Specific Assay for the Quantitation of lndocyanine Green in Rat Plasma Using High-Performance Liquid Chromatography with Fluorescence Detection

Specific Assay for the Quantitation of lndocyanine Green in Rat Plasma Using High-Performance Liquid Chromatography with Fluorescence Detection MARYBETHDORR**AND GARYM. POLLACK*’ Received March 8, 1988, from the *Division of Pharmaceutics, School of Pharmacy, C.B. #7360, Beard Hall 200 H, University of North Carolina, *Present address: Parke-Davis Pharmaceutical Research Accepted for publication October 11, 1988. Chapel Hi//,NC 27599-7360. Division, Warner-Lambert Company, 2800 Plymouth Road, Ann Arbor, MI 48105. Abstract 0 A rapid and sensitive reversed-phase HPLC assay employ-

ing fluorescence detection was developed for quantitating indocyanine green (ICG) in rat plasma. Sample preparation entailed precipitation of plasma proteins with acetonitrile prior to injection on the column. The assay was linear from 0.4 to 200 pg/mL, with a detection limit of 3 ng on column. The plasma concentration-time profile of ICG was characterized by this HPLC method and compared with the traditional spectrophotometric assay following iv bolus and iv infusion administration of 5 mg/kg of ICG to rats. Concentrations of ICG obtained using the spectrophotometric assay were consistently higher than those determined by HPLC. In animals receiving ICG by infusion, the maximum difference between the two assays was observed 1 min post-infusion and became negligible by 5 min post-infusion. The calculated pharmacokinetic parameters for ICG, systemic clearance and apparent volume of distribution, were higher using the HPLC assay as compared with the spectrophotometric procedure. The data suggest that a biotransformation or degradation product of ICG is formed in the rat and interferes with the determination of ICG by the spectrophotometric assay. Since the HPLC assay is specific for the parent dye, it is suggested that this assay method be used when determining pharmacokinetic parameters of ICG in rats.

formed in vivo either by hepatic biotransformation or by enzymatic conversion in the circulating blood. The purpose of the present investigation was to develop a sensitive HPLC assay for ICG to be used to evaluate the disposition kinetics of the dye in the rat. Since future studies would involve estimating the hepatic extraction of ICG following intraportal infusion of very low doses of the dye (<0.5mgkg), and due to the limited sample volume obtainable from rats (
Experimental Section M a t e r i a l d t e r i l e lyophilized indocyanine green (ICG) powder was generously supplied by Hynson, Westcott, and Dunning (HWD, Baltimore, MD). Nonsterile ICG powder was purchased from Sigma Chemical (St. Louis, MO). The internal standard, 1-naphthalene acetic acid, was purchased from Kodak (Rochester, NY). Heparin sodium injection, USP (10,000 d m L ) was manufactured by ElkinsSinn (Cherry Hill, NJ). All solvents used were of HPLC grade and were obtained from Fisher Scientific (Fair Lawn, NJ). High-Performance Liquid Chromatography Instrumentation and Chromatographic Conditions-The HPLC system consisted of a high pressure pump (Glenco Scientific, Houston, TX) and a spectrofluorometer (Schoeffel model SF-970, Ramsey, NJ). Samples were introduced into the system via a Rheodyne injector (model 7125, Cotati, CA), and detector response was monitored with a recording integrator (Shimadzu model C-R3A, Kyoto, Japan). Chromatographic separation was achieved using a 25-cm by 4.6-mm i.d. analytical column (octylsilane, 5p particles, Alltech Associates, Deerfield, IL) and a mobile phase of 0.05 M phosphate buffer containing 1%triethylamine (pH 4.0):acetonitrile (100:50, v/v) delivered a t a flow rate of 1.33 mL/min. A 5-cm by 4.6-mm i.d. precolumn packed with an octadecylsilane stationary phase (30-40 p particles, Vydac, Hesperia, CA) was employed to protect the analytical column from debris contained in plasma samples. The eluant was monitored at a n excitation wavelength of 214 nm, with a 370 nm emission cutoff filter. Preparation of StandardePreliminary studies indicated that ICG was unstable in aqueous solution, resulting in a n unacceptable degree of day-to-day variability in standard curve construction. To circumvent this problem, a methanolic solution of ICG powder (1

