International Journal of Pharmaceutics, 41 (1988) 105-113
`Elsevier
`
`UP 01378
`
`Light degradation of ketorolac tromethamine
`
`Leo Gu, Hi-Shi Chiang and David Johnson
`Institute of Pharmaceutical Sciences, Syntex Research, Palo Alto, CA 94304 (U.S.A.)
`
`(Received 2 December 1986)
`(Accepted 20 July 1987)
`
`Key words: Ketorolac tromethamine; Stability; Autoxidation; Light degradation; Mechanism
`
`Summary
`
`Aqueous and ethanol solutions of ketorolac tromethamine were found to decompose rapidly under laboratory black light (350
`nm) to yield CO2, decarboxylation product 4 and 3 oxidation products, 1, 2, and 3. Complete material balance of these 4 products in
`ethanol was found while the material balance in aqueous solutions was poor and decreased with the extent of the reaction. A
`mechanism which involves an initial decarboxylation of the triplet excited state of ketorolac, followed by oxidation. is proposed to
`account for the observed oxygen concentration-dependent kinetics and the product distribution of the reaction.
`
`Introduction
`
`Drug substances stored either as pure raw
`material, in the solid or liquid dosage forms, or
`during the manufacturing processes, are subject to
`various degrees of irradiation by sunlight and
`fluorescent
`room light. Drugs with absorption
`greater than 280 nm have the potential for decom-
`position in sunlight and drugs with absorption
`maxima greater than 400 nm have the potential
`for degradation in both sunlight and room light.
`Direct sunlight and room light experiments are
`time-consurning and often result
`in inconsistent
`data due to the day-to-day variation in the light
`intensity. Different light model systems have been
`used to simulate the light degradation of drug
`substances (Lachman et al., 1960; Lin and Lach-
`man, 1969; Griinert and Wollman, 1978). One of
`
`Correspondence: L. Gu, Institute of Pharmaceutical Sciences,
`Syntex Research, Palo Alto, CA 94304, U.S.A.
`
`these consists of a Rayonet Photochemical Reac-
`tor equipped with laboratory black lights (350
`nm). This system accelerates the degradation by
`~ 150 times compared to typical north window
`sunlight and by 2 75,000 times compared to typi-
`cal laboratory fluorescent light. It has been used
`successfully in evaluating the light reactivities of a
`series of compounds with different functionalities
`in this laboratory. This paper reports the light
`stability studies of ketorolac trometharnine,
`a
`potent non-narcotic analgesic agent (Muchowski
`et al., 1985; Bloomfield et al., 1984; Yee et al.,
`1984, 1985), in various aqueous and ethanol solu-
`tions in a Rayonet Photochemical Reactor (see
`Scheme 1).
`
`Materials and Methods
`
`Materials
`Ketorolac tromethamine was obtained from the
`
`Institute of Organic Chemistry, Syntex Research.
`
`0378-5173/88/$03.50 © 1988 Elsevier Science Publishers BV. (Biomedical Division)
`
`IPR2015-01099
`IPR2015-01097
`IPR2015-01100
`IPR2015-01105
`
`Lupin EX1171
`Page 1
`
`

`
`106
`
`Ethanol was USP grade and buffers were reagent
`grade. High-performance liquid chromatography
`(HPLC) grade acetonitrile and nanopure water
`were used to prepare the mobile phase.
`
`Kinetics
`
`Photolysis of ketorolac tromethamine was per-
`formed in a Rayonet model RPR 1000 Photo-
`chemical Reactor equipped with sixteen 350—nm
`black-light
`lamps. Sample solutions of ketorolac
`tromethamine in various media were prepared and
`transferred into a set of clear pyrex culture tubes
`(i.d. = 15 i 1 mm) before photolysis. For irradia-
`tion under oxygen or argon, sample solutions in
`culture tubes were purged with solvent-saturated
`gas for at least 10 min before sealing with Teflon-
`lined caps. A stability-specific HPLC method (see
`below) was used to follow the extent of the reac-
`tion.
