Br. J. clin. Pharmac. (1983), 16,271-280
`
`EFFECTS OF TOBACCO SMOKING AND ORAL
`CONTRACEPTIVE USE ON THEOPHYLLINE DISPOSITION
`
`M.J. GARDNER, K.M. TORNATORE & W.J. JUSKO
`Department of Pharmaceutics, School of Pharmacy, State University of New York at Buffalo,
`New York 14260, USA
`R. KANARKOWSKI
`Department of Biopharmaceutics, Medical Centre of Postgraduate Education, Faculty of Pharmacy,
`Bydgoszcz, Poland
`
`1 The independent as well as interactive effects of chronic (> 6 months) oral contraceptive (OC) use
`and cigarette smoking on single-dose (4 mg/kg) theophylline disposition were assessed in 49 young,
`healthy women.
`Significant elevations (40%) in theophylline plasma clearance were found in women who smoked.
`2
`OC use resulted in decreases in clearance of a similar magnitude (28%). These factors do not appear
`to interact with respect to theophylline disposition. The combination of main effects tended to cancel
`one another (clearance of49.1 ml h-I kg-'ideal body weight for OC non-user, non-smoker, vs 49.7 ml
`h-' kg-'for OC user-smoker).
`Single dose exposure to OC in non-users did not significantly alter theophylline pharmacokinetics
`3
`for the group as a whole. However, in the subgroup of smoking subjects, significant decreases in
`clearance were evident (P < 0.05). Analogous results were found for half-life. Volume of distribution
`was slightly diminished in smokers, but was unaffected in OC users.
`Areas under the serum concentration-time (AUC) profiles of norgestrel and ethinyloestradiol
`4
`were examined in 27 women as indices of OC exposure. The smallest values of theophylline clearance
`were found in the subjects with largest AUC of both OC steroids.
`Appropriate statistical analyses of data which are influenced by multiple factors are discussed.
`5
`Special concern is needed when the factor partitioning process yields subgroups of unequal sizes.
`Keywords smoking theophylline
`oral contraceptives
`
`Introduction
`
`Oral contraceptives (OC) are likely to alter drug dis-
`position and clinical effects because of their frequent
`and protracted use and known capability to affect
`biotransformation rates. Many oxidative pathways
`such as for antipyrine (Carter et al., 1974; Teunissen
`et al., 1982), and theophylline (Tornatore et al., 1982)
`and reductive processes such as for nitrazepam
`(Jochemsen et al., 1982) exhibit decreased metabo-
`lism rates. On the other hand, the conjugation of
`lorazepam and oxazepam is increased in OC users
`(Patwardhan et al., 1981).
`Another ubiquitous source of altered drug elimina-
`tion is tobacco smoking which seems to cause selec-
`tive enzyme induction and increased biotransforma-
`tion rates of some drugs (Jusko, 1978). Marked
`increases in theophylline clearances are exhibited in
`some smokers (Jusko et al., 1978).
`
`The dual effects of smoking and oral contraceptive
`use are of special interest. The combined use of
`tobacco and OC is now well recognized as contribu-
`tary to increased risk of cardiovascular disease and
`mortality (Beral & Kay, 1977; Stadel, 1981). The
`nature of any combined effects of these factors on
`hepatic function and drug metabolism has not been
`elucidated.
`Theophylline has a low therapeutic index and is
`subject to extensive hepatic metabolism (Cornish &
`Christman, 1957), the rate of which is sensitive
`to numerous drug/disease/physiologic/environmental
`factors (Jusko et al., 1979). With OC and tobacco use
`causing opposite effects on theophylline clearances in
`man, it is of clinical and pharmacological interest to
`examine the net effects of these factors. Another
`purpose of this report is to consider the appropriate
`
`271
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`272
`
`M.J. GARDNER ETAL.
`
`statistical methods used in assessing the influences of
`both independent and interactive factors on drug dis-
`position.
`
`Methods
`
`Subjects
`
`The data collected from 49 young, healthy women
`who had been studied at the Millard Fillmore Hos-
`pital of Buffalo were employed in this analysis. A
`portion of these data (22 subjects) was extracted from
`a data base constructed between 1975-1979 whereas
`the remainder was collected subsequently through
`1981. These were essentially part of continuous studies
`of factors affecting theophylline disposition in man
`and identical drug, clinical, record, assay, and phar-
`macokinetic procedures were involved. Protocols
`were approved by university and hospital Human
`Investigation Committees and all subjects provided
`informed consent in participating.