Indocyanine green (ICG) is a tricarbocyanine dye that has been used extensively in human and animal studies to assess hepatic function,’* estimate hepatic blood flow,= and determine cardiac output.6 The dye is rapidly removed from plasma by transport across the hepatic sinusoidal membrane via an active anion transport system,7 with subsequent excretion into bile.2 There does not appear to be any significant extrahepatic elimination or enterohepatic recirculation of 1CG.3 These properties have led to the frequent use of ICG in investigations of the effects of various disease states (e.g., acute8 and chronic9 renal failure, chronic liver disease,1° aging’l) on liver function and hepatic blood flow. The traditional method for quantitating ICG in biological samples is by absorbance spectrophotometry at 775-800 nrn.25 The method is simple, rapid, and relatively sensitive. However, this assay is not specific, and quantitation of ICG by this method requires the assumption that the dye is neither degraded nor biotransformed prior to analysis. Recently, several HPLC asays have been developed to circumvent the possible nonspecificity of the spectrophotometric a~say.1”’~ These assays have been used to quantitate ICG in the plasma of rabbits123’5 and h ~ m a n s , ~ 3 Jand 4 in rat liver perfusate.l6 In several s t ~ d i e s , ~ ~the ,~~ spectrophotometric J5 assay consistently yielded higher concentrations of ICG than those measured by HPLC. This observation suggests the presence of an interfering substance that absorbs light at the same wavelength as the parent dye. Heintz and c o - ~ o r k e r s ~ ~ proposed that the interfering substance is a contaminant in the commercial product. Alternatively, the substance may be 328 / Journal of Pharmaceutical Sciences Vol. 78, No. 4,April 1989

