| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
1
Serveis de Bioquímica i Medicina Interna, Hospital General Universitari "Vall d'Hebron," P. Vall d'Hebron 119–129, 08035-Barcelona, Spain.
a Author for correspondence. Fax +34-3-4280443.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Key Words: indexing terms: interference, source of • bacterial infections • D-lactate acidosis • ultrafiltration • sample preparation
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The presence of D-lactate in sterile body fluids has been reported to indicate local bacterial infection or absorption from a site containing a high density of bacteria (8). Many bacterial pathogens produce D-lactic acid under anaerobic conditions that resemble those found at sites of infection (9). Several reports suggest that D-lactate assessment in ascitic, pleural, cerebrospinal, and synovial fluid could be a highly specific and sensitive test for the early diagnosis of bacterial infection, as compared with the Gram stain, bacterial culture, and pH or L-lactate determination (10)(11)(12)(13).
Except in short bowel syndrome patients, the concentrations of D-lactate in body fluids are on the order of micromoles per liter, and L-lactate is measured in millimoles per liter. Among the methods described to assess D-lactate (1)(2)(14)(15)(16), the most widely used is the enzymatic–spectrophotometric technique based on the conversion of D-lactate to pyruvate in the presence of NAD+ and D-lactate dehydrogenase (D-LD), in both the end-point version (15) and the kinetic version (16).
Usually both D-lactate and L-lactate are required in the same specimen. To avoid the formation of L-lactate in vitro, the samples are either deproteinized by acid precipitation or collected in sodium fluoride to inhibit glycolysis. Deproteinization results in a 3- to 10-fold sample dilution, which implies measuring D-lactate at concentrations close to the method limit of detection in many cases. Nondeproteinized samples collected in NaF are exposed to interference by the endogenous L-LD/L-lactate content, because NAD+ is the cosubstrate of both L-LD and D-LD.
In this work we studied the interference of L-lactate and L-LD in the determination of D-lactate in nondeproteinized samples with a spectrophotometric method, and we describe a sample deproteinization treatment that avoids dilution and permits measurement of this component with commonly used laboratory instrumentation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
When indicated, samples were made protein-free by ultrafiltering 300 µL through a cellulose membrane (Ultrafree-MC; Millipore Corp., Bedford, MA) and centrifuging (12 000g) at 4 °C for 90 min.
d-lactate assay
We used an enzymatic–spectrophotometric method based on the
oxidation of D-lactate to pyruvate by
NAD+ in the presence of bacterial D-LD
(EC 1.1.1.28, from Lactobacillus leichmannii). Pyruvate is
transformed in a subsequent reaction catalyzed by alanine
aminotransferase (ALT; EC 2.6.1.2) in the presence of
L-glutamate. All reagents were purchased from Boehringer
Mannheim (Mannheim, Germany) except L-glutamic acid, which
was from Sigma Chemical Co. (St. Louis, MO).
The method was automated with a Cobas-Mira Plus analyzer (Roche, Branchburg, NJ) under the following conditions: final concentrations of the reagents: glycylglycine/L-glutamate buffer, pH 10.0, 206 and 34.3 mmol/L, respectively; NAD+ 3.44 mmol/L; ALT 22.0 kU/L; and D-LD 57.7 kU/L. The reaction was started by adding D-LD, and the ratio of final volume to sample volume was 13. Absorbance (340 nm) was measured at 0 and 20 min after starting the reaction; the difference between the two absorbances was used in the calculations.
A calibration curve with solutions of D-lactate (monolithium salt, Boehringer Mannheim) in water was processed in triplicate in each batch. A reagent blank to compensate for NAD+ stability was analyzed in each run and subtracted from the calibrators and unknowns. A sample blank was also processed in the nondeproteinized samples to assess the nonspecific NAD+ transformation.
