(Clinical Chemistry. 1998;44:1680-1684.)
© 1998 American Association for Clinical Chemistry, Inc.
|
General Clinical Chemistry |
Early detection of peritonitis in continuous ambulatory peritoneal dialysis patients by use of chemiluminescence: evaluation of diagnostic accuracy by receiver-operating characteristic curve analysis
Gönül Dalaman1,
Goncagül Haklar1,
Armand Sipahiu2,
Çetin Özener2,
Emel Ako
lu2,
and A. Süha Yalçin1,a
Departments of
1
Biochemistry and
2
Nephrology, School of Medicine, Marmara University
stanbul, 81326 Haydarpa
sa-
stanbul, Turkey.
a Author for correspondence. Fax 90-216-418-1047; e-mail ayalcin{at}marun.edu.tr.
 |
Abstract
|
|---|
Continuous ambulatory peritoneal dialysis (CAPD) is now a widely
accepted treatment for end-stage renal disease. However, the high
incidence of peritonitis is a major complication of CAPD.
Polymorphonuclear leukocytes (PMNs) play a major role in antimicrobial
response of the host. During phagocytosis, the PMNs undergo a striking
increase in oxidative metabolism, known as the respiratory burst, and
emit light as chemiluminescence (CL). CL is thus a sensitive measure of
PMN oxidative potential and correlates well with antimicrobial
activity. In view of the observation of increased susceptibility to
infection in CAPD patients, we have studied lucigenin- and
luminol-enhanced CL in peritoneal fluids of these patients and assessed
the diagnostic accuracy of these tests by ROC curve analysis. ROC
curves showed diagnostic accuracies for both tests that were superior
to counts of PMNs in the dialysis fluid (P <0.001). At
selected cutoff values of 150 000 cpm/vial for lucigenin CL and
600 000 cpm/vial for luminol CL, sensitivities were 100%.
Specificities for lucigenin and luminol CL were 89% and 80%,
respectively. Our results suggest that CL measurements can be used as
an early marker for the presence of infection in CAPD patients.
 |
Introduction
|
|---|
Continuous ambulatory peritoneal dialysis
(CAPD)1
is a
widely accepted treatment for end-stage renal disease that may be
caused by chronic glomerulonephritis, pyelonephritis, hypertension,
some immunological diseases, and toxic or ischemic damage to the kidney
(1)(2)(3). However, the frequent occurrence of peritonitis,
which is associated with high risk of mortality and morbidity, is a
major complication of CAPD (1)(4)(5)(6)(7). The
diagnosis and effective treatment of peritonitis depends on clinical
evaluation of the patient and correlation of this with laboratory
examination of the dialysate. The latter routinely includes the
determination of total leukocyte count and the recovery and
identification of microorganisms (7)(8)(9). Previous reports
have demonstrated problems associated with the diagnosis of peritonitis
based solely on these indicators
(4)(7)(10). Various techniques have
been used to facilitate the recovery of microorganisms from dialysate,
among them are the use of selected broth media, processing of large
volumes of dialysis effluent by concentration techniques or total
volume culture, chemical or physical disruption of phagocytes in
dialysate sediment for recovery of sequestered organisms, and the
removal of antibiotics from dialysate (7). However,
microorganisms are not always recovered from dialysate during
peritonitis, and up to one-third of reported cases of peritonitis have
been culture-negative (4)(7)(10).
The major role of phagocytes in antimicrobial response by the host has
encouraged the study of phagocyte function in CAPD patients. During
phagocytosis, the polymorphonuclear leukocytes (PMNs) undergo a
striking increase in oxidative metabolism, which is known as the
respiratory burst. Superoxide anion
(O2-) is produced during the
burst and is bactericidal either directly or by conversion to other
reactive metabolites via metal catalyzed reactions (11).
Production of O2- is accompanied by light emission
(chemiluminescence, CL) (12)(13). Therefore, CL
is a sensitive measure of the oxidative potential of phagocytes
and correlates well with antimicrobial activity.
