`
`Medication Safety
`
`Occurrence and Impact of Unanticipated Variation in
`
`Intravenous Methotrexate Dosing
`
`Christopher S Parshuram. L Lee Dupuis, Teresa To. Sheila S Weitzman, Gideon Koren. and Andreas Laupacis
`
`
`
`BACKGROUND: Studies using direct measurement sugest that the doses of up to 65% of drug infusions are outside industry
`standards. These preparation-associated errors occur despite routine safety procedures. As of April 5. 2006. the clinical impact of
`these errors had not been evaluated.
`
`OBJECTIVE: To measure the occurrence and associated clinicat outcomes of variations in intravenous methotrexate dosing.
`METHODS: A prospective observational study was performed on 47 methotrexate infusions of 800 mglrn" that were administered to
`19 children with acute lymphoblestic leukemia. Serum methotrexate concentrations were measured at the end of the infusions.
`which were administered over 24 hours. The total methotrexate dose was determined by direct measurement of the concentration
`and the volume of each infusion.
`
`RESULTS: Dosing errors greater than or equal to 10% occurred in 11 (23%) infusions and ranged trom —61% to 55% of the ideal
`dose. Repeated measures regression analysis found the measured total methotrexate'dose was not significantly associated with
`the serum methotrexate concentration (p = 0.58) or with clinical toxicities. The methotrexate dose administered over the last hours
`of infusion (p = 0.006} and the serum creatinine level at diagnosis fp = 0.05) were the most significant predictors of the rnsthotrexata
`concentration. Hih methotrsxete concentrations were significantly associated with increased hepatic aminotransferase levels;
`however, the degree of elevation was of limited clinical relevance.
`
`CONCLUSIONS: While unexpected errors in drug dosing are more common than is suggested by other methods, the clinical impact
`observed in this model of methotrexate infusion was not demonstrably greater than medication errors described by other methods.
`Subsequent studies in this model of dosing error will require larger sample sizes, and other drugs should be evaluated.
`KEY wanes: adverse drug event. medication error, methotrexater
`Ann Pharmacofher 2006;40:605-11.
`Published Online, 25 Apr 2006. wwnttheannaiscom. DOI 10.1345iaph.1(3334
`
`espitc the accumulating evidence of unexpected vari-
`Dability between ordered and measured concentrations
`in up to 65% of drug infusions and concern about the as-
`sumption of accurate dosing made in pharmacokinctic
`studies, the investigation of the clinical impact of these
`dosing errors has been limited.“5 These preparation-associ-
`ated errors occur despite routine safety procedures. There
`is great potential for clinically significant effects arising
`from dosing errors in medications with a low therapeutic
`index such as methoucxate.‘
`
`Mctlmtrcxatc, which is routinely administered to chil-
`dren with malignancy, causes dose- and serum concentra-
`
`
`Author information provided at the end of the text.
`
`tion~related nephrotoxicity, hepatotoxicity, and myelosup-
`pression."" Scrum drug concentrations are measured as part
`of routine care; significant but unexplained variability in the
`levels with mctltotrcxate infusions exists.”“‘ Preparation-
`associated error may explain much of this variability. We
`performed a prospective observational study of directly
`measured drug doses and the alasociated clinical outcomes.
`
`Methods
`
`Children receiving 800 mgim'.24 hour method-este infusions as part
`of deem-rent for acute lwnphohlastic leukemia on the current Children's On—
`cology Group (COG) protocols 9904 and 9935 arms A or C were eligibte
`for study. We excluded children receiving infusions who had congenital
`leukemia or Down's syndrome. who was elder than 18 years at diagnosis,
`or who were scheduled to receive nonstandard method-exam dosing.
`
`WJheannafscom
`
`The Annals ofPharmacoriterapy I 2006 May, Volume 40 I 805
`
`Medac Exhibit 2031
`
`Koios Pharmaceuticals v. Medac
`
`IPR2016-01370
`
`Page 00001
`
`Medac Exhibit 2031
`Koios Pharmaceuticals v. Medac
`IPR2016-01370
`Page 00001
`
`
`
`CS Paulina-an: et al.