0022-3549/89/O400-0328$0 1.0010 0 1989, American Pharmaceutical Association

mg/mL) was prepared, and aliquots (100 pL) were pipetted into 1.5mL polypropylene tubes (Sarstedt, Princeton, NJ). This solution was evaporateid to dryness under nitrogen at room temperature and stored at -20°C. Immediately prior to use, the ICG residue was reconstituted with 50 pL of 95% ethanol and diluted to a final concentration of 150 pg/mL with blank rat plasma. Serial dilutions of this plasma solution were made to obtain the desired range of concentrations for construction of standard curves. High-Performance Liquid Chromatography Sample Preparation-Saniples were stored a t 5 "C until the time of sample preparation, whic:h in all cases was within 6 h of collection. Preliminary studies indicated that ICG was stable for this period of time under these storage conditions. Each sample (25 pL) was prepared by precipitating plasma proteins with acetonitrile (30 pL) containing 1naphthalene acetic acid as an internal standard (1 pg/mL). The samples were vortexed for 10 s and centrifuged at 13,000 x g for 30 s. Aliquots of the supernatant (5 to 20 pL) were injected into the chromatographic system within 5 min of preparation. Standards and samples were treated identically and run in parallel. Spectrophotometric A n a l y s i d t a n d a r d s and samples were prepared immediately prior to analysis by diluting 50 pL of plasma with 450 pL of double distilled water. Absorbance of visible light was measured a t a wavelength of 795 nm using a Beckman (Irvine, CA) model DZMO spectrophotometer. Stability of Indocyanine Green (ICG) in Prepared SamplesThe accurate determination of ICG in plasma is potentially complicated by degradation of the dye following sample preparation. The present experiment was performed to assess the stability of ICG in plasma following protein precipitation with acetonitrile. Standards of ICG in plasma (12.5 pg/mL) were prepared for analysis by HPLC as described above. Aliquots of the prepared samples were injected repeatedly over the ensuing 72 h. The peak area of ICG was determined in each 20-pL aliquot. For comparison with the results in precipitated plasma, solutions of ICG in normal saline were analyzed at various times following preparation by both HPLC and spectrophotometric methods. Purity of Commercial Sources of Indocyanine Green (1CG)Three prominent peaks were consistently observed following injection of methanolic solutions of ICG into the chromatographic system. The percentage of the total amount of material injected onto the column representing authentic ICG was estimated as follows. Solutions of ICG prepared from different lots of the sterile lyophilized product were injected into the HPLC system and the areas of each of the three primary peaks were measured. The largest peak was assumed to be ICG. This analysis was also performed on solutions prepared from two lots of nonsterile ICG powder. The percentage of the amount of dye represented by each peak in the chromatogram was estimated by comparing the appropriate peak area with the sum of the three peak areas. This method of analysis assumes similar fluore,sceiit properties for all three chemical species, an assumption that has yet to be verified. However, in the absence of authentic contaminants, this procedure allows a reasonable estimation of the chromatographic purity of the dye. In Vivo Studies-Five male Sprague-Dawley rats, weighing 350 to 460 g, were anesthetized with a n intraperitoneal injection of urethane (1gm/kg, Sigma Chemical, St. Louis, MO). Rectal temperature was maintained at 37 "C using a rectal probe interfaced with a temperature controller (Simpson Electric, YSI model 73ATD, Chicago, IL) and heating pad. A silicone rubber cannula (0.020 i.d. by 0.037 o.d., Dow Corning, Midland, MI) was inserted into the right jugular vein using the procedure of Upton.'7 Three animals received ICG (5 mg/kg) by rapid iv bolus through the jugular vein cannula, which was subsequently flushed with 0.2 mL of heparinized saline (20 unitts/mL). Blood samples (150 pL) were withdrawn from this cannula a t 0.5, 1, 1.5,2,3,5,7.5,10, and 15 min. The remaining two animals received a 2-min iv infusion (2.5 mg/kg/min) through the jugular vein cannula. A 16-gauge needle attached to polyethylene tubing (F'E-50, Clay Adams, Parsippany, NJ) was inserted into the tail artery and blood samples (150 pL) were collected from this cannula a t 0.33, 0.67, 1, 1.33, 1.67, 2, 2.33, 2.66, 3, 5, and 7.5 min following initiation of the infusion. Plasma was harvested and prepared for HPLC and spectrophotometric analysis as described above. In Vitro Incubation Study-An aqueous solution of ICG (800 pL, 1mg/mL) was added to 8 mL of heparinized (50 units/mL) rat whole blood and placed in a shaking water bath a t 37 "C. Aliquots (250 pL) were withdrawn at times corresponding to the in vivo experiments.

Plasma was harvested and assayed by both analytical methods within 12 h of collection. Calculations and Statistical Analysis-Pharmacokinetic parameters were calculated using noncompartmental analysis1* of the concentration-time data following both iv bolus and iv infusion administration. The elimination rate constant (/3) was calculated as the slope of the terminal portion of the concentration-time profile using log-linear regression. The area under the curve (AUC) and the area under the first moment curve (AUMC) for each concentrationtime profile were calculated using the trapezoidal rule with extrapolation to infinity. Systemic clearance (CL)and steady-state volume of distribution (Vd,,) were calculated using the following equations:

Dose AUC

CL = Vd,,

=

Dose x AUMC (AUC)'

(2)

The dose of ICG employed in these calculations reflects the actual content of the dye in the dosing solution, measured with the HPLC assay. Differences between pharmacokinetic parameters generated using the two different assay methodologies for ICG were assessed using a one-tailed paired t test. A probability c0.05 was considered statistically significant.