In view of the observation of increased susceptibility to infection in
CAPD patients and on seeing the low sensitivity and specificity of the
methods available for the diagnosis of peritonitis, we have attempted
to study CL formation in dialysates of CAPD patients. We have evaluated
the diagnostic accuracy of CL in peritonitis by using ROC analysis of
data.
 |
Patients and Methods
|
|---|
patients
Patients (n = 55) attending the nephrology outpatient clinic
of Marmara University Hospital were included in the study. These were
27 women and 28 men, with an age distribution of 472 years,
established on CAPD for 2 months or longer. All patients gave informed
consent for the study, which was approved by the ethical committee.
Patients who had an underlying disease known to be associated with
abnormal PMN function, such as diabetes mellitus, tuberculosis, or
systemic lupus erythematosus, were excluded. Plasma creatinine
concentrations of the patients were 628 ± 221 µmol/L (7.1
± 2.5 mg/dL), and blood urea nitrogen concentrations were 31.7 ±
11.8 mmol urea/L (88.9 ± 33.0 mg/dL). The dialysis fluids
contained NaCl (5.7 g), sodium lactate (3.9 g),
CaCl2 (257 mg), MgCl2 (152 mg), and
glucose (13.6, 22.7, and 38.6 g/L) with an osmolality range of 275494
mOsm/L. All patients were trained in aseptic technique, and none had
infection at the time of study or had received antibiotics for 1 month
before the study. Dialysate fluids (n = 112) were collected from
55 patients, which included both noninfected patients and those
presenting with acute peritonitis. When symptoms such as cloudy fluid,
fever, abdominal pain, and rebound tenderness were present, diagnosis
of peritonitis was suspected; manual PMN counts >100/mm1
in the peritoneal dialysis effluent and positive culture proved the
diagnosis.
procedures
CL measurements.
Fresh specimens (3 mL) were transferred to
counting vials. CL was measured using a scintillation counter (Tricarb
1500, Packard Instruments) in out-of-coincidence mode with a single
active photomultiplier tube. CL was recorded after the addition of
luminol or lucigenin at a final concentration of 0.2 mmol/L to the
vials at room temperature. Each vial was observed continually for 60
min at 1-min intervals. After subtracting the appropriate control
values, peak CL was expressed as cpm per vial.
Additional quantitation and contribution of hypochlorite, hydroxyl
radical (·OH), and hydrogen peroxide
(H2O2) to luminol CL was determined in
samples with high CL (arbitrarily defined as >20 000 cpm/vial).
Inhibitors, enzymes, or scavengers affecting luminol-enhanced CL were
added to the vials, and the samples were recounted. Effects of
inhibitors and scavengers were expressed as the percentage of
inhibition of CL.
Spectrophotometric measurements.
Measurement of
O2- and myeloperoxidase (MPO) in dialysis effluents
was made by spectrophotometric methods. O2-
determination was based on reduction of ferricytochrome C
(14). MPO activity was determined by using
H2O2-dependent oxidation of
3,3',5,5'-tetramethylbenzidine as substrate (15).
Bacteriological methods.
The injection ports were disinfected
with methanol and allowed to dry for 2 min. Dialysate (50 mL) was then
aspirated into a sterile container and centrifuged at 2000g
at room temperature. The sediment was used for gram-stained film and
inoculated to the following media: blood agar, MacConkey agar,
chocolate agar, thioglyconate broth, and blood culture medium. All
cultures were incubated and examined daily for 7 days. Identification
of isolates was determined by standard methods.
Statistical analysis.
Numerical data were expressed as
mean ± SD. The significance of differences between the
experimental groups was estimated by the MannWhitney
U-test, and the differences were considered significant when
the probability was P <0.05.
ROC curves.
To assess the diagnostic accuracy of lucigenin-
and luminol-enhanced CL, we used ROC curves (16) and
calculated the areas under curves (AUCs) for comparison. ROC curves
were generated by plotting the relationship of the true positivity
(sensitivity) and false positivity (1 - specificity) at various
cutoff points of the tests. An AUC of 1.0 is characteristic of an ideal
test, whereas 0.5 indicates a test of no diagnostic value
(17). Cutoff values of lucigenin- and luminol-enhanced CL
for diagnosis of peritonitis were selected from experimental data and
were the values that maximized the sum of sensitivity and specificity.
 |
Results
|
|---|
peritoneal dialysate lucigenin and luminol cl concentrations in
patients with and without peritonitis
As shown in Fig. 1
, the mean lucigenin CL was 16 179 ± 14 916 cpm/vial in
patients without peritonitis (n = 81) and 1 389 554 ±
806,046 cpm/vial (P <0.05) in patients with peritonitis
(n = 41), whereas that of luminol CL was 11 780 ± 7111 and
5 501 879 ± 2 839 440 cpm/vial (P <0.05),
respectively.

