Sol S. Zimmerman, MD.
`Associate Professor of Clinical Pediatrics
`New York University School of Medicine
`Assistant Director of Pediatrics
`Director, Pediatric Intensive Care Unit
`University Hospital
`New York University Medical Center
`New York, New York
`
`ASSOCIATE EDITOR
`
`Joan Holter Gildea, R.N., B.S., M.A.
`Clinical Assistant Director of Nursing
`New York University Medical Center
`New York, New York
`
`
`
`
`
`CRITICAL CARE
`PEDIATRICS
`
`
`
`
`
`1985
`W B. SAUNDERS COMPANY
`Philadelphia London Toronto MexicoCity RiodeJaneiro Sydney Tokyo HongKong
`
`SUN - IPR2017-01929, Ex. 1021, p. 1 of 7
`SUN - |PR2017-01929, Ex. 1021, p. 1 of 7
`
`
`

`

`Philadelphia, PA 19105
`
`Library of Congress Cataloging in Publication Data
`Main entry under title:
`
`Critical care pediatrics.
`
`1. Pediatric intensive care. I. Zimmerman, Sol 3.
`II. Gildea, Joan Holter. [DNLM: 1. Critical Care—in
`infancy & childhood. WS 366 C93332]
`RI370.C747 1985
`618920028
`84-10570
`ISBN 0—7216-1143-5
`
`Critical Care Pediatrics
`
` W. B. Saunders Company: West Washington Square
`
`
`
`g
`
`':1):-
`
`leg—4
`
`
`
`
`
`“hm—.-i-__‘—v-;--_-==I:‘_-nh1w”...‘‘w...a”;
`
`ISBN 0—7216—1143-5
`
`the Uniform Copyright Conven-
`© 1985 by W. B. Saunders Company. Copyright under
`tion. Simultaneously published in Canada. All rights reserved. This book is protected by copy-
`right. No part of it may be reproduced, stored in a retrieval system, or transmitted in any form or
`by any means, electronic, mechanical, photocopying, recording, or otherwise, without written permission
`from the publisher. Made in the United States of America. Press of W. B. Saunders Com-
`pany. Library of Congress catalog card number 84—10570.
`‘
`
`Last digit is the print number:
`
`9
`
`8
`
`7
`
`6
`
`5
`
`4
`
`3
`
`2
`
`1
`
`
`
`SUN - IPR2017-01929, Ex. 1021, p. 2 of 7
`SUN - |PR2017-01929, Ex. 1021, p. 2 of 7
`
`SUN - IPR2017-01929, Ex. 1021, p. 2 of 7
`
`

