| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology, University of Gent, B-9000 Ghent; Department of Physiology, University of Louvain Medical School, B-1200 Brussels; and Cardiovascular Center of O. L. Vrouw Ziekenhuis, B-9300 Aalst, Belgium
 |
ABSTRACT |
|---|
 TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
The aim of this study was 1) to investigate the validity of repeated estimations of blood flow using colored microspheres (CMS) and 2) to develop and validate a method that permits four consecutive estimations in the same animal using nonradiolabeled microspheres (NRMS). Several mixtures of different types of microspheres were injected in dogs, with each mixture containing the radiolabeled microspheres (RMS; labeled with 113Sn) with either three CMS, four CMS, or three CMS and one type of fluorescent (crimson labeled) microsphere (FMS). The blood flows estimated with the use of any of the injected microspheres were compared with those measured using the RMS as the "gold standard." The results were analyzed by 1) regression analysis, 2) variance analysis (ANOVA I), and 3) estimation of the limits of agreement between RMS and NRMS flow rates. The results indicate that simultaneous estimations of blood flow obtained with the use of more than three CMS lack accuracy and reliability. A combination of three types of CMS with crimson-labeled FMS, however, offers the possibility to estimate consecutively four different flow rates in the same animal in an accurate way and with relatively high precision.
radioactive microspheres; fluorimetry; absorptiometry; gamma spectrometry
| |
INTRODUCTION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
COLORED (CMS) as well as fluorescent microspheres (FMS) are used for measuring regional blood distribution in animals (7, 9, 13). The method in which either CMS or FMS are used has been validated (7, 8, 11) indirectly by comparing the results with those obtained via the method in which radioactive microspheres (RMS) are used, which are regarded as a "reference standard" for the measurement of the blood flow in experimental research.
Frequently, repeated blood flow measurements in the same animal (same tissue or organ) must be performed (1, 16). For this purpose, microspheres labeled with different color or fluorescent components are injected, and, thereafter, the components are measured simultaneously. The determination of the amounts of the different components may be disturbed by interference between the different absorption or fluorescent spectra or by (negative or positive) background effects due to incompletely digested substances from blood or tissue.
For several years, we have conducted experiments in dogs in which regional blood flow measurements were performed. In these studies (3, 10), microspheres labeled with 113Sn, 60Co, 86Rb, or 85Sr were used for repeated (usually four) estimations of the blood flow in the same dog under different experimental conditions. The disadvantages (manipulation, radioactive waste, costs) inherent in the use of RMS have incited us to replace them, when possible, with nonradiolabeled microspheres (NRMS). The need for the use of CMS was obvious, because no apparatus for simultaneous fluorimetric measurement of the different fluorescent components currently used was available in our laboratory. Preliminary experiments, however, showed that the required accuracy and reliability were not attained in a number of experiments in which more than three CMS were injected. Because our experimental protocol necessitates four repeated estimations in the same animal, we undertook an analytic study with respect to the validation of a method for obtaining four repeated blood flow estimations using a combination of different CMS with one FMS. This was performed by comparing the results of in vitro as well as in vivo experiments.
| |
MATERIALS AND METHODS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Materials
CMS [diameter (mean ± SD): 15.5 ± 0.2 µm], solubilized in saline + 0.1% Tween 80 + thimerosal and containing 60 million spheres/20 ml suspension, were purchased from Triton Technology. They were labeled with one of the following color components: white (Blancophor) with maximum absorbance at 370 nm; yellow (Resolin Brillantgelb) with maximum absorbance at 780 nm; red (Rouge Terasil E-BST) with maximum absorbance at 530 nm; blue (Resolin Brillantblau) with maximum absorbance at 672 nm; or violet (Resolin Rotviolet) with maximum absorbance at 740 nm.The pure white, yellow, blue, and violet components were from Bayer (Leverkusen, Germany), and the pure red component was from Ciba-Geigy (Zurich, Switzerland). Microspheres labeled with the crimson fluorescent component from Molecular Probes (Eugene, OR) were also used in a series of experiments. The RMS labeled with 113Sn [diameter (mean ± SD): 15 ± 0.5 µm; solubilized in saline + 10% dextran + 0.01% Tween 80; specific activity = 12.5 mCi/g] were from NEN (Boston, MA). The following products or reagents were used for the isolation of the microspheres from tissue or blood: 4 M KOH solution, Tween 80, and N, N-dimethylformamide (DMF) from Aldrich (Milwaukee, WI); and a 22-mm drain disk polyethylene (PE) support filter (SN 1800614) from Nucleopore (Costar, Bodenheim, Germany).
Methods
Digestion procedure for tissue and blood. Seven milliliters of KOH solution (4 M for tissue; 12 M for blood) containing 2% Tween 80 (freshly prepared) were added to a tube (with glass stop) containing either 500-600 mg of tissue or 2 ml of blood; the tube was shaken continuously in a water bath at 70°C for 24 h. After digestion, the hot liquid in the tube was filtered, under light vacuum, through a 25-mm PE membrane (8.0 µm) laying on a PE drain disk filter (22-mm diameter), with both membrane and drain disk supported by a metallic grate of 25 mm in diameter and 0.5 mm in thickness; the filter, membrane, and grate were clamped between two round clamps of stainless steel, and the underclamp was connected to a vacuum pump. The tube and membrane filter were washed three times with 5 ml of 4 M KOH solution containing 2% Tween. This filtration system has been shown to give a 100% recovery of any of the microspheres used.
After filtration, the 25-mm membrane (8.0 µm) containing the microspheres was taken with a tweezer, transferred onto the bottom of a glass tube, and allowed to dry at room temperature (overnight); 0.3 ml DMF was then added, and the tube was vortexed for 10 s. After centrifugation, the supernatant DMF extract was transferred to a small tube (0.5 ml); the tube was closed until measurement was performed by absorptiometry and fluorimetry.Gamma spectrometry. Before the tissue or blood was digested with KOH solution, the tube containing the tissue or blood was put into a plastic tube (inner diameter: 2 cm; height: 20 cm) with a round bottom, and gamma emission was counted (counts/min) for at least 5 min in a Berthold gamma spectrometer with settings corresponding to the total energy of the 113Sn radioisotope. All measurements of radioactivity were performed under the same conditions. Count ranges of tissues were >2,000 counts/min and were corrected for the background count range (40 counts/min).
Absorption spectrophotometry. The absorbance (A) from 325 to 750 nm of the DMF extract was measured with a Beckman (UV/VIS 7000) diode array spectrophotometer (wavelength accuracy: 1.25 nm; wavelength repeatability: 0.05 nm; spectral bandwidth: 2 nm; photometric accuracy: 0.005 absorption units). The software program of the spectrophotometer makes it possible to calculate the concentration (in µg/ml) of each color component in the DMF extract by comparing the measured absorbance with that of an adequate standard curve constructed using the instrument for either one color component or a mixture of three or four color components; the standard curves were calculated by Fourier analysis on the basis of 12 different concentrations of one color component or 12 different mixtures of color components. The number of microspheres per milliliter of DMF extract was calculated by multiplying the concentration of the color component of the extract by the number of microspheres per microgram of color component, which was determined experimentally by measuring the amount (in µg) of color component per milliliter of microsphere suspension for which the number of microspheres per milliliter of suspension was known.
Fluorimetry.
Fluorimetry of the DMF extract was performed immediately after
absorptiometry using a Farrand K2 spectrofluorometer (bandwidth: 3 nm).
DMF extracts (200 µl) were diluted to 1 ml with 2-ethoxyethyl acetate
(Aldrich), and the relative fluorescence at 640 nm was measured using
excitation light of 605 nm. The number of microspheres in the DMF
extract (0.3 ml) was calculated from a standard curve plotting the
relative fluorescence versus the number of microspheres; the number of
microspheres was determined experimentally from a DMF extract of the
original crimson-labeled microsphere suspension for which the number of
microspheres per milliliter of suspension was known. The fluorescence
standard curve was linear up to 5,000 microspheres, and the sensitivity
was 
30 microspheres.
Calculation of blood flow.
Tissue blood flow (F; expressed in
ml · g
1 · min1)
was calculated from the reference blood flow (7.64 ml/min) and the
number of microspheres in tissue and reference blood according to
|
Experimental protocol
In vivo experiments. In this series of experiments blood flow rates in different organs (heart, lung, kidney) were simultaneously measured using different types of CMS and then compared with measurements obtained using RMS, which are accepted as the gold standard. A total of 10 mongrel dogs (weighing between 18 and 24 kg) were anesthetized with pentobarbital sodium (Nembutal; 30 mg/kg iv). A left thoracotomy was performed in the fourth intercostal space. The pericardium was opened and the heart exposed. Catheters were inserted into the descending thoracic aorta and the left atrial appendage. The left circumflex coronary artery was carefully dissected and instrumented with a snare occluder. Simultaneous injections of 3 or 4 NRMS (CMS or CMS + FMS) together with RMS were performed in basal conditions (3 dogs), during maximal coronary hyperemia with dipyridamole (1.12 mg/kg iv; 4 dogs), and during left circumflex coronary artery occlusion (3 dogs). NRMS and RMS were sonicated and vortexed before and after mixing, and the mixture of microspheres was injected into the left atrial appendage. Simultaneously, a blood sample was withdrawn from the descending thoracic aorta at a speed of 7.64 ml/min for 150 s. At the end of the experiment, the dogs were killed with a lethal dose of pentobarbital sodium (50 mg/kg). Tissue samples were taken from the left ventricular free wall, the lungs, and the kidneys. The samples were weighed, placed in tubes, and counted for gamma emission (see METHODS). Thereafter, the samples were processed for digestion, followed by absorptiometry and/or fluorimetry.
Three groups of dogs were distinguished according to the composition of the injected mixture of NRMS with RMS. Group I (6 dogs) comprised white, yellow, and red CMS plus RMS; group II (2 dogs) comprised white, yellow, red, and blue CMS plus RMS; and group III (2 dogs) comprised white, yellow, and red CMS plus crimson FMS and RMS. The mixture for group I contained the following numbers of microspheres: 2 × 106 RMS, 6 × 106 white, 4.5 × 106 yellow, and 5 × 106 red. In experiments with groups II and III, the number of microspheres differed by dog and by type of microsphere as follows: RMS, 2.0 × 106 (all dogs); white, 6.0 × 106 (dogs G, H, K, and L), 4.5 × 106 (dogs I and J), or 5.0 × 106 (dogs M and N); yellow, 4.5 × 106 (dogs G, H, K, and L), 5.5 × 106 (dogs I and J), or 4.8 × 106 (dogs M and N); red, 5.0 × 106 (dogs G, H, K, and L), 5.6 × 106 (dogs I and J), or 5.8 × 106 (dogs M and N); blue, 5.0 × 106 (dogs G and H), 4.0 × 106 (dogs I and J); and FMS, 2.0 × 106 (dogs K and L) or 3.0 × 106 (dogs M and N).In vitro experiments. This series of experiments was performed to examine the accuracy and precision of the analytic technique (isolation procedure and measurement of color or fluorescent components). First, each type of microsphere was subjected to the isolation procedure, either individually or as a mixture of different microspheres (either white, yellow, and red CMS + crimson FMS or white, yellow, red, and blue CMS), and extracted with 0.3 ml DMF; the recovery value for each type of microsphere was calculated by comparing the absorbance or fluorescence of the extract with those of the same number of microspheres directly extracted with 0.3 ml DMF and measured either individually or as a mixture.
Second, to examine whether absorptiometry or fluorimetry is disturbed by interfering substances from tissue, different samples of noncontaminated fresh heart tissue (n = 10, ranging from 1,155 to 1,235 mg) were dissolved, extracted with 0.6 ml DMF, and divided in two parts. Each part was added to a reference mixture of either white, yellow, and red CMS and crimson FMS or white, yellow, red, and blue CMS; their absorbance and fluorescence values were compared with those of the pure reference mixture. Third, to obtain an idea about the precision of a simultaneous measurement of the flow rate in a dog by using microspheres labeled with white, yellow, red, and crimson components, duplicate samples (N = 30 samples) were prepared containing different amounts (from 578 to 612 mg) of fresh heart tissue and a mixture of a variable number of (white, yellow, red, and crimson) microspheres corresponding with flow rates between 0.85 and 4.65 ml · g1 · min1.
The series of duplicate samples was digested and extracted as described
in Digestion procedure for tissue and
blood. The flow rates were calculated using the blood
reference standard of dog I.
All experimental procedures were carried out in accordance with the
National Institutes of Health Guide for the Care and
Use of Laboratory Animals.
Statistical Analysis
To establish whether flow rates estimated using CMS or FMS agreed with those measured using RMS, the results were analyzed with the use of three different types of statistical analysis: regression analysis, variance analysis, and estimation of the precision of the limits of agreement.Regression analysis.
The curve that best fits the points when the CMS or FMS flow rates
(y-axis) of the tissues of all dogs of the in
vivo experiments (groups I-III) are
plotted versus the RMS flow rates
(x-axis) was calculated by linear
regression analysis. The strength of the relation between both
variables was expressed by the correlation coefficient
(R). The 95% confidence interval
(CI) of each regression curve was calculated, and the slope of the
regression curve, or regression coefficient
(b), was compared with unity (see
Table 1).
|
Variance analysis. The flow rates of all analyzed tissues of each dog (groups I-III) measured using the same type of microsphere were considered as a set of data. Differences between the means of the different sets (different types of microspheres) were analyzed by the method of repeated measures of variance (ANOVA I, with Dunnett's multiple-comparison test), with the flow rates of each tissue item measured using the different types of microspheres being linked.
Estimation of precision of agreement.
The precision of the agreement between the RMS flow rates (as
reference) and those measured using each of the CMS (i.e., white, yellow, red, and blue) or the FMS was calculated according to the
method of Bland and Altman (2). Therefore, the RMS flow rates of each
type (group I,
II, or
III) of the in vivo experiments were
divided into classes (see Tables 5-7) and the following parameters calculated: 1) the differences (D)
between RMS and either CMS (white, yellow, red, or blue) or FMS flow
rates; 2) the mean difference (
) and standard deviation (s) of each flow
rate class; 3) the limits of
agreement ( 
2s and
+ 2s);
4) the 95% CI extreme lower (LL)
and upper precision limits (UL) of 2s and + 2s, respectively; and
5) the
values as well as the LL and the UL values of each class expressed as
percentages of the RMS flow rate mean (
) of
the class (i.e.,
10
/,
100LL/, and
100UL/, respectively). With respect to the
in vitro experiments, an estimate of the standard deviation was
calculated from the differences (d) between the two results of the
duplicates using the formula s = 
, and the
CI was calculated using the formula CI = 2.042s/
(
= 0.05).
| |
RESULTS |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
Reciprocal Interference of Color and Fluorescent Components
The reciprocal interference of the absorption spectra of the color components was derived from measurements of the absorption spectra of equal concentrations of the components in DMF. The results (Table 1) show 1) no spectral overlap of white or yellow with red, violet, or blue component; 2) an important overlap of red with yellow component; 3) a very high overlap of violet with red component; 4) an important overlap of blue with white component; and 5) a yet higher overlap of a combination of blue and red or of blue and yellow with the white component.Fluorescence measurements have shown that the relative fluorescence (at 640 nm, using excitation light of 605 nm) of amounts of crimson component, corresponding with 400 crimson-labeled microspheres, is not disturbed by comparable amounts of the color components used.
In Vivo Experiments
Regression analysis.
Table 2 summarizes the parameters of the
regression curves between CMS (or FMS) flow rates
(y-axis) and RMS flow rates
(x-axis) measured in all analyzed
tissues of each group (groups I-III) in
the in vivo experiments. When white, yellow, and red color components
with or without the crimson fluorescent component (groups I and III) are
estimated simultaneously, a strong linear relationship is seen between
CMS or FMS flow rates and RMS flow rates. Table 2 shows that
1) the correlation coefficients
R, being >0.980 (a somewhat lower
R = 0.974 was found for the curve for
red CMS), express a high correlation between any of the CMS or FMS flow rates and the RMS flow rates, and 2)
the regression coefficients b (between
0.960 and 0.982; a somewhat lower value was found for the curve for red
CMS of group III) are close to 1 and
have narrow 95% CI that include 1, indicating that, in the mean, the
change of y equals that of
x.
|
Variance analysis.
The flow rates measured by a specific type of microsphere in the
tissues of a given dog were considered as a set of data (see MATERIALS AND METHODS). The mean
values and standard deviations of these sets of flow rates are
represented in Tables 3 and
4 for each dog and each group
(I-III) of the in vivo experiments. Statistical
analysis performed to assess a possible difference between the means of
these sets, using ANOVA I with repeated measures of variance, indicates
first that when dogs were treated with three (white, yellow, and red)
CMS (group I), only small or nonsignificant differences were seen between the mean flow rates measured by CMS and
RMS (Table 3). Indeed, the mean flow rates of dogs
A-F as measured by white CMS do not differ
significantly (P > 0.05) from the
RMS values. The same holds for the mean flow rates measured by yellow
CMS, with the exception of dog A,
because its mean flow rate was 9% lower
(P < 0.01) than the mean RMS flow
rate. The flow rate means for red CMS are also in good agreement with
the RMS values, with the exception of dogs
A and F; the flow rate means for red CMS for these dogs are ~10% (dog
A) or 9% (dog F) lower (P < 0.05) than the RMS flow
rate mean.
|
|
Estimation of precision of agreement. The degree of the precision of the agreement between RMS flow rates on the one hand and those measured by each of the CMS or the FMS on the other hand was calculated for the in vivo experiments (see Statistical Analysis).
With regard to the experiments with three CMS (group I), the results (Table 5) show that 1) the mean differences (), as such or expressed as percentages of
the mean RMS flow rates of the classes
(10/), are
almost all positive and low, indicating that somewhat lower flow rates
were measured with white (<2.8%), yellow (<5.4%), and red CMS
(<8.3%) than with RMS, and 2) the values of 100LL/ and
100UL/ vary from 22.4 to +23.4 for
white CMS, from 22.0 to +23.8 for yellow CMS, and from
20.0 to +23.7 for red CMS. Thus, when the blood flow in a dog is
measured simultaneously using white, yellow, and red CMS, the mean
percentage error of any of the CMS flow rates with respect to the RMS
value is low. The 100LL/ and
100UL/ values, however, indicate that these
mean percentage errors may vary between about 22 and +22%.
|
/ values are
also low and negative (for white and yellow CMS) or positive (for red CMS and crimson FMS), indicating that, in the mean, with white and
yellow CMS a somewhat higher blood flow, and with red CMS and FMS a
somewhat lower blood flow, than with RMS may be anticipated. The ranges
of these relative percentage errors, as expressed by the
100LL/ and 100UL/
values, are comparable with those seen for the dogs of
group I.
|
and
10/ values were
found (Table 7) than for the dogs of
group I or
III; the values are negative for white
CMS and positive for yellow, red, and blue CMS, indicating that higher
flow rates are obtained with white CMS (mean: 8.7%) and lower
flow rates are obtained with yellow (mean: 7.7%), red (mean: 19.3%),
and blue CMS (mean: 9.6%) than with RMS. Moreover, the
10/ values vary
within a considerably larger range as indicated by the extreme
100LL/ and 100UL/
limits, with the latter values lying between 81.2 and +54.5 for
white CMS, between 29.8 and 41.6 for yellow CMS, between
22.8 and 62.4 for red CMS, and between 40.4 and +58.4 for
blue CMS, respectively. The higher range of the differences between RMS
and CMS flow rates in the experiments of group
II compared with those of group
III is also illustrated in Figs.
1 and
2.
|
|
|
In Vitro Experiments
The purposes of the in vitro experiments were 1) to obtain an idea about the accuracy and precision of the analytic procedure in the absence of tissue and 2) to examine whether the absorptiometry or fluorimetry was disturbed due to components from incompletely digested tissue. The results with respect to the accuracy and precision (Table 8) show that microspheres, when treated separately through the analytic procedure, were almost completely recovered (~98.3-99.3%) with a good precision, with the coefficients of variation (CV) lying between 1.6 and 2.4%.When a mixture of white, yellow, and red CMS and crimson FMS was treated in the same way, comparable recoveries (between 97.8 and 98%) and CV values (between 1.9 and 2.9%) were found. For a mixture of white, yellow, red, and blue CMS, somewhat lower recovery (between 95.8 and 97.1%) with higher CV values (between 3.6 and 4.7%) were found.
The results of the experiments concerning the suitability of the
digestion procedure (Table 8, labeling
component data) show that 1) the
absorbance or fluorescence of white, yellow, red, and crimson labeling
components, when measured as a mixture, is recovered almost completely
(97.6-100.6%) and with high precision (CV: between 1.1 and 2.4%)
in the presence of tissue extract; and
2) somewhat lower recoveries
(92.1-96.2%) with lower precision (CV: between 3.8 and 5.7%) are
found when the four color components (white, yellow, red, and blue)
were measured in the same conditions.
|
Finally, from the series (N = 30) of
duplicate determinations of white, yellow, and red CMS and crimson FMS
added to heart tissues (see MATERIALS AND
METHODS), the mean flow rates, standard deviations of
the duplicates, and confidence intervals (CI = t · s/
,
using t statistic with = 0.05) for
one determination were calculated for each type of
microsphere. The values (M ± s) were as follows: 1.29 ± 0.0911 for white CMS, 3.41 ± 0.264 for yellow CMS, 1.45 ± 0.101 for red CMS, and 1.87 ± 0.088 for crimson FMS. Thus, when blood flow is estimated repeatedly in a dog with the use of these four
types of microspheres, there is a 95% chance that any true flow rate X
(in
ml · g1 · min1)
will lie between X ± t · s/, i.e., X ± 0.186 for white CMS, X ± 0.538 for yellow CSM, X ± 0.204 for red CMS, and X ± 0.176 for crimson FMS. When the
confidence intervals are expressed as percentages of the flow rate
means of the series of duplicate determinations, the following
percentage errors for one determination were found: 14.4% for white
CMS, 12.7% for yellow CMS, 14.0% for red CMS, and 9.4% for crimson FMS.
| |
DISCUSSION |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
It was the aim of this study to examine the possibility of using nonradioactive (colored and/or fluorescent) microspheres for repeated estimations of blood flow in the same animal. This requires an accurate and reliable determination of several color (and/or fluorescent) components in the same tissue. The validation of the method using NRMS was approached in two ways: first, in a comparative way, i.e., by comparing the results found using NRMS with those obtained using RMS as a "gold standard," and second, in an analytic way by checking the criteria of accuracy and precision of the first method. Because blood flow to any organ may vary with time, it was considered essential to inject the different microsphere types at the same time to assess possible differences between the flow rates. To ensure that the analyzed tissues should contain at least 200 microspheres of each type, between 3 × 106 and 6 × 106 microspheres (according to type) were injected. In this way, an adequate spectrophotometric or fluorimetric estimation was made possible (8) and the error due to statistical variation (4-6, 8, 12, 14, 15) was reduced to a minimum. Because several types of microspheres were injected simultaneously, the total number of injected microspheres was rather high but was lower than the number needed to result in permanent hemodynamic alterations (1). Moreover, it is justifiable to assume that the different microsphere types are distributed to the organs in proportion to their concentration of the injected mixture so that the flow rates measured by them should be equal for each tissue.
The simultaneous injection of different types of microspheres implies that, in case of a complete agreement between the estimated flow rates, the proportion between the number of different microspheres in any analyzed tissue should be the same as in the injected mixture. However, the proportion between the numbers of different microspheres is not a determining factor for concluding whether there is agreement or disagreement between flow rates. Nevertheless, we found it suitable to compare the flow rates in experiments for groups II and III (four NRMS) on the basis of two different mixtures (see MATERIALS AND METHODS).
The results of the comparative study (in vivo experiments) were analyzed statistically in three ways to conclude whether there was agreement between RMS and CMS or FMS flow rates. Regression analysis shows a very high correlation between these flow rates and each type (groups I-III) of experiment. Correlation coefficients, however, although highly significant (P < 0.01), only express an extent of association between pairs of x,y variables and do not permit us to conclude whether there is equality between a pair of x,y variables. The estimated derivative of a regression curve or regression coefficient (b), expressing the change of y with x, provides more information about the agreement between two variables having the same dimension; b will indeed be 1 when the two variables of each x,y pair are equal.
The 95% CI of the curves of the experiments with three CMS (group I) or with three CMS and one FMS (group III) include or nearly include 1, indicating that the flow rates measured by any of them do not (or nearly not) differ from those measured by RMS. On the contrary, the b coefficients of the curves corresponding to the experiments with four CMS (group II) are considerably <1, except for the curve for white CSM, and their 95% CI do not include 1. Thus regression analysis reveals a lack of agreement between the blood flow rates measured by yellow, red, and blue CMS and those measured by RMS when repeated blood flow measurements in the same animal are performed using white, yellow, red, and blue CMS.
The method of repeated measurements of analysis of variance, with the number of analyzed tissues taken as replicates of blood flow estimation in different experimental conditions (different types of microspheres), leads to an analogous conclusion. In the experiments in which three types (white, yellow, and red) of CMS were used with or without FMS, the differences between the mean flow rates measured by RMS and NRMS are either small or not significant, respectively. In the experiments in which four (white, yellow, red, and blue) CMS were used, however, very important and statistically significant differences are seen between RMS and CMS flow rates.
When the differences between the flow rates measured by RMS and CMS or FMS were analyzed in a more direct way, as proposed by Bland and Altman (2), considerably less agreement was found when blue CMS, in addition to the three other CMS, were measured. More importantly, larger 95% CI precision limits of the differences were found under these conditions, indicating a high degree of unreliability of the CMS flow rates. On the contrary, the agreement between RMS and CMS or FMS flow rates and the reliability were much better when blue CMS were omitted or replaced by FMS.
The lack of reliability of blood flow estimations for which a combination of white, yellow, red, and blue CMS were used (group II) is not likely due to technical shortcomings: tissues of all experiments were analyzed by the same person in the same way and according to a rigorous analytic procedure. The in vitro experiments (see Table 8), however, indicated that the addition of a mixture of colored components to tissue extract immediately before spectrophotometry yielded lower mean recovery values and a larger range of individual recovery values when the mixture contained blue component. A still lower precision may thus be expected for the blood flow estimations of in vivo group II, because each estimated value incorporates the sum of errors of the different steps of the analytic procedure. Moreover, the lack of reliability of estimates for group II may also be due to the fact that the blue absorption spectrum, having a maximum peak at 672 nm, shows a slight, yet increasing, absorption in the wavelength zone from 500 to 325 nm, i.e., in the region of the absorption peaks of white, yellow, and red components. Indeed, this overlapping zone increases the risks for 1) interference between the absorption spectra of the different color components and 2) disturbances from aspecific components of digested tissue, both negatively affecting the precision of the flow estimations.
In conclusion, the results of our study demonstrate that 1) three repeated blood flow measurements in the same dog with the use of three CMS (white, yellow, and red) can be performed in an accurate way and with the necessary precision; 2) a quadruple estimation of the blood flow in the same dog with the use of four CMS (white, yellow, red, and blue) leads to unreliable results that are difficult to interpret, especially in those studies in which changes (increases or decreases) of <25% of the blood flow are to be measured; and 3) four repeated estimations of the blood flow in the same animal with a high degree of precision and accuracy are made possible only with the use of the three CMS labeled with white, yellow, and red components, respectively, in combination with FMS labeled with the crimson component.
| |
ACKNOWLEDGEMENTS |
|---|
We express gratitude to E. Steenhoudt for experienced help during the analytic work. We also acknowledge the generous gifts of the color components Blancophor, Resolin Brillantgelb, Resolin Rotviolet, and Resolin Brillantblau from the Bayer Company, Leverkusen, Germany, and of Rouge Terasil E-BST from Ciba-Geigy, Zurich, Switzerland. Finally, we acknowledge the much appreciated secretarial assistance of J. Cano in the preparation of this manuscript.
| |
FOOTNOTES |
|---|
This work was supported in part by grant no. 9.0034.91 from the Nationaal Fonds voor Wetenschappelijk Onderzoek and by grant no. 3.4572.96 from the Fonds National de la Recherche Scientifique Medicale.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact.
Address for reprint requests and other correspondence: G. R. Heyndrickx, Cardiovascular Center, OLV Ziekenhuis, Moorselbaan 164, B-9300 Aalst, Belgium (E-mail: guy.heyndrickx{at}olvz-aalst.be).
Received 12 March 1998; accepted in final form 17 November 1998.
| |
REFERENCES |
|---|
|
TOP
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
REFERENCES
|
|---|
1.
Baer, R. W.,
B. D. Payne,
E. D. Verrier,
G. J. Vlahakes,
P. N. Molodowitch,
P. N. Uhlig,
and
J. I. E. Hoffman.
Increased number of myocardial blood flow measurements with radionuclide-labeled microspheres.
Am. J. Physiol.
246 (Heart Circ. Physiol. 15):
H418-H434,
1984.
2.
Bland, J. M.,
and
D. G. Altman.
Statistical methods for assessing agreement between two methods of clinical measurement.
Lancet
8:
307-311,
1986.
3.
Bol, A.,
J. A. Melin,
J. L. Vanoverschelde,
T. Baudhuin,
D. Vogelaers,
M. De Pauw,
C. Michel,
A. Luxen,
D. Labar,
M. Cogneau,
A. Robert,
G. R. Heyndrickx,
and
W. Wijns.
Direct comparison of [13N]ammonia and [15O]water estimates of perfusion with quantification of regional myocardial blood flow by microspheres.
Circulation
87:
512-525,
1993
4.
Buckberg, G. D.,
J. C. Luck,
D. B. Payne,
J. I. E. Hoffman,
J. P. Archie,
and
D. E. Fixler.
Some sources of error in measuring regional blood flow with radioactive microspheres.
J. Appl. Physiol.
31:
598-604,
1971
5.
Dole, W. P.,
D. L. Jackson,
J. I. Rosenblatt,
and
W. L. Thompson.
Relative error and variability in blood flow measurements with radiolabeled microspheres.
Am. J. Physiol.
243 (Heart Circ. Physiol. 12):
H371-H378,
1982.
6.
Flameng, W.,
B. Winkler,
B. Wusten,
and
W. Schaper.
Minimum requirements for the measurements of regional myocardial flow using tracer microspheres.
Bibl. Anat.
15:
24-29,
1977.
7.
Glenny, R. W.,
S. Bernard,
and
M. Brinkley.
Validation of fluorescent-labeled microspheres for measurement of regional organ perfusion.
J. Appl. Physiol.
74:
2585-2597,
1993
8.
Hakkinen, J. P.,
M. W. Miller,
A. H. Smith,
and
D. R. Knight.
Measurement of organ blood flow with colored microspheres in the rat.
Cardiovasc. Res.
29:
74-79,
1995[Medline].
9.
Hale, S. L.,
K. J. Alker,
and
R. A. Kloner.
Evaluation of nonradioactive, colored microspheres for the measurement of regional myocardial blood flow in dogs.
Circulation
78:
428-434,
1988
10.
Heyndrickx, G. R.,
W. Wijns,
D. Vogelaers,
Y. Degrieck,
L. Vandeplassche,
and
J. A. Melin.
Recovery of regional contractile function and oxidative metabolism in stunned myocardium induced by 1-hour circumflex coronary artery stenosis in chronically instrumented dogs.
Circ. Res.
72:
901-913,
1993
11.
Kowallik, P.,
R. Schulz,
B. D. Guth,
A. Schade,
W. Paffhausen,
R. Gross,
and
G. Heusch.
Measurement of regional myocardial blood flow with multiple colored microspheres.
Circulation
83:
974-982,
1991
12.
Nose, Y.,
T. Nakamura,
and
M. Nakamura.
The microsphere method facilitates statistical assessment of regional blood flow.
Basic Res. Cardiol.
80:
417-429,
1985[Medline].
13.
Prinzen, F. W.
Gradients in myocardial blood flow, metabolism and mechanics across the ischemic left ventricle wall (PhD thesis). Maastricht, The Netherlands: University of Limburg, 1982.
14.
Prinzen, F. W.,
and
R. W. Glenny.
Developments in non-radioactive microsphere techniques for blood flow measurement.
Cardiovasc. Res.
28:
1467-1475,
1994[Medline].
15.
Reneman, R. S.,
A. H. M. Jageneau,
W. Van Gerven,
J. Dony,
and
P. Beirnaert.
The radioactive microsphere method for the assessment of regional blood flow after coronary artery occlusion.
Pflügers Arch.
353:
337-347,
1975[Medline].
16.
Wieland, W.,
P. F. Wouters,
H. Van Aken,
and
W. Flameng.
Measurement of organ blood flow with colored microspheres: a first time-saving improvement using automated spectrophotometry.
In: Computers in Cardiology 1993: Proceedings September 5-8, 1993, London, UK. Los Alamitos, CA: IEEE Comput. Soc., 1993, p. 691-694.
This article has been cited by other articles:
|
|
 |
|
  J. G. Abraldes, Y. Iwakiri, M. Loureiro-Silva, O. Haq, W. C. Sessa, and R. J. Groszmann Mild increases in portal pressure upregulate vascular endothelial growth factor and endothelial nitric oxide synthase in the intestinal microcirculatory bed, leading to a hyperdynamic state Am J Physiol Gastrointest Liver Physiol, May 1, 2006; 290(5): G980 - G987. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. Ahmet, Y. Sawa, T. Yamaguchi, and H. Matsuda Gene transfer of hepatocyte growth factor improves angiogenesis and function of chronic ischemic myocardium in canine heart Ann. Thorac. Surg., April 1, 2003; 75(4): 1283 - 1287. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. L. Collier, Y. Hwang, R. Ramasamy, R. R. Sciacca, K. T. Hickey, N. R. Simpson, and S. R. Bergmann Synthesis and Initial Evaluation of 17-11C-Heptadecanoic Acid for Measurement of Myocardial Fatty Acid Metabolism J. Nucl. Med., December 1, 2002; 43(12): 1707 - 1714. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. A. Berg, A. B. Sanders, K. B. Kern, R. W. Hilwig, J. W. Heidenreich, M. E. Porter, and G. A. Ewy Adverse Hemodynamic Effects of Interrupting Chest Compressions for Rescue Breathing During Cardiopulmonary Resuscitation for Ventricular Fibrillation Cardiac Arrest Circulation, November 13, 2001; 104(20): 2465 - 2470. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
