Fig. 1. Chromatogram showing peak of GTN (RF=0.47) extracted from urine.

Fig.2. (a) Peak purity spectra for GTN, extracted from urine sample, scanned at the peak start, peak apex and peak end positions of the spot (correlation, r_start,_apex = 0.997, r_apex,end = 0.996); (b) Comparison of spectra of GTN extracted form urine with that of standard GTN (correlation > 0.99).

Table 1: Precision of proposed HPTLC method for estimation of GTN spiked in human urine

Concentration of GTN (ng/spot)	Area Mean (n=3)	Intra-day Precision %RSD	Area Mean (n=3)	Inter-day precision %RSD
400	401.90	1.10	402.21	0.38
600	604.05	1.14	595.85	1.21
800	799.73	1.27	798.57	0.92

Pharmacokinetics of Gatifloxacin are linear and time-independent at doses ranging from 200 to 800 mg administered over a period of up to 14 days. Steady-state concentrations are achieved by the third daily oral or intravenous dose of gatifloxacin Serum protein binding of gatifloxacin is approximately 20% and is concentration independent. Gatifloxacin is excreted as unchanged drug primarily by the kidney. More than 70% of an administered dose recovers as unchanged drug in the urine within 48 hours following oral and intravenous administration².

Table 2: Accuracy data for proposed HPTLC method for GTN in urine

Drug	Initial Amount (ng/µL)	Amount added (%)	% Recovered	% R.S.D.
GTN	600	0	100.89	1.39
	600	80	101.08	1.10
	600	100	100.53	0.54
	600	120	100.22	1.85

It is well documented that in a typical bioavail- ability or bioequivalence study, samples of an assessible biological fluids, such as blood or urine, are analyzed for drug and/or its metabolite concentrations³.Urine provides a non-invasive sample collection method and determination of drug levels in urine is comparatively less complex than plasma and other body fluids^4,5. Several reports indicate that urinary excretion data can be used to arrive at bioequivalence decision of different drug formulations^6-9. HPLC methods have been reported for estimation of GTN in biological fluids (plasma, urine)^10-13. Due to its speed and versatility, it was thought of interest to develop HPTLC method for estimation of GTN in urine. The present study describes development, validation and application of a simple and specific HPTLC method for estimation of GTN in urine. The urinary excretion data, thus obtained, was successfully utilized to compare bioavailability of GTN after administration of two market GTN tablet formulations in twelve healthy human volunteers for assessment of bioequivalence.

2. MATERIAL AND METHOD:

2.1. Instruments:

A HPTLC system consisting of Camag Linomat IV semiautomatic spotting device, Camag glass twin-trough chamber (20/10 cm2), Camag TLC Scanner 3,Camag CATS 4 software (Camag Sonnenmattstr., Muttenz, Switzerland) and a 100 ml HPTLC syringe (Hamilton Company, Reno, NV) were used for chromatographic analysis.

2.2. Preparation of standard solutions:

A 1 mg/ml stock solution of GTN was prepared in methanol (Solution S1). An appropriate volume of stock solution was further diluted with methanol to obtain a standard solution of GTN having a final concentration of 100 µg /mL (Solution S2).

2.3. Preparation of phosphate buffer (pH 7.0):

A phosphate buffer of pH 7.0 was prepared by mixing 41.3 ml of 1/15 M solution of potassium dihydrogen phosphate with 58.7 ml of 1/15 M solution of disodium hydrogen phosphate.

Table 3: Summary of validation parameters for the proposed HPTLC method for the estimation of GTN in Urine.

No.	Parameter	Result for GTN
1	Linearity range
	(a) Standard Gatifloxacin (b) Gatifloxacin spiked in urine	200-1200 ng/spot ( y = 9.9779x + 5802.7 r = 0.9994) 200-1200 ng/spot (200-1200 µg/ml) (y = 9.7865x + 5858.8 R2 = 0.9997)
2	Limit of detection	25.64 Ng/spot
3	Limit of quantitation	77.70 Ng/spot
4	Precesion (RSD) (a) Repeatability of sample application (b) Intra-day (c) Inter-day	1.13 1.27 1.10 0.92 0.38
5	Accuracy (%)	100.22 101.08
6	Specificity	specific