`The ages of the subjects ranged from 19 to 30 years.
`None of the women examined were considered to be
`obese (i.e., total body weight did not exceed esti-
`mated ideal body weight (Diem & Lentner, 1970) by
`more than 15%), abusers of alcohol, excessively
`habituated to caffeine, or suffering from any under-
`lying hepatic/renal disease. Sixteen of the subjects
`admitted to smoking marihuana in the months prior
`to participation. In such instances, however, the
`amount of the substance smoked did not exceed that
`normally associated with social use (i.e., less than one
`joint per week). Twenty-two subjects used OC on a
`chronic basis (i.e., longer than 6 months). Of these,
`14 were using Ovral (Wyeth Laboratories: 0.05 mg of
`ethinyloestradiol/0.50 mg norgestrel) at the time
`of study, while two were using Ovulen (Searle Co.:
`0.10 mg mestranol/1.0 mg ethynodiol diacetate). Six
`subjects in the former group had been taking this
`product routinely, whereas eight were using various
`other commercially available combinations and
`switched to Ovral only for the monthly cycle prior to
`study. Information concerning the OC type used by
`the remaining six subjects was not available.
`Twenty-seven subjects used tobacco daily. The
`average daily use exceeded 20 cigarettes per day,
`while the duration of use spanned a period in excess
`of 7 years.
`Table 1 presents the demographic data for the sub-
`jects considered. The group has been partitioned into
`four disjoint subgroups based upon tobacco and OC
`use.
`
`Posology and assays
`
`Subjects were requested to fast overnight prior to
`participation in the investigation. In addition, all
`
`xanthine containing foods/beverages were deleted
`from their normal diets both during the study day and
`for the 24 h period preceding examination. In the
`morning of the study day, theophylline was adminis-
`tered orally as an aqueous solution of its ethylene-
`diamine salt, aminophylline. A weight adjusted dose
`(4 mg'kg) of this salt was placed in 200 ml of orange
`juice, which was then consumed by the subject.
`Venous blood samples (7 ml) were serially collected
`from an indwelling catheter at times: 0 (just before
`administration of the drug), 0.5, 1, 2, 3, 5, 7, 11 and
`23 h after the dose was given. The patency of this
`catheter was maintained using a 20 unit/ml solution of
`heparin in normal saline. When slight deviations from
`this sampling protocol occurred, such sampling times
`provided for adequate estimations of both area under
`the serum concentration versus time profile and the
`slope of the terminal disposition phase. The samples
`were allowed to clot and then were centrifuged.
`Serum was harvested and frozen pending analysis via
`high performance liquid chromatography (Jusko &
`Poliszczuk, 1976). For 27 subjects, a single Ovral
`tablet was ingested 60 min before the test dose of
`aminophylline. For 11 of these women who were
`using OC as a method of birth control, this tablet was
`their routine daily intake of contraceptive steroids.
`For the remaining 16, such treatment represented
`single dose exposure. In these cases, two additional
`blood samples were collected, corresponding to the
`zero time with respect to OC administration and 0.5 h
`post-administration. For these 27 subjects, all serum
`samples were assayed for both ethinyloestradiol and
`norgestrel (Cook et al., 1974; Stanczyk et al., 1975).
`Appropriate time corrections were implemented
`where necessary. Contraceptive steroid serum con-
`centrations were not measured for the remaining 22
`subjects.
`
`Pharmacokinetics
`
`Area under the serum concentration vs time curve
`(AUC) was estimated either by numerical inte-
`gration using the trapezoidal rule or via mathematical
`integration of fitted interpolating polynomials (Yeh
`& Kwan, 1978) in combination with log-trapezoidal
`approximation. Both approaches provided for inclu-
`sion of the terminal area extrapolated to infinite time.
`Plasma clearance was obtained from the relationship,
`CL = Dose/AUC. Since it has been shown that theo-
`phylline in solution is completely absorbed from the
`gastrointestinal tract following oral administration
`(Hendeles et al., 1977), bioavailability was assumed
`to be unity. Clearance was normalized for both total
`body weight (TBW) and ideal body weight (IBW),
`the latter estimated from body frame size and height
`(Diem & Lentner, 1970). The slope of the post-
`absorptive decline of serum concentrations (A) was
`obtained via log-linear least-squares regression analy-
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`SMOKING AND OC STEROIDS AFFECT THEOPHYLLINE CLEARANCE
`
`273
`
`Table I
`
`Characteristics of subjects
`
`Characteristics
`
`Group:
`
`Non-smokers
`I
`Non-users
`
`II
`OC users
`
`Smokers
`
`III
`Non-users
`
`IV
`OC users
`
`15
`19-30
`24.1
`(3.3)
`59.9
`(7.9)
`0
`
`1.2
`(0.4)
`7.5
`(3.6;15)
`0.47
`0.93
`4.8
`(2.4)
`0.49
`(O. 19;15)
`14.1
`(6.0;14)
`20.7
`(7.5;15)
`58.7
`(15.6;15)
`
`12
`21-29
`24.7
`(2.5)
`55.7
`(6.9)
`6.8
`(2.8;3)
`1.1
`(0.4)
`5.9
`(3.2;10)
`0.50
`0.92
`2.0
`(0.1)
`0.45
`(0. 18;11)
`11.9
`(9.4;8)
`23.3
`(9.8;11)
`50.3
`(12.2;11)
`
`10
`19-28
`24.2
`(2.6)
`59.3
`(8.2)
`4.5
`(2.5;8)
`
`0 0
`
`0.20
`1.0
`2.1
`(1.3)
`0.42
`(0.23;10)
`19.0
`(9.3;9)
`21.6
`(12.9;10)
`47.7
`(15.0;10)
`
`Number of subjects
`Age range (years)
`Mean age (years)
`(s.d.)a
`Total body weight (kg)
`(s.d)
`Duration of OC use (years)
`(s.d. ;n)b
`Tobacco use (pk day)
`(s.d.)
`Duration of tobacco use (years)
`(s.d. ;n)
`Marihuana usec
`Alcohol usec
`Caffeine used
`(s.d.)
`Bilirubin (0.1-1.1 mg%)e
`(s.d. ;n)
`SGPT (0-43 u,l)e
`(s.d. ;n)
`SGOT (11-46 u,I)e
`(s.d. ;n)
`Alkaline phosphatase (34-133 ul)e
`(s.d. ;n)
`a Standard deviation
`b Standard deviation; number of subjects
`c Coded as 0 = none, 1 = social, 2 = daily use
`d Number of cups of coffee or tea per day
`e Normal range in parentheses
`
`12
`22-28
`24.4
`(2.3)
`54.5
`(4.7)
`
`0 0 0
`
`0.17
`1.0
`2.0
`(1.8)
`0.71
`(0.20;12)
`17.8
`(7.3;12)
`18.3
`(5.6;12)
`53.7
`(16.7;12)
`
`sis. The apparent volume of distribution (VD) was
`then generated as VD = CLA and subsequently
`normalized for TBW.
`
`Statistics
`
`The disparate backgrounds of the subjects with re-
`spect to factors known to affect theophylline disposi-
`tion such as caffeine (Mitoma et al., 1969) and
`marihuana (Jusko et al., 1978) precluded a direct
`assessment of OC and smoking effects on theophyl-
`line disposition via a two-way analysis of variance. An
`additional factor lacking adequate control was re-
`lated to the fact that 16 of the subjects examined were
`not chronic users of OC, yet were exposed to a single
`OC dose shortly before aminophylline administra-
`tion. Consequently, a four-way analysis of covariance
`was deemed the most appropriate statistical approach
`to pursue. The main effects were defined as cigarette
`smoking, OC use, marihuana use, and acute ex-
`
`posure to contraceptive steroids. Daily caffeine
`intake served as the covariate. This method of analysis
`was chosen over a two factor approach (cigarettes and
`OC) with a more stringent subject selection process,
`since the technique of subject matching is not always
`optimal and often results in a dramatic reduction in
`the number of data available for consideration
`(McKinlay, 1977). In addition, confirmation of
`various factor effects on theophylline disposition
`tends to enhance the credibility of the particular an-
`alysis conducted.
`Since the four-factor partitioning process resulted
`in subgroups of unequal sizes, the main effects were
`rendered dependent upon one another. As a conse-
`quence, additivity of individually determined sums of
`squares did not hold. This necessitated pursuing a
`hierarchial approach, whereby all main effects were
`assigned status values relative to one another. With
`this form of analysis, sums of squares were adjusted
`only for factors possessing higher priorities, those
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`sidered are presented in Figure 1. In each instance,
`absorption is rapid, peak concentrations being attained
`within 2 h following administration of the dose. Post-
`absorptive declines of serum concentration reflected
`apparent first order processes.
`
`Non-smoker
`Non-OC user O
`OC user o
`1
`
`Smoker
`8
`
`10 _
`
`5
`
`2X
`0
`
`0.2
`
`E 0
`
`.
`
`5
`
`0.11
`0
`
`III
`10
`15
`25
`20
`Time (h)
`Figure 1
`Representative serum concentration vs time
`profiles of theophylline for subjects in each of the four
`groups examined. The symbols are defined in the key.
`Solid lines are the results of least-squares fittings of data
`in terminal phases.
`
`The results of the statistical analysis, conducted
`with theophylline plasma clearance normalized for
`IBW as the dependent variable, are exhibited in
`Table 3. The same type of analysis, performed with
`clearance values adjusted for TBW, resulted in quali-
`tatively similar conclusions (i.e., the rank order of
`significant main effects was maintained but the levels
`of significance were found to be somewhat lower). As
`indicated in Table 3, cigarette smoking, chronic OC
`ingestion, and daily caffeine intake all significantly
`alter theophylline clearing processes in the subjects
`examined. On the other hand, the social use of canna-
`bis and acute exposure to contraceptive steroids both
`failed to induce remarkable changes in this para-
`meter. The latter finding suggests.,at aktheclearance
`data presented for smoking and nonsmoking groups
`of non-OC users (groups 1 and 3) have not been
`rendered misrepresentative as a consequence of the
`data blending process.
`Figure 2 illustrates the effects that cigarette
`smoking and OC use have upon theophylline clear-
`
`274
`
`M.J. GARDNER ETAL.
`
`with lower statuses having no effect (Overall &
`Spiegel, 1969). Cigarette smoking was given the
`highest priority due to its well documented effect on
`theophylline disposition (Hunt et al., 1976; Grygiel &
`Birkett, 1981). This was followed by daily caffeine
`intake, OC use, use of cannabis, and acute exposure
`to contraceptive steroids. This ordering of main
`effects reflects our suspicions with regard to the abili-
`ties of these factors to alter theophylline disposition.
`Although the net effect of acute contraceptive steroid
`exposure was unknown, its effect, if any, was not
`anticipated to be as dramatic as those induced by the
`other factors, and consequently was assigned the
`lowest priority. The covariate, caffeine use, was
`introduced into the analysis at the second stage. This
`provided for the removal of extraneous caffeine re-
`lated influences prior to making assessments regard-
`ing the contributions of OC use, marihuana smoking,
`and acute exposure to OC. Such an ordering did
`permit caffeine effects to confound the analysis of
`tobacco related influences. It was felt, however, that
`tobacco effects would be considerably stronger than
`those of caffeine and hence any contamination of this
`nature would be minimal. Generally, it is desirable to
`feather out any covariate influences prior to assess-
`ment of all main effects. However, interchanging the
`positions of tobacco use and caffeine intake in the
`analysis could result in an exaggerated estimate of
`caffeine effect, since it would be entered into the
`analysis first and was shown to be significantly cor-
`related with the magnitude of the subject's tobacco
`habit.
`All two-way interactions were evaluated. Each re-
`spective sum of squares was corrected for the in-
`fluences of all main effects and all other two-way
`interactions. The question regarding higher order in-
`teractions was not addressed.
`Since a pre-written digital computer program
`capable of performing the type of analysis desired was
`unavailable,
`appropriate sums of squares were
`generated following the fitting of coded data to
`various linear regression models (Kim & Kohout,
`1975).
`This approach to data analysis was employed in
`order to assess the effects of these factors on theo-
`phylline clearance, volume of distribution, and half-
`life (tJ2).
`Least-squares fitting of serum concentration-time
`data was conducted employing conventional log-
`linear regression theory (Yeh & Kwan, 1978). All
`other bivariate regression analyses were performed
`recognizing the fact that both variables involved were
`subject to random error (Riggs et al., 1978).
`
`Results
`Serum concentration vs time profiles for individual
`subjects representing each of the four groups con-
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`SMOKING AND OC STEROIDS AFFECT THEOPHYLLINE CLEARANCE
`
`275
`
`Table 2 Summary of theophylline pharmacokinetics in smoking and non-smoking
`female subjects who were users/non-users of oral contraceptives
`
`Parameter
`(s.d.)
`
`Grc
`oup:
`
`Non-smokers
`I
`Non-users
`
`II
`OC users
`
`Smokers
`
`III
`Non-users
`
`IV
`OC users
`
`9.57
`(1.34)
`0.55
`(0.13)
`39.4
`(10. 1)
`37.2
`(11.8)
`
`5.13
`(1.09)
`0.51
`(0.06)
`74.3
`(17.6)
`68.7
`(17.8)
`
`6.22
`(1.53)
`0.46
`(0.10)
`49.7
`(22.3)
`54.0
`(21.8)
`
`vations were confirmed statistically. In addition,
`visual inspection of the data in this figure suggests that
`perhaps a slight interactive effect occurs between
`cigarette smoking and OC use. The average clearance
`of group 4 (smokers, OC users) approaches control
`values (49.7 ml/h/kg IBW vs 49.1 ml/h/kg IBW) in
`spite of the apparent differences in the magnitudes of
`the individual main effects. However, when the
`underlying effects of caffeine and other factors were
`removed, this interaction was shown not to attain
`statistical significance (P > 0.3). This suggestion of
`
`0 S 00
`
`0
`
`a U *
`
`.a
`
`U U
`
`U
`
`U U
`
`000
`
`000
`
`0 00
`
`0 0
`
`0 0 o
`
`n
`
`0
`
`b
`
`_
`
`t112 (h)
`VD (I/kg TBWa)
`CL (ml h-' kg- IBWb)
`
`7.76
`(2.39)
`0.55
`(0.09)
`49.1
`(14.3)
`51.0
`(12.7)
`a Normalized for total body weight
`b Normalized for ideal body weight
`
`CL (ml h-' kg-' TBWa)
`
`ance in the entire group of subjects. These data have
`been normalized for both TBW and IBW and are
`slightly contaminated by the disparate use of caffeine.
`Regardless of the normalization approach, the data
`suggest that OC use exerts a significant inhibitory
`effect on theophylline clearance, whereas cigarette
`smoke exposure markedly enhances theophylline's
`hepatic metabolism. The effect of tobacco use
`appears. to be more pronounced than that of OC
`exposure when clearance values are normalized for
`IBW vs TBW. As indicated previously, these obser-
`a120r
`
`0
`
`0
`-A0-
`0
`
`0
`
`a
`
`an
`
`U U
`
`I U1
`
`mu
`
`U
`
`o
`
`00
`
`0
`
`0
`
`00 0
`
`OC
`00
`
`100
`
`80
`
`60 -
`
`40 -
`
`20-
`
`CD
`
`_
`
`- 0
`
`c
`)
`C.)
`a
`
`0.
`
`c 0-
`
`c
`I-
`
`olI
`OC: NU
`Smoking:
`S
`NS
`NS
`S
`Plasma clearances of theophylline in the four designated subject groups. Clearances were normalized for
`Figure 2
`a) total body weight and b) ideal body weight. Horizontal lines denote mean values for each group. U user, NU
`non-user, S smoker and NS non-smoker.
`
`U
`
`NU
`
`I
`U
`
`I
`
`NU
`
`I
`U
`
`NU
`
`I
`U
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`0 0 0
`
`0
`
`0
`
`ocgOCP0
`t~~~
`*
`-
`0~~~~~~~~00
`:S
`oa
`es
`
`0
`
`0
`
`NU
`
`U
`
`NU
`
`U
`
`NS
`
`S
`
`276
`
`M.J. GARDNER ETAL.
`
`14r
`
`121-
`
`10
`
`8
`
`6
`
`4
`
`_)
`1-
`
`I
`
`0
`
`I
`
`U
`
`0
`
`8
`
`0
`
`a
`
`*
`
`*
`*
`
`0
`
`0 o
`
`u 0%
`
`0
`
`0,
`
`0
`
`.0
`
`in
`4-
`
`._
`
`1.0r
`
`0.81-
`
`0.61
`
`0.4
`
`0.2
`
`I
`OC: NU
`Smoking:
`
`U
`
`NS
`
`I
`
`NU
`
`S
`
`Figure 3 Volume of distribution and half-life of theophylline in the four subject groups. Horizontal lines denote
`mean values. U user, NU non-user, S smoker and NS non-smoker.
`
`potential interaction is not as evident when normal-
`ization is conducted using TBW. In this case, the
`main effects of OC and tobacco use are significant but
`appear to be simply additive.
`The apparent volume of distribution of theophyl-
`line was unaffected by chronic OC exposure (Figure
`3). However, a small but significant reduction in this
`parameter was detected in the cigarette smoking
`group (P < 0.05). This was the only factor considered
`in the model which exerted such an effect on the
`distributional space. Figure 3 also presents the effect
`of these two main factors on theophylline half-life.
`As with plasma clearance, both cigarette smoking
`(shorter t112) and OC (longer t%) effects were deter-
`mined to be significant. The effect induced by
`cigarette use on t12 is the combined result of both
`increased clearance and reduction of distribution
`volume, whereas the OC effect is primarily mediated
`via inhibition of theophylline elimination. In parallel
`with the results of the clearance data analysis, both
`caffeine intake and the interaction between cigarettes
`and acute exposure to OC were shown to have signifi-
`cant effects. A summary of the pharmacokinetic
`parameters for each group is contained in Table 2.
`The effects of OC components on theophylline
`clearance is displayed in an exposure-response format
`in Figure 4. Areas under the serum concentration-
`time profiles were employed as indices of OC ex-
`posure. The smallest values of theophylline clearance
`were noted in the subjects with largest AUC of both
`norgestrel and ethinylestradiol.
`
`Discussion
`
`The present investigation assessed the independent
`effects of daily OC ingestion and cigarette smoking as
`well as the potential interactive effect of these and
`other factors on theophylline disposition in young
`women. Both cigarette smoke exposure and daily OC
`intake significantly alter theophylline disposition. In
`particular, tobacco use enhances the metabolism of
`theophylline, while chronic OC exposure inhibits the
`clearance of this bronchodilator. Together, the two
`factors offset each other.
`
`Primary smoking and OC effects
`
`The marked enhancement of theophylline clearance
`associated with cigarette use confirms the original
`findings of Jenne et al. (1975) and Hunt and co-
`workers (1976). Although this increase in metabolic
`activity is evident regardless of the normalization
`approach taken (IBW vs TBW), Gal and associates
`(1978) found the use of the subject's IBW more
`appropriate. Examination of the clearance data
`revealed that an increase of roughly 40% was
`characteristic of the cigarette smoking group. These
`values are in reasonable agreement with those pre-
`viously reported from this laboratory (Jusko et al.,
`1978), although we now have more selective inclusion
`of subjects as the original report did not control OC
`and caffeine use. Increases in CL are likely due to the
`inductive effects of cigarette smoke components such
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`SMOKING AND OC STEROIDS AFFECT THEOPHYLLINE CLEARANCE
`
`277
`
`Table 3 Summary of statistical analysis conducted to assess the effects of various factors on theophyl-
`line plasma clearance normalized for IBW
`
`Source of variation
`
`Sum of
`squares
`
`d.f
`
`Mean
`square
`
`Covariates
`Caffeine (Caf)
`1408.62
`7303.88
`Main effects
`Cigarettes (65.2;44.7)b
`5129.13
`OC use (47.3;63.1)
`1909.96
`Marijuana use (MJ)
`186.02
`Acute exposure (AE)
`78.77
`3569.47
`Two-way interactions
`0.26
`Cig-Caf
`Cig-OC
`152.28
`20.30
`Cig-MJ
`2305.26
`Cig-AE (59.8;53.4)
`10.36
`Caf-OC
`Caf-MJ
`51.77
`389.72
`Caf-AE
`199.40
`OC-MJ
`67.23
`MJ-AE
`12281.96
`Explained
`7038.09
`Residual
`19320.05
`Total
`a Given as P-value for F-distribution with appropnate degrees of freedom
`b (Average CL for group indicated; average CL for complementary group)
`cP>0.05
`
`1408.62
`1825.97
`5129.13
`1909.96
`186.02
`78.77
`396.61
`0.26
`152.28
`20.30
`2305.26
`10.36
`51.77
`389.72
`199.40
`67.23
`877.28
`207.00
`
`1
`4
`1
`1
`1
`1
`9
`1
`1
`1
`1
`1
`1
`1
`1
`1
`14
`34
`
`F
`
`6.80
`8.82
`24.78
`9.23
`0.90
`0.38
`1.92
`0.01
`0.74
`0.10
`11.14
`0.05
`0.25
`1.88
`0.96
`0.33
`4.24
`
`Significance
`ofP
`
`0.013
`< 0.001
`< 0.001
`< 0.005
`NSC
`NS
`0.082
`NS
`NS
`NS
`0.002
`NS
`NS
`NS
`NS
`NS
`< 0.001
`
`as the polycyclic aromatic hydrocarbons on distinct
`forms of cytochrome P-450 (Jusko, 1978).
`Daily OC exposure was shown to markedly inhibit
`the clearing processes responsible for theophylline
`removal from plasma, although the magnitude of this
`effect was not as striking as that elicited by the use of
`tobacco. Subjects employing this form of birth con-
`trol were shown to have clearance values which aver-
`aged approximately 28% below those exhibited by
`non-OC users. This difference is similar to that
`reported previously in a subgroup of non-smoking
`(Tornatore
`1982).
`normal
`volunteers
`al.,
`et
`Patwardhan and associates (1980) presented results
`which indicate that similar metabolic inhibition
`occurs following the single-dose administration of
`caffeine which is generally sensitive to the same
`affecting theophylline
`factors
`disposition.
`Daily
`exposure to OC for periods longer than 6 months in
`nonsmoking subjects resulted in a 40% reduction in
`caffeine plasma clearance. Also, examination of
`blood concentration data obtained from women using
`OC have led others to note concomitant elevations in
`plasma concentrations of caffeine and ethinylestra-
`diol (Ahluwalia et al., 1977; Kaul et al., 1981). Since
`theophylline is almost exclusively eliminated via
`hepatic biotransformation pathways, it is reasonable
`to conclude that such inhibitory behavior reflects
`either specific competition for metabolic sites or
`
`generalized depression of hepatic function secondary
`to contraceptive steroid exposure. The primary re-
`sults of the present analysis should be considered
`qualitative in nature, and the numerical values
`presented should serve only as an approximation of
`the magnitude of any metabolic perturbations which
`might occur during chronic dosing of patients with
`theophylline.
`Crawford and co-workers (1981) reported that
`cigarette smoking had no effect on plasma concentra-
`tions of either ethinyloestradiol or norgestrel when
`compared to nonsmoking subjects. Similar observa-
`tions have been made in our laboratory with respect
`to the plasma clearances of these compounds (to be
`published). The results of the present investigation as
`well as those of others have shown that the primary
`metabolic pathways responsible for theophylline
`degradation are very sensitive to such exposure, thus
`suggesting minimal overlap of metabolic mechanisms
`or pathways. The observation that acute exposure to
`OC agents apparently has no significant effect on
`theophylline clearance may partly support this con-
`tention. However, the possibility does exist that a
`competitive effect occurs, but that single dose
`exposure to these agents does not produce sufficient
`concentrations in the milieu of the metabolic sites to
`elicit a discernible effect.
`The most likely explanation for this inhibitory
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`278
`
`M.J. GARDNER ETAL.
`
`100
`
`80
`
`60
`
`40
`
`s 20-0
`
`0
`
`00
`
`*
`0'm
`z
`
`a
`
`r-
`
`*
`
`\
`
`°0
`
`\
`°
`
`0
`
`*
`
`-19
`
`c
`
`<1> 100 -
`('
`100
`5
`O
`
`80
`
`40 -
`
`20
`
`0.o2
`0.5
`5
`2
`Total norgestrel AUC x 10-2 (ng ml' h)
`
`,
`
`.,
`
`1
`
`10
`
`*
`
`0
`
`0
`
`0
`
`*
`*
`
`0
`
`* \XX
`
`0
`
`0
`
`0
`
`00
`
`60
`
`0
`\
`
`0
`
`20
`
`o 0
`0.5
`5
`2
`1
`Total ethinyloestradiol AUC x 10-3 (pg mi-' h)
`Figure 4
`Relationships between theophylline clear-
`ance and the areas under the curves (AUC) for nor-
`gestrel and ethinyloestradiol for 26 women who received
`single or multiple doses of Ovral. A log transformation
`ofthe AUC data was performed in adherence to classical
`dose-response theory. Symbols are defined in Figure 1.
`Norgestrel results from one subject were excluded as
`Regression
`grossly
`misrepresentative.
`lines
`were
`generated
`employing
`an
`orthogonal
`least-squares
`approach. Significant negative correlations were found
`for both norgestrel (r = -0.53, P < 0.01) and ethinyl-
`oestradiol (r = -0.46, P < 0.02) exposure.
`
`phenomenon is that chronic exposure to these
`steroidal agents results in a reduction of the non-
`conjugative activities of the liver. This is strongly
`suggested by the numerous observations that contra-
`ceptive steroids are capable of markedly inhibiting
`oxidative pathways that are mediated through dis-
`tinct cytochromic apoproteins (i.e., antipyrine via
`cytochrome P450 and theophylline via P-448). More
`direct evidence in this regard has also been reported
`(MacKinnon and co-workers, 1977), as rats subjected
`to a 5 day exposure to ethinyloestradiol (5 mg kg-'
`day-'), exhibit a significant reduction in the activity of
`
`hepatic ethylmorphine N-demethylase and of cyto-
`chrome P-450, cytochrome b5, and NADPH cyto-
`chrome-c-reductase. The strong correlation between
`theophylline CL and AUC of the OC steroids supports
`this hypothesis.
`
`Secondary factors
`
`The social or occasional use of marihuana was shown
`not to significantly alter theophylline disposition.
`Jusko and co-workers (1978) had concluded that
`regular marihuana use results in an enhancement of
`theophylline metabolism. In this earlier instance, the
`marihuana users were subjects whose use of cannabis
`greatly exceeded that of current subjects. A more
`rigorous analysis of this phenomenon showed that
`occasional use of marihuana was not a primary factor
`influencing theophylline CL (Jusko et al., 1979) . In
`fact, the analysis suggested that the social use of
`marihuana results in a reduction of metabolic acti-
`vity. 'iThis finding, however, appeared as a low priority
`possibility in the statistical cascade of influential
`*factors. Minimal use of tobacco has also been re-
`ported not to affect theophylline clearance (Piafsky et
`al., 1977; Grygiel et al., 1979).
`The use of caffeine has been implicated in markedly
`altering theophylline clearing processes. Both en-
`hancement (Mitoma et al., 1969) and inhibition of
`theophylline metabolism (Monks et al., 1979) have
`been reported. In the subjects examined in the present
`investigation, it appears that a significant, but rather
`small enhancement of clearance occurred in caffeine
`users.
`The apparent volume of distribution of theophyl-
`line was unaffected by OC use in the subjects studied.
`Similar findings have been reported for caffeine
`vPatwardhan e al., 1980). However, a modest reduc-
`t.
`. * '
`.
`ton in VD was evident in the cigarette smoking group.
`This phenomenon was noted earlier (Hunt et al.,
`1976) but cannot be explained.
`A significant interaction between cigarette smoking
`and acute exposure to OC was noted. As previously
`discussed, it is possible that a portion of the inhibitory
`effect exhibited by these agents may be ascribable to
`competitive binding to metabolic sites, and hence
`could manifest itself acutely. Although such acute
`exposure was deemed insignificant when considered
`alone, a notable effect was present when super-
`imposed upon smoking. This finding is consistent
`with those of Conrad & Nyman (1980) who reported
`that the inhibitory effects of both propranolol and
`metoprolol are most pronounced in subjects whose
`theophylline plasma clearance values were initially
`elevated (i.e. in smokers). Qualitatively similar find-
`ings were presented by LaForce and co-workers
`(1981) who reported that subjects with the highest
`initial clearance rates of theophylline, tended to have
`the largest reductions in clearance subsequent to
`
`AMN1073
`Amneal Pharmaceuticals LLC v. Alkermes Pharma Ireland Limited
`IPR2018-00943
`
`

`

`SMOKING AND OC STEROIDS AFFECT THEOPHYLLINE CLEARANCE
`
`279
`
`treatment with erythromycin ethylsuccinate. Carter
`and co-workers (1975) reported that only subjects
`with the shortest antipyrine half-lives exhibited signi-
`ficant increases in this parameter following a 3 month
`exposure to 2 mg norethindrone/0. 1 mg mestranol. In
`addition, it has been apparent that the enzyme induc-
`tive effects of smoking on disposition of several drugs
`does not occur in patients with cirrhosis (Farrell et al.,
`1978) or in the elderly whose metabolic capabilities
`are intrinsically low (Vestal et al., 1980). It thus
`appears that a higher degree of metabolic activity may
`be associated with greater sensitivity to metabolic
`inhibition.
`These findings suggest that perturbations in theo-
`phylline elimination may occur during pregnancy, a
`period when circulating concentrations of endo-
`genously produced gonadotropins are relatively high.
`Preliminary data do not appear to support this
`(Sutton et al., 1978). Presently,we are engaged in a
`more comprehensive examination of the effects of
`pregnancy on theophylline disposition. To date our
`results are in agreement with those of Sutten and
`co-workers (i.e., there is a trend towards a reduction
`in theophylline clearance during pregnancy albeit
`statistically insignificant).
`
`Statistical complexities
`
`Evaluation of data upon which multiple factors have
`exerted diverse effects is difficult. Such difficulty is
`compounded when partitioning of the study group by
`these factors results in disjoint subgroups of unequal
`sizes. Attempts are generally made to avoid such a
`situation by carefully matching experimental units
`with regard to extraneous factors known to