The reaction was linear between 0 and 2000 µmol/L (A = 2.6 x 10-4 x [D-lactate]µmol/L + 2.42 x 10-3; r = 0.999). The detection limit of the method (mean + 3 SD reagent blank) was 15 µmol/L.
assessment of d-ld stereospecificity
To check D-LD stereospecificity, we determined
D-lactate in solutions (0–50 mmol/L) of
L-lactate (Sigma). Because L-lactate
preparations can be contaminated with the D- isomer, we
proceeded as follows: A solution of 50 mmol/L L-lactate was
prepared in glycylglycine and glutamate buffer, to which we added
NAD+ and ALT. The solution was separated into two
aliquots. One was supplemented with D-LD (to completely
transform the D-lactate) and the other with an equal volume
of water. The final concentrations of all reagents were the same as
those used in the D-lactate determination. We confirmed
that in the first aliquot the D-lactate content was
completely transformed by monitoring the appearance of NADH at 340 nm
until constant absorbance was reached. The D-lactate
contamination of the L-lactate solution, calculated with
the NADH absorptivity molar coefficient at 340 nm (6.299 x
106 cm2/mol) and the absorbance at 20
min, was 125 µmol/L D-lactate in a 50 mmol/L
L-lactate solution, i.e., 0.25% contamination of the
L isomer by the D isomer.
Dilutions were then prepared with the two aliquots and D-lactate and L-lactate were determined. L-Lactate was analyzed with a commercially available kit in a Hitachi 917 autoanalyzer (Boehringer Mannheim).
estimation of apparent l-ld and d-ld km for nad+
To determine the KM value of
L-LD for NAD+ in our experimental
conditions, we performed the main reaction (lactate to pyruvate)
without coupling the auxiliary reaction with L-glutamate
and ALT. Various NAD+ concentrations (0–5 mmol/L)
were tested at several fixed L-lactate concentrations
(2–80 mmol/L). The buffer/L-lactate was incubated with
purified L-LD for 3.3 min, after which time
NAD+ was added and the reaction monitored over its
linear part. Purified L-LD5 (from human placenta; Sigma)
was chosen because it is the form predominantly found in
bacteria-infected fluids.
We conducted similar experiments with the D-LD used in the determination of D-lactate. In this case the buffered D-lactate solution (D-lactate ranged from 2 to 10 mmol/L) was first incubated with D-LD, and the reaction was started by adding the NAD+ solution at different concentrations (0–10 mmol/L).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
).
|
To study the influence of the
L-lactate/L-LD content of the sample on
interference, we added similar concentrations of D-lactate
and various concentrations of L-lactate to plasma samples
with constant concentrations of L-LD. The results show that
the value of the sample blank depended on the L-LD and
L-lactate concentrations in the samples (Table 1
). With similar L-LD catalytic concentrations, the
value of the sample blank increased when L-lactate
increased and, similarly, with approximately equal
L-lactate concentrations the sample blank increased when
L-LD increased.
|
Interference of the L-lactate concentration on
D-lactate determination: D-LD
stereospecificity.
As shown in Fig. 2
, the D-LD used in the D-lactate
determination did not catalyze the transformation of
L-lactate. At all the points where D-lactate
was eliminated (line A), the signal in the D-lactate
determination was below the detection limit, whereas in those not
previously treated with D-LD (line B), the slope of the
line confirmed the degree of contamination calculated above for the
L-lactate preparation, 0.25%.
|
Affinities of L-LD and D-LD for
NAD+ in the D-lactate analysis reaction
conditions.
Plotting the reciprocal of initial velocity vs the
reciprocal of NAD+ concentration gave a series of
straight lines that intercepted at a single point. This is the pattern
expected for a sequential mechanism in a bireactant system. The rate
equation describing this mechanism is:
 | (1) |
and to estimate the constant
Ka. Similar kinetic experiments were carried out with D-LD and D-lactate.
