Cefuroxime Hydrolysis Kinetics and Stability Predictions in Aqueous Solution

DAN WANG AND ROBERT E. NOTARI~ Received April 15, 1993, from the The Department of Pharmaceutical Sciences, College of Pharmacy, Medical University of South Carolina, 171 Ashley Avenue, Charleston, SC 29425. Accepted for publication September 20, 1993".

Abstract Cefuroxime hydrolysis rate constants (k) were determined to predict the degradation rate of cefuroxime in aqueous solution as a function of pH, temperature, and buffer. At constant temperature, the pH-rate expression was: k = kH(aH+) + kslfi + ks2f2 + koH(aow), where f l is the fraction of cefuroxime in the undissociated form and f 2 is the anionic fraction, k~ and kw are the catalytic rate constants for hydrogen activity (a,++) and hydroxyl ion activity (aoK), and ksl and ks2 are first-order rate constants for spontaneous hydrolysis. Formate, acetate, phosphate, and borate buffers did not catalyze degradation. Temperature dependencies for kH, ksl, ks2, and kon were described with values for A (pre-exponential term) and E (energy of activation) calculated from k = Ae-€IflT (where R is 1.987 cal/moldeg and T is absolute temperature). Combining the pH and temperature equations allowed predictions for cefuroxime hydrolysis rates in aqueous solutions at any pH and temperature. Results were validated by predicting the observed rate constants for every experimental condition and also for a reconstituted commercial product storedat 30 'C. Maximum stability was observed in the pH-independent region from pH 4 to 7, where the time during which cefuroxime concentration exceeded 90% of its initial concentration at 25 O C was 1.2 days. Rate constants employed in predictions were based on stability-indicating HPLC assays. For selected conditions, additional rate constants were calculated from changes in cefuroxime UV absorbance. These UV constants did not consistently agree with the kvalues obtained with HPLC assays or with either of the rate constants determined for loss of cefuroxime to descarbamoyl cefuroxime or to the competing route, which is primarily P-lactam hydrolysis.

Cefuroxime is a semisynthetic cephalosporin antibiotic that is poorly absorbed following oral administration.' It is marketed for parenteral administration as sterile crystalline cefuroxime sodium for reconstituting before injection and also as a frozen iso-osmotic solution.2

Cefuroxime axetil is a prodrug that is marketed for oral administration. Although there are no liquid oral products, the kinetics of hydrolysis of cefuroxime axetil have been quantified as a function of pH, temperature, buffers, and ionic strength: and sufficient data were obtained to predict the shelf-life for aqueous solutions of cefuroxime axetil under potential storage conditions.

Cefuroxime is administered in aqueous parenteral solutions, but kinetic data are not available for cefuroxime hydrolysis. Based on previous studies on aqueous cephalosporin solutions,4.5 cefuroxime would be expected to degrade by several routes, with principal loss via hydrolysis of the 0-lactam and the carbamate group at position 3 as illustrated in Scheme 1.

Cephalosporin hydrolysis kinetics based on UV spectropho- tometric changes have also been The observed rate constants were attributed to loss via routes competing with the hydrolysis of various cephalosporin substituents a t position 3. For cefuroxime, these are represented by kl in Scheme 1. However, this interpretation is questionable because the UV changes are nonspecific.

Therefore the specific aims of this study were (1) to provide sufficient data to reliably predict the rate constant for cefuroxime degradation in aqueous solution as a function of pH, temperature, and buffer and (2) to assess the meaning of the UV kinetic data

~~~

Abstract published in Advance ACS Abstracts, January 1, 1994.

Lactone Cefuroxime Descafbarnoyl

R - N H J q J CHzOCONH2 R-NHp>cH20HA yn&yH2 COOH

0 0 COOH

i

I k'

0-lactam hydrolysls 0-laclam hydrolysis and other products and other products

6CH3

Scheme 1

by comparing the rate constants obtained from changes in UV absorbance with those obtained for kl (Scheme 1) based on HPLC assays.

Experimental Section Materials-Cefuroxime sodium (sterile, for injection) and descar-

bamoyl cefuroxime (reference standard) were used as received from Glaxo Inc. (Research Triangle Park, NC). All other chemicals were analytical or HPLC grade.