View larger version (11K):
[in this window]
[in a new window]
|
Figure 1. The distribution of luminol and lucigenin CL in dialysate
fluids of patients without peritonitis (n = 81) and patients with
peritonitis (n = 41).
|
|
percentage of inhibition of luminol cl
Luminol CL was 98.5% inhibited by 0.1 mmol/L sodium azide, which
is an inhibitor of phagocyte MPO. The inhibition with 1570 U/vial
catalase (an enzyme that breaks down H2O2)
was 98.3% and that of 50 mL/L dimethyl sulfoxide (a scavenger of ·OH
radicals) was 63.2%.
spectrophotometric measurements
Table 1
shows the results obtained from
O2- and MPO determinations.
These indicators were significantly increased in patients with
peritonitis compared with noninfected patients.
roc curves
Figure 2
shows that the ability of peritoneal dialysate lucigenin and
luminol CL to differentiate patients with peritonitis from noninfected
cases exceeds that of PMN counts (AUC, 0.937 vs 0.774 for lucigenin CL
vs PMN count, and 0.928 vs 0.774 for luminol CL vs PMN count,
P <0.001). There was no significant difference between
lucigenin and luminol CL (AUC, 0.937 vs 0.928). The selected cutoff
values for diagnosis of peritonitis were 150 000 and 600 000 cpm/vial
(lucigenin and luminol CL, respectively). Table 2
shows the sensitivities and specificities of each test for the
diagnosis of peritonitis.

View larger version (16K):
[in this window]
[in a new window]
|
Figure 2. ROC curves for lucigenin CL, luminol CL, and PMN count of
peritoneal fluids to discriminate patients with peritonitis from those
without peritonitis.
The AUC values for lucigenin CL, luminol CL, and PMN count are 0.937,
0.928, and 0.774, respectively.
|
|
View this table:
[in this window]
[in a new window]
|
Table 2. Diagnostic performance of lucigenin and luminol CL at
selected cutoff points in patients with
peritonitis.
|
|
 |
Discussion
|
|---|
CAPD, brilliantly conceived by Popovich et al. (18),
should be an ideal form of dialysis treatment for acute and chronic
renal failure because it provides continuous, steady renal substitution
with good biochemical control, more liberal dietary and fluid intake,
improvement in anemia, and freedom from machines (19).
Although it can be easily performed at home, the relatively high
incidence of peritonitis is the major problem of CAPD. The incidence of
peritonitis was 5.27.5 episodes of per patient per year of dialysis
in late 1968 (20). The risk has reduced to 1 case in 2436
months with improvement of the technique (21), which was
later confirmed by a multicenter trial in Canada (22). The
most commonly encountered microorganism is coagulase-negative
Staphylococcus (3040%), whereas S. aureus
(20%), Streptococcus sp. (1015%), and other organisms,
including fungi, have also been isolated
(1)(23)(24)(25). Every peritonitis attack damages
the peritoneum and decreases peritoneal diffusion capacity
(25).
The diagnosis of peritonitis in CAPD patients is established when any
two of the following three criteria are present: cloudy or turbid
effluent containing >100 leukocyte/mm1
, symptoms and
signs of peritoneal inflammation, and a positive fluid culture
(26). In the absence of infection or disease, the peritoneal
cell population consists of 90% macrophages, 510% lymphocytes, and
<5% PMNs (24). On inflammation, rapid migration of PMNs
occurs, causing cloudiness of the fluid. However, an increase in
dialysate leukocyte count for the diagnosis of peritonitis is an
unreliable criterion, because the number of cells may vary within a few
hours (1). Drugs such as manidipine hydrochloride, a
dihydropyridine-type calcium channel blocker, may also cause turbidity
in peritoneal dialysate (27). In addition, symptoms and
signs of peritoneal irritation may precede the development of a turbid
effluent (1). Despite the use of diverse laboratory
procedures, including sedimentation of large volume of effluent, lysis
of phagocytes sequestering bacteria, and removal of antibiotics,
specimens of peritoneal fluid from patients with clinical peritonitis
have yielded negative cultures in 1028% of episodes
(28)(29). Conversely, positive cultures have
been recorded in as many as 37% of patients without other evidence of
peritonitis (30).