From these experiments the following estimates were obtained:
D-LD KM (NAD+) = 12.34 ± 0.44 mmol/L
L-LD KM (NAD+) = 0.741 ± 0.065 mmol/L
Effect of increasing the NAD+
concentration in the reaction mixture
. The effect produced with
different concentrations of NAD+ in the reaction mixture is
shown in Fig. 3
for a plasma sample with a high L-lactate
concentration and moderate D-lactate and L-LD
contents. The plots produced in monitoring the reaction show that the
final slopes of the sample blank and the unknown became more similar as
the concentration of NAD+ was increased.
|
determination of d-lactate in ultrafiltered specimen
Because sample blanks could not be measured properly in
nondeproteinized samples, we ultrafiltered the specimen before
D-lactate determination.
To evaluate the effectiveness of our ultrafiltration system in
eliminating endogenous L-LD, we determined the
catalytic concentrations of L-LD in a set of plasma pool
aliquots with different additions of purified L-LD (see
Table 1
), before and after ultrafiltration. The initial
L-LD catalytic concentrations were those shown in Table 1
,
whereas after ultrafiltration, L-LD was undetectable in all
samples (Deutsche Gesellschaft für Klinische
Chemie method, 37 °C, Hitachi 917 automated analyzer). To check
the influence of the initial protein concentration on the effectiveness
of protein removal, we prepared three samples (a pool of plasma and two
different dilutions with NaCl 9 g/L) to obtain protein concentrations
between 18.4 and 72.4 g/L; L-lactate was added to the pool
and to both dilutions to obtain L-lactate concentrations
(endogenous plus added) of between 3.5 and 7.6 mmol/L. In the
ultrafiltered samples there was no detectable total protein (biuret
method, Hitachi 917 analyzer), and L-lactate concentrations
were only slightly higher (mean concentration factor 1.04).
The imprecision of the procedure as a whole (ultrafiltration plus determination) was studied in plasma pools. The within-run CVs at D-lactate concentrations of 31, 65, and 107 µmol/L were 4.2% (n = 12), 2.5% (n = 14), and 3.0% (n = 30), respectively. The between-run imprecision was determined at D-lactate concentrations of 49, 70, and 115 µmol/L and corresponding CVs were 7.3% (n = 13), 4.4% (n = 15), and 3.1% (n = 14).
Recovery experiments implemented with the ultrafiltration procedure were conducted in triplicate by adding 20, 40, 60, and 80 µmol/L D-lactate to a plasma pool. Mean ± SD recovery was 108.3% ± 2.4%.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Eynard et al. (12) reported that end-point enzymatic determination of D-lactate is not influenced by L-lactate in concentrations up to 8 mmol/L. They measured D-lactate in nondeproteinized plasma and cerebrospinal fluid and did not describe the L-LD content of the samples or whether a sample blank was processed in the analyses. Ludvigsen et al. (16) used the enzymatic method with two-point kinetic calibration and a 20-min preincubation of the sample with all reagents except D-LD, and reported that this process eliminated the need for specimen-blanking or protein-precipitating pretreatment (8).
In contrast, we found that D-lactate results were
clearly affected in the nondeproteinized samples, even when sample
blanks were processed. Fig. 1
shows that preincubating the sample with
NAD+ for as long as 20 min did not eliminate
unspecific NAD+ reduction, and illustrates the need for
processing sample blanks even in kinetic measurements. The difficulty
in correct determination of sample blank values when nondeproteinized
specimens are used is also evident.
Our experiments show that L-lactate alone does not account for interferences in D-lactate assay because D-LD is unable to catalyze L-lactate oxidation. The dissimilar slopes of the sample and the sample blank after completion of D-lactate oxidation would be explained if D-LD had a higher affinity than L-LD for NAD+. The L-LD5 KM for NAD+ estimated in our assay conditions is about threefold that reported for the same isoenzyme by Bais and Philcox (17) (0.74 ± 0.07 vs 0.22 ± 0.06) in N-methyl-D-glucamine buffer (pH 9.4). However, our estimate of D-LD KM for NAD+ (12.34 ± 0.44) is ~17-fold that of L-LD, indicating that the former has a lower affinity for NAD+. Therefore, neither the absence of stereospecificity of the D-LD preparation nor its affinity for the coenzyme justifies the false results observed in D-lactate determination in the nondeproteinized specimens.