2.4. Chromatographic conditions:

TLC plates (20×10 cm²) were activated by pre- washing with methanol followed by drying in oven for 5 min (50 ± 1⁰C) and bringing down to room temperature. Chromatographic estimations were performed using activated TLC plates under following conditions-

Mobile phase: - n -butanol /methanol / ammonia (5:1:2, v/v/v)

Volume of mobile phase: 8 ml

Chamber saturation time: 30 min

Temperature: 25 ± 1 ⁰C,

Relative humidity: 35 -40%

Migration distance: 80 mm Wavelength of detection: 292 nm Band width:6 mm

Space between two bands: 4 mm

Spraying rate: 150nL/sec

2.5 Extraction of GTN from urine:

One milliliter of urine (drug-free or drug-spiked or volunteer urine sample) was transferred into 10 ml volumetric flask. Volume was adjusted to 10 ml with phosphate buffer (pH 7.0). One millilitre of this solution was extracted with chloroform (2×1 ml) by vortexing for 5 min. at high speed, followed by centrifugation (5 min, 1200-1300 rpm). 0.8 ml of chloroform layer was collected on each extraction and combined (total of 1.6 ml).

2.6. Chromatographic separation:

Appropriate volumes of the combined extract or standard solution of GTN (Solution S2) were spotted on the TLC plate 10 mm from bottom edge using Camag Linomat IV semiautomatic spotting device. TLC plate was developed in ascending mode in twin-trough chamber previously saturated for 30 min with mobile phase, n -butanol /methanol / ammonia (5:1:2, v/v/v). The plate was removed from chamber, dried in air and scanned in absorbance/ reflectance mode using Camag TLC Scanner 3 at 292 nm. Data of peak area was recorded using Camag CATS 4 software.

2.7. Preparation of calibration curve:

2.7.1. Calibration curve for standard GTN:

The standard solution (0.2-1.2 µL) was applied on TLC plate with the help of microlitre syringe, using Linomat V sample applicator. The plate was developed, dried and scanned as described in Section 2.6. A plot of peak area versus corresponding GTN concentration was constructed.

2.7.2. Calibration curve of GTN spiked in urine:

One millilitre of drug-free urine was transferred in six different 10-ml volumetric flasks. Aliquots of 0.2, 0.4, 0.6, 0.8, 1.0 and 1.2 ml of S2 were added separately in the volumetric flasks.

The solutions were diluted upto the mark with phosphate buffer (pH 7.0). One millilitre of the solution was extracted and analyzed as described under Section 2.6 and Section 2.7. Quantitative determination was performed by fitting areas of the peaks corresponding to GTN from the chromatograms into corresponding calibration curve equation.

2.8. Validation of the method:

2.8.1. Linearity:

The linearity of response for GTN was assessed in the range of 200-1200 ng/spot for standard GTN and GTN spiked in urine after extraction.

2.8.2. Limit of quantitation and limit of detection:

Limit of quantitation of GTN was taken as the lowest concentration of GTN in the calibration range. For limit of detection, concentrations of GTN lower than the limit of quantitation were spotted and the minimum concentration detected under given chromatographic conditions was considered as limit of detection.

2.8.3. Precision:

Precision of the proposed method in terms of intra-day variation (RSD) was determined by analyzing urine samples spiked with GTN at different concentrations (200-1200 ng/spot) for 3 times on the same day and inter-day precision (RSD) was assessed by analyzing urine samples spiked with different concentrations of GTN (200-1200 ng/spot) on 7 different days over a period of 1 week.

2.8.3.1. Repeatability of measurement of peak area:

Ten microlitres of S2 (100 µg/ml) were spotted on TLC plate, developed, dried and the spot was scanned for seven times without changing the plate position and RSD for measurement of peak area was determined.

2.8.3.2. Repeatability of sample application:

Ten microlitres of S2 (100 µg/ml) were applied seven times on TLC plate by semiautomatic spotting device. The plate was developed and analyzed as described under Section 3.6 and RSD for peak area for different peaks was calculated.

2.8.4. Accuracy:

The accuracy was determined by standard addition method at different concentration levels of GTN. Different volumes of GTN were added to urine samples spiked with GTN (600 ng/µl). The samples were extracted with chloroform and analyzed as described under Section 3.6. Amount of total GTN in urine was determined by fitting the corresponding peak area into the calibration curve equation for GTN spiked in urine. Accuracy of determination of GTN in urine was computed using the formula: (Total amount of GTN found in urine=Total amount of GTN spiked in urine)×100

Table 4: Average % cumulative LMF excreted rate of LMF excretion and log-transformed rate of LMF excretion data after administration of Formulation R and Formulation T.