`
`Preparation of degradation products I -4
`A stock solution containing 100 mg ketorolac
`tromethamine in EtOH/H20 (1/9, v/v) was pre-
`pared and transferred into eight 20 ml culture
`tubes. The solutions were purged with oxygen for
`5 min, sealed with Teflon-lined caps and irradia-
`ted with 350 nm lamps in a Rayonet Photochem-
`ical Reactor at 0°C (maintained using an ice
`bath) to ~ 80% completion. The reaction mixtures
`were combined, evaporated to dryness. and stored
`in a freezer before analysis. Samples were then
`dissolved in mobile phase and separated by semi-
`preparative HPLC (see HPLC methods below) to
`give ~ 4-6 mg of each of the degradation prod-
`ucts 1-4.
`
`Compound 1 was identified to be (:)-5-ben-
`zoyl-2,3-dihydro-1-hydroxy-3H-pyrrolo[1,2a]pyr—
`role:
`‘H NMR (300 MHz, CDCI3) 8 2.53-2.91
`(2H, m. C—CH2—C), 4.52 (2H, m, N-CH2), 5.27
`(1H, dd, -CH-OH), 6.18-6.85 (2H, dd, pyrrolic),
`7.47-7.83 (5H, m, Ph); EIMS (70 eV. m/e) 227
`(m/b).
`210.
`105,
`77; HRMS, Calcd.
`for
`CMHBNO2: 227.0946. Found: 227.0946. Anal.
`Calcd.: C, 73.99; H, 5.77; N, 6.16. Found: C,
`73.71; H, 5.82; N, 6.45.
`Compound 2 was identified to be (i)—5-ben-
`zoyl-2,3-dihydro-1-hydroperoxy-3H-pyrrolo[1,2a]-
`pyrrole: 1H NMR (300 MHZ, CDCI3) oz 2.73-2.85
`
`(2H, m, C-CH2—C), 4.52 (2H, m, N-CH3). 5.47
`(1H, dd, —CH—OOH), 6.27-6.99 (2H, dd, pyr-
`rolic), 7.47-7.83 (5H, m, Ph). 7.95 (1H, bs, -OOH);
`EIMS (80 eV. m/e) 243(m), 227, 210, 105(b), 77.
`Anal. Calcd. for CMHUNO3: C, 69.12; H, 5.39;
`N, 5.76. Found: C, 68.92; H, 5.17; N. 5.91.
`Compound 3 was identified to be 5-benzoy1-
`2,3-dihydro-1-oxopyrrolo[1.2a]pyrrole. The
`‘H
`NMR, MS and CHN analysis data have been
`reported elsewhere (Gu et al., 1987).
`Compound 4 was identified to be 5—benzoyl-
`1,2-dihydro-3H-pyrrolo[1,2a]pyrrole:
`‘H NMR
`(300 MHz. CDCI3) 8 2.57 (2H. p. -C-CH3-C).
`2.90 (2H,
`t, —CH2-C), 4.44 (2H,
`t. N-CH3-).
`5.94-6.81 (2H, dd, pyrrolic), 7.44-7.82 (5H. m,
`Ph); EIMS (80 eV.m/e)211(m/b).210,182. 134.
`105, 77. Anal. Calcd. for CMHBNO: C. 79.59. H.
`6.20: N, 6.63. Found: C. 79.59; H. 6.36: N. 6.52.
`
`Yield of C03
`A titration method (Niewenberg and Hegge,
`1951) was used in selected runs to determine the
`yield of the gaseous product CO2. First.
`the cul-
`ture tubes were fitted with a gas inlet-outlet de-
`vice. Sample solutions were then purged with CO3
`free (trapped by concentrated NaOH solution)
`argon gas for 10 min before irradiation. Argon gas
`was continuously passed through the reaction mix-
`ture into a CO2-free Ba(OH)2 solution. The
`trapped CO2 was then quantitated by titration of
`the Ba(OH)2 solution with standard HCl solution
`after photolysis was complete.
`
`Analytical methods
`The details of the reverse phase HPLC methods
`are described elsewhere (Gu et al.. 1987). Method
`A (used mainly for kinetic analysis) employed a
`C8 Ultrasphere (Altex) 5-it column (14.6 mm X 250
`mm) and a mobile phase of CH3CN/H20/HOAc
`(45/55/0.5). The flow rate was controlled at 1.0
`ml/min and the wavelength of detection at 314
`nm. Excellent linearities by area integration were
`obtained for ketorolac tromethamine and com-
`
`pounds 1. 3 and 4 (isolated pure materials were
`used as authentic materials) with injection sizes in
`the range of 0.015-1.0 pg. The correlation coeffi-
`cients and relative molar response factors thus
`obtained using 8 sample solutions are summarized
`
`Page 2
`
`

`
`in Table 1. Hydroperoxide 2 was unstable in mo-
`bile phase and degraded to 52% remaining after 24
`h at room temperature. Thus, the molar response
`factor for 2 (Table 1) was established using only
`two sample solutions (5.2 and 10.4 pg/ml, respec-
`tively).
