| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Clinical Chemistry, Academic Hospital Vrije Universiteit Brussel (AZ-VUB), Laarbeeklaan 101, B-1090 Brussels, Belgium
aauthor for correspondence: fax 32-2-477-5047, e-mail lchmgsf{at}az.vub.ac.be
The classification of pleural effusions as transudates or exudates often is an early step in their evaluation (1)(2)(3)(4). Exudates but not transudates require additional and often invasive diagnostic procedures. The criteria of Light et al. (5) for classification of pleural effusions have been reported to lead to unwarranted invasive diagnostic procedures in 20–30% of patients with a transudate (1)(3) and to misclassify some exudates as transudates (3)(4).
Protein analysis has been proposed to improve classification into transudates or exudates based on the assumption that large proteins are present only in exudates because of the increased capillary permeability (1). Agarose gel electrophoresis (AGE) has been the standard procedure for serum protein fractionation for >30 years, but it is laborious and methods do not agree (6)(7)(8). Capillary zone electrophoresis (CZE) is an attractive alternative (6)(9)(10) and avoids between-protein differences in dye affinity.
We evaluated the Beckman Paragon CZETM 2000 system, a multicapillary instrument for automated serum protein electrophoresis (7)(8), for quantitative fractionation of proteins in body fluids.
For method comparison, leftover samples from 49 pleural and 11 ascitic fluids from 47 inpatients were used. There were no preset selection criteria regarding age, sex, or type of disease (11); however, only 37 pleural effusions were classified as transudates or exudates after exclusion of patients lacking definitive clinical diagnosis (n = 5), patients with multiple diagnoses (n = 3), or in case of interference by radioopaque agents (n = 4). A serum sample was collected simultaneously with the pleural fluid (n = 30) or within 24 h before (n = 13) or after (n = 17) the puncture. Final diagnosis was retrieved from the patient’s file based on the conclusions reached by physicians from the Department of Internal Medicine, who were unaware of the results of the method comparison, and after review of all available clinical, anatomicopathological, and microbiological reports.
VitrosTM slides (Ortho Clinical Diagnostics) were used for total protein and lactate dehydrogenase (LD) measurement. Albumin in fluids was measured immunonephelometrically on a BNATM analyzer (Dade Behring).
AGE was performed on the REPTM system (Helena
Laboratories) using Ponceau S staining. CZE was carried out with
Paragon CZE 2000. When necessary for CZE, fluids were concentrated
(Centricon filters; cutoff, Mr
30 000; Millipore Corporation) to total protein concentrations 
35
g/L. All electrophoretograms were independently interpreted by two
laboratory investigators blinded for the final clinical diagnosis. The
study was performed in accordance with the current revision of the
Helsinki Declaration of 1975 (12).
The selected 37 pleural fluids (see above) were classified as exudates
(n = 26) or transudates (n = 11) according to the patient’s
clinical diagnosis [for criteria, see Ref. (1)] and
compared with their classification (a) according to Light et
al. (5) when at least one of the following criteria was met:
(i) pleural fluid/serum total protein ratio >0.5,
(ii) pleural fluid/serum LD ratio >0.6, and/or
(iii) pleural fluid LD more than two-thirds of the upper
reference limit of serum LD; or (b) according to Chen and
Lam (1) when any of the 
2-globulin,
ß-lipoprotein, and 
-globulin bands were detected in the
electrophoretogram (serum-like pattern on AGE).
Paired Wilcoxon tests and unpaired Mann–Whitney U-tests were performed two-tailed with Analyze-It for Microsoft Excel (significance threshold <0.05). Linear or Passing-Bablok regression analysis was used for method comparisons. ROC curve analysis was performed with MedCalc.
Forty-nine pleural and 11 ascitic fluids were analyzed with CZE and
AGE. In four pleural and one ascitic fluid samples, excluded from
statistical analysis, radioopaque agents caused a spurious peak in the
2-globulin fraction in the CZE
electrophoretograms, as has been observed in serum (13).
The median total protein concentration (range) was 32.0 g/L (1.0–65.0
g/L) for the remaining 45 pleural fluids and 17.0 g/L (9.0–30.0 g/L)
for the 10 ascitic fluids. Results obtained by CZE and AGE were
statistically different for albumin and 1-,
2-, ß-, and -globulins (P
0.001) but were very significantly correlated (see Table 1
). Albumin concentrations determined electrophoretically
correlated significantly with nephelometric data (Table 1
); values
obtained by AGE, but not by CZE, were higher than those obtained by the
BNA method (P <0.0001). As in serum, values obtained by CZE
for the 1- and ß-globulin fractions were
30% and 15% higher, respectively, because of detection of
1-acid glycoprotein and
C3-complement, which were not detected by AGE
(7)(8)(10). In 20 samples, a
pseudomonoclonal fibrinogen peak was detected in the -globulin
fraction by AGE, but not by CZE. -Globulin fractions measured by AGE
were higher in samples with (P <0.006) and without
fibrinogen (P = 0.03; n = 35). In one pleural
effusion, a monoclonal (IgG 
) peak was detected in the -globulin
fraction in the AGE (11.7 g/L) and CZE (12.0 g/L) electrophoretograms.
|
Thirty-seven pleural fluids were clinically classified as exudates (n = 26; 10 patients with malignancies, 13 with pneumonic and parapneumonic effusions, 3 with Dressler syndrome) or as transudates [n = 11; 10 patients with congestive heart failure and 1 with (pseudo)Meigs syndrome].