Time (mid-point) (h)	% cumulative excreted		Rate of excretion (dXu/dt)		(log (dXu/dt) (mg/h)
Time (mid-point) (h)	Formulation R	Formulation T	Formulation R	Formulation T	Formulation R	Formulation T
0.5	4.55±1.43	4.52± 1.28	18.23±5.75	18.09±5.14	1.23±0.18	1.23±0.15
1.5	10.11±2.74	10.21 ±1.86	22.22±5.31	22.75±2.66	1.33±0.11	1.35±0.05
2.5	16.14±3.63	16.55± 2.02	24.11±4.10	25.38±3.75	1.37±0.08	1.40±0.06
3.5	21.83±4.36	22.94± 2.03	22.76±4.47	25.53±4.97	1.34±0.09	1.39±0.09
4.5	28.09±5.02	28.92 ±2.67	24.66±4.22	23.91±4.58	1.38±0.07	1.37±0.08
5.5	36.12±3.56	35.44± 3.71	32.47±7.39	26.09±5.95	1.50±0.09	1.40±0.12
7	45.41±3.42	43.56± 4.76	18.57±3.19	16.23±3.71	1.26±0.08	1.20±0.09
9	54.42±2.73	51.85± 4.34	18.02±2.81	16.57±2.39	1.25±0.07	1.21±0.06
11	61.99±2.24	58.66± 5.28	15.15±2.47	13.63±3.34	1.17±0.07	1.11±0.15
18	68.56±2.57	65.68± 2.74	2.18±0.80	2.34±1.18	0.31±0.14	0.34±0.14
30	72.92±2.78	70.83± 3.60	1.45±0.20	1.71±1.06	0.15±0.06	0.18±0.18
42	76.08±2.52	73.63± 3.52	1.05±0.27	0.93±0.27	0.00±0.14	0.00±0.12

2.8.5. Specificity:

The specificity of the method was ascertained by analyzing standard GTN, drug-free urine and urine spiked with GTN. The spot for GTN spiked in urine was confirmed by comparing its RF and absorbance/reflectance spectrum with that of standard GTN. The peak purity of GTN spiked in urine was assessed by comparing the spectra at peak start, peak apex and peak end positions of the GTN spot.

2.8.6. Recovery studies:

Recovery of GTN from urine was calculated as the ratio of area of GTN peak after extraction from urine to the area of standard GTN at respective concentrations followed by application of the correction factor.

2.9. Bioavailability study:

2.9.1. Study protocol:

A pilot bioequivalence study was performed on 12 healthy human volunteers (20 - 28 yr, 55-70 kg) employing a single dose two treatment, two period, open randomized crossover design with a wash out period of minimum 7 days between the treatments. The written informed consent was obtained from all the Volunteers. Healthy status of the volunteers was assessed by history physical examination and laboratory investigations. Volunteers with a history of major Kidney, liver and heart disease were excluded from the study. Renal functionality of the volunteers was assessed on the basis of creatinine clearance test. Individuals with known gastrointestinal disease that might affect absorption of the drug, history of adverse reaction and hypersensitivity to fluoroquinolones were excluded from the study. An only non-smoking and non alcoholic individual with no clinically significant abnormal findings during medical history, physical examination and laboratory evaluations was allowed to participate in the study. The study protocol was submitted to the local ethical committee. None of the volunteers was received any other drug at least 2 weeks prior to day 1 of the study and during the study. The volunteers was abstained from consumption of xanthene containing food and beverages (chocolate, tea, coffee or coke) for 24 h before administration of the dose and will be fasted overnight (at least 10 h). A standard breakfast and standard lunch was provided after 4 and 6 h of sampling, respectively. Each volunteer received a formulation (R or T) along with 200 ml of potable water. Water consumption was restricted up to 4 h after administration and was allowed ad libitum there after. Urine samples was collected before administration and at 1,2,3,4,5,6,8,10,12,24,36 and 48 h after administration of the formulation. The volume of the urine collected during sampling time from each volunteer will be measured.