`HPLC method B was used mainly for collection
`of degradation products. It employs a semi-pre-
`parative C8 Ultrasphere 5-‘u column (10 mm X 250
`mm) and a mobile phase identical to that used in
`method A.
`
`Results
`
`Material balance studies
`
`tromethamine was
`Photolysis of ketorolac
`studied in H20, H20/EtOH (9/1) and EtOH
`with laboratory black light (350 nm). Four decom-
`position products, 1-4. were isolated from HPLC
`(method B). Their structures were identified as
`shown in Scheme 1.
`
`O /
`
`O
`
`\
`
`N
`
`o
`
`..
`
`+
`
`C-0‘ NH3C(CH20H)3
`
`Ketorolac tromethamine
`
`107
`
`The material balance of the reaction as a func-
`tion of solvent and extent of reaction was de-
`
`termined using HPLC method A and the response
`factors of each degradation product shown in
`Table 1. The results are summarized in Table 2.
`
`When the photolysis was conducted in EtOH,
`94—l01% of the reacted ketorolac was accounted
`
`for regardless of the initial drug concentration, the
`extent of the reaction or the oxygen concentration.
`In deoxygenated samples, decarboxylation prod-
`uct 4 was the only observed product whereas
`under aerobic conditions, oxidation products 1-3
`accounted for ~ 30% of the degraded ketorolac
`(see Fig. la and Table 2). The relative yield of
`hydroperoxide 2 decreased with the extent of the
`reaction while the relative yields of 1 and 3 in-
`creased suggesting that 2 decomposed under the
`reaction conditions to 1 and 3.
`
`When the photolysis was conducted in H20,
`the distribution and the material balance of prod-
`ucts 1-4 were pH—dependent. At pH 2.0, where
`ketorolac exists mainly as a neutral species (free
`acid).
`the decarboxylated product, compound 4
`was the predominant product regardless of whether
`oxygen was present or not and > 80% of reacted
`ketorolac was accounted for by compounds 1-4.
`At pH 7.0 where ketorolac exists mainly as an
`anion, ondation products 1-3 became more pre-
`dominant as
`the reaction proceeded in air or
`oxygen saturated solutions. The yields of products
`1-4 at this pH decreased with the extent of the
`
`TABLE 1
`
`Linearity and molar response factors for ketorolac tromethamine
`and its degradation products at 3/ 4 nm
`
`Compound
`
`Ketorolac
`tromethamine
`
`Linearity
`correlation
`coefficient “
`
`0.9999
`
`0.9998
`—
`0.9999
`0.9999
`
`Molar
`response
`factor b
`1.00
`
`0.91
`0.98
`1.42
`1.02
`
`“ Eight concentrations and triplet injections.
`b Relative to ketorolac tromethamine.
`° Not determined due to the instability of the compound.
`
`Page 3
`
`

`
`108
`
`TABLE 2
`
`Results of laboratory black light (350 nm) photolysis of 10 pg/ ml ketorolac tromethamine solutions at various kinetic time points
`
`Solvent
`
`Atmosphere "
`
`Remaining
`
`Products distribution (96)
`
`(96)
`
`85
`10
`92
`10
`89
`13
`
`90
`11
`86
`23
`90
`55
`88
`7
`91
`
`pH 7.0 H20 °
`
`‘
`
`0.010 N 1-IC1 d
`
`Air
`
`Argon
`
`.
`
`.
`
`3
`
`4
`9
`11
`13
`—
`
`t
`
`51
`70
`41
`67
`6.5
`—
`10
`11
`t
`
`85
`85
`70
`73
`99
`98
`
`23
`8
`47
`20
`59
`100
`80
`77
`99
`
`Material
`
`balance
`(96)
`100
`101
`100
`96
`94
`95
`
`76
`46
`60
`60
`55
`48
`83
`80
`85
`
`" The desired atmosphere was purged into the photolysis media for at least 10 minutes prior to irradiation.