Results and Discussion High-Performance Liquid Chromatography Assay PerformanceSample chromatograms for ICG in rat plasma are displayed in Figure 1. Baseline resolution was attained for ICG and the internal standard with a total run time of 9 min. The assay was linear throughout the entire range of ICG concentrations studied (i.e., through 200 pglmL). The limit of sensitivity of the assay was 3 ng on column (twofold peak-to-noise ratio), corresponding to a plasma concentration of 0.4 pglmL. This degree of sensitivity is adequate for in vivo

= 1 I

I

I

- - -

J

0 3 6 9

0 3 6 9

0 3 6 9

Run Time (min.)

Figure 1-Chromatograms of blank rat plasma following precipitationof plasma proteins with acetonitrile containing the internal standard (left), plasma obtained from a rat 2 rnin following bolus injection of a 5-mg/kg dose of ICG, evidencing an ICG concentration of 18.6 pg/mL (center), and blank rat plasma spiked with the detection limit (0.4 pg/mL) of ICG (right). Key: (I) plasma contaminant; (11) internal standard; (111) ICG. Defector set at range of 0.05 pA in all cases. Journal of Pharmaceutical Sciences / 329 Vol. 78, No. 4, April 1989

studies in rats involving small volume plasma samples (25 pL) and doses as low as 0.5 mg/kg, and exceeds that for previously reported HPLC assays employing UV absorbance detection.lz14 The results of the assay validation procedures are summarized in Tables I and 11. The assay was associated with a high degree of precision, with intraday coefficients of variation ranging from 2.46% (100 pg/mL) to 5.12% (1 pg/mL). Interday variability was approximately twofold higher than intraday variability. Furthermore, calculated ICG concentrations were statistically indistinguishable from the known values, indicating acceptable accuracy in the range of concentrations employed. Standard curves were reproducible from day to day, with coefficients of variation of 4.82 and 19.2% for the slopes and intercepts, respectively (data tabulated in Table 11). In all cases, the correlation coefficient for linear regression of the peak area ratio versus the known ICG concentration exceeded 0.99, suggesting that the assay was linear within the concentration range examined. Stability of Indocyanine Green ( I C G b T h e change in ICG peak area with time following plasma protein precipitation is displayed in Figure 2. The fluorescence response to the dye increased linearly over the first 2 h postprecipitation, a t a rate of -53%/h. Thereafter, the change in detector response appeared to conform to a monoexponential approach to a constant value, which was achieved at -48 h postprecipitation. Overall, the apparent fluorescence increased threefold over the 72-h interval. During this same time interval, the peak area of the internal standard remained constant, suggesting that the matrix itself was not changing with time. The time-dependent change in fluorescence of ICG in saline is contrasted with the change in absorbance of visible light by the dye in Figure 3. The degree of absorbance of 800 nm light increased from 0.12 to 0.27 over the first 45 min, and remained stable thereafter. This time-dependent change in absorbitivity of ICG in aqueous solution has been described previously,1321 and is presumably due to stabiliza-

300

250

I-

/

0

r

200

L

150 Y 0

al

100

a c3 0

50

LL

0

20

10

30

40

50

70

60

Time (hr) Figure 2-Time course of change in peak area of ICG (12.5 pg/mL) fol/owingprotein precipitation with acetonitrile (inset: data during first 4 h postprecipitation). n

rn

120 rL

1 4

E

0

-

UJ

3

- -/4

7

1

/.//

60

4

Y

0 Q)

Concentration, mg/L

lntradaya

0

~

1 .o

0.964 f 0.0494' (5.12%) 10.5f 0.266 (2.53%) 101.4f 2.49 (2.46%)

10

100

lnterdayb ~

~

1.05 ? 0.0990 (9.40%) 9.70 f 0.616 (6.35%) 98.0 f 2.38 (2.43%)

a Results of seven replicate determinations for each concentration. Results of replicate determination for each concentration, once daily for six consecutive days. =Data presented as mean t SD.dCoefficient of variation.