HPLC Assays-The system consisted of a Beckman llOB Solvent Delivery Module, a Waters M-441 280-nm fixed wavelength detector, and a Hewlett Packard HP3394A integrator. Chromatographic sepa- rations (Figure l) were achieved on a 150 X 4.6 mm, 5 - ~ m spherisorb C6 column (Keystone Scientific Inc., Bellefonte, PA) with a flow rate of 1 mL/min. The injection volume was 20 wL, and the column temperature was 20-30 'C. Mobile phases were filtered through a type HA 0.45-jtm membrane filter (Millipore Corporation) and deaerated under reduced pressure.

The quantitative separation of cefuroxime and descarbamoyl ce- furoxime from the degradation products that formed in the acidic and basic regions required two different mobile phases. At pH < 7.5, the mobile phase was 9% methanol, 0.1% triethylamine as an ion-pairing agent and 90.9 % 0.1 M aqueous acetate buffer (pH 3.4). The cefuroxime retention time was -13 min and that of descarbamoyl cefuroxime was -9 min. In the pH region 7.6 to 9.2, the mobile phase was 5% methanol, 0.02% triethylamine and 94.98% 0.1 M acetate buffer (pH 5.1), which provided retention times of -15 min for cefuroxime and -10 min for its descarbamoyl product.

Kinetics of formation of the descarbamoyl product were analyzed when its concentrations were >lo% of the initial cefuroxime concen- trations. Linear calibration plots were conducted for 0.8-9 X l@ M cefuroxime and for 0.1-7.5 X 10-5 M descarbamoyl cefuroxime. The CV was 0.893 for cefuroxime analysis and 1.50 for the descarbamoyl.

Kinetics-The temperatures, composition of the solutions, and the pH values are given in Tables 1 and 2. Temperatures were stable to fO.l OC. The pH values for buffered reactions and the reconstituted productwere determined before and after the reaction at the experimental temperatures. Those for the HCI reactions were calculated with activity coefficients.6

Reactions were initiated by placing 1 mL of an aqueous solution of cefuroxime (or descarbamoyl cefuroxime) into 20 mL of HCl or buffer at the temperature of the study to provide -8 X l(r M substrate. A sample was removed as a function of time and cooled, and a 0.5-mL aliquot was diluted with 5 mL of water or buffer to provide a pH of 4-5. No loss of cefuroxime or descarbamoyl cefuroxime was detected by HPLC when these dilutions were stored for 24 h in the refrigerator.

Spectrophotometric Analysis-The changes in the UV absorbance spectrum for -8 X 1(r M cefuroxime reactions were measured as a function of time with a UV/vis diode-array spectrophotometer (model 8452A, Hewlett Packard). The absorption maximum was 280 nm in

0 1994, American Chemical Society and American Pharmaceutical Association

0022-3549/94/ 1200-577$04.50/0 Journal of pharmaceutical Sciences / 577 Voi. 83, No. 4, Aprll 1994

1

55 HOURS

2

0 4 8 1 2 1 6 0 4 8 1 2 1 6 0 4 8 1 2

RETENTION TIME (MIN)

Figure 1-HPLC chromatograms as a function of time during cefuroxime hydrolysis at 40.1 OC in 0.194 M acetate buffer at pH 4.53. Key: (1) cefuroxime peak; (2) descarbamoyl cefuroxime (other peaks were not identified).

Table 1-First-Order Rate Constants for Hydrolysis of 8.0 X lo-' M Cefuroxime In HCP and the Percentage Differences (Dlf) from Predicted Valuesb

temperature, OC pH HCI, M NaCI, M k, min-' Dif, %

40.1 0.44 0.4869 0.0131 0.00418 1.0 0.60 0.3408 0.1592 0.00382 2.1 0.84 0.1996 0.3004 0.00348 2.0 1.44 0.0506 0.4494 0.00285 0.0

50.1 0.45 0.4869 0.0131 0.0131 0.0 0.84 0.1996 0.3004 0.0114 -0.6 1.44 0.0506 0.4494 0.00923 0.4

60.1 0.45 0.4869 0.0131 0.0379 0.8 0.85 0.1996 0.3004 0.0321 5.0 1.44 0.0506 0.4494 0.0281 0.0

pH values were calculated with activity coefficients from Harned and Owens (p = 0.5). bBased on eq 7.

acid and 274 nm in neutral and alkaline pH. Under all conditions, the absorbance at 274 nm was linear with concentration, and the molar absorptivity in water was 17 466.