`b t means trace.
`° 0.025 M phosphate buffer.
`d pH = 2.05.
`
`Ketorolac
`
`Ketorolac
`
`B
`Time.min
`
`12
`
`16
`
`12
`
`15
`
`Fig. 1. HPLC chromatograms of photodegraded samples of ketorolac tromethamine in air-saturated (a) EtOH (30% remaining) and
`(b) H20 at pH 7.0 (36% remaining).
`
`Page 4
`
`

`
`reaction with as low as 46% accounted for after
`
`89% degradation in air. At pH 7.0, HPLC analysis
`of the degraded samples showed that additional
`degradation products which eluted near the solvent
`front were also found (Fig. lb). These apparently
`polar products were not identified in this study.
`It was noted that small amounts of 1 and 3
`
`were formed in argon-purged aqueous solutions at
`pH 7.0, presumably because argon purging of the
`aqueous solution did not remove all the available
`oxygen. After residual oxygen was removed by
`several freeze-thaw cycles, compound 4 was the
`only observed product in aqueous solutions.
`
`Yield of C02
`From the products isolated (Scheme 1), it was
`apparent that the decarboxylation of ketorolac is a
`major photoreaction pathway. Quantitative de-
`termination of the expected product CO2 was con-
`ducted in deaerated EtOH and H20 solutions,
`and the results are summarized in Table 3. In both
`
`solutions approximately 67% of the expected CO2
`was accounted for.
`
`Kinetics
`
`When EtOH was used as the solvent, an ap-
`parent first—order reaction was observed when the
`concentration of the trometharnine salt was s 2.0
`
`pg/ml (see Fig. 2a). However, at concentrations
`2 10 pg/ml where the photolysis solution was no
`longer optically thin (o.d. of the solution is < 0.03),
`the kinetics were non-first-order (Fig. 2b) and the
`t./,0 (time to reach 90% remaining) increased with
`increasing drug concentration. This concentration
`effect on the rate of photolysis agrees qualitatively
`with theoretical predictions (Mendenhall, 1984).
`In EtOH, oxygen appeared to quench the reac-
`
`TABLE 3
`
`C02 formation from photolysis of ketorolac tromethamine
`
`Solvent 3
`EtOH
`pH 7.0 H20 °
`
`Conc.
`1.0 mg/ml
`0.10 mg/ml
`
`% Remaining
`50
`78
`
`% of CO2 b
`67
`66
`
`“ Purged with CO2-free argon.
`b Based on the loss of ketorolac.
`° Phosphate buffer.
`
`
`
`Conc.uglmL
`
`-h0
`
`
`
`Conc.uglmL
`
`irradiation time, min.
`
`I
`16
`
`.1
`20
`
`Irradiation time, min.
`
`Fig. 2. Photolysis kinetics of ketorolac tromethamine in EtOH
`under argon (ID), air (0) or oxygen (0) atmosphere at (a) 2.0
`,u.g/ml and (b) 100 pig/ml drug concentrations.
`
`tion significantly at all concentrations studied, as
`evidenced by the decreasing rate obtained in going
`from argon to air to oxygen saturated solutions
`(Fig. 2).
`'
`The effect of oxygen concentration on the rate
`of the reaction in H20 was markedly different
`from that observed in EtOH. For example, at pH
`7.0 the initial photolysis rate in air or oxygen
`saturated solutions was similar to that in argon-
`saturated solution (see Fig. 3). However, at < 90%
`remaining, an autocatalysis reaction was observed
`in the air-saturated solutions which was not ap-
`parent in oxygen or argon-saturated solutions. The
`autocatalytic kinetic behavior was also observed in
`air-saturated aqueous solutions at pH 2.0.