`

`
`
`
`CLINICAL PHARMACOLOGY IN THE CRITICALLY ILL CHILD / 91
`
`
`
`
`
`concentration; in 3.3 half—lives, it will be 90
`R X t‘l/
`C55 = —'—— 14.4
`percent; and in five half-lives,
`it will be 97
`0.7 X IV,l >< Wt.
`(
`)
`
`percent of the plateau level.
`Dosage recommendations for continuous in—
`
`fusions are designed to produce appropriate
`
`plasma concentrations at equilibrium. The phe—
`
`nomenon just described, which is often termed
`drug accumuIa/ion, entails a delay in achieving this
`
`concentration. The magnitude of the delay is
`
`related to the half-life of the drug, whereas the
`ultimate concentration is determined by the Vd,
`the half—life, and the rate of administration.
`
`
`
`
`
`
`I’1KUI
`
`HHCDI
`
`I,Ir
`
`r.
`
`
`
`where R is expressed in mg/hr and CSS is ex-
`pressed in mg/L.
`Equation 14.4 indicates that equilibrium drug
`concentration is related to three variables: half-
`
`life (fl/2), volume of distribution (Vd), and rate
`of administration (R, or D/T). Thus, doubling
`Vd has the same effect upon steady state con—
`centration as halving tVz. Either alteration will
`lead to a 50 percent reduction in drug concen-
`tration that can be exactly offset by doubling
`the dosage.
`V
`
`Plasma Drug Clearance
`
`The plasma clearance (C1) of a drug is of
`primary importance in appreciating the rela—
`tionship between rate of drug administration
`and consequent drug concentration. Drug
`clearance, like creatinine or inulin clearances,
`is determined by relating the rate of elimination
`(E) to the plasma concentration at equilibrium
`(C55):
`
`3I:
`C.
`
`C
`
`At equilibrium, E = R.
`
`hus,
`
`R
`C1 = —-
`V
`Css
`
`or, with rearrangement,
`
`R
`C5. = a
`
`(14.3)
`
`R = Cl >< Cs.
`
`'
`
`(14314)
`
`Where R, = rate of administration.
`Equation 14.3A emphasizes that it is drug clear—
`ance, rather than half-life, which determines
`the rate of administration (R) or dosage per
`interval (D/T) necessary to achieve a specific
`concentration. Ultimately, clearance is related
`to half—life and volume of distribution:
`
`C1 = 0.7 X Vd X Wt.
`tl/z
`
`when V,l is expressed in L/kg, half-life (tyz) in
`hours, and clearance (C1) in L/hr.
`With substitution into equation 14.3, this be-
`comes:
`
`Multiple Dose Kinetics
`
`The reader has been introduced to the phe—
`nomenon of drug accumulation as it occurs
`during continuous infusion. Drug accumula—
`tion also occurs with intermittent dosage
`schedules.
`
`Consider a drug that is given by intermittent
`IV injection. When the first dose is‘ given, the
`concentration of drug is zero. Immediately after
`the dose, a peak concentration is recorded. The.
`concentration then declines at a rate deter—
`mined by the drug’s half-life. If the next dose
`is given before the concentration has once again
`reached zero, the second peak will be higher
`than the first. As this process continues, the
`peak (CW) and trough (Cmin) levels rise toward
`plateau values, as will the average concentra-
`tion, Cave. This process is illustrated in Figure
`14—4. Drug accumulation occurs during inter-
`mittent administration when a second, or nth,
`dose is administered before all the previous
`dose has been eliminated. For most clinical pur-
`poses, this condition is satisfied when the dos—
`ing interval is less than twice the half-life of
`the drug. As with continuous infusions, 50 per-
`cent of a plateau concentration is achieved in
`one half—life,- 97 percent is achieved after five
`half-lives.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Cave is analogous to the equilibrium concen—
`tration (C55) that develops during continuous
`infusion. Thus, its value is determined only by
`the relationship between clearance (Cl) and rate
`of administration (R; see Equation 14.3). The
`peak (Cmax) and trough (Cmin) concentrations
`
`fluctuate around the Cave in a manner that is
`
`determined by the size of the dose and the
`length of the dosing interval. For example,
`
`theophylline may be administered by inter-
`
`mittent IV injection. In an adult, a standard
`
`
`regimen calls for 300 mg every 6 hours (1200
`
`
`SUN - IPR2017-01929, Ex. 1021, p. 3 of 7
`
`SUN - |PR2017-01929, Ex. 1021, p. 3 of 7
`SUN - IPR2017-01929, Ex. 1021, p. 3 of 7
`
`

`

`
`
`
`
`
`
`92 / GENERAL PROBLEMS AND TECHNIQUES OF CRITICAL CARE
`
`(mg/L)
`
`
`
`TheophyllineConcentration
`
`_i O
`MOlllilllllll
`
`
`
`_A
`
`_U!
`
`(J1
`
`
`
`I
`2
`
`I
`4
`1
`
`.!
`6
`
`!
`8
`2
`
`
`l
`..
`. _
`L_.l_l|]lli|
`10 121416 18 2O 22 24 26 28 30 Hours
`3
`4
`5
`6
`7 Number of
`Half-lives
`
`
`
`
`
`Figure 14—4.
`Concentration of theophylline during intermittent IV administration. Note that CW, Oman. and Cave increase
`to equilibrium values. Cave = CSS that obtains during continuous administration if the total daily dosages are identical.
`
` mg/day). Alternatively, one may administer
`
`remain within the respective therapeutic ranges
`throughout the dosing interval.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`I
`[
`
`150 mg every 3 hours (1200 mg/day). Finally,
`some physicians administer a continuous in-
`fusion of 50 mg/hr (1200 mg/day). Both in—
`termittent regimens produce the same (Cave),
`which is equal to the CSS during the continuous
`infusion. This does not mean that the regimens
`are equivalent. The 6—hour schedule produces
`much greater fluctuation around Cave than the
`3-hour schedule. With the 3-hour regimen, the
`peaks and troughs lie closer to Cm. This in-
`creases
`the
`likelihood of remaining within the therapeutic
`range throughout the dosing interval (Fig. 14—
`5).
`'
`_
`In this regard, the half—life of a drug is a
`watershed. When a drug is given at an interval
`that is equal to its half—life (T = ti/z), CmaX/Cmin
`is approximately 2. During more frequent
`administration (T < ti/z), me/Cmin is less than
`2, and. during less frequent administration
`(T > tn): C,m,X/Cmin is greater than 2.
`
`Thus, drugs with a long half—life, such as
`
`digoxin or phenobarbital, are often given once
`
`daily, because even with this schedule, T is less
`
`than t1/2 and the plasma concentration remains
`
`within the relatively narrow therapeutic range
`
`of these agents. Conversely, theophylline and
`
`quinidine have relatively short half-lives (3 to
`
`6 hours in children). When conventional fOr-
`
`
`mulations of these agents are administered,
`
`they require relatively frequent dosing (every
`
`3 to 6 hours) if the plasma concentration is to
`
`
`Nonlinear Kinetics
`
`To this point, the discussion has concerned
`first—order kinetic behavior in which a fixed pru—
`porfinn of drug is eliminated per unit time. Zero—
`order kinetics occurs under some conditions, no-
`tably when plasma drug concentrations are rel-
`
`
`
`
`
`Concentration(arbitraryunits)
`
`
`
`‘
`
`Dosage = DI2
`
`Interval = T/2 = t.,,/2
`
`Time
`
`Effect of varying both dosage and dose
`Figure 14—5.
`interval upon peak (CW) and trough (Cmin) concentrations
`during steady state. The solid saw tooth line indicates the
`time concentration curve that results with intermittent IV
`administration of dosage D at an interval T equal to the
`drug t1/2. Note that CW/Cmin = 2. The interrupted line
`indicates the curve that results when dosage (0/2) and
`interval (T/2) are halved. Own/Cm;n = 1.5. The straight
`solid line indicates Cm, which is thesame during both
`conditions. and is equal to CSS, which results when the
`same total daily dosage is administered by continuous lV
`infusion.
`
`'
`
`SUN - IPR2017-01929, Ex. 1021, p. 4 of 7
`
`SUN - |PR2017-01929, Ex. 1021, p. 4 of 7
`SUN - IPR2017-01929, Ex. 1021, p. 4 of 7
`
`
`

`

`
`
`CLlNiCAL PHARMACOLOGY IN THE CRITICALLY ILL CHILD / 93
`
`atively large. With zero order or nonlinear ki-
`netics, a fixed amount of drug is eliminated per
`unit time. Ethanol is an extreme example, be-
`cause the usual dosage is large (gm amounts)
`relative to other drugs (mg amounts). Within
`the usual range of blood ethanol concentra-
`tions, humans eliminate about 120 mg per kg
`per hr of the substance. Because the volume of
`distribution of ethanol is about 0.5 L per kg,
`blood ethanol levels decline at a fixed rate of
`20 to 25 mg/dl per hr. The rate of elimination
`does not change with increases in concentra-
`tion. Consequently, increments in dosage pro-
`duce much greater changes in concentration
`than would be the case for a drug eliminated
`in accordance with first-order kinetics.
`Many substances follow a first—order model
`at low plasma concentrations but a zero-order
`model at higher concentrations. When the tran-
`sition from first— to zero—order elimination oc-
`curs at concentrations appreciably higher than
`the usual therapeutic range, the pharmacoki—
`netic treatment of the drug is uncomplicated
`and a first-order kinetic model will be suffi-
`ciently accurate for most clinical purposes.
`Unfortunately, a few commonly used drugs,
`such as phenytoin and salicylate, exhibit this
`transition at concentrations within the thera-
`peutic range.
`A change from first"- to zero—order kinetics
`as concentration increases is typical of an en—
`zyme-mediated process. This change is due to
`saturation of the enzyme system that is re-
`sponsible for metabolic transformation of the
`drug. There is a limited amount of enzyme at
`the metabolic site; therefore, there is a maxi-
`mum rate at which transformation can occur
`(Vmax). At concentrations that are low relative
`to Vmax, first-order behavior predominates. As
`concentration increases, Vmax is approached.
`After Vmax has been achieved, further increases
`in concentration cannot augment the metabolic
`rate. Thus, a fixed amount of drug is metab-
`olized per unit time. This amount, of course, is
`equal to Vmax. Mathematically, this process is
`described by the Michaelis—Menton equation:
`
`_ Vmax x C
`E —
`Km + C
`
`14.5
`
`(
`
`)
`
`Where E is the rate of elimination or metabo—
`lism; Vmax is the maximum rate of metabolism;
`Km is the Michaelis—Menton constant, which
`defines the affinity of the enzyme for the drug;
`and C is plasma drug concentration.
`Note that when C is much less than Km, E
`Varies directly with C. This resembles a first-
`Order process. When C is greater than Km, E
`
`approaches Vmax, and zero—order behavior oc—
`curs.
`
`There are two important consequences of
`this kinetic behavior. The first is that an in—
`crease in dosage produces an exponential, rather
`than a linear rise in concentration. This occurs
`very often when treating patients with phen—
`ytoin (Fig. 14—6); on occasion, this phenomenon
`is recognized during treatment with theophyl—
`line. It requires that dosage adjustments must
`be made cautiously and in small amounts. The
`second major consequence of Michaelis—Men—
`ton kinetics is that the apparent plasma half—life
`increases with the plasma concentration. The
`greater the plasma concentration, the slower is
`the relative rate of elimination. Using repre—
`sentative values of Km and Vmax for phenytoin,
`one can estimate that at a concentration of 10
`mg per L, the apparent tVz of phenytoin is 24
`hours; at a concentration of 25 mg per L, the
`apparent t1,2
`is 42 hours. This means: 1) in—
`creases in dosage cause lengthening of the ap-
`parent tyz (thus, Michaelis—Menton kinetics is
`sometimes referred to as dose dependent kinetics),
`2) small increments in dosage can produce huge
`increases in drug concentration, and 3) intox-
`ication with phenytoin will be prolonged, be—
`cause, at high concentrations, elimination is ex-
`tremely slow relative to the amount of drug in
`the body.
`
`MAINTENANCE DOSE
`
`The maintenance dose (MD) is the amount
`of drug (R for continuous infusion, D/T for an
`intermittent schedule) that is administered dur-
`ing equilibrium. Thus, from Equation 14.3A,
`maintenance dose, MD, is equal to the product
`of clearance, Cl, and desired steady state plasma
`concentration, C55,
`(MD 2 Cl
`>< C55). The
`maintenance dose is often determined by con—
`sulting standard reference material. In patients
`
`Phenytoin
`(Michaelis-Menton)
`
` Gentamicin
`(First-order)
`
`‘
`
`Concentration
`
`
`
`Daily Dosage
`
`Effect of dosage upon plasma concentra-
`Figure 14-6.
`tion for drugs following first-order vs. Michaelis-Menton
`kinetics.
`
`
`
`SUN - IPR2017-01929, Ex. 1021, p. 5 of 7
`SUN - |PR2017-01929, Ex. 1021, p. 5 of 7
`
`

`

`94 / GENERAL PROBLEMS AND TECHNIQUES OF CRITICAL CARE
`
`with abnormal drug disposition, better indi-
`vidualization of therapy is achieved if one ap-
`preciates the relationship between changes in
`CI and consequent changes in C55. Knowledge
`of a patient’s Cl can be used to calculate dosage
`requirements. This procedure is described in
`the case study at the conclusion of this chapter.
`It is not always appropriate to initiate therapy
`with the maintenance dose. Both continuous
`
`and intermittent schedules produce a gradual
`rise from the initial (usually zero) to the equi-
`librium concentration. Recall that 75 percent of
`the plateau is reached in two half—lives, and 97
`percent is reached in five half-lives. Thus, for
`drugs that have a long half-life, there will be
`a substantial delay in acquisition of the plateau
`concentration. Because the plateau concentra-
`tion may be close to the minimum effective
`concentration, this delay may be unacceptable
`in acutely ill patients. For example, an asth-
`matic who is simply placed on a theophylline
`infusion will not begin to experience relief for
`about 8 hours.
`
`LOADING DOSE
`
`The solution to the problem of delay in
`achieving adequate levels is to administer a
`loading dose (LD). The loading dose is the
`amount of drug that will rapidly produce a
`therapeutic plasma concentration.
`If one is emphasizing a target concentration
`strategy, calculating the loading dose is simple,
`because the loading dose and the desired con—
`centration (C55) are related through the volume
`of distribution (see Equation 14.2). With ap-
`propriate modifications,
`this expression be—
`comes:
`
`LD = Vd(L/l<g) >< Wt(kg) >< Css(mg/L)
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`PlasmaConcentration
`
`t1/2
`
`4 l1/2
`
`Time
`
`Figure 14—7. Administration of an appropriate loading
`dose eliminates the delay (4t.,,) in achieving equilibrium
`concentration, C... Solid line-infusion alone, beginning at
`T = 0. Interrupted line—loading dose at T = 0, followed
`by continuous infusion.
`
`individual product information when designing
`the loading dose. Readers interested in a the—
`oretical approach to this issue should consult
`the suggested reading by Rowland and Tozer.
`When intravenous therapy is indicated, the
`loading dose is often given as a single rela-
`tively brief infusion. In the case of a drug with
`a narrow therapeutic range or a prolonged
`phase of distribution, the physician may choose
`to divide the loading dose, as is commonly done
`with digoxin. In general, a loading dose is not
`indicated when the half-life is much less than
`
`(i.e., drug accumulation
`the dosing interval
`does not occur) or when the therapeutic range
`is wide. Thus, penicillin therapy does not begin
`with a loading dose. Of course, a loading dose
`is not indicated when there is no urgency in
`achieving the equilibrium drug concentration.
`There is also no point in administering a load-
`ing close when the half-life of a drug is very
`short, as with most pressor agents, because
`equilibrium conditions are reached in a matter
`of minutes during continuous maintenance in-
`fusion.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`where CSS is the desired equilibrium concen-
`
`
`tration. Thus, for a child weighing 10 kg in
`tend the analysis by taking into account the
`
`
`rate and extent of absorption. Drugs that are
`whom one wishes to achieve a plasma theo—
`
`
`completely and efficiently absorbed after in—
`phylline (Vd = 0.5 L/kg) concentration of 12
`
`
`tramuscular injection should maintain a similar
`mg per L, the correct loading dose is 60 mg.
`
`
`This amount of drug should be administered
`Cave, although peak and trough levels may lie
`
`
`slowly, over about a 15-minute period. Im-
`closer to Cave (lower peak, higher trough). After
`
`
`oral administration, many drugs either are not
`mediately thereafter, the appropriate mainte-
`
`
`nance dose is initiated. The effect of a loading
`completely absorbed from the gastrointestinal
`
`
`tract or once absorbed are efficiently extracted
`dose is shown in Figure 14-7.
`
`
`If one is using an empirically derived main—
`and then biotransformed on the first pass
`
`
`through the liver. This process effectively re-
`tenance dose and is not attempting to achieve
`
`
`duces the dosage of drug that reaches the sys-
`a specific drug concentration (target effect
`
`
`temic circulation. Equations 14.2 through 14.4
`strategy), the problem is less straightforward.
`
`
`In such cases, it is probably best to consult
`are modified by multiplying the dosage ad-
`
`
`
`
`SUN - IPR2017-01929, Ex. 1021, p. 6 of 7
`SUN - |PR2017-01929, Ex. 1021, p. 6 of 7
`
`(14.6)
`
`The foregoing kinetic description applies to
`intravenous administration. Following intra-
`muscular or oral administration one must ex-
`
`
`
`SUN - IPR2017-01929, Ex. 1021, p. 6 of 7
`
`

`

`
`
`CLINICAL PHARMACOLOGY IN THE CRITICALLY ILL CHILD / 95
`
`ministered (R, or D/T) by fowl, the fraction of
`orally administered drug that reaches the sys—
`temic circulation. Several of the suggested read—
`ings contain a table of representative foral values.
`
`KINETIC VARIATION—CLINICAL
`APPLICATION
`
`In the foregoing discussion, three variables
`that affect the plasma drug concentration have
`been identified: 1) the rate of administration
`(R, or D/T), 2) the volume of distribution (Vd),
`and 3) the half-life (ti/z). These three variables
`are related to one another through Equation
`14.4. Of these, only the rate of administration
`is under the physician’s control. Thus, the art
`of pharmacokinetics consists of altering this
`parameter in order to compensate for individual
`differences in elimination or distribution.
`
`Altered Distribution
`
`Critical illness and the age of the patient af—
`fect the distribution of many drugs. Several
`mechanisms may be
`involved,
`including
`changes in the content or distribution of body
`water, alterations in plasma protein binding,
`perturbations in regional blood flow, and dif-
`ferences in body fat content.
`
`AGE—RELATED CHANGES
`
`Table 14—2. Selected Drugs With Altered Volume
`of Distribution (Vd) in Neonates and Children
`
`
` Drug Effect on Vd
`
`l N
`Diazepam
`T N, T C
`Digoxin
`T N,
`Furosemide
`T N,
`Gentamicin
`T N
`Lidocaine
`T N
`Phenobarbital
`T N
`Phenytoin
`
`
`Theophylline T N
`
`T C
`
`T = V, larger; 1 = Vd smaller; N : neonate; and C =
`children older than 1 month. The effect of changes in Va
`may be modified by alterations in protein binding and
`elimination rate,
`
`ance. Clearance (Cl) is related to both Vd and
`half-life. In infants, drug half-life may be pro-
`longed (vide infra), and this may offset the
`increase in Vd (recall that C1 = 0.7 X Vd/t%).
`The net effect is frequently a reduced dosage
`requirement or an increased loading dose, fol—
`lowed by a reduced maintenance dose.
`
`DISEASE-RELATED CHANGES
`
`Several processes affect Vd by altering body
`water content, body fat content, or the degree
`of protein binding. Notable examples are ure-
`mia, chronic liver disease, and congestive heart
`failure.
`In uremia, the water content of the body is
`frequently greater than normal. This factor, to-
`gether with disturbed protein binding, causes
`the volume of distribution (Vd) of several drugs
`to increase (e.g., gentamicin) or decrease (e.g.,
`digoxin). As in newborns, the larger Vd is fre-
`quently accompanied by prolongation of drug
`half—life.
`Chronic liver disease is associated with de-
`creased levels of plasma proteins and with fluid
`accumulation. Thus, it is not surprising that the
`Vd of several drugs increases in the presence of
`cirrhosis. Other conditions associated with ex-
`tracellular fluid expansion also increase the dis-
`tribution of certain drugs; for example, ami-
`noglycoside antibiotics distribute into ascitic
`fluid. Thus, in the presence of ascites, the Vd
`of these agents may be substantially increased.
`In cystic fibrosis, the Vol of several of the ami-
`noglycosides seems to be higher than average.
`Unless daily dosage is increased, this may lead
`to subtherapeutic drug concentrations.
`Abnormalities of regional or global blood
`flow may reduce distribution by limiting per-
`fusion to sites of uptake In patients with low-
`output states (congestive heart failure, CHF;
`shock), the rate of distribution is likely to be
`
`The water content of the body changes dra-
`matically with age. At 28 weeks gestation, the
`water content of the body is 85 percent. This
`figure decreases to 70 percent at term and to
`60 percent in adults. There is a concurrent in-
`crease in the amount of body fat from 1 percent
`Of body weight at 28 weeks to 15 percent at
`term, as well as altered binding to protein. Dis-
`ease (tachypnea, dehydration),
`the environ-
`ment in which the infant is nursed, and the
`Volume and composition of administered fluids
`Produce fluctuations in body water. Thus, it is
`anticipated that drugs that are distributed
`mainly in the body water have a different, usu—
`ally greater, volume of distribution (Vd) in in-
`fants and young children than in adults. Table
`14—2 lists several drugs for which the Vd is
`knQWn to differ between newborn infants and
`adults. This information provides the rationale
`for many empirically determined dosage mod—
`ifications A larger Vd does not necessarily in-
`VO1Ve a larger dosage. The actual determinant
`0f the maintenance dosage requirement is clear-
`SUN - IPR2017-01929, Ex. 1021, p. 7 of 7
`SUN - |PR2017-01929, Ex. 1021, p. 7 of 7
`
`
`
`