First-order rate constants were calculated from the change in absorbance at 274 nm with eq 1:

(1)

In eq 1, A, is the initial absorbance, Af is the final absorbance, At is the absorbance a t any time t , and k,is the apparent first-order rate constant. First-order plots were linear for more than two half-lives, although the decrease in absorbance was generally -30% of the initial value.

ln(A, - A,) = ln(A, - A,) - k,t

Results and Discussion Determination of First-Order Rate Constants-At a

constant temperature, pH, and ionic strength (p ) , the loss of -8 x 10-4 M cefuroxime (or descarbamoyl cefuroxime) in the presence of HCl or aqueous buffers was described by pseudo- first-order kinetics.7 All reactions were followed for more than two half-lives.

The influence of buffers on the rate of cefuroxime hydrolysis was evaluated as follows. Apparent first-order rate constants were determined with two dilutions for each constant acid-base

ratio of formate, acetate, phosphate (40.1 "C), and borate (60.1 "C) buffers. There was no observable buffer catalysis on cefuroxime hydrolysis (Table 2).

Selection of the pH-Rate Expression-Apparent first- order rate constants were determined as a function of pH at three temperatures. A t constant temperature, the rate constant for cefuroxime loss as a function of pH was described by two kinetically equivalent pH-rate expressions. The two equations differed only with respect to the terms used to describe the inflection near pH 1 to 2 in the pH-rate profiles (Figure 2). The following discussion presents the rationale for selecting eq 2.

In eq 2, the inflection was attributed to an increase in the spontaneous hydrolysis rate of the cefuroxime carboxylic acid relative to its carboxylate:

(2) In eq 2, kH and k o H are the catalytic rate constants (M-I min-l) for hydrogen ion activity (uH+) and hydroxyl ion activity (UOH- = K,/ [ a s+] ) and ksl and ksz are first-order rate constants (mi&) for spontaneous hydrolysis. The values for K, as a function of temperature were taken from Harned and Owen.6 The ce- furoxime fractions in the uncharged (f1) and the anionic cfi) forms were described by eqs 3 and 4, respectively:

(3)

(4)

In eqs 2 and 3, K, is the dissociation constant. In eq 5, the inflection was ascribed to an increase in the specific

acid-catalyzed rate of the cefuroxime carboxylate relative to its carboxylic acid

k = &(aH+) + kslfi + k , j 2 + koH(a0H.I

f1 = (aH+)/[(aH+) + K,I

f 2 = Ka/ [ (aH+) + K a 1

k = k , i ( a ~ + ) f i + k ~ z ( a ~ + ) f , 4- ks -I- ~ o H ( ~ , H - ) (5) With this equation, the (uH+) catalytic constants at 40.1 "C were lZ~1 = 0.00255 and k H 2 = 0.352 (min-l-mol-l). This provided a k ~ S / k ~ l ratio of 138. The increase in (UH+)-CatalyZed reactivity for the cefuroxime carboxylate relative to its carboxylic acid is consistent with previously reported values for penicillin car- boxylates relative to their carboxylic a c i d ~ . ~ ~ 9 However, the 138- fold kHP/kH1 ratio is larger than those reported for penicillins, which were only -9- to -20-fold. Moreover, the electron- donating ability of the carboxylate would be expected to reduce

578 /Journal of Pharmaceutical Sciences Vol. 83, No. 4. April 1994

Table 2-Experlmental Conditionsd and Flrst-Order Rate Constants for Hydrolysis of 8.0 X lo4 M Cefuroxlme and the Percentage Differences (Dlf) From Predicted Valuesb

temp, Buffer Concentration, M

"C pH HCOOH HCOONa NaCl k, min-1 Dif, YO 40.1 2.49 0.1993 0.01993 0.48007 0.00123 0.8

2.43 0.7971 0.07971 0.42029 0.00128 3.1 3.47 0.1107 0.11070 0.38930 0.000592 2.4 3.53 0.4429 0.44290 0.05710 0.000578 2.6

50.1 2.43 0.7971 0.07971 0.42029 0.00443 0.5 60.1 2.44 0.7971 0.07971 0.42029 0.0139 0.0

temp, OC pH

30.0 5.39 40.1 4.53

4.55 5.38 5.46

50.1 4.51 60.1 4.49

Buffer Concentration, M

CH3COOH CH3COONa NaCl k, min-I Dif, %

0.09714 0.09714 0.40286 0.000505 2.0 0.38857 0.38857 0.11 140 0.000501 2.8 0.00971 0.07771 0.42229 0.000503 1.0 0.03886 0.31086 0.18914 0.000497 2.2

0.09714 0.09714 0.40286 0.00629 4.0

0.00971 0.07771 0.42229 0.000125 -0.8

0.09714 0.09714 0.40286 0.00193 -1.0

~~ ~~ ~ ~ ~~

Buffer Concentration, M temp,

"C pH NaH2P04 Na2HP04 NaCl k, min-l Dif, %

30.0 6.47 40.1 6.48

6.66 7.33 7.46

50.1 6.43 60.1 6.40

0.03000 0.03000 0.3800 0.000127 0.03000 0.03000 0.3800 0.000518 0.12000 0.12000 0.0200 0.000517 0.003000 0.02400 0.4250 0.000538 0.012000 0.09600 0.2000 0.000546 0.30000 0.03000 0.3800 0.00196 0.030000 0.03000 0.3800 0.00630

temp,

-1.6 -1.2 -0.4 2.0 3.1 -3.1 3.5

Buffer Concentration, M

0.1 F I

h 0.01

!= .r

.- E v

Y

0.001

t \ 30 "C , o m

0.0001 I I 4 I I I 1

0 1 2 3 4 5 6 7 8 9 10

PH Figure 2-The curves at 60.1, 50.1, and 40.1 OC were obtained by simultaneous nonlinear regression with eq 7 to fit the observed cefuroxime hydrolysis rate constants (0). The values in Table 3 were employed in eq 7 to generate the curve at 30 OC, which predicted the observed rate constants for the reconstituted commercial product (m) and for acetate and phosphate buffered solutions (0).

Table 3-Parameter Values to Predlct the Cefuroxlme Hydrolysis Rate Constant with Equation 7

constant €, cai/mol A, min-1 ~~~~~ ~~

kH 19 097 5.5649 X loio ks 1 23 755 1.2530 X l O I 4 ks2 26 389 1.3076 X loi5 kOH 15 585 5.0357 X 10l2

the hydrolysis rate of the carbamate group at position 3, thus "C pH H3B03 Na2B407 NaCl k, min-I Dif, %

40.1 8.63 0.00640 0.00640 0.4872 0.00132 2.3 favoring the observed increase in ksl over ksz. At 40.1 "C, the 50.1 8.58 0.00640 0.00640 0.4872 0.00483 2.3 ksllksz ratio in eq 2 is 6.7. The magnitude of this ratio was 60.1 8.52 0.00640 0.00640 0.4872 0.0168 -4.2 considered more reasonable than the 138-fold kH21kH1 ratio.

8.52 0.02000 0.02000 0.4600 0.0178 -9.6 Therefore, eq 2 was selected to describe the first-order rate constants as a function of pH.

Rate C o n s t a n t as a F u n c t i o n of pH a n d Tem- perature-The Arrhenius equation was used to define kH, ksl,

Buffer Concentration, M temp,

"C pH Na~B407 NaOH NaCl k, min-' Dif, % kS9. and knn as a function of temrrerature: 40.1 9.16 0.02000 0.01333 0.4467 0.00340 -0.9 50.1 9.14 0.02000 0.01333 0.4467 0.0125 4.0 60.1 9.07 0.02000 0.01333 0.4467 0.0396 3.0 30.0 6.61 ReconstitutedC 0.000124 0.8

The range of pH values determined before and after the reaction was f0.03 pH units; temperature ranges were f0.1 "C. Based on eq 7. Sterile cefuroxime sodium (750 mg) was dissolved in 100 mL of deionized distilled water.

the hydrolysis rate of the carbamate group at position 3, thus favoring kHl over k ~ 2 for that route of cefuroxime degradation (see k2 in Scheme 1).

In eq 2, the difference in reactivity was assigned to the hydrolysis term. Previous reports indicate that the P-lactam degradation in the pH-independent region is predominantly via nucleophilic attack on the carbonyl carbon by the neighboring side-chain amido group.4 For this process, a larger degradation rate would be expected for cefuroxime carboxylic acid (ksl) relative to the carboxylate (ksz), which would increase electron density at the carbonyl carbon, thus hindering nucleophilic attack. The electron-donating carboxylate would also reduce

In eq 6, A is the pre-exponential term, E is the energy of activation in calimol, R is 1.987 cal/mol-deg, and T is absolute temperature? Substitution in eq 2 provides the following:

Simultaneous nonlinear regression was applied to all of the data at 40.1, 50.1, and 60.1 "C with eq 7, where f1 and f2 were described by eqs 3 and 4, respectively. The adjustable parameters were E, A, and K,. Although K, was allowed to vary with temperature, a single kinetic estimate of K , of 9.406 X 10-3 provided excellent results for all three temperatures (Figure 2).