Representative samples of urine (10 ml) were stored, in glass test tubes sealed aluminum foil, at -20^oc until analysis. The urine samples, after bringing to room temperature, were analyzed for Gatifloxacin content by the proposed HPTLC method and urinary excretion profile was used to determine various pharmacokinetic parameters.

2.9.2. Pharmacokinetic analysis:

The peak excretion rate (dXu/dt)_max and peak excretion time (t _max) values were obtained from the urinary excretion rate (dXu/dt ) versus time curves obtained for each volunteer after administration of Formulation R and Formulation T. Various other pharmacokinetic parameters such as overall elimination rate constant (K), terminal elimination rate constant (k _el), elimination half-life (t_1/2) were obtained from log-transformed urinary excretion rate log(dXu/dt),mg/h versus mid-point of time (h) curves. Both, AUC _0–48and AUC _0-∞ (mg) were calculated using untransformed (dXu/dt). AUC _0–48data was calculated using linear trapezoidal rule and was extrapolated to infinite time, AUC _0-∞. K_el was calculated from the slope of terminal linear portion of log (dXu/dt) versus mid-point of time curve. K was obtained as quotient of intercept of terminal linear line extrapolated to Y-axis and dose (mg). The elimination half life (t_1/2) was calculated using the formula, t_1/2 = 0.693/kel. Cumulative amount of GTN excreted in 48 h (C_total, mg) and % dose of GTN excreted were also estimated

Table 5: Summary of pharmacokinetic parameters of LMF after administration of GTN tablet formulations

Pharmacokinetic parameter	Formulation R	Formulation T
t _max (h)	4.7 ± 1.4	4.1 ± 1.2
(dXu/dt) _max(mg/h)	33 ± 7	29 ± 2.4
AUC _{0 - 48} (mg)	329.3 ± 16.13	315.2 ± 13.50
AUC _0-∞ (mg)	354 ± 15	335 ± 16
t_1/2(h)	5.3 ± 1.7	4.8 ± 0.72
Cumulative amount of GTN excreted (C_total, mg)	304 ± 11	295 ± 14
% cumulative amount of GTN excreted ^a	76 ± 2.6	74 ± 3.5
K (1/h)	2.534 ± 1.056	2.769 ± 0.5255
k_el (1/h)	0.1404 ± 0.03216	0.1487 ± 0.02358

^a % cumulative amount of GTN excreted with respect to the administered dose (400 mg). Note: Each value indicates mean ± S.D. for data from twelve volunteers. Formulation R, Gaity, Dr Reddy’s Ltd, India; Formulation T, Zigat, FDC Ltd, India.

Table 6: Relative bioavailability (%) ^a

^a Calculated as (ratio of pharmacokinetic parameter of Formulation T to corresponding parameter of Formulation R) ×100.

AUC _{0 - 48}	95.71
AUC_0-∞	94.63
C_total	97.03

2.9.3. Statistical analysis:

The relative bioavailability was Calculated as ratio of pharmacokinetic parameter of Formulation T to corresponding parameter of Formulation R ×100 and cumulative amount of GTN excreted (C_total)from Formulation T relative to Formulation R. Classical 90% confidence intervals were estimated for AUC _0-48, AUC _0-∞, (dXu/dt)_maxand C_total.

3. RESULTS AND DISCUSSION:

3.1. HPTLC method development and validation:

Due to its versatility and speed of analysis, HPTLC technique was found suitable for estimation of GTN levels in urine. At pH 7.14, GTN is easily extracted from urine by means of organic solvents. Various solvents viz. Chloroform, dichloromethane, ethyl acetate were tried for quan titative extraction of GTN from urine. Use of chloroform could provide better clean-up and recovery of GTN. Adjustment of pH of urine to 7.0, followed by two times extraction with chloroform could improve extraction efficiency to more than 90%, which was satisfactory. It was observed that about 5 times volume of phosphate buffer (pH 7.0), as compared to that of the volume of urine sample, was required to make the pH of urine sample to 7.0. Different compositions of n -butanol, methanol and ammonia were tried to obtain optimum RF and separation of GTN from urine components on the TLC plate. Various modifiers like triethyla mine, diethylamine, ammonia solution were tried to achieve sharp band of GTN. A mixture of n-butanol: methanol: ammonia (5:1:2 v/v/v), could provide sharp peak of GTN well resolved from other urine components at RF of 0.47 (Fig. 1).

