in period 1 which were not cleared of bacteria within three days of
therapy, and their respective susceptibility to gentamicin and clindamycin, are shown in
Table IV.

Bacteriological clearance of peritonitis episodes (Table IV). Of the 19 episodes of bac-

teriologically verified peritonitis treated during period 1, 53% were culture negative
 Vancomydn and netihnidn in CAPD peritonitis 755

Table IV. Results of serial bacteriological cultures during therapy of CAPD peritonitis (number of
episodes cleared bactenologically*)

Duration Period 1 Bacterial isolate Period 2

(days) (episodes) & suscept. to GM/Clif (episodes)

1 6 28
2 2 — 5
3 2 — 2
5 E.coli S/R —
7 Staph. epidermidis S/R —
8 Staph. aureus S/R —
9 Str.faecalis I/R —
10 Staph. epidermidis I/R —

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

13 E. coli, Staph. epidermidis S/R, I/R —
14 Str faecalis I/R —
15 2 Staph. epidermidis I/R —
Staph. epidermidis, Klebsiella spp S/R, S/R
19 35

*The Candida albicans infections (Table HI) were late supennfections in all instances and were not included
in 'bacteriological clearance'.
fSuspecL = susceptibility, GM = gentamicin, Cli"=clindamycin, Gentanuan S>=MIC<4mg/l &
R = MIC>16mg/l;Clindamycin:S = M I C O mg/l& R = MIC>4mg/l

within three days, the remainder requiring 5-15 days of therapy. Of the 35 peritonitis
episodes treated with vancomycin/netilmicin, 80% were culture negative after one day of
therapy and all were negative after three days. The difference is statistically significant
-0005; Wilcoxon signed rank test).

Vancomycinjnetilmicin serum levels. The individual minimum and maximum netilmicin

and vancomycin blood levels attained during therapy with the maintenance dose
regimens (Table I) are shown in Figure 1.

Side effects. A 26-year-old woman (not previously exposed to vancomycin) with

known hypersensitivity to X-ray contrast medium and ampicillin experienced an attack
of asthma and a rise in blood pressure to 180/100 mm Hg after a rapid iv injection of
vancomycin. The reaction occurred immediately after first exposure to vancomycin.
Twenty minutes after treatment with theophylline, hydrocortisone and adrenaline all
symptoms abated and despite sustained high vancomycin/netilmicin blood levels the
reaction did not recur.
Audiometry was not performed regularly. None of the patients complained of
decreased hearing or vestibular malfunction.

Discussion
The drop in the incidence of peritonitis in our CAPD patients, from one episode/40
patient/months to one episode/8-9 patient/months, corresponds to the reduction
recorded by Gokal et al. (1982). It was probably the combined result of the development
of more adequate facilities, more elaborate training of patients and nursing staff and the
756 L. Brenner et aL

(b)

32
•
TT
•
•
TT
24 - T
T T
_ _T_ _
- •
•
TT •

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

16 • TT
T T
TTT TT
TTT
-•w- TTT
TT TT
TTTTT
TTT
e TT
TT

Mm Mo i Mm Mo»
Figure 1. NeUlmicin (a) and vancomyan (b) minimum and maximum concentrations in serum during
therapy with the maintenance dose regimen described in Table I. Horizontal bars — mean values.

institution of rigorous aseptic procedures. These improvements followed the increased

awareness of the role of Staph. epidermidis as a pathogen in CAPD peritonitis and of the
role of nosocomial dissemination of these strains between patients and nursing staff.
With the change from gentamicin/clindamycin to netilmicin/vancomycin for first-line
therapy several advantages were gained. Relative to the susceptibility pattern of current
infecting organisms a more potent antibiotic combination was obtained. The number of
days of hospitalization was reduced. Furthermore a reduction in the amounts and in
the variety of antibiotics used in the ward was achieved. As seen in Table III, micro-
organisms frequently associated with superinfection after prolonged antibiotic therapy
(Klebsiella pneumoniae and Candida albicans) were more frequently isolated in period 1.
Since therapeutic failures were more common in the first period, treatment was often
prolonged and often several antibiotic combinations were tried in the individual cases. In
the second period the rapid clinical effect of the vancomycin/netilmicin combination
obviated the long and varied courses.
So far we have seen no development of resistance to vancomycin or netilmicin among
isolates from the dialysis department. Since the introduction of the vancomycin/netil-
micin regimen patients and nursing staff have been bacteriologically screened at regular
intervals. The multiply resistant Staph. epidermidis isolates have not disappeared.
Between August 1982 and May 1984, ten multiply resistant Staph. epidermidis, three of
which were from the same patient, have been isolated. Plate dilution MIC determi-
nations showed that all strains were sensitive to vancomycin (MIC 0-5—1 Omg/1), eight
were sensitive (MIC ^ 4 mg/1) and two intermediately sensitive (MIC 8 mg/1) to netil-
micin, while nine of the strains were resistant to gentamicin and tobramycin (MIC
16—> 64 mg/1). The importance of continued screening for resistance development
should be underlined.
 Vancomycin and netilmiriii in CAPD peritonitis 757
The predominance of Gram-positive bacteria in general, and Staph. epidermidis in
particular, in peritonitis in CAPD patients is well documented (Nolph & Sorkin, 1981;
Oreopoulos, Vas & Khanna, 1981; Oreopoulos et al., 1981; Gokal et al., 1982; Grefberg,
Danielsson & Nilsson, 1984). However, as recently pointed out by Fenton (1984),
peritonitis in these patients may be caused by a 'bewildering array' of micro-organisms.
In the present study, isolated cases of peritonitis with pneumococci, Clostridium per-
fringens, Veillonella parvula and Bacillus cereus occurred. Others have found B. cereus,
Haemophilus influenzae (Gokal et al., 1982), Nocardia asteroides and Aspergillus sp.
(Nolph & Sorkin, 1981) as occasional causative agents. Gokal et al. (1982) described 13
patients who developed Cl. difficile colitis, a problem also noted by Borriello & Honour
(1983). In this respect, antibiotic courses should be kept as short as possible and
cephalosporins should perhaps be avoided as first line therapy (Gokal et al., 1982;
Fenton, 1984).

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

The vancomycin dose regimen used was approximately half of that recommended by
Oreoupoulos et al. (1981). However, the concentrations of netilmicin and vancomycin
attained in serum with the maintenance regimen indicate that concentrations in relevant
tissues would surpass the MICs of susceptible bacteria more or less continuously. At the
same time vancomycin serum concentrations were well below the recommended maxi-
mum levels (Kucers, 1982) and the concentrations of netilmicin were never above 8 mg/1
(Mawer, 1979). Vancomycin and netilmicin both have oto- and nephrotoxic properties.
Although the choice of netilmicin was not primarily based on toxicity aspects there is
reason to believe that netilmicin, if anything, is less oto- and nephrotoxic than the other
aminoglycosides (Kahlmeter & Dahlager, 1984). Audiometric investigations were only
performed in a few patients while the rest were evaluated clinically. The fact that no
patient complained of loss of hearing or of vestibular problems does not guarantee the
absence of eighth nerve side effects.
The only side effect seen, probably caused by vancomycin, was an acute attack of
asthma immediately following iv administration of 500 mg. The reaction abated despite
sustained vancomycin serum levels. Generalized flushing and pruritus related to fast
infusion of vancomycin (<15min) has been described (Masur et al., 1983). These
findings would suggest that rapid iv infusion of vancomycin should be avoided.
In conclusion, a combination of vancomycin and netilmicin was used successfully as
first line therapy of peritonitis in CAPD patients. The combination covers relevant
bacteria, it offers rapid bactericidal action, both drugs are stable in dialysis concentrates
and to our knowledge do not interact. A disadvantage is their potential oto- and nephro-
toxicity and the need for careful monitoring of therapy with regular determination of
drug serum levels.

Acknowledgements
We gratefully acknowledge the expertise of Professor Folke Nordbring, Department of
Infectious Diseases, and of Associate Professor Carl Kamme, Infection Control Nurse
Elwy Ekman and Laboratory Technician Anita Steen, Department of Hospital Infection
Control.

References
Borrelho, S. P. & Honour, P. (1983). CAPD peritonitis (Letter to the Editor). Lancet i, 348.
Ericsson, H & Sherris, J. C. (1971) Antibiotic sensitivity testing. Report of an international
collaborative study. Ada Pathologica et Microbiologica Scandinavica (B), Suppl., 217.
758 L. Branner et al.

Fenton, P. (1984). Treatment of peritonitis complicating continuous ambulatory peritoneal

dialysis. Journal of Antimicrobial Chemotherapy 13,411-12.
Gokal, R., Francis, D. M. A., Goodship, T. H. J., Bint, A. J., Ramos, J. M., Ferner, R E., Proud,
G. & Ward, M. K. (1982). Peritonitis in continuous ambulatory peritoneal dialysis. Lancet ii,
1388-91.
Glew, R. H. & Pavuk, R. A. (1981). Stability of vancomycin and aminoglycoside antibiotics in
peritoneal dialysis concentrate. Nephron 28,241-3.
Grefberg, N., Danielson, B. G. & Nilsson, P. (1984). Peritonitis in patients on continuous
ambulatory peritoneal dialysis Scandinavian Journal of Infectious Diseases 16,187-93.
Kahlmeter, G. (1980). Some causes for variation in aminoglycoside plate assays. Journal of
Antimicrobial Chemotherapy 6,43-52.
Kahlmeter, G. & Dahlager, J. I. (1984). Aminoglycoside toxicity—a review of clinical studies
published between 1975 and 1982. Journal of Antimicrobial Chemotherapy 13, Suppl. A, 9-22.
Kucers, A. (1982). Chloramphenicol, erythromycin, vancomycin & tetracyclines. Lancet ii, 425-8
Masur, H., Francioli, P., Ruddy, M. & Murray, H. W. (1983). Vancomycin serum levels and

Downloaded from http://jac.oxfordjournals.org/ at Carleton University on June 27, 2015

toxicity in chronic hemodialysis patients with Staphylococcus aureus bacteremia. Clinical
Nephrology 2, 85-8.
Mawer, G. E. (1979). Aminoglycoside pharmacology. In Recent Advances in Infection (Reeves, D.
& Geddes, A., Eds), pp. 121-35. Churchill Livingstone, Edinburgh.
Nolph, K. D. & Sorkin, M. I. (1981). Diagnosis and treatment of peritonitis. CAPD Update, pp.
265-72. Masson Publishing U.S.A., Inc.
Oreopoulos, D. G,, Robson, M., Izatt, S., Clayton, S. & deVeber, G. A. (1978). A simple and safe
technique for continuous ambulatory peritoneal dialysis (CAPD). Transcripts of American
Society of Artificial Internal Organs 24,484-9.
Oreopoulos, D. G., Vas, S. & Khanna, R (1981) Treatment of peritonitis (in patients on CAPD).
CAPD Update, pp. 259-64. Masson Publishing U.S.A., Inc.
Oreopoulos, D. G., Williams, P., Khanna, R. & Vas, S. (1981). Peritonitis in patients on CAPD.
Peritoneal Dialysis Bulletin, Suppl. 6, 5, 17-18.
Popovich, R. P., Moncrief, J W., Decherd, J. F., Bomar, J. B & Pyle, W. K. (1976). The definition
of a novel portable/wearable equilibrium peritoneal dialysis technique. Abstracts of American
Society of Artificial Internal Organs 5, 64.
Stevens, P. & Young, L. S. (1977). Simple method for elimination of aminoglycosides from serum
to permit bioassay of other antimicrobial agents. Antimicrobial Agents and Chemotherapy 12,
286-9.
The Swedish Reference Group for Antibiotics (1981). A revised system for antibiotic sensitivity
testing. Scandinavian Journal of Infectious Diseases 13, 148-52.
(Manuscript accepted 14 November 1984)