`
`METHOTFIEXATE INFUSION
`
`The current practice of our institution is to accept up to a 10% differ-
`ence between the ordered dose and the ideal {protocol} dese. Sodium bi-
`carbonate {30 mmolfl.) is added to the infusion solution of dextrose 5%
`and sodium chloride 02%. and additional intravenous sodium bicarbon-
`ate boluses are administered to maintain alkaline urine. Folinic acid (leu—
`oovorin) rescue therapy (IO my’m‘) is started 42 hours after the initiation
`of the methotrertate infusion. These therapies are discontinued at the dis-
`crelion of the treating physician after the methotrexnle concentration is
`below 02 pmollL.
`infusions were prepared using 25 mg’mL methotrexate stock solution
`(Faulding. Quebec. Canada). This was diluted to the prescribed volume
`and concentration by addition of the stock solution to the diluent. One to
`3 sterile intravenous bags were used depending upon the total volume re—
`quired. Following standardized prchydration. children received the
`methouexate infusion. Each infusion began with a 200 mg/m‘ bolus of
`mclhotrexatc over 20 minutes and was followed by a separate 800
`mgr‘m’ infusion administered over the remaining 23 hours and 40 min-
`utes. Bags were administered in labeled order.
`
`PRIMARY OUTCOMES
`
`SECONDARY OUTCOMES
`
`The secondary outcomes were rates of hepatotorticity, nephrotoxicity,
`and myelotoxicity over the 4 weeks following a methonexate infusion.
`The maximum values of creatinine. aspartate aminotransferase. and ala—
`nine aminotransferase were recorded. Myelosuppression was assessed
`by the minimum neuu-ophil. platelet. and hemoglobin counts.
`
`POTEN11AL FACTORS AFFECTING THE 24 HOUFI
`METHOTFIEXATE CONCENTRATION
`
`Seven factors that could explain variability in serum methotrcxate
`conCentrations were records-(12 (I) the total doser'm’. (2) the dosea’m1 in
`the last bag administered. (3) the amount of methotrcxalc administered
`in the last hours of the infusion. (4) giomerular function at diagnosis. [5}
`
`806
`
`I The Annals ofPharmacother-apy I
`
`2006 May, Volume 40
`
`mrheammlscom
`
`There were 2 primary outcomes for this study: the total dose adminis-
`tered and the 24 hour serum methotrexate concentration. The total dose
`administered was calculated from the volume and concentration of the
`infused mediotrexate solution. The methotrexate concentration was de-
`termined by direct measurement using HPLC with ultraviolet detection.
`The volume inhised was measured by weighing the bag with prescribed
`chemotherapy before and after administration (CAI 1274888. Mcttler.
`Toledo. OH; accuracy 2 0.01 g). Con'cction for the infusion density
`(4054.8 g/L) was made ignoring variations due to methotrexate (as
`sumed <0.06%). The missing weights of emptied bags were imputed.
`The percentage error of the total dose administered was also calculated
`versus the ideal (protocol) infusion dose (800 mg/m'). the prescribed
`dose. and the stated dispensed dose.
`Methotroxate concentrations were measured at the completion of a 24
`hour infusion. Merl-rotrertate concentrations are a surrogate for toxicity.
`are collected routinely as a part of standard care. and are not affected by
`the site of venous sampling."""5 The serum methotrexate concentration
`was measured using the T91 Mediotrexatc ll assay system (Abbott Lub-
`oratories, Chicago, 1L). This fluorescence polarization immunoassay is
`linear over the range 0.05— l .0 umoUL. is calibrated regularly. and has a
`coefficient of variation (CV) of 5% (product information). A 1:100 dilu-
`tion that is required for analysis introduces less than 1% additional error.
`Independent evaluation in our laboratory found CV values between 5%
`and 10%.
`
`glomerular function within 48 hours before the start of the infirsion,(6)
`the time that the methotrexate concentration was determined relative to
`the documented completion of the methotrexate infusion, and (7) the
`study protocol on which a patient was treated. The contributions of ac-
`tive tubular secretion and concomitant medication administration were
`not evaluated.