Table Il-Summary of Dally lndocyanlne Green Standard Curves Utilizing the High-Performance Liquid Chromatography Assay' Day

Slope

Intercept

P

1

0.0483 0.0451 0.0449 0.0445 0.0504 0.0462

0.0170 0.0298 0.0244 0.0276 0.0210 0.0234

0.9999 0.9823 0.9996 0.9999 0.9998 0.9992

2 3 4 5 6

aConstructedfrom eight standards (0.5-200mg/L) on six consecutive days. 330 /Journal of Pharmaceutical Sciences Vol. 78, No. 4, April 1989

n

E

0.25 S

0 0

W

al 0

0.15 S

0

0.10

i

2oye

Actual lndocyanine G~~~~(ICG) Calculated ICG Concentration, mglL

0.30

00

a Table I-lntra- and lnterday Reproducibility of the HlghPerformance Liquid Chromatography Assay for lndocyanlne Green in Rat Plasma

80

*

I

I

50

I

'

"

"

100

'

'' I 150 200

"

"

-P0 UJ

n

Q

0.05

I

250

Time (min) Figure &-Time course of change in fluorescence (0) or absorbance (0) of ICG in normal saline (72.5 pg/mL).

tion of the electronic structure of the dye secondary to aggregation of ICG molecules. Such stabilization would decrease the energy difference between the excited and ground states of the pi electrons in the molecule, thus making absorption of an incident photon more likely. In addition, the absorbance maximum would be expected to shift to longer wavelength (lower energy) photons, as has been noted previously for ICG.20 A similar mechanism may well be responsible for the time-dependent increase in the fluorescence of ICG. However, the divergent time courses between the changes in absorbance and fluorescence (Figure 3) suggest that increased absorption of incident radiation is not entirely responsible for the increased fluorescence. Further investigation is required to determine the precise mechanism responsible for this novel instability of ICG. From a practical standpoint, however, the present data indicate that samples containing ICG must be analyzed immediately following preparation, regardless of the analytical method (fluorescence HPLC versus spectrophotometry) employed. Purity of Commercial Sources of Indocyanine Green (ICG)-Chromatograms of two methanolic solutions of ICG,

one representing the sterile lyophilized product and the other unsterilized ICG powder, are displayed in Figure 4. Two contaminant peaks eluted later than ICG, at -10.5 and 14.5 min, respectively. Both contaminant peaks were more prominent in the sterilized product than in the raw powder. The results of the purity assessment of different sources of ICG are summarized in Table 111. The percentage of material injected on column represented by the most prominent peak was relatively consistent between lots from the same source. However, there was a notable difference between the apparent purity of the sterile lyophilized product and ICG powder. The powder form appeared to be more pure, with ICG representing -86% of the total in a freshly opened bottle. This value compares favorably with the labeled purity of 85%. Following HPLC analysis of the sterile lyophilized product, the peak corresponding to ICG represented only -75% of the total. These results suggest that some of the dye

may have degraded during the sterilization and/or lyophilization procedure. In light of these observations, when using ICG in animal studies, it would be prudent to use the more pure source. In human studies, however, it is necessary to use the sterile product. In this situation, it may be best to use the same lot throughout the entire study and to use only freshly opened vials. In Vivo Studies-A representative plasma concentrationtime profile obtained following bolus administration of the dye is illustrated in Figure 5 . The concentrations measured by the spectrophotometric assay were greater than those determined by HPLC at almost every time point sampled, whereas the difference was relatively small (<10%) in the dosing solution. Furthermore, the between-assay difference was not constant across the disposition profile. The maximal difference appeared a t -3 min postinjection, and decreased markedly thereafter. It is also noteworthy that the shape of the disposition profile was markedly different between the two assays. The plasma concentrations of ICG determined by spectrophotometry appeared to decline in a mixed zeroordedfirst-order fashion, while the decline of ICG determined by HPLC appeared to be first-order and was easily fit with a nonlinear least-squares regression programz2 using a twocompartment model. The apparent nonlinear disposition of ICG in various species, including the rat, has been widely r e p ~ r t e d . ~The ~ , ~nonlinearity ~,~~ is most prominent after high doses (>5 mg/kg) of the dye and has been proposed to be due to the presence of a saturable uptake process in the li~er.~3.25.26 The fact that nonlinear decline was observed in the present investigation only when concentrations were determined with the nonspecific spectrophotometric assay suggests that a biotransformation or degradation product may contribute to the appearance of nonlinearity. A comparison of the concentrations obtained by each assay method during infusion of ICG is shown in Figure 6. The concentration of ICG in the diluted dose solution was almost identical (60.2 versus 64.4 pg/mL) when measured by both methods. As with the bolus dose study, time-dependent differences between the HPLC and spectrophotometric assays were observed, resulting in a marked counterclockwise hysteresis. The difference in concentration between the assays was negligible during the first minute of the infusion period. After this point, the spectrophotometric assay yielded

---

1

0

10

5

15

20 A

T i m e (min.)

-1

Figure 4--Chromatograms of raw ICG powder (solid line) and sterilized, lyophilized product (broken line) dissolved in methanol. Detector setting was 0.1 pA for both determinations. Key: (111) ICG; (IV) Contaminant I; (V) Conta,minant2. Table Ill-Purity of Commercial Sources of lndocyanine Green as Determined by High-Peformance Liquid Chromatography Source

Lot #

I-c

E

100

, 60

Percent of Total

Opened

lndocyanine Contaminant Contaminant 1 2 Green HWDa 798 HWD 363 HWD 622 Sigmab 45F3663 Sigma 64F3447

sameday sameday 5months same day 2 years

74.7 76.5 70.1 86.0 82.0

13.3 9.7 23.7 8.4

8.0

12.0 13.8 6.2 5.6 10.0

a Hynson, Westcott, and Dunning (Baltimore, MD). *Sigma Chemical (St. Louis, MO).

0

2

4

6

8

10

12

14

16

Time (min) Figure 5-Concentration-time profile for ICG following a S-mg/kg iv bolus dose to a rat as measured by the HPLC (0) orspectrophotometric (0) methods (inset: time course of difference in concentration between the two assays). Journal of Pharmaceutical Sciences / 331 Vol. 78, No. 4, April 1989

120

r

Table IV-Pharmacokinetic Parameters Obtalned following Intravenous Administration of lndocyanine Green (5 mglkg) to Rats

I)/

Assay

CL, mUmin/kg

Spectrophotometric

11.8 3.0a 7.40 2 1.17'

80

6o

HPLC

r

*

Vd,,, rnUkg 64.0 2 2.7 50.5 5 16.7

p, min-' 0.141 0.204

* 0.03 5

0.10

'Data presented as mean ? SD for five animals. bSignificantly different from HPLC assay, p < 0.05.

UJ

' l l l " l , ' l ' , , ' l , , . l , L

20

40

60

80

100

120

Plasma ICG (mg/L) HPLC Assay Figure 6-Comparison of the dosing solution dilution (0) and ICG plasma concentrations (0)obtained following Pmin iv infusion of 5 mg/kg to a rat as determined by spectrophotometricand HPLC analysis. The solid line represents identity. The arrows along the dashed line indicate the direction of time, from 0 through 7.5 min.

higher concentrations, with the difference becoming maximal at 3 min (1 min postinfusion) and again negligible by 7 min. Heintz and co-workers16reported that differences between a UV absorbance HPLC assay and the classical spectrophotometric method for ICG in rabbit plasma was due to a contaminant in the ICG itself. This contaminant was cleared more slowly than the authentic dye, leading to an increase in the between-assay concentration difference with time. Such a phenomenon was not observed in the present investigation. No change in the contaminant-to-ICG peak area ratio was observed following injection of the ICG solution into rats; all three peaks decreased in parallel. The difference between the HPLC and spectrophotometric assays could not be accounted for by monitoring the contaminant peaks, suggesting that an additional degradation (or biotransformation) product was formed afcer administration of the dye to the animals. Attempts to isolate a degradationhiotransformation product formed in vivo have as yet been unsuccessful. No additional Le., nonICG, noncontaminant) peaks appeared on chromatograms, suggesting that the substance may elute with the solvent front. Alternatively, the product may not fluoresce, making it transparent to the present HPLC assay. The temporal pattern in the difference between the two assay methods in the present study is important in that it addresses the potential source of the substance that interferes with the spectrophotometric assay. If the substance was present in the injectate as a contaminant, the difference in apparent ICG concentration between assays would be expected to evidence a monotonic function with time. For example, if the contaminant was cleared from the systemic circulation more slowly than the dye, as reported in the rabbit,l5 the difference between assay results would increase with time. Conversely, if the contaminant was cleared more rapidly than ICG, the difference between assays would decrease with time. The time course of the assay differences shown in Figures 5 and 6 are consistent with the generation of an interfering substance in vivo. The pharmacokinetic parameters determined following both rapid iv bolus and iv infusion were essentially identical and are grouped together for comparison. These values are 332 /Journal of Pharmaceutical Sciences Vol. 78, No. 4, April 7989

summarized in Table IV. Due to large differences in the calculated AUC, the systemic clearance (CL) of ICG was much greater following HPLC analysis than spectrophotometric quantitation (p < 0.05). Likewise, due to higher apparent concentrations of ICG, the volume of distribution of the dye was lower using the spectrophotometric data as compared with that generated with the HPLC assay, although the difference did not reach statistical significance (p < 0.1). Since the CL of ICG has been shown to increase with decreasing dose of the it is important to consider the dose of ICG employed when comparing values from different studies. The CL of 7.40 1.17 ml/min/kg obtained using the spectrophotometric assay compares favorably with values reported in the literature (6.9 to 8.2 mumidkg) for doses of -5 mg/kg in the rat.8,9,27No such determinations using a specific HPLC assay for ICG have been reported in the literature t o date. The difference in calculated pharmacokinetic parameters obtained between these two assay methods underlines the importance of noting the assay method employed when comparing literature results. Source of the Interfering Substance-The concentration of ICG did not decline following a 20-min incubation in vitro at 37 "C in whole rat blood. This result was consistent for both assays. Indocyanine green has been reported to be stable in plasma for prolonged periods of time, albeit at low temperatures.'* Thus, the substance interfering with the spectrophotometric assay is probably not formed by simple chemical degradation or by enzymatic processes occurring in the systemic circulation. Furthermore, when considering the results of the in vivo study, the apparent systemic appearance of the product as early as 1 min after infusion of the dye rules out the possibility of degradation in the gastrointestinal tract. The processes of biliary secretion and absorption would probably delay the circulation of the product in the plasma by considerably longer than 1min. Since the elimination of ICG is exclusively via the liver,3 the present data are suggestive of a hepatic origin for the interfering substance.

*

Conclusions The HPLC assay described herein is of sufficient sensitivity to be used to quantitate ICG in small volume samples following administration of low doses of the dye to rats. In addition, the data presented suggest that a degradation product of ICG, probably of hepatic origin, is formed in the rat. The apparent degradation product absorbs light of wavelength 795 nm, and therefore cannot be discerned from authentic ICG using the traditional spectrophotometric assay. Disease-inducedchanges in the formation or elimination of this product may lead to erroneous conclusions regarding the influence of disease on the disposition of ICG per se if the spectrophotometric assay is employed for measuring plasma concentrations of the dye in this animal species. As the HPLC assay is specific for the parent dye, changes in the systemic clearance derived from this method should only reflect changes in the clearance of the parent compound.

References and Notes 1. Hunton, D. B.; Bollman, J. L.; Hoffman, H . N . Gastroenterology 1960,39, 713-724. 2. Cherrick, G. R.; Stein, S. W.; Leevy, C. M.; Davidson, C . S. J . Clin. .Invest. 1960, 39, 592-600. 3. Caesar, J.; Shaldon, S.; Chiandussai, L.; Guevara, L.; Sherlock, S. 171in. Sci. 1961,21, 43-57. 4. Leevy, C. M.; Mendenhall, C. L.; Lesko, W.; Howard, M. M. J . Clin. Znuest. 1962,41, 1169-1179. 5. Villeneuve, J. P.; Huot, R.; Marleau, D.; Huet, P. M. Am. J . Gastnoenterol. 1982, 77, 233-237. 6. Sekelj, P.; Shankar, K. R.; Doman, J.; Sukumar, I. P.; Palmer, W. H. J . Appl. Physiol. 1970,29, 249-253. 7. Scharschmidt, B. F.; Waggoqer, J. G.; Berk, P. D. J . Clin. Znuest. 1975,56,1280-1292. 8. YateE:, M . S.; Emmerson, J.; Bowmer, C. J . Bwchem. Pharmacol. 1983,32, 31093114. 9. Yates, M. S.; Emmerson, J.; Bowmer, C. J. J . Pharm. PharmaC O ~ .1983,35, 593-594. 10. Branch, R. A.; James, J. A.; Read, A. E. Clin. Pharmacol. Ther. 1976,20, 81-89. 11. Wood, A. J. J.; Vestal, R. E.; Wilkinson, G. R.; Branch, R. A.; Shantl, D. G. Clin. Pharmacol. Ther. 1979,26, 16-20. 12. RappiapOrt, P. L.; Thiessen, J. J. J . Pharm. Sci. 1982, 71, 157161. 13. Svensson, C. K.; Edwards, D. J.; Lalka, D.; Mauriello, P. M.; Middlleton, E. J . Pharm. Sci. 1982, 71, 1305-1306. 14. Donn, K. H.; Powell, J. R.; Rogers, J. F.;Plachetka, J . R. J . Clin. Pharmacol. 1984,24,360-370. 15. Heintz, R.; Svensson, C. K.; Stoeckel, K.; Powers, G. J.;Lalka, D. J..Fharm.Sci. 1986, 75, 398-402.

16. Roberts, R. K.; Heath, C. A,; Johnson, R. F.; Speeg, K. V.; Schenker, S. J . Lab. Clin. Med. 1986,107, 112-117. 17. Upton, R. A. J . Pharm. Sci. 1975, 64,112-114. 18. Gibaldi, M.; Perrier, D. Phurmacokinetics, 2nd Ed.; Marcel Dekker: New York, 1982. 19. Michie, D.; Goldsmith, R.; Mason, A. Proc. Soc. Exp. Biol. Med. 1962,111,540. 20. Sutterer, W.; Hardin, S.; Benson, R.; Krovetz, L.; Schiebler, G. Am. Heart J. 1966. 72. 345. 21. Landsman, M.; Kwant, G.; Mook, G.; Zijlstra, W. J. Appl. Physiol. 1976, 40, 575. 22. Mitzler, C.; Elfring, G.; McEwen, A. Biometrics 1974, 30, 562563. 23. Stoeckel, K.; McNamara, P. J.; McLean, A. J.; duSouich, P.; Lalka, D.; Gibaldi, M. J . Pharmacokinet. Bwphnrm. 1980, 8, 483-496. 24. Watkins, J. B.; Noda, H. J . Pharmacol. Exp. Ther. 1986, 239, 467-473. 25. Paum artner, G.; Probst, P.; Kraines, R.; Leevy, C. M. Ann. N.Y. fcad. Sci. 1970,170, 134-147. 26. Thiessen, J. J.; Rappaport, P. L.; Eppel, J . G. Can. J . Physiol. Pharmacol. 1984, 62, 1078-1085. 27. Iga, T.; Klaassen, C. D. J . Pharmacol. Exp. Ther. 1979,211,690697.

Acknowledgments The authors would like to thank Mr. Paul M. Savina for his ex ert advice during development of the assay and his careful review ofthe manuscript.

Journal of Pharmaceutical Sciences / 333 Vol. 78, No. 4, April 1989