The finding that increasing the NAD+
concentration of the reaction mixture reduces the difference between
the final slopes of the unknown and the sample blank (Fig. 3
) suggests
that a substantial amount of the coenzyme is complexed to the high
D-LD concentration in the unknown, thereby decreasing the
availability of the coenzyme for L-LD. This would result in
the lower slope observed in some unknowns, when compared with
corresponding blanks at the end of the reaction period when
D-lactate is completely oxidized: The enzyme could be
complexed with the coenzyme but there is no specific substrate
(D-lactate) to transform. Because the slope of the
unspecific reaction is masked in the unknown, proper measurement of
sample blanks in nondeproteinized samples is extremely difficult.
Although increasing NAD+ concentrations reduces the
difference in the slopes of the sample blank and the unknown, this
would result in high initial and final absorbances that, in many cases,
are above those of the highest D-lactate calibrator (data
not shown). Also, proper measurement of the sample blank is not
guaranteed.
The interferent effect observed in enzymatic D-lactate determination is produced only in the presence of L-LD and L-lactate together, and although it is especially striking when both are in high concentrations, it is noticeable even at moderate concentrations. This circumstance is common in infected fluids. We have found L-lactate and L-LD concentrations from 0.76 to 10.32 mmol/L and 153 to 3648 U/L, respectively, in noncomplicated metapneumonic or tuberculous effusions, and from 0.94 to 49.9 mmol/L and 593 to 50 000 U/L in empyema (data not shown). This is noteworthy because D-lactate assessment has been proposed as an efficient tool to differentiate infected from noninfected pleural fluids (10)(11). In body fluids with lower L-LD catalytic concentrations such as cerebrospinal fluid, the interference is expected to be less pronounced; nevertheless, it is advisable to treat the sample before D-lactate determination.
It is likely that the interference we observed may account for the higher values found when nondeproteinized human plasma and no sample blank are used to assess D-lactate than when protein-free samples are used.
All the facts so far demonstrate the advantages of using protein-free samples for assessment of D-lactate. Generally samples are collected in perchloric acid to avoid in vitro L-lactate production, but because this can influence the pH of the final reaction, several authors recommend neutralizing the sample to keep the reaction pH at optimum (2)(14). These two steps imply considerable dilution, so that in many samples D-lactate is measured at concentrations equal to or below the method detection limit (8). In all the reports in which D-lactate is determined in several body fluids, including normal plasma, nondetectable concentrations are found in noninfected, but also in many infected, fluids (10)(11)(12)(13)(18). In addition, the procedure can be influenced by the rapid conversion of S-D-lactoylglutathione to D-lactate, especially in such fluids as whole blood, because of the high erythrocyte content of this D-lactate precursor (2). We recommend this ultrafiltration system to eliminate proteins because it is simple, convenient, and causes almost no alteration of the lactate content of the sample. Furthermore, the imprecision of the procedure as a whole, ultrafiltration plus determination, is low. The high recovery ranges found, with values >100%, can be explained, at least in part, by the concentration effect of the ultrafiltration process on low-molecular-mass solutes.
The detection limit and the measurable range of the enzymatic method we propose to assess D-lactate make it suitable for detecting bacterial infection in body fluids with cutoff values between 50 and 150 µmol/L(10)(11)(12)(13)(18). The method can also be applied to D-lactate determination in plasma of patients with suspected D-lactate acidosis.
In conclusion, samples for D-lactate determination by enzymatic methods should be deproteinized. The use of ultrafiltration avoids dilution, requires a moderate volume of sample, and results in good imprecision and acceptable recovery. Moreover, it eliminates interference by the in vitro rapid conversion of S-D-lactoylglutathione to D-lactate, observed when deproteinized whole blood is used.
|
Acknowledgments |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  D. J. Herrera, K. Morris, C. Johnston, and P. Griffiths Automated assay for plasma D-lactate by enzymatic spectrophotometric analysis with sample blank correction Ann Clin Biochem, March 1, 2008; 45(2): 177 - 183. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, sample blank.