`
`Glomerular function was assesaed twice by using the serum creati—
`nine level and the calculated creatinine clearance from the formula of
`Schwartz ct til." due to concerns about the validity of the calculated val-
`ue in this population and changes in muscle mass during therapy." The
`dosing rate of methotrexate in the last hours of the metl'rotrexate infusion
`was included because administration rates may be altered to ensure infu-
`sion completion close to the prespecified time. The dosing rate in the last
`hour of infusion was calculated as the rate of administration determined
`from the last recorded volumes of mcthotrexate infusion administered di-
`
`vided by the time interval over which it was given. The dosing rate per
`meter squared was then calculated using the child's surface area and the
`methotrertale concentration measured from the bag being administered.
`
`VALIDATION OF METHOTFIEXATE INFUSION SAMPLING
`
`We validated the mediotrexate infusion sampling method by taking 3
`samples from each of 2 bags made with a concentration similar to that of
`standard methotrexate infusions (0.33 pmolfL). Samples were taken
`from the full bag, the half-emptied bag, and when the bag was nearly
`empty. The methotrexatc concentration was measured. and the CV was
`calculated for the set of measurements from each bag.
`
`ANALYSIS
`
`The percentage error in volume. concentration. and dose Were deter-
`mined for each bag prepared. The total dose per ureter squared was cal-
`culated using the most recent measurements, and the percentage enor
`between ideal (800 mgr‘m‘). prescribed. dispensed. and administered in-
`fusions was determined. The body surface area was calculated as BSA
`(m3) = s/ [(height in cm x weight in kg) {3600}.
`Regression analyses were used to evatuate the relationship between
`predictive factors and serum methotrexate concentrations and the rela—
`tionships between late toxicity outcomes and the 2 potentially predictive
`fattors: the serum methotrexate concentrations and the total dose adminis-
`
`tered. The proportion of variability in the methotrexate concentration that
`was explained by the predictive factors was estimated using the r‘ value.
`A repeated measures regression analysis was used to accommodate
`the inclusion of multiple melhotrertate infusions from one patient and the
`potential impact of the order of infusions. The regression analysis began
`with all 7 variables. These were sequentially removed using a back-
`wards—stepwise method until only variables significant at the p : 0.05
`level remained.
`
`The secondary outcomes were divided into quartiles of the total mea-
`sured dose per meter squared and the methotrexatc concentration; the
`mean values From each quartile were compared using linear regression.
`
`SAMPLE SIZE
`
`The sample size was determined using the methctrexate concentra-
`tions from a historical cohort. Forty infusions were required to describe
`differences among concentrations of 2 pmolfL with 90% power and a
`risk of a type one error of 2.5%. This difference was suggested to be the
`smallest clinically important difference by local content experts. The ap-
`proach was selected recognizing that variations in infusion rinse (volume
`and concentration) were not precisely known. and toxicities were sec-
`ondary outcomes.
`
`Page 00002
`
`
`
`This study was reviewed and approved by the institutional research
`ethics board. The need for consent was waived due to the lack of direct
`
`patient contact and the observational design.
`
`Unanir'cr‘mted Vomit in Intravenous Methanexare Dosing
`
`2.6%). Eleven (23%) of the measured total doses infused
`had errors of 10% or more. The errors ranged from —61 %
`to 55% (mean ~i 3%; Figure 1).
`
`Results
`
`SERUM METHOTFIEXATE CONCENTHAHONS
`
`A total of 47 infusions administered to 19 patients over
`8 months were studied. There were 8 children on the COG
`
`9904 protocol (20 infusions), and 11 children on the COG
`9905 protocol (27 infusions). The maximum number of in-
`fusions per child was 4. Each infusion was comprised of 1
`(n : 22), 2 (n = 19), or 3 (n = 6) bags of methotrexate.
`The prescribed doses were identical to the dispensed
`doses, suggesting that the infused concentration should be
`between 0.323 and 0.346 rig/ml... The oncology pharma-
`cists did not change physicians’ orders. Differences be-
`tween documented and recalculated BSA were minimal
`
`(—0.01 to 0.02 mi).
`The sampling method appeared to be reliable. Two bags
`of methotrexate were made. The 3 concentrations obtained
`
`DOSlNG ERRORS
`
`Seventy-eight bags of methotrexate solution were pre-
`pared for the 47 infusions studied. All bags were weighed
`when full, and 62 (79%) were weighed after the infusion
`was administered. Seven (11%) of the 62 bags with initial
`
`and final weights had errors of 5% or more between the
`stated dispensed volume and the administered volume
`(range —6.6% to 1.4%).