The major role played by phagocytes in the host antimicrobial response
largely depends on oxidative reactions known as the respiratory burst
(31). The respiratory burst consists of "four
increases": increase in oxygen uptake, O2-
production, H2O2 production, and hexose
monophosphate shunt activity. The initial radical
(O2-) is dismutated to H2O2,
which is then converted to hypochlorus acid by the action of phagocyte
MPO when Cl- is available, or to ·OH in the presence of
metals (32). CL is a phenomenon that is related to
respiratory burst and light emission by phagocytes and has proved to be
useful as a test of phagocyte function
(13)(32)(33)(34).
Luminol and lucigenin are widely used CL probes that differ in
selectivity and enhance the sensitivity of native CL. Luminol
quantifies a group of reactive species, including
H2O2, ·OH, hypochlorite, and
peroxynitrite, but lucigenin is mostly selective for
O2- (35). We have observed increases in
both lucigenin and luminol CL in effluents of patients having
peritonitis. We have added inhibitors and scavengers of free radicals
and determined that the phagocytes were functionally intact by all
means. The functional well-being of dialysate-elicited peritoneal
macrophages was also reported by others (36). Oxygen
concentrations within the peritoneal cavity were sufficient to maintain
adequate respiratory burst, even during acute peritonitis
(37).
We have examined additional indicators to validate CL increases. In
agreement with lucigenin CL, spectrophotometric determination showed
that O2- production was higher
in patients with peritonitis compared with the noninfected group. MPO,
which catalyzes the conversion of H2O2 to
hypochlorus acid for bactericidal action, was again significantly
increased (P <0.05).
To compare the performance of CL tests with other diagnostic measures
of peritonitis, we have used ROC plots, which provide pure indices of
accuracy (16). Both lucigenin and luminol CL exhibited
greater observed accuracy than PMN count for the diagnosis of
peritonitis. Statistical comparison of the AUCs showed that the
differences in accuracy were highly significant. The ROC plot AUC is
the most convenient global way to quantify the diagnostic accuracy of a
test (17). The AUC values for lucigenin CL indicated that in
93.7% of the cases, a randomly selected patient with peritonitis will
have a higher lucigenin CL than will a randomly selected noninfected
patient. This value is 92.8% for luminol CL. Because positive culture
was used for the estimation of peritonitis in constructing the plot,
comparison of diagnostic accuracy of CL with microbiologic techniques
was not possible. However, we had two patients with negative culture
results but with above-cutoff CL values, and they were later diagnosed
with fungal peritonitis.
The use of ROC curve analysis to assess the diagnostic accuracy of a
biological marker for use in patient management requires the selection
of a decision threshold. The large number of peritoneal fluids used in
our study allowed us to calculate with precision the cutoff values of
lucigenin and luminol CL as 150 000 and 600 000 cpm/vial,
respectively. We have selected cutoff values that maximized the sum of
sensitivity and specificity assuming that, at present, both sensitivity
and specificity are equally important in diagnosing patients with
peritonitis.
CAPD has proved to be a safe and effective alternative to hemodialysis
for patients with end-stage renal disease. Despite improvements in
technology, a major complication associated with this procedure is
peritonitis. Detection of etiologic agent relies heavily on culture
techniques, but the results have not been satisfactory. Our results
show that CL measurements can be used as an early marker for the
presence of infection in CAPD patients. In addition, decisions on
treatment effectivity and duration can be easily made if CL values are
considered.
 |
Footnotes
|
|---|
1 Nonstandard abbreviations: CAPD, continuous ambulatory
peritoneal dialysis; PMN, polymorphonuclear leukocyte; CL,
chemiluminescence; MPO, myeloperoxidase; and AUC, area under the
curve. 
 |
References
|
|---|
-
Keane WF, Vas SI. Peritonitis. Gokal R Nolph KD eds. The textbook of peritoneal dialysis 1994:473-501 Kluwer Academic Publishers The Netherlands. .
-
Oreopoulos DG, Rosbon M, Izatt S, Clayton S, DeVeber GA. A simple and safe technique for continuous ambulatory peritoneal dialysis. Trans Am Soc Artif Intern Organs 1987;24:484-487.
-
Nissenson AR, Gentile DE, Soderblom RE, Oliver DF, Brax C. Morbidity and mortality of continuous ambulatory peritoneal dialysis: regional experience and long-term prospects. Am J Kidney Dis 1986;7:229-234.
[ISI][Medline]
[Order article via Infotrieve]
-
Gould IM, Casewell MW. The laboratory diagnosis of peritonitis during continuous ambulatory peritoneal dialysis. J Hosp Infect 1986;7:155-160.
[ISI][Medline]
[Order article via Infotrieve]
-
Korzets Z, Korzets A, Golan E, Zevin D, Bernheim J. CAPD peritonitisinitial presentation as an acute abdomen with a clear peritoneal effluent. Clin Nephrol 1992;37:155-157.
[ISI][Medline]
[Order article via Infotrieve]
-
Morduchowicz G, van Dyk DJ, Wittenberg C, Winkler J, Boner G. Bacteremia complicating peritonitis in peritoneal dialysis patients [Case report]. Am J Nephrol 1993;13:278-280.
[ISI][Medline]
[Order article via Infotrieve]
-
Males BM, Walshe JJ, Amsterdam D. Laboratory indices of clinical peritonitis: total leukocyte count, microscopy, and microbiologic culture of peritoneal dialysis effluent. J Clin Microbiol 1987;25:2367-2371.
[Abstract/Free Full Text]
-
Taylor PC, Poole-Warren L, Grundy RE. Increased microbial yield from continuous ambulatory peritoneal dialysis peritonitis effluent after chemical or physical disruption of phagocytes. J Clin Microbiol 1987;25:580-583.
[Abstract/Free Full Text]
-
Peer G, Serban I, Blum M, Cabili S, Iaina A. Early diagnosis of gram-negative peritonitis in continuous ambulatory peritoneal dialysis with the limulus amebocyte lysate assay. Am J Nephrol 1992;12:19-21.
[ISI][Medline]
[Order article via Infotrieve]
-
Ludlam HA, Price TN, Berry AJ, Phillips I. Laboratory diagnosis of peritonitis on continuous ambulatory peritoneal dialysis. J Clin Microbiol 1988;26:1757-1762.
[Abstract/Free Full Text]
-
Babior BM. Oxygen-dependent microbial killing by phagocytes [Review]. N Engl J Med 1978;298:659-668.
[ISI][Medline]
[Order article via Infotrieve]
-
Ritchey EE, Wallin JD, Shah SV. Chemiluminescence and superoxide anion production by leukocytes from chronic hemodialysis patients. Kidney Int 1981;19:349-358.
[ISI][Medline]
[Order article via Infotrieve]
-
Allen RC, Stejernholm R, Steele R. Evidence for the generation of an electronic excitation state(s) in human polymorphonuclear leukocytes and its participation in bactericidal activity. Biochem Biophys Res Commun 1972;47:679-684.
[ISI][Medline]
[Order article via Infotrieve]
-
Bhuyan DK, Bhuyan KC. Assessment of oxidative stress to eye in animal model for cataract. Methods Enzymol 1994;233:630-639.
[ISI][Medline]
[Order article via Infotrieve]
-
Grisham MB, Jefferson MM, Melton DF, Thomas EL. Chlorination of endogenous amines by isolated neutrophils. J Biol Chem 1984;259:10404-10408.
[Abstract/Free Full Text]
-
Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine [Review]. Clin Chem 1993;39:561-577.
[Abstract/Free Full Text]
-
Huguet J, Castineiras MJ, Fuentes-Arderiu X. Diagnostic accuracy evaluation using ROC curve analysis. Scand J Clin Lab Investig 1993;53:693-699.
[ISI][Medline]
[Order article via Infotrieve]
-
Popovich RP, Moncrief JW, Nolph KD, et al. The definition of a novel portable/wearable equilibrium peritoneal dialysis technique [Abstract]. Am Soc Artif Intern Organs 1976;5:64.
-
Rae A, Pendray M. Advantages of peritoneal dialysis in chronic renal failure. JAMA 1973;225:937-941.