Electrophoretic patterns of unconcentrated pleural effusions are shown
in Fig. 1
. By visual inspection of AGE electrophoretograms, all exudates
(n = 26) but only 7 of 11 transudates were correctly classified vs
24 of 26 exudates and 10 of 11 transudates correctly classified with
the criteria of Light et al. (5). Consistent with other data
(3)(4), the criteria of Light et al.
(5) achieved 92% diagnostic sensitivity [95% confidence
interval (CI), 75–99%] and 91% specificity (95% CI, 59–100%) for
exudates and correctly classified 92% of the fluids. Visual
interpretation of AGE electrophoretograms achieved 100% sensitivity
(95% CI, 87–100%) and 64% specificity (95% CI, 31–89%) with an
overall 89% diagnostic efficiency, consistent with Chen and Lam
(1). Visual inspection of AGE electrophoretograms
complemented the criteria of Light et al. (5) in diagnosing
two otherwise undetected malignant exudative effusions; conversely, the
criteria of Light et al. identified three transudates attributable to
congestive heart failure [valvular dysfunction (n = 2) and cor
pulmonale (n = 1)] in addition to those detected by visual
inspection of AGE electrophoretograms. The pleural effusion caused by
the Meigs syndrome was assigned as exudate by both criteria.
|
Qualitative evaluation of CZE electrophoretograms was not performed
because concentration steps before analysis (see above) could visualize
small 2- and -globulin bands in
transudates, whereas ß-lipoprotein bands are not observed in CZE 2000
electrophoretograms (10). Quantitative analysis was
performed on 26 exudates and 11 transudates, applied at similar
concentrations in both methods. Electrophoretically fractionated (not
shown) and total protein concentrations were significantly higher
(P <0.05) in exudates [median (range), 35.0 g/L
(11.0–65.0 g/L) vs 18.0 g/L (1.0–31.0 g/L) for transudates]. The
2:albumin ratio was lower (P
<0.001) in transudates [median ratio (range), 0.10 (0.09–0.22) vs
0.19 (0.06–0.59) for exudates with CZE; 0.06 (0.05–0.11) vs 0.12
(0.05–0.40) for exudates with AGE]. Cutoff values for optimal
diagnostic efficiency determined by ROC curve analysis (Fig. 1
) were
>0.14 for CZE and >0.11 for AGE; the areas under the ROC curves were
0.85 (95% CI, 0.70–0.95) and 0.83 (95% CI, 0.68–0.94),
respectively. The established cutoff values for
2:albumin ratios are lower than those
calculated (using Beckman SPE gels) by Chen and Lam (1), who
classified effusions by electrophoretic instead of clinical criteria.
Clinical classification, however, is limited by the fact that some
pleural effusions are mixtures of transudates and exudates
(1).
For CZE the established cutoff value for the
2:albumin ratio achieved a diagnostic
sensitivity and specificity for exudates of 81% (95% CI, 61–93) and
91% (95% CI, 59–99), respectively, with 84% correctly classified
fluids. Likewise, AGE achieved a diagnostic sensitivity of 62% (95%
CI, 44–80) and specificity of 100% (95% CI, 100–100) with 73%
efficiency, which is not significantly different from all other
criteria used because of overlapping CIs. Nevertheless, compliance with
the criteria of Light et al. (5) for exudates or with CZE
2:albumin ratio >0.14 correctly classified
all exudates, but with a specificity of 82% (two transudates still
being classified as exudates). With an AGE
2:albumin ratio >0.11 as criterion, two
malignant effusions were still erroneously diagnosed as transudates.
Thus, combined use of the criteria of Light et al. and CZE
quantification can achieve 100% sensitivity for exudates, but may
still lead to unwarranted investigations in a limited number of
patients because of their suboptimal specificity.
In conclusion, the Beckman Paragon CZE 2000 analyzer can be used for
the fractionation of proteins in pleural and ascitic fluids. The
results correlate significantly with those obtained by REP AGE. Albumin
concentrations measured by CZE agree better with
immunonephelometrically determined values on the BNA analyzer than do
results obtained by AGE. Protein zone electrophoresis could complement
the criteria of Light et al. (5) by increasing the
sensitivity of the detection of exudative pleural effusions. The cutoff
for the 2:albumin ratio on CZE achieved a
slightly, albeit not significantly, higher sensitivity for exudates
than the same ratio on REP.
Acknowledgments
We gratefully acknowledge the invaluable technical assistance of V. Baeten and S. Exterbille.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