`Errors in the measured concentration of each bag of 10%
`
`or more were found in 24 (31%) of the 78 bags comprising
`the 47 infusions. Measured concentrations ranged from
`0.08 to 0.54 yg/mL, with a mean concentration of 0.34
`
`{mg/ml... The percentage errorranged from 46% to 63%.
`The measured dose in each bag differed by 10% or more
`from the stated dispensed dose in 20 (26%) of the bags. Er-
`
`rors ranged from 47% to 56%. with a mean percentage er-
`ror of —l 3%. When the last bag infused was evaluated in
`isolation. errors of 10% or more were found in 14 (30%)
`bags from the 47 infusions. and the magnitude of the er—
`rors ranged from -52% to 57% (mean —051%).
`The ideal (800 mg/m‘) and the prescribed total doses
`were similar (mean difference —0.5%, range w3.8% to
`
`www.maannaiscom
`
`from each bag were similar (0.339, 0.351, and 0.355
`,ugimL {CV 2.4%]; and 0.309. 0.332. and 0.338 pglmL
`[CV 4.8%1), suggesting that the method used was repre-
`sentative of the true concentration in the bag. One infusion
`was excluded from analysis on the basis of assumed in-
`complete mixing. It had a measured concentration and vol—
`ume suggesting that an extremely high dose of 3391
`trig/m2 (vs 300) would be administered. This dose resulted
`in an unexpectediy low 24 hour serum methotrexate con-
`centration of 3.9 ymolfL.
`
`The Annals ofPhannacorherapy I 2006 May. Volume 40 I 807
`
`Serum methotrexate concentrations were obtained for all
`infusions. Concentrations were drawn at the documented
`
`time of infusion completion in 24 (51%) infusions and var-
`ied from 15 minutes before to 140 minutes after (mean 16,
`median 0) the documented infusion completion time. The
`mean steady-state serum concentration was 7.9 JurnoliL
`(range 3—16, median 7.8).
`
`PATIENT FACTORS
`
`Serum creatinine level measurements were available for
`
`all children at the time of diagnosis and within 48 hours
`before infusion started in 46 of the 47 infusions. The mean
`
`calculated creatinine clearance values using the Schwartz
`et a1.” formula and serum creatinine levels were 141
`
`mUminJl .73 m2 and 0.46 mgde. respectively, at diagno-
`sis. and 171 mJJminf] .73 m1 and 0.39 mgde. respective-
`ly, at the start of each infusion (Table I).
`
`ASSOCIATIONS WITH SERUM METHOTHEXATE
`CDNCENTRAHON
`
`Analysis of isolated predictor variables with linear re-
`gression suggested that lower serum creatinine values were
`
`
`
`No.infusions
`
`U1
`
`Jo
`
`-5o
`
`-3o
`
`-to
`
`+10
`
`«so
`
`«50
`
`um «no
`
`Error between ordered and measared dose (%)
`Figure 1. Histogram oi the percentage difleranee hehveen the methotrexala
`dose administered and the ordered dose in 4'." methotraxate infusions at 300
`mgirn‘z administered to 19 children with acute lymphomasth leukemia. Doe-
`os were calculated using concentration measurements by HPLC and volume
`assessments of bags before and sitar infusion. The area bounded by the
`dashed lines represents the acceptable range of dose variation (t1fl%).
`
`Page 00003
`
`
`
`CS Parchment at :11.
`
`significantly associated with lower methotrexate concen-
`trations. The dose administered in the last hours of the
`
`methotrexatc infusion was also significantly associated
`with its serum concentration. The mean interval between
`the last and second-last times that volumes of methottexate
`
`infusion were charted was 1.13 hours (range 0.41—4.22).
`Regression analysis suggested that changes in the rate of
`methotrexate infusion in the last hours of infusion affected
`
`SECONDARY OUTCOMES
`
`assessed. The serum methotrexate concentration was asso-
`
`ciated with modest elevations in both aspartate and alanine
`aminotransferases, but not postinfusion renal or hemat0~
`logic measurements (Table 2).
`
`Discussion
`
`We found errