Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Furosemide Concentrations in Serum and Urine, and Its Binding by zez16524


									CLIN.   CHEM.      20/2,    152-158      (1974)

Furosemide Concentrations in Serum and Urine, and Its Binding
by Serum Proteins as Measured Fluorometrically

A. W. Forrey, B. Kimpel, A. D. Blair, and R. E. Cutler1

We describe a modification of the fluorometric meth-                                         method,         not only for convenience,                   but also to mini-
od of H#{228}ussler          [Arzneim. Forsch. 14, 704
                 and Hajd#{252}                                                              mize systematic      errors.  We have also studied    the
and 709 (1964)]           for assay of furosemide                  in either                 binding   of the drug to serum proteins,  in order to in-
serum or urine. A 1-mi sample,                   acidified     to pH 2, is                   terpret its pharmacokinetics.
extracted      with 5 ml of diethyl ether; 4 ml of the ether
is back-extracted          into 1 ml of phosphate               buffer (pH                   Materials and Methods
7.0, 0.1 mol/liter),         and finally acidified           with 1 ml of
                                                                                                35S-labeled     and chemically     pure unlabeled                                furo-
dilute    HCI (0.6 mcI/liter).          A procedure        for estimating
                                                                                             semide    were obtained     by courtesy   of Hoechst                               Phar-
blanks     in urine was derived               to correct       for dilution
caused      by diuresis.      Internal     standards       are used, and                     maceuticals    Inc., Somerville,  N.J. 08876.
the “effective”       extraction       ratio is used to correct             for
the effects      of quenching          and extraction         differences.
 In equilibrium,       93% of the drug is bound to serum                                       Absorption  spectra    were recorded     with a “Spec-
proteins;     65% is tightly bound. Erythrocytes                     contain                 tronic 505”  spectrophotometer       (Bausch    & Lomb,
less than 5% of the drug. Quantum                       yield of fluores-                    Rochester, N. Y. 14603) in 0.1 mol/liter          sodium
cence      at pH 1 is 0.0496                for furosemide            and is                 phosphate            and      acetate      buffers   at selected           pH’s.        Un-
0.0163       for     4-chloro-5-sulfamoylanthr#{224}nilic               acid.                corrected      measurements          of fluorescence   excitation
Furosemide        fluorescence         diminishes       with increasing                      and emission        spectra       were made      in a Model      430
pH, while that of 4-chloro-5-sulfamoylanthranilic                                            spectrofluorometer          fitted    with a scanning      attach-
acid (a degradation          product)      increases.                                        ment      (G. K. Turner          Associates,    Palo Alto, Calif.
                                                                                             94303).     For this, samples           were prepared    in these
Additional       Keyphrases: pharmaco                     kinetics   #{149}drug bind-
ing by serum proteins                   colorimetry
                                   #{149}                     compared                       same buffers,but 10-fold more dilute than the con-
                                                                                             centrations          used      for absorption        spectra.

   Hajd#{252} H#{228}ussler and H#{228}ussler Hajd#{252}
             and                  (1)                     and                                         of furosemide
                                                                                                Quantum              yields(4-chloro-N-furyl-
(2) reported       a fluorometric        assay for furosemide         (a methyl-5-sulfamylanthranilic               acid) and CSA2 emis-
nonmercurial        diuretic    of the sulfonamide         group)   for
                                                                         sion were measured           in the same instrument,           with use
                                                                         of quinine      bisulfate     standards,      by the procedure          of
use in measuring             concentrations         of the drug       in
serum. We find that this method is subject to several                    Parker and Rees (3). The samples,                 in buffers of vari-
                                                                         ous pH, and the quinine                standard,      in H2S04       (0.1
sources of error not previously             noted and have there-
fore modified       the technique.        In earlier studies of the      mol/liter)     at the same absorbance               (within     0.05) as
urinary    excretion      of furosemide,                  and Hajd#{252} the drug at the exciting
                                               H#{228}ussler                                                wavelength,       were excited at
(2) used the colorimetric             determination        of urinary    345 nm and the emission              spectra     were recorded.       Ef-
furosemide       concentrations        by employing         the Brat-    fective quantum          yields were calculated          from the rel-
ton-Marshall        reaction     for the detection       of aromatic     ative areas of the intensity         vs. frequency      spectra.
amines. We have also applied the fluorometric          meth-                                 Assay
od to urine samples        because a direct comparison       of
the concentrations      in serum and urine is required      for                                 The method                of    H#{228}ussler            (2)
                                                                                                                                             and Hajd#{252} was modi-
data on plasma       and renal clearances   of the drug; it                                  fied as follows:               1   ml of serum or urine was acidified
seemed    preferable      to use the same      fluorometric                                  with about 50               sl     of concentrated    HC1 and extracted
                                                                                             for 2 mm with                  5    ml of diethyl   ether in a 16 X 150
   Department       of Medicine,        University        of Washington         at Harbor-   mm screw-capped                     culture tube by shaking rapidly on
view Medical      Center,    Seattle,    Wash. 98104.
   1 Address     reprint    requests      to: R. E. Cutler,             M.D.,     325-9th
Ave., Seattle,    Wash. 98104.                                                                  2   Nonstandard          abbreviation     used:   CSA,       4-chloro-5-sulfamoyl-
   Received    July 30, 1973; accepted            Sept.    10, 1973.                         anthranilic     acid.

152     CLINICAL      CHEMISTRY,          Vol. 20, No.2.         1974
a wrist-action      shaker, and then centrifuged.     A 4-ml                     to predict         the concentration                from the least-squares
aliquot of the ether phase was removed and re-ex-                                fit to the straight line of the unextracted aqueous
tracted for the same time interval with 1 ml of 0.1                              standards.     These calculations    and further   computa-
mol/liter     potassium     phosphate    buffer, pH 7.0, fol-                    tions, which express the final results in terms of the
lowed by centrifugation,        and the ether was then re-                       percentage     of injected   dose present   in each liter of
moved. One milliliter of 0.5 mol/liter          HC1 was then                     serum,     were done on a Burroughs-5500           computer
added and the fluorescence            read in a Turner 430                       using a program written in the ALGOL               program-
spectrofluorometer       set at 345 nm excitation    and 417                     ming language.
nm emission, with 60-nm bandpass.
   A control serum was obtained        immediately      before                   Binding
furosemide   administration.      Normal     sera containing                        We used the method   of Hummel     and Dreyer (4) to
1, 5, and 10 mg of furosemide      per liter, internal    stan-                  measure   equilibrium binding  by collecting    samples
dards, and aqueous      standards   of 0.5, 1, 2, 3, 5, 7, 10,                   from a 0.9 X 35 cm column      containing    “Sephadex
and 15 mg/liter   were taken through        the same proce-                      G-50”        (Pharmacia,         Piscataway,                  N. J. 08854), which
dure. The working standards        were prepared     from the                    was eluted at 12 ml/h with a special “binding                 buff-
same lot as the drug administered.            These were, in                     er” having      the same ionic concentration           as serum
turn, standardized against a pure furosemide stan-                               (NaHCO3,        16; NaCl,    138; NaH2PO4,          4; KC1, 4;
dard    solution.        This    procedure          provides   a control    of   CaCI2, 2.5; and MgCl2, 1.5 mmol/liter),             and adjust-
the amount     administered,                   because     furosemide      de-   ed to pH 7.40 with HC1. The column was equilibrat-
grades slightly to a relatively                 nonfluorescent      product,     ed with buffer plus the desired concentration             of furo-
CSA, during storage,             injection,        and analysis.                 semide and any potential        competitors     such as salicy-
   Urine controls     were prepared       from a timed basal                     lates. A 0.5-ml aliquot of serum was carefully            layered
urine sample obtained       from the study subject          imme-                at the top of the resin bed, under the buffer, with a
diately   before   furosemide       was administered.          This              15-cm     16-gauge    needle   and a 2-ml syringe.           Forty
pre-drug   urine was diluted       in proportion     to the ratio                2-ml fractions     were collected     and analyzed      in one of
between the pre-drug and post-drug urine flow rate.                              two ways: (a) chemically,       either by the direct acidifi-
To demonstrate      the validity     of this correction      proce-              cation of 1-ml aliquots      of the nonprotein       containing
dure, we compared        a blank curve of serial dilutions                       fractions    or by ether extraction       of the protein-con.
of the pre-drug      urine    in one normal        subject     with              taming         fractions        and      several         adjacent       nonprotein
timed samples obtained during an ethacrynic acid-                                containing     fractions;    or (b) radiochemically,       when
induced diuresis. Ethacrynic acid was chosen                      be-            35S-labeled     furosemide     was used, by pipetting       1 ml
cause it closely     mimics     the diuretic     response       seen             into a 20-ml vial for liquid scintillation        counting    as
with furosemide,     but has no measurable          fluorescence                 noted below. The net area under the peak, expressed
at concentrations      as high as 100 mg/liter.         The rela-                in milligrams,      and that contained    in the trough were
tionship   between    the measured       blank value and the                     calculated and the furosemide bound to a literof
urine flow rate is described     by the equation:                                serum protein       was calculated   from the average area.
                                                                                 The percentage       of drug bound per liter of serum is:
            B,      =
                                       yr      E                                                      Binding(%)                                100
                                                                                                                             \      L’.-’eql

where B = blank urine sample,                   i; B0   = blank       of pre-
drug urine sample; F0 = flow rate of pre-drug                          urine;    where    A = mg of drug bound per liter of serum, and
V = volume of urine sample i; T1 = collection                            time    Ceq      equilibrium
                                                                                          =                 concentration     of unbound       drug, in
or urine sample            i, in minutes;         B1     =    blank urine        milligrams     per liter.
sample      immediately          preceding       sample      i; and      E =         An identical      unequilibrated      separation      on the col-
“effective”      extraction      ratio.                                          umn was also conducted             with a 0.5-ml serum sample
   This estimated           blank3 value is a ratio of the drug-                 to which furosemide           was added to give a total serum
induced urine flow rate to the pre-drug flow rate,                               concentration       identical    to that of the free drug in the
with an adjustment              made for a urinary tract (renal                  binding     buffer of the equilibration           column     (10 mg/
pelves, ureters)       hold-up volume of about 35 ml.                            liter). The samples          from this column         were analyzed
   We controlled          the “effective”         extraction       ratio by      as above. Binding           of furosemide     by serum proteins
adding      internal      standards       to both the serum                and   was also measured          with “Centriflo”      filter cotes (Am-
urine samples,         and compared          the properly         corrected      icon,        Lexington,         Mass.           01273),         which    had     been
net signal caused by the internal                  standard      to that of      shown to give a constant           ratio of furosemide        in the
the equivalent          unextracted        acidified      aqueous       stan-              to
                                                                                 filtrate that in the nonfiltrable             material    upon re-
dard.     The sample           extraction      ratio was calculated              peated     trials,  with use of furosemide           dissolved     in
and then used in conjunction                       with the sample     values    binding     buffer. Binding    to erythrocytes       was estimat-
                                                                                 ed on aliquots      of heparinized       whole blood to which
   3This is an estimated     value based on a post-micturition volume            [35S]furosemide      at a concentration       of 10 and 15 mg/
of residual    urine in upper and lower tracts of 20-50 ml. Residual
volumes     may be even larger during diuresis   when urine flow rates           liter of whole blood had been added and equilibrated
are high.                                                                        for 15 mm. The packed           erthrocytes    were isolated      by

                                                                                                      CLINICAL         CHEMISTRY,          Vol. 20, No. 2, 1974       153
centrifugation       and the plasma       fractions    saved for
analysis.     The cells were then twice re-suspended              in
two plasma       volumes     of isotonic saline and re-centri-                                            S
fuged. The supernatant          solutions  were decanted        and                                                                                furosem,de

saved, and the cells were finally resuspended                in one
plasma      volume of saline. Furosemide         concentrations                                           F

were determined         in all fractions,  both by fluoromet-
nc       analysis     and      by     liquid         scintillation       counting        of
                                                                                                          #{188}                                   4chIoro-       5- sulfamoylanthronilic    acid
the serum,    by the assay procedures      described.  The                                                0
chemical   methods   were controlled   by adding internal
standards   to the aliquots   of serum    and wash solu-

Thin-Layer            Chromatography                                                                                            Wavelength         -   nm

       Thin-layer      chromatograms        run on “Q-1” sili-
                                                       were                                   Fig.   1. Emission    spectra   of 4-chloro-5-sulfamoylanthranil-
ca gel plates    (Quantum      Industries,     Fairfield,  N. J.                              ic acid    and furosemide,       excitedat 340 nm in 0.1 mol/
07006),    using   the solvent      system      of Kindt     and                              liter HCI and recorded       immediately    after dilution
                                                                                              The samples are of equal absorbance at the exciting wavelength. The
Schmidt     (5) and that of Diamantstein           and Ehrhart                                quinine spectrum is scaled by a factor of 0.1 with respect to the other
(6). The plates       were examined        under     long wave-                               two spectra and was recorded from a sample in H2S04 (50 mmol/liter)
length   (366 nm) ultraviolet       light and sprayed       with
Bratton-Marshall       reagent   as described     by Kirschner
(7).                                                                                                                     o 4- ddo,o- 5.   $uIf,nOyIOflth,OnhI,c         id
                                                                                                                             A,thonihC   ocd

Liquid        Scintillation         Counting                                                                   05

   One-milliliter serum  samples     were mixed with 1
ml of “Protosol”     (New England      Nuclear,     Boston,                                              b
Mass. 02118), on a vortex-type mixer, afterwhich 15                                                            03
ml of “Aquasol”    (New England Nuclear) was added
and again vortex mixed to give a clear homogenous                                                              02

solution.A seriesof quench standards       were prepared
from 100 &l of aqueous   [35S]furosemide     standard,      to
which         1 ml of water            or 1 ml of serum                 and    various
                                                                                                                         3           4         5       6            7        8        9        I0
concentrations       of erythrocyte       hemolysate      was added
                                                                                                                                          Butter       pH
as quencher.       The samples        were counted       on a Model
3315 liquid      scintillation      spectrometer       (Packard    In-                        Fig.2. pH dependence                        of furosemide   and 4-chloro-5-sul-
                                                                                              famoylanthranilic acid,                    with that of anthranilic acid shown
strument       Co., Inc., Downers          Grove,    Ill. 60515) at                           for comparison
19% gain and window             settings    of 50-1000 and 250-                               The   samples   are of equal absorbance    and are diluted shortly before
1000. The AES channel was set at 2.5% gain 300-                                               reading to diminish acid catalyzed degradation of turosemide
1000. The counts, recorded on punched                    paper tape,
were quench corrected          off-line as described       by Forrey                          change   column      chromatography                                    is used to separate
(8).                                                                                          these compounds         before their                                concentration  is mea-
                                                                                              sured fluorometrically.
Results                                                                                       Assay
Spectra                                                                                          When we first used the assay reported             by H#{228}ussler
   The excitationand emission spectra of furosemide                                                         (2),                         in
                                                                                              and Hajd#{252} we had difficulty re-dissolving                  the
exhibited maxima  at 342 nm and 417 nm, respective-                                           residue after evaporation         of the ether extract.      Back-
ly,      in agreement               with       the      values       of HajthI      and       extraction   of furosemide      into neutral    phosphate     buff-
            (1). The excitation
H#{228}ussler                     and emission  spectra    of                                 er circumvented       this difficulty    in all cases, while si-
CSA in 0.1 mol/liter HC1 is shown in Figure 1. The                                            multaneously       diminishing       the background       fluores-
quantum      yields of furosemide    and CSA at pH 1.0                                        cence in urine samples by 40%.
were 0.0496 and 0.0163, respectively.        The relation-                                       The first factor       of importance       in adopting      this
ship between            pH and fluorescence                      intensity    for both        modification was to assess how reproducible the ex-
furosemide     and CSA is shown in Figure                2. These                             traction             and   back-extraction                            processes               were.   The
values demonstrated          that the fluorescence       emission                             amount     of drug remaining     in the aqueous      phase was
of furosemide     is enhanced     at acid pH’s compared           to                          measured      after ether extraction     of acidified      water,
that of the CSA, which is present as a contaminating                                          serum, and urine samples.       The furosemide       remaining
breakdown     product    in all preparations      of furosemide.                              in the ether phase of these same samples            after back-
Likewise, the relativesensitivity the fluorescence
                                            of                                                extraction     into neutral   phophate      buffer     was also
of these compounds         to changes     in solvent pH is ap-                                measured. Two approaches             were taken:         (a) the
parent,         and     is    especially             important        when     ion-ex-        slopes and intercepts of the fitted straight lines to

154       CLINICAL     CHEMISTRY,          Vol. 20, No. 2, 1974
                                                  Table 1. Characteristics of Furosemide Assay
                                                                 Regression ilnea                                                                                                                               extraction
                                      No.                                              Slope,                                                                                                                   calculated
                                                                                mg/liter/division           Blank reading,b divisions                                        Effective                         from slop.
         Sample type              samples         intercept      mg/liter              X 10’                                                                            extraction     ratio5                     ratios#{176}

      Unextracted                     63          -0.102        ±    0.020        5.89 ±        0.13                                     0                                                                        0.93
       aqueous std.
      Ether-extracted                 63          -0.298        ±   0.022         5.47      ±   0.13                                     0                                0.88 ± 0.01
        aqueous std.                                                                                                                                                  (range 0.69-1.10)
      Ether-extracted                 53          -0.120        ±    0.093        7.60 ± 0.99                               10.00 ± 2.32                                 0.71 ± 0.02                               0.72
        serum and                                                                                                                                                     (range 0.41-1.00)
        internal std.                                                                                            (range                0.02-22.04)
      Ether-extracted                 46          -0.105        ±    0.104        7.75 ± 0.35                                    7.48     ±      1.23                     0.56 ± 0.02                             0.71
        urineand                                                                                                                                                      (range 0.20-0.88)
      Mean   ± SD.
      Mean   ± SEM.
      Calculated     from ratio of slope of unextracted              aqueous      standards       to slope of extracted                                sample.

the      fluorometric          measurements                of the            extracted
standards in water, serum, and urine were compared                                                                                                       diluted p.ediurCic
                                                                                                                                                       #{149}                 uoine
                                                                                                                                                       #{149} dor.nQ ethocrynic oc,d     di,o,.wl
to those of the unextracted     drug in acidified water;                                                                                           -     calculated d,lotiOfl
and (b) the measured        radioactivity   appearing   in                                                            ‘.0

each of the fractions before fluorometry         was com-                                                             0
pared. The back-extraction      process significantly  di-
minished           the variable       blank    values         for urine       samples
but, because furosemide is a diuretic, the fluorescent
impurities  were variously diluted in each urine sam-
ple. Because a true blank value is unmeasurable     dur-
ing furosemide              diuresis, an estimated    value is re-
quired so that            the sample   can be corrected    for such                                                                               Flow Rate Ratio              (Vs    / Vo I
dilution.4  We estimated      the true blank values by
                                                                                                  Fig. 3. Dilution                            effect of ethacrynic   acid-induced                                 diuresis
assaying urine samples obtained at various flow rates                                             on fluorescence                              of urine blanks carried    through                              the entire
during an ethacrynic     acid diuresis (see “Methods”                                             analytic                  proced ure
section).  Dilutions calculated   from the ratio of pre-
drug to post-drug urine flow rates gave a good fit to
the observed data (Figure 3), whereas calculations
made from changes in osmolality        gave a poor fit.
This result  might   have been predicted    from                                     the
known effect of ethracrynic  acid and furosemide                                        on
renal        concentration            and     dilution          mechanisms.              It
should        be noted,       however,         that      this       model       for the
urinary               of furosemide assumes that the
              measurement                                                                                        II
drug has no effect on either the reabsorption or the
secretion of fluorescent                 impurities,        but only on the                                a)

urine flow rate, thus                  producing         only a dilution  of                           \
these substances.    With the aforementioned     correc-                                                   E
tions, our assay results for furosemide extracted from
water, serum,            or urine       samples        are respectively              lin-
ear (Table 1).
      Over     the      calibration         range      0-15         mg/liter,        the
blank     corrected    values for the extracted     aqueous                                                                      10               50                                         150
                                                                                                                                                             Meter      divisions
standards      do not fit the same regression   line as the
unextracted       aqueous standards, as shown by the pa-                                          Fig.4. Standard  curves of extracted                                                   and unextracted                     fu-
                                                                                                  rosemide are shown as a regression                                                     of concentration                    on
rameters           of the fit of these lines (Figure                 4). The ratio                meter reading
                                                                                                  Extracted                 (-    -)    slope,    in mg/liter/div.: 0.0612 ± 0.00077:      intercept                              In
     The blank value is the signal that would be produced by the                                  mg/liter:          0.085   ± 0.260;                standard error ot tlt: 0.135. Unextracted                             (--)
instrument   if it were measuring a sample produced by a compa-                                   slope:         0.0559    ± 0.00047;                    intercept:    0.225     ± 0.114;           standard     error    of fit:
rable urine flow rate in the absence of the diuretic furosemide.                                  0.071

                                                                                                                                        CLINICAL            CHEMISTRY,               Vol. 20, No. 2, 1974                   155
of these slopes does agree reasonably           well, however,                            Likewise,    only small amounts    of drug                                   could be de-
with the “effective”      extraction     ratios for both aque-                            tected    chemically   in the ether phase                                     of extracted
ous and serum samples. Moreover, the radiochemi-                                          aqueous or serum standards     after the back-extrac-
cally measured     partition     coefficient for serum sam-                               tion into neutral buffer, but urine samples   behaved
ples agrees closely with the extraction            ratios mea-                            differently.   Nonetheless, although    most of the drug
sured fluorometrically.                                                                   in samples was back-extracted (as measured radio-
   The differences noted above between the slopes of                                      chemically    and judged    by the absence     of either flu-
linesfitting the aqueous extracted standards and the                                      orescent   spots or Bratton-Marshall      reacting    materi-
internal standards     that  are added  to urine      and                                 als from the ether phase after thin-layer        chromatog-
serum is related   to a value expressed  as “effective”                                   raphy), the measured fluorescence intensity attribut-
extraction ratio. This value is composed, in part, of a                                   able to the drug was lower, particularly   with urine
partition  coefficient of the drug between the ether                                      samples, than for the same                              concentration           of furosem-
phase and the back-extraction     buffer. In urine, this                                  ide in the neutral   buffer                              measured           directly     after
partitionmay vary because  each patient                               will excrete        acidification.
differingconcentrations of the various                               solutes   that             We     thought            that        possibly        the      acid     catalyzed               a
are present             in urine. These substances                  would tend to         transformation    of furosemide      to CSA, producing   a
increase the solubility      of the drug in the ether phase,                              lower fluorescence.    Additionally,    the quenching  ef-
particularly      during the back-extraction,    owing to the                             fect       of either          dissolved          organic          solvent      or peroxide
affinity of the uncharged        but somewhat    polar solutes                            could  be responsible. The possible quenching effect
extracted      into the ether for the furosemide      and the                             of peroxide was studied by warming the neutral buff-
affinity     of the aromatic    compounds     for the organic                             er solutionsat 37 #{176}C dark
                                                                                                               in the           for 20 mm and then
phase.           This partitioning       effect is demonstrated                  when     adding mercaptoethanol,       ascorbic acid, or stannous
1.0 mol/liter  vs. 0.1 mol/liter     of neutral                        phosphate          chloride as reducing agents to counteract        the possi-
buffer  is used to back-extract        aqueous                          standards         ble quenching  effects  of peroxide.    Such measures   did
(Table 2). In this example,       the partition                        coefficient        not      change         the     results,       however,        suggesting            that      per-
as measured       radiochemically       between                           aqueous,        oxides are not involved.  Possibly other extracted                                               in-
serum, and urine standards       is in the direction pre-                                 terfering substances  may be operating as internal                                               fil-
dicted, but it is not as dramatic as the change in the                                    ters and (or) chemical       quenchers.
“effective” extraction ratio measured fluorometrical-                                         Because     all of the major     errors     in furosemide      esti-
ly.While the “effective”extraction ratiobetween ex-                                       mation      in clinical  samples,     particularly       in urine,   are
tracted    internal    standards      and unextracted        aqueous                      reflected in the “effective” extraction ratio, attempts
standards      can vary between         assays, the standard       de-                    were made to control this potential      source of diffi-
viation of the within-run          error ratio remains      constant                      culty by adding internal standards     to selected urine
at about     3%. It was also observed-by             radiochemical                        samples.     Typically,                 the samples    used were early or
counting      and by thin-layer          chromatography        of the                     pre-administration                     samples,   with little or no furo-
concentrated        aqueous      phase     of extracted      aqueous                      semide, respectively,                    in order to avoid errors of mea-
standards            and urine samples-that                little     residual     fu-    suring        small           differences          in     relatively         large          values
rosemide or CSA could be detected                             in the acidified            that result from                   the peak             urine collections               in oral
aqueous  phase, either on thin-layer                           plates sprayed             studies  or early                  collections           in intravenous                studies.
after development                 with Bratton-Marshall                 reagent      or   The calculated                   furosemide             concentrations               for these
measured              directly     as radioactivity.          These      findings         samples           are     related           by the        extraction          ratio         of the
agree with those                 noted   by Kindt          and      Schmidt (5).          urine       sample            to theaqueous  standard.
                                                                                          The extraction ratiosfor these early samples    may                                               be
                                                                                          further    used for subsequent       dilute     urine     samples.
     Table 2. Extractable Radioactivity When                                              The results      (Table  1) demonstrate       that “effective”
  [S]Furosemide is Added to Serum and Urine                                               extraction    ratios for urine samples       are substantially
and Back-Extracted into Different Phosphate                                               lower than those of comparable          saline standards,         al-
                      Buffers                                                             though the radiochemically       measured        partition    coef-
                                                       Effective extraction
             Sample type           No. samples                  ratio’                    ficients   are similar.  Also, urinary      concentration        are
0.1 mol/liter                                                                             generally         subject          to considerably                greater     variation              in
neutral phosphate                                                                         a given experiment         series. Internal   standards    added
            buffer                                                                        to serum samples         also give somewhat         lower “effec-
   serum                                 27                  0.71 ± 0.02                  tive” extraction     ratios, but effect the serum concen-
   urine                                  6                  0.80 ± 0.02                  tration curve much less.
                                                                                              Interesting  results     relevant   to “effective”    extrac-
1.0 rnoi/liter                                                                            tion was discovered          in the urine     of patients     with
neutral         phosphate                                                                 renal insufficiency who were receiving large quan-
            buffer                                                                        tities (2 mg/kg per day) of furosemide     for therapy.
   serum                                 12                  0.29 ± 0.03                  When 240 mg of [35S]furosemide (specific activity, 80
      Mean ± SD.
   #{176}                                                                                 nCi/mg) was given intravenously     to those patients,
                                                                                          as much    as 36% of the urinary  radioactivity    could

156          CLINICAL       CHEMISTRY,    Vol. 20, No.2,     1974
not be extracted      by organic     solvents     from acidified
samples,    even    after   repeated      trials.    These     com-
pounds   would not be detected           by the present       assay
and would explain, in part, the lower urinary excre-
                                                                                                                 0                          (2,606,782)
tion of un-transformed        furosemide      observed     in these
patients as compared      to normal subjects        (9, 10).

   The nature and extent of binding of furosemide     to
serum proteins   was examined     by the equilibrated
column technique    of Hummel     and Dreyer (4); we
used     both     [35S]furosemide        and    fluorometric                   analysis
to measure the concentration.      Although     less precise,                                                                                    10                     20                 30
                                                                                                                                                 Fraction   number
the chemical   determinations       substantially      agreed
with the more precise radiochemical        determinations.                                       Fig.5a.Chromatography fresh serum equilibrated
                                                                                                                      of                                                                              with
                                                                                                 [355]furosemide         at 10 mg/liter      in binding      buffer (condi-
The extent of equilibrium     binding, as determined         by
                                                                                                 tions described       in the text)
this technique    in two concentrations     of ligand, is                                        Numbers adjacent to the peak and trough are total Counts bound per mil-
shown in Table 3. To obtain        some gross measure-                                           liliter of serum: figure on the plateau is the concentration of the free hg-
                                                                                                 and,    in cpm/ml
ment of the dynamics     of binding in whole serum, fu-
rosemide    was also added to serum and the mixture
was separatedon the same column of G-50 Sephadex                                                                                                    #{149}-.
                                                                                                                                                         Furosemide        Serum
used     above,            was equilibrated
                         which                      with binding                                         l8                                              Furosemide

buffer only. As shown in Figure 5, the well-separated                                                                                I’
unbound      furosemide    peak contained        only 35% of the
total drug while, at the time of elution (20 mm), the
protein fraction still had 65% of the added furosem-
ide, indicating      that the half-time       for dissociation   of                                                                  10                     20                 30                    40
one class of slow binding        sites probably         exceeds  30                                                                          Fraction        number
mm, while that of another class of fast sites is proba-                                          Fig. 5b. Separation of free and rapidly equilibrating furo-
bly less than 1 to 2 mm. The extent of equilibrium                                               semide from strongly bound drug, at 10 mg of total furo-
                                                                                                 semide per liter
binding,    as measured     by ultrafiltration       with Amicon                                 The area of the protein peak is 65% of the total count.                              The columns         are
Centriflo       filter     cones    (Table     3), was 95%, in good                              run under the identical   conditions   depicted     in Figure 5a but without                        higand
                                                                                                 in the buffer. Adsorption    phenomena        probably    explains the later                       elution
agreement     with measurements        by the technique   of                                     time of the tree ligand compared     to the figure above
Hummel      and Dreyer. Because     studies with free furo-
semide    in binding   buffer indicated     that some furo-
semide was bound to the filter       cones, the precision of                                     serum and washes (Table 3) was 96% of the dose
this estimate    is somewhat    lower, owing to both ran-                                        added, indicatingthat the extent of erythrocytese-
dom and systematic      error.                                                                   questration                 of furosemide,                  as        determined              by     this
   The extent            of furosemide       binding        to erythrocytes                      method,              is certainly        considerably                 less   than 10%.
was estimated     by adding   [35S]furosemide                                 to whole
heparinized   blood. Recovery    of furosemide                                from the           Discussion
                                                                                                    Need for a convenient    and relatively     precise meth-
                                                                                                 od in which common         laboratory      instruments     are
  Table 3. Binding Properties of [S]Furosemide                                                   used to determine   furosemide    in biological     fluids has
                                                                                                 led to the use of the fluorometric  technique.                                                      This
            Type of experiment                    ments           Percent          bounda        method is adequately sensitive and reasonably                                                      selec-
Serum                                                                                            tive for use with                 clinical            samples;          its main        drawback
   “Centriflo” filter cones                         12                 0.95    ±       2         is that             there     may be interferences                           from other             fluo-
   “G-50 Sephadex” column                              2               0.91    ±       1         rescent             substances           that      may          also be present               in these
       equilibrated   with                                                                       samples. With care, the use of solvent extraction
       [Slfurosemide,      5 mg/liter                                                            procedures   can lead to reasonably      reliable measure-
   Binding      buffer      only                       2               0.68    ±       1         ments of un-transformed         drug. CSA, a contaminat-
     using serum with                                                                            ing or biotransformed       product,  causes    insignificant
     [S]furosemide,   10 mg/liter                                                                fluorescent   interference,    owing to its low quantum
   Equilibrated with                                   1               0.84     ±      1         yield under the conditions of this assay.
     [t5Sjfurosemide, 5 mg/liter,                                                                       This         simple       assay           is      unsuitable            for     measuring
     and sodium salicylate,                                                                      metabolic              products.                   parti-
                                                                                                                                           The effects               of pH, solvent
     100mg/liter                                                                                 tioning, different quantum yields, superimposed       ab-
Erythrocytes                                           2               0.040       ±       0.1   sorption peaks, fluorescing impurities, and quenchers
  u    Mean ± SD.                                                                                of fluorescence   dictate a preliminary separation    be-
                                                    ________________                             fore          measurement.                 Recently,              a     method           in     which

                                                                                                                             CLINICAL      CHEMISTRY,                Vol. 20, No.2,       1974            157
high-pressure       liquid    chromatography       is used   has               substances    produced   by the drug could create an un-
been developed          in our laboratory     to separate   and                detectable    systematic  error in urinary    blanks by this
measure      furosemide     in unextracted    serum and urine                  method,    and therefore   in the derived drug concentra-
in effluents     of ion-exchange      columns   by both fluoro-                tion.
metric     and spectrophotometric            detection       techniques,          In spite of these well-recognized     drawbacks,   use of
with comparable     sensitivity.                                               this assay in clinical       studies  of furosemide    phar-
   The present     method      for measuring     furosemide              in    macokinetics has produced                     reasonably        accurate mea-
serum is straightforward,         requires  only common               lab-     surements of concentrations                     in serum        and urine (10)
oratory  instrumentation,          and can be adapted                   for    that compare             favorably  to similar     measurements        of
clinical use in monitoring these concentrations  in pa-                        [35S]furosemide              in normal   subjects.      Until   other,
tients. However, in patients   with normal renal func-                                      and convenient methods of clinical
                                                                               more reliable,
tion, serum concentrations   of only 1-2 mg/liter pro-                         assay are available,     this method                   will provide   a rea-
duce effective       diuresis    and the drug is rapidly elimi-                sonable  solution    to the problem                    of measurement      of
nated (serum half-life,           about 30 mm), with 85-95%                    furosemide         concentration            in biological     fluids.
of injected      furosemide        appearing       in the urine un-
changed.      Although        the plasma      half-life   in anephric
                                                                                  This study     was supported      by Hoechst   Pharmaceuticals,   Inc.,
patients    is only prolonged          about three-fold        (9), little     by contract   NIH NIAMD         72-2219, and by a grant (RR-133)     from
information       is available      concerning       accumulation         of   the General Clinical Research          Centers  Branch,   Division of Re-
drug metabolites.          Nevertheless,       current     clinical    ex-     search Resources,    NIH.
perience     suggests     that large doses in cases of renal
failure appear       to produce       only transient      hypoacusia.          References
Thus, monitoring of concentrations       in plasma may                                                             A.,
                                                                               1. HajdO, V. P., and H#{228}ussler, Untersuchungen               mit dem        Sali-
           c      s
be of littlelinical ignificance,   although    crucial for                     diureticum      4-Chloro-N-(2-furylmethyl)-5-sulfamyl-anthranil-
toxicological       and pharmacological          investigations.               sAure.Arzneim.  Forsch. 14,709(1964).
                                                                               2.              A.,
                                                                                   H#{228}ussler, and HajdO, V. P., Untersuchungen                   mit dem Sali-
   The main disadvantage       in the use of the fluoro-
                                                                               diureticum          4-Chloro-N-(2-furylmethyl)-5-sulfamyl-anthranil-
metric method is that several sources of error are dif-                        sAure.Arzneim.      Forsch. 14,710(1964).
ficult to control. The first is the presence of extracta-                      3. Parker, C. A., and Rees, W. T., Correction                      of fluorescence
ble, spectrally interfering,  and quenching impurities.                        spectra and measurement          of fluorescence      quantum        efficiency.   An-
                                                                               alyst (London) 85, 587 (1960).
Second, although the need to back-extract       from the                       4. Hummel,       J. P., and Dreyer, W. J., Measurement                     of protein-
ether phase into an aqueous medium           does avoid                        binding phenomena by gel filtration. Biochim.                  Biophys.       Acta 63,
some adsorption           phenomena      on glassware      and does            530(1962).
promote     the removal        of some interfering     compounds,              5. Kindt,  H., and        Schmidt,    E.,    Urinary  excretion    of furosemide
                                                                               in healthy  subjects       and patients      with liver cirrhosis.    Pharmacol.
it is also a source of variability.          Although    within-run
                                                                               Gun. 2, 221 (1960).
variation is not great (about 5%), gross between-run                           6. Diamantstein,       T., and      Ehrhart,    H., Thin-layer      chromato-
variation      is an annoying        source of error, which is                 graphic   separation     of a metabolic       product    of tryptophane.      Z.
only partly       controlled     by the use of added internal                  Physiol. Chem. 326, 131 (1961).
                                                                               7. Kirschner,     J. G., Thin-layer     chromatography.     In Technique      of
standards.       In the presence       of high drug concentra-
                                                                               Organic Chemistry,        12, A. Weissberger,     Ed., J. Wiley, New York,
tions in serum or of spectrally            interfering   drugs such            N.Y., l96’7,p 166.
as salicylate,      these measures        are even less reliable.              8. Forrey,A. W., An ALGOL computer program for the computa-
Fluorometry   of furosemide in urine is even more dif-                         tion of quench correction by remote terminal using  punched
                                                                               paper    tape.    In Organic Scintillators and Liquid Scintillation
ficult than in serum, but the alternative   (colorimet-                        Counting,      D. L. Horrocks and C. T. Peng, Eds. Academic    Press,
nc) assay       also has     many of these           same difficulties,        New York, N. Y., 1970, p 835.
especially      adequacy       of estimation           of the urinary          9. Cutler,  R. E., Forrey, A., Christopher,          T. G., and Kimpel, B.
blank value in the face of varying diuresis. Our esti-                         M., Pharmacokinetics          of furosemide   in normal subjects     and func-
                                                                               tionally anephric    patients.     Gun. Pharmacol.    Ther. (in press).
mation of the effect of diuresis as a function of urine
                                                                               10. Kelly, M.      R.,    Cutler,  R. E., Forrey      A., and     Kimpel,     B. M.,
flow is only a convenient    device, and any superim-                          Pharmacokinetics         of orally administered        furosemide.      Clin. Phar-
posed changes in the elimination     rate of interfering                       macol. Titer. (in press).

158      CLINICAL   CHEMISTRY,      Vol. 20, No.2,    1974

To top