`
`Page 5
`
`

`
`110
`
`Discussion
`
`Photodecarboxylation of aryl acetic acids and
`their salts has been reviewed extensively (Epling
`and Lopes, 1977; Coyle, 1978; Givens and Levi,
`1979). The primary photoprocess for the acid is
`believed to involve the singlet excited state of the
`acid which undergoes an a—bond (C—C(O)) clea-
`vage to yield an alkyl radical:
`
`Am}gcooH33(Am}nCooH)‘
`
`~ArCH;+'CO2H
`
`(1)
`
`Evidence for the radical mechanism include the
`
`detection of both alkyl and -COZH radicals from
`flash photolysis studies (Mittal et al., 1973; Meiggs
`et al., 1972) and the formation of radical coupling
`products (Epling and Lopes, 1977; Meiggs et al..
`1972)
`Although the singlet excited state of the salts
`was also believed to be involved,
`the primary
`photoreaction for the salts is less clear. Meiggs et
`al. (1972) have detected solvated electrons from
`flash photolysis of phenylacetic acid in water at
`pH 8.4 and thus indicated a radical pathway:
`
`mmxgcoo-33phcH,-+Co,+e*
`
`0)
`
`However, when sodium phenylacetate was photo-
`lyzed in the absence of oxygen in H20/(CH3),
`CDOH (99/1), 95% of the quantitative product,
`toluene, was not deuterium-incorporated. Since
`hydrogen abstraction by benzyl radical would oc-
`cur primarily at
`the C-D bond of
`the
`(CH3)2CDOH an ionic mechanism involving a
`benzyl anion followed by protonation appeared to
`be the dominating pathway (Epling and Lopes,
`1977).
`
`HK}fiCOO‘E:PhCH;—+CO2
`
`(3
`
`PhCH2"
`
`_._ _j——%>
`
`H20/(CH3)2CDOH
`(99/1)
`
`PhCH3 + PhCH2D
`95%
`
`(4)
`
`Photolysis of ketorolac tromethamine showed
`significant differences from simple aryl acetic acids
`and their salts in many ways. First,
`the benze-
`ylpyrrole structure in ketorolac has a similar con-
`jugation system to that of benzophenone which
`has a quantum yield of unity for the intersystem
`crossing (cbst)
`from singlet-excited state
`to
`triplet-excited state (Turro, 1979). The triplet en-
`ergy (ET)
`for benzophenone is 69 kcal/mol
`(Turro, 1979). If we assume that the ET for keto-
`rolac is close to 69 kcal/mol, then this energy is
`capable of breaking the C—C(O) (~ 68 kcal/mol)
`bond of ketorolac but not
`the C(O)—O (~ 106
`kcal/mol) or 0-H bond ( ~ 104 kcal/mol) (Weast,
`1972-1973) (Scheme 2).
`Thus, unlike most simple aryl acetic acids and
`salts the decarboxylation of ketorolac probably
`results from the triplet—excited state. This sugges-
`tion is supported by the experimental observation
`that
`the initial photolysis rate of ketorolac in
`EtOH and aqueous solutions was slower in the
`presence of oxygen (Fig. 2), which is an efficient
`triplet quencher (Foote, 1968).
`To test if the photolysis products of ketorolac,
`compounds 1—3 were formed from the secondary
`decomposition of the decarboxylated product 4
`(Crosby and Tang, 1969), compound 4 was sub-
`jected to the identical photolysis conditions as
`those used for ketorolac. Table 4 summarizes the
`
`results. No degradation could be found after pho-
`tolysis of 4 in EtOH for 10 min. Under identical
`conditions ketorolac tromethamine decomposed to
`25% remaining. Less than 2% degradation of 4 was
`observed after photolysis in H20 for 21 min while
`ketorolac tromethamine decomposed to 88% re-
`maining. We therefore conclude that compound 4
`is not an intermediate in the formation of the
`
`oxidation products 1-3. A mechanism which is
`consistent with all
`the experimental observations
`is outlined in Scheme 3.
`
`Page 6
`
`

`
`111
`
`ionic pathways. Oxygen has a dichotomous effect
`on the photolysis rate. Oxygen can slow down the
`reaction by quenching the triplet-excited state of
`ketorolac anion, but it can also react with alkyl
`radical
`I
`to form peroxy radical III (route b).
`When a good hydrogen donor solvent such as
`EtOH is used, peroxy radical III can abstract a
`hydrogen from EtOH to give hydroperoxide 2
`which can decompose thermally or photochem-
`ically (Lundberg 1961) to alcohol 1 and ketone 3.
`When a poor hydrogen donor solvent such as H 20
`is used peroxy radical III aggregates and eventu-
`ally initiates the free radical chain oxidation of
`ketorolac (route c). This thermal chain free radical
`oxidation is different
`from that base-catalyzed
`ionic
`autoxidation of ketorolac observed at
`
`elevated temperatures (Gu et al., 1987) and yields
`mainly 1—3 and some unidentified polar products.
`Thus, autocatalysis kinetics were observed in air-
`saturated aqueous solutions but not in EtOH.
`Quenching of triplet ketorolac by oxygen in
`oxygen-saturated aqueous solutions should be 5
`times more efficient
`than those in air—saturated
`
`solutions. Therefore,
`
`the accumulation of
`
`free
`
`TABLE 4
`
`Results of photolysis of ketorolac tromethamine and decarboxy/a-
`tion product 4 "
`
`Compound
`
`Solvent
`
`Ketorolac
`tromethamine EtOH
`EtOH
`
`4
`Ketorolac
`
`tromethamine pH 7.0 H20 b
`pH 7.0 H20 b
`
`4
`
`Photolysis Remaining ( %)
`time (min)
`
`10
`11
`
`21
`21
`
`25
`100
`
`88
`99
`
`" Substrate concentration = 100 pg/ml; ambient air.
`" Phosphate buffer.
`
`The discharge of an electron from the triplet-
`excited state of ketorolac anion followed by the
`cleavage of ot(C—C(O)) bond leads to the forma-
`tion of CO2 and alkyl radical I. In the absence of
`oxygen, the solvated electron can recombine with
`radical
`I
`to yield carbanion II (route a) which
`protonates rapidly in the presence of H20 or
`EtOH to give 4 as the major product. Thus, the
`proposed mechanism involves both radical and
`
`C
`
`Free radical oxidation
`
`/
`
`\
`
`Scheme 3
`
`I
`
`R
`
`\
`
`/
`
`N
`
`2
`
`v
`
`/\
`RN
`
`00H 3-
`
`or hi-
`
`\
`
`l
`
`R/41%/O0 HOH
`
`.
`
`go
`
`Page 7
`
`

`
`-§3
`
`MO
`
`
`
`Cone.uglmL
`
`40
`
`60
`
`Irradiation time, min.
`
`Fig. 3. Photolysis kinetics of 100 pig/ml ketorolac trometha-
`mine in H20 at pH 7.0 under argon (0) air (0) or oxygen
`(0) atmosphere.
`
`radicals in solutions is expected to be slower and
`autocatalysis was observed only with prolonged
`photolysis time (Fig. 3).
`Finally, it can be suggested that protonation of
`carboanion II is more favorable in acidic solutions
`
`(Crosby and Tang, 1969) and compound 4 re-
`mains to be the major product at pH 2 in the
`presence of oxygen (Table 2).
`
`Conclusions
`
`The pronounced effect of oxygen on the pho-
`tolysis kinetics and product distribution of l<eto—
`rolac tromethamine has led to the identification of
`
`both radical and ionic reaction pathways. The
`results presented in this study may be applied to
`other
`acidic non-steroidal
`analgesic /anti-in-
`flammatory agents containing substituted arylace-
`tic acid structure (e.g., naproxen,
`indomethacin,
`ibuprofen, ketoprofen. etc.).
`
`Acknowledgements
`
`The authors thank Mrs. J . Nelson for obtaining
`the ‘H NMR spectra, Dr. L. Partridge for obtain-
`ing the mass data and Mr. M. Chung for obtain-
`ing the CHN analysis data. Helpful discussions
`from Drs. W. Lee, and M. Powell are appreciated.
`
`References
`
`Bloomfield, S.S., Mitchell, J., Cissell, G. and Borden, T.P.,
`RS-37619 and aspirin analgesia for post-partum uterine
`cramps. Clin. Pharmacol. Ther.. 35 (1984) 228.
`Coyle, J.D., Photochemistry of carboxylic acid derivatives.
`Chem. Rev., 78 (1978) 97-123.
`Crosby, D.G. and Tang, C.-S., Photodecomposition of l-naph-
`thalene acetic acid.
`J. Agr. Food Chem.,
`17
`(1969)
`1291-1293.
`Epling, G.A. and Lopes, A., Fragmentation pathways in the
`photolysis of phenylacetic acid. J. Am. Chem. Soc, 99
`(1977) 2700-2704.
`Foote, C.S., Photosensitized oxygenations and the role of sing-
`let oxygen. Acc. Chem. Res, 1 (1968) 104-110.
`Givens. R.S. and Levi, N., The photochemistry of organic
`acids, esters. anhydrides, lactons, and imides. In S. Patai
`(Ed.), The Chemistry of Acid Derivatives, Part
`I, Wiley.
`1979. pp. 641-753.
`Granert. R. and Wollman, H., Experimental exposure of
`pharmaceutical solutions to light. Zbl. Pharm.. 117 (1978)
`993-997.
`Gu, L., Chiang, H.-S., and Becker, A. Kinetics and mecha-
`nisms of the autoxidation of ketorolac tromethamine in
`
`aqueous solution. Int. J. Pharm., 41 (1988) 95-104.
`Lachman, L.. Swartz, C1. and Cooper, J., A comprehensive
`pharmaceutical stability testing laboratory 111. A light sta-
`bility cabinet for evaluating the photosensitivity of phar-
`maceuticals. J. Am. Pharm. Assoc. 49 (1960) 213-218.
`Lin, S.-L. and Lachman. L., Photochemical considerations of
`parenteral products. Bull. Parent. Drug Assoc. 23 (1969)
`149-165.
`Lundberg. W.D., Autoxidation and Antioxidants. Vol. I. Wiley,
`1961.
`
`Ind.
`
`Meiggs, T.D. Grossweiner, L.I. and Miller, S.l., Extinction
`coefficient and recombination rate of benzyl radicals.
`l.
`Photolysis of sodium phenylacetate. J. Am. Chem. Soc, 94
`(1972) 7981-7986.
`Mendenhall, D.W., Stability of parenterals. Drug Dev.
`Pharm.. 10 (1984) 1297-1342.
`Mittal, L.J., Mittal, JP. and Hayon, E., Photo-induced de-
`carboxylation of aliphatic acids and esters in solution.
`Dependence upon state of protonation of carboxyl group.
`J. Phys. Chem, 77 (1973) 1482-1487.
`Muchowski, J.M., Unger, S.H.. Ackrell, 1., Cheung, P., Cooper.
`G.F., Cook, J., Callegra. P., Halpern. 0.. Koehler. R..
`Kluge, A.F., Van Horn, A.R., Antonio. Y., Carpio.
`1-1,.
`Franco, F., Galeazzi, E., Garcia, 1., Greenhouse. R., Guz-
`man, A., lriate, J., Leon, A., Pena, A., Perez. V.. Valdez, D.,
`Ackerman, N.. Ballaron, S.A.. Murthy, D.V.K., Rovito,
`J.R.. Tomolonis, A.J., Young, J.M., and Rooks ll. W.l-l..
`Syntheses and anti-inflammatory and analgesic activity of
`5-aryl-1,2—dihydro-3H-pyrrolo[1.2-a]pyrrole-1-carboxylic
`acids and related compounds. J. Med. Chem. 28 (1985)
`1037-1049.
`
`Niewenberg, C.J.V. and Hegge, L.A., Absorption of carbon
`dioxide by baryta in an organic solvent. Anal. Chem. Acta,
`5 (1951) 68-70.
`
`Page 8
`
`

`
`Turro, N.J., Modern Molecular Photochemistry, Benjamin, New
`York, 1979, 352 pp.
`Weast, R.C., CRC Handbook of Chemistry and Physics, 53rd
`edn. Chemical Rubber, OH 1972—1973, F189.
`Yee, J., Brown, C.R., Allison, C. and Wildman, K., Analgesia
`from intramuscular ketorolac tromethamine compared to
`
`morphine (MS) in “severe” pain following “major” surgery.
`Clin. Pharmacol. That, 37 (1985), 239.
`Yee, J., Brown, C.R., Sevelius, H. and Wild, R.N., The analgesic
`Efficacy of (1-)-5-Benzoyl-1,2-Dihydro-3H-Pyrrol[l,2a]Pyr-
`role—l-Carboxylic Acid, Tromethamine Salt
`(B)
`in Post-
`Operalive Pain. Clin. Pharmacol. Ther., 35 (1984) 285.
`
`113
`
`Page 9