Validation of Predictions-Because buffer catalysis was not observed for formate, acetate, phosphate, and borate buffers, the rate constant can be predicted in the presence or absence of these buffers as a function of pH and temperature with eq 7 and the E and A values from Table 3. Validation was carried out in two ways. Predicted rate constants were compared with the experimentally observed values at 40.1, 50.1, and 60.1 "C.

Journal of Pharmaceutical Sciences / 579 Vol. 83, No. 4, April 1994

h rv z 0 I- < c c z W 0 Z 0 0

0.0000 1

0.000001 0 10 20 30 40 50

HOURS Flgure 3-Semilog plots (60.1 OC, pH 6.4 in 0.06 M phosphate buffer) showing: (A) the loss of cefuroxime; (B) the formation and loss of descarbamoyl cefuroxime from cefuroxime; and (C) the independently determined loss from a descarbamoyl cefuroxime solution. The negative slope of (A) is kand of (C) is k'. The (B) curve was obtained by nonlinear regression with eq 9, where k2 was the adjustable parameter.

These differences generally varied from 0 to 5% (Tables 1 and 2). The rate constants were also experimentally determined for three solutions prepared where stability is at its maximum, in the pH-independent region 4 to 7, and stored a t 30 "C. One solution was prepared by reconstituting a commercial product in accordance with the instructions for a 750-mg intravenous infusion pack. This 0.0168 M solution contained 750mgof sterile cefuroxime sodium in 100 mL of deionized distilled water. Solutions containing 8 X lo4 M cefuroxime in acetate and phosphate buffers were also studied. The rate constants experimentally observed a t 30 "C differed by 0.8-1.6% from those predicted with eq 7 (Table 2).

Descarbamoyl Cefuroxime Formation and Loss-The formation and loss of descarbamoyl cefuroxime was also studied to compare the total rate of loss of cefuroxime with the rate of hydrolysis of its alkyl carbamate group to form the descarbamoyl product (Scheme 1). Rate constants for the formation and loss of descarbamoyl cefuroxime were calculated when the concen- trations of the descarbamoyl were sufficiently high for reliable assessments. In accordance with Scheme 1, the apparent first- order rate constant for loss of cefuroxime was defined as k = kl + k2 and the rate constant for loss of descarbamoyl cefuroxime was defined as k' = k3 + k4. Thus, the concentrations of cefuroxime ([CEF]) and descarbamoyl cefuroxime ([DES]) were described as a function of time by eqs 8 and 9, respectively:

[CEF] = [CEF,]e-kt

[DES] = kz[CEF,](e-kt - e-k't)/(k'- k ) Hydrolysis of cefuroxime and of descarbamoyl cefuroxime

were studied independently at identical conditions with each as starting materials. These independently determined values for k and k' were used in eq 9 to fit the [DES] data by nonlinear regression, where kz was the adjustable parameter (Figure 3). The competing rate constant was calculated from kl = k - k2. The resulting k l values and the percentages for cefuroxime loss via this route are reported in Table 4.

Rate Constants Determined by UV Studies-It was reported for other cephalosporins that the rate constants determined from UV measurements were similar to the kl values determined by HPLC.495 This would require similar absorptivity

Table 4-Values (103.min-') ?or k, (Scheme 1) and Thelr Percentage Contribution to k (%k,)'; Rate Constants From UV Measurements (kuv) and Thelr Percentage Differences (Dlf) from the k Values Determined by HPLCb

O C PH ki % kja kuv Dif, %=

40 0.44 -d 1.82 -56 2.31 -40 0.60 2.03 -42 0.84 2.86 0 1.44

2.43 0.682 53.3 2.49 0.594 48.3 1.29 5 3.47 0.368 62.2

60 4.49 3.23 51.4 6.40 3.42 54.3 7.11 13

4.32 -1 1 50 8.58 9.57 -23 9.14

- - - - - - -

- -

- - - -

- - - -

a %kl = 100(kdk). * kvalues inTables 1 and 2. Dif (%) = 100(kuv - k)/k. -, Not available.