[ISI][Medline]
[Order article via Infotrieve]
-
Cohen SL, Percival A. Prolonged peritoneal dialysis in patients awaiting renal transplantation. Br Med J 1968;1:409-413.
-
Fellin G, Gentile MG, Manna GM. Peritonitis prevention: Y connector and sodium hypochlorite 1987:114-118 Khanna R, Nolph KD, Prowart B, Twardowski ZJ, Oreopoulos DG, eds. Advances in peritoneal dialysis. Toronto: Peritoneal Dialysis International, Inc. Three years' experience. Report of the Italian study group. In. .
-
Canadian CAPD, . Clinical Trials Group. Peritonitis in continuous ambulatory peritoneal dialysis: a multicenter randomized clinical trial comparing the Y connector disinfectant system to standard systems. Perit Dial Int 1989;9:159-163.
-
Blagg CR, Scribner BH. Long-term dialysis: current problems and future prospects [Editorial]. Am J Med 1980;68:633-635.
[ISI][Medline]
[Order article via Infotrieve]
-
Berger HG, Oettinger N. Bacterial peritonitis in surgery; pathogenesis, symptoms, surgical therapy, peritoneal lavage and progress. Augistin R eds. Peritonitis in CAPD 1987:3-9 Karger Press Basel. .
-
Lamperi S, Carozzi S, Nasini MG. Peritoneal membrane defence mechanism in CAPD. Augistin R eds. Peritonitis in CAPD 1987:69-72 Karger Press Basel. .
-
Peterson PK, Matzke G, Keane WF. Current concepts in the management of peritonitis in patients undergoing continuous ambulatory peritoneal dialysis. Rev Infect Dis 1987;9:604-612.
[ISI][Medline]
[Order article via Infotrieve]
-
Yoshimoto K, Saima S, Echizen H, Nakamura Y, Ishizaki T. A drug induced turbid peritoneal dialysate in five patients treated with continuous ambulatory peritoneal dialysis. Clin Nephrol 1993;40:114-117.
[ISI][Medline]
[Order article via Infotrieve]
-
Fenton P. Laboratory diagnosis of peritonitis undergoing continuous ambulatory peritoneal dialysis. J Clin Pathol 1982;35:1181-1184.
[Abstract/Free Full Text]
-
Gokal R, Francis DMA, Goodship THJ, Bint AJ, Romas JM, Ferner RE, et al. Peritonitis in continuous ambulatory peritoneal dialysis. Lancet 1982;ii:1388-1391.
-
Knight KR, Polak A, Crump J, Maskell R. Laboratory diagnosis and oral treatment of continuous ambulatory peritoneal dialysis peritonitis. Lancet 1982;ii:1301-1304.
-
Paul JL, Roh-Alliver M, Man NK, Luong N, Moatti N, Raichvarg D. Influence of uremia on polymorphonuclear leukocytes oxidative metabolism in end-stage renal disease and dialysed patients. Nephron 1991;57:428-432.
[ISI][Medline]
[Order article via Infotrieve]
-
Babior BM. Oxidants from phagocytes: agents of defense and destruction[Review]. Blood 1984;5:959-966.
-
Sagone AL, Jr, King GW, Metz EN. A comparison of the metabolic response to phagocytosis in human granulocytes and monocytes. J Clin Investig 1976;57:1352-1358.
-
Webb LS, Keele BB, Jr, Johnston RB, Jr. Inhibition of phagocytosis-associated chemiluminescence by superoxide dismutase. Infect Immun 1974;9:1051-1056.
[Abstract/Free Full Text]
-
Davies GR, Simmond NJ, Stevens TRJ, Sheaf MT, Bnavala N, Laurensen IF, et al. Helicobacter pylori stimulates antral mucosal reactive oxygen metabolite production in vivo. Gut 1994;35:179-185.
[Abstract/Free Full Text]
-
Coles GA, Lewis SL, Williams JD. Host defense and effects of solutions on peritoneal cells. Gokal R Nolph KD eds. The textbook of peritoneal dialysis 1994:503-528 Kluwer Academic Publishers The Netherlands. .
-
Topley N, Alobaidi HMM, Davies M, Coles GA, Williams JD, Lloyd D. The effect of dialysate on peritoneal phagocyte oxidative metabolism. Kidney Int 1988;34:404-411.
[ISI][Medline]
[Order article via Infotrieve]