It was observed that activation of TLC plates (pre-washing with methanol followed by drying at 50⁰C) and pre-saturation of TLC chamber with mobile phase for 45 min ensures good reproducibility and peak shape of GTN.

Densitometric evaluation was performed at 292 nm, the wavelength of maximum absorbance of GTN (λ _max), in absorbance/reflectance mode.

3.1.1. Validation:

Using the optimized extraction method and chromatographic conditions, developed HPTLC method was validated in terms of linearity, limit of detection, limit of quantitation, precision, accuracy and specificity.

3.1.1.1. Linearity:

Peak areas of standard GTN were found to be linear in the range of 200-1200 ng/ spot (i.e. 200-1200 µg/ml, n = 6) with correlation coefficient of 0.9994. Peak areas of GTN spiked in urine were found to be linear in the range of 200-1200 ng/spot (i.e. 200-1200 µg/ml, n = 6) with correlation coefficient of 0.9997.The average linear regressed equations for the corresponding curves were y = 9.9779x + 5802.7 (slope=9.9779, intercept=5802.7) and y = 9.7865x + 5858.8 (slope=9.7865, intercept=5858.8), where ‘y ’ is the concentration of GTN in ng/spot and ‘x ’ is the corresponding peak area.

3.1.1.2. Limit of quantitation and limit of detection.

The limit of quantitation was 77.70 ng/spot for GTN spiked in urine, while minimum detectable quantity of GTN was found to be 25.64 ng/spot.

3.1.1.3. Precision:

The intra-day variation for determination of GTN in urine was in the range of 1.27 to 1.10 %, while inter-day variation was ranging from 0.92 to 0.38 % (Table 1). Incase of repeatability of sample application, peak area values showed a RSD of 1.13 (Table 3).These values were found to be satisfactory.

3.1.1.4. Accuracy:

The percentage accuracy for estimation of GTN in urine, determined using standard addition method, was found to be between 100.22 and 101.08 % over the concentration range studied (Table 2).

3.1.1.5. Specificity:

Comparison of chromatograms of urine spiked with GTN and blank (drug-free) urine, showed no interference from the urine components in the separation of GTN. Peak purity check showed high degree of correlation between spectra scanned at peak start, peak apex and peak end positions (r_start,apex = 0.9997 and r_apex,end= 0.996) of GTN peak which confirmed that the peak represents a pure single component i.e. LMF (Fig. 2a). This was further supported by equally good correlation (r = 0.99) between spectrum of standard GTN and the spectrum of GTN spiked in urine (Fig. 2b). Average recovery of GTN, from urine, over the range of spiked concentration of 200-1200 ng/spot, was found to be 100.68 %. GTN in urine was found to be stable over a period of 7 days at – 20 ⁰C. Different validation parameters for the proposed HPTLC method are summarized in Table 3. Thus, the proposed method is simple, sensitive, specific, precise and accurate and can be utilized for estimation of GTN excreted in human urine.

3.2. Bioavailability study:

The average values (± S.D.) for % cumulative amount of GTN excreted, rate of excretion (dXU/dt) and log-transformed rate of excretion (log (dXU/dt)) for both the formulations with respect to mid-point of time are given in Table 4. The plots of average % cumulative of GTN excreted (± S.D.) over a period of 48 h versus mid-point of time and average (log(dXU/dt )) (± S.D.) versus mid-point of time, are shown in Fig. 4 and Fig. 5, respectively. From these figures it is evident that both the formulations show similar excretion behavior, which in turn, indicates similarity in their bioavailability.

About 304 ± 11mg (82.32 ±/2.75 % of dose) and 315.2 ± 13.50 mg (78.8 ± 3.37 % of dose) of GTN is excreted in 48 h (C _total) after oral administration of Formulation R and Formulation T, respectively. It was observed that both the formulations showed maximum excretion rates in the interval of 3.5- 4.5 h (t _max) in terms of mid-point of time (Fig.5).Average values (± S.D.) of various pharmacokinetic parameters are reported in Table 5.

AUC_0-48 value for Formulation R was found to be 329.3 ± 16.13 mg and that for Formulation T was 315.2 ± 13.50 mg, which are comparable. Maximum amounts excreted in corresponding time intervals (dXu/dt)_max from Formulation R and Formulation T were 33 ± 7 and 29 ± 2.4 mg/h, respectively (Table 5).

Fig.3. Average % cumulative GTN excreted, with respect to GTN dose, versus mid-point of time plots after administration of GTN tablet formulations. (Note: The vertical lines indicate S.D. in log (dXU/dt) at corresponding mid-point of time).

Fig.4. Average log excretion rate (log (dXU/dt) versus mid-point of time plots for GTN after administration of GTN tablet formulations. (Note: The vertical lines indicate S.D. in log (dX_u/dt) at corresponding mid-point of time).

Values of other pharmacokinetic parameters of Formulation T were also comparable with that of reference formulation (Formulation R). Comparison of all these parameters indicated similar bioavailability of GTN from Formulation T as compared to Formulation R. The relative bioavailability of GTN from Formulation T, in terms of AUC _{0 - 48}, AUC _0-∞ and C_total were found to be 95.71, 94.63 and 97.03 %, respectively (Table 6).

For bioequivalence decision, AUC_0-48, AUC_0-∞, C_total and peak excretion rate (dXU/dt) max values were subjected to statistical analysis like 90% confidence interval.

ANOVA revealed that there is no significant difference between Formulation T and Formulation R, since the calculated F-values were found to be less than the corresponding critical values at given degrees of freedom.

The estimates of 90% confidence interval (P = 0.05) for the ratio of these four parameters were found to be within the specified limits of 80–120% for untransformed data as per US FDA requirement for bioequivalence. Thus, it was observed that Formulation T is bioequivalent to Formulation R.

CONCLUSIONS:

A HPTLC method was developed for estimation of GTN excreted in urine. The method was validated and found to be simple, sensitive, specific, accurate and precise. The proposed method was successfully used to obtain urinary excretion data for GTN after administration of GTN tablet formulations in twelve healthy male volunteers. Statistical analysis of various pharmacokinetic parameters calculated using urinary excretion data of GTN revealed that Formulation T is bioequivalent with Formulation R. This method can further be use for the bioanalysis of gatifloxacin to compare the excretion profiles of different groups of healthy human subjects such as male-female, smokers-nonsmokers as well as alcoholics and non-alcoholics.

ACKNOWLEDGEMENT:

We would like to thank especially Principal, R.C.Patel Institute of Pharmaceutical Education and Research, Shirpur for his encouragement and patience throughout the duration of this project and providing all necessary facilities for our research work.

REFERENCES:

1. Wayne PA, (2000) Methods for Dilution Antimicrobial Susceptibility Tests for Bacteria That Grows Aerobically. NCCLS 5:12-14, 20.

2. Easton PA, (2000) USP24, NF19, Mack Printing Company, 2056-2061.

3. Shargel BL. et al. Applied Biopharmaceutics and Pharmacokinetics.Prentice-Hall International, 1999; USA 4: 247-279.

4. Shah S, Rathod S I, Savale B, Patel J, (2002) Chromatogr.–B 767:83-91.

5. Straughan AA, Melikan M, Meyer J, (1979) Pharm. Sci 68:1099-1102.

6. Martin MB, Uihlein R, Ings L, Pharm Sci 1984;73:437-441.

7. Shishoo SC, Shah I, Rathod S, Savale M, Vora, Int. J Pharm. 2001; 228:53-67.

8. Adriana MD, Raul H, Marcela, Drug Dev Ind Pharm, 2000;26:821-828.

9. Zhang X, Brian R, Michael B, So win ski, Kevin M, Gatifloxacin Pharmacokinetics in Healthy Men and Women. Journal of Clinical Pharmacology, 2006; 46(10), 1154-1162.

10. Lockyer M, Al-Dgither S, Al-Gaai E, Yousuf A, (2005).

11. Mornerl HK, Harlwigg H, Determination of Gatifloxacin in Human Serum and Urine by HPLC. 2000; 52:105-107.

12. Brian R, Overholser, Michael B, Kays, Kevin M, Sowinski, Determination of gatifloxacin in human serum and urine by high-performance liquid chromatography with ultraviolet detection, 2003.

13. Al-Dgither S, Naseeruddin Alvi S, Muhammad M, Hammami, Development and validation of an HPLC method for the determination of gatifloxacin stability in human plasma, 2003.