0 0 5 10 15 20 25 30

"C Figure 4-The predicted shelf-life (To, = 0.105/k) as a function of temperature for an aqueous cefuroxime solution in the pH range 4-7 in the presence or absence of acetate or phosphate buffers.

for the three compounds shown in Scheme 1 and similar values for their P-lactam hydrolysis rate constants. Therefore, the change in the cefuroxime UV absorbance was studied to compare the resultant first-order rate constants (kUJ with those obtained by HPLC. Values fork,, and the difference between k,, and the values obtained by HPLC measurements are listed in Table 4.

Contrary to the suggested similarity with kl, the k , values at pH 1.44,2.49,6.40, and 8.58 agreed with the k values determined from HPLC not with kl values. The corresponding kl values at pH 2.49 and 6.4 were only -47% of the k,, values. However, at pH 9.14 and at pH values below 0.84, the k , values were smaller than 12. Under these conditions, the descarbamoyl concentrations during cefuroxime hydrolysis were too small t o assign values to kz. This precluded a comparison between k , and k , because kl was calculated from k - k2.

According to previous studies, the formation of the descar- bamoyl in acidic conditions was accompanied by the rapid formation of a lactone with the carboxylic acid group at the 4 position (Scheme l).4,5 This would explain the low concentrations of this alcohol even if significant formation occurred during cefuroxime hydrolysis.

580 /Journal of Pharmaceutical Sciences Vol. 83, No. 4, April 1994

Thus, the k,, values determined from cefuroxime as a function of pH did not consistently estimate any single rate constant associated with Scheme 1 over the pH range studied. Such values are not useful for stability prediction.

Shelf-Life Predictions-Shelf-life was defined as the time during which the concentration of cefuroxime equalled or exceeded 90% of its initial concentration (2'90).

The observed rate constants for the 0.0168 M reconstituted produce (pH 6.61) and for 8 X 10-4 M cefuroxime in buffers (pH 5.39 and 6.47) were all similar (Table 2, 30 "C). This result demonstrated that the constants were independent of the pH and the initial cefuroxime concentration in this range. The rate constant was also independent of ionic strength, which was 0.5 for the buffers but only 0.02 for the reconstituted product.

The maximum storage stability for a solution occurred in the pH range 4-7, The acetate and phosphate buffers used in this pH region did not accelerate cefuroxime hydrolysis. Therefore, hydrolysis rate constants in buffers are the same as those in water. Based on these observations, the shelf-life was predicted from Tw = 0.105/k, with eq 7 to calculate k . The predicted shelf-life for an aqueous cefuroxime solution in the pH range 4-7 as a function of temperature (Figure 4) shows values of 1.2 (25 "C), 5.8 (15 "C), and 30.3 days (5 "0.

References and Notes 1. Wozniak, T. J.; Hicks, J. R. In Analytical Profiles of Drug

Substances, Vol. 20; Florey, K. T.,Ed., Academic: New York, 1991;

2. Physicians' Desk Reference, 46th ed.; Medical Economics Data: Montvale, NJ, 1992; p 1066.

3. Nguyen, N-A. T. Pharm. Res. 1991,8,893-898, 4. Yamana, T.; Tsuji, A. J. Pharm. Sci. 1976,65, 1564-1573. 5. Fabre, H.; Eddine, N. H.; Berge, G. J. Pharm. Sci. 1984,73,611-

618. 6. Harned, H. S.; Owen, B. B. The Physical Chemistry of Electrolytic

Solutions, 3rd ed.; Reinhold New York, 1958; pp 638, 748. 7. Connors, K. A. Chemical Kinetics; VCH: New York, 1990; pp 26,

245. 8. Connors, K. A.; Amidon, G. L.; Stella, V. J. Chemical Stability of

Pharmaceuticals, 2nd ed.; Wiley: New York, 1986; pp 650-654, 290-294.

9. Nguyen, N-A. T.; Mortada, L. M.; Notari, R. E. Pharm. Res. 1988, 5, 288-296.

pp 209-236.

Acknowledgments We thank Glaxo Inc. for supplying cefuroxime sodium and descarb-

moyl cefuroxime and for supporting this work through Glaxo's Partners in Research program.

Journal of Pharmaceutical Sciences / 581 Vol. 83, No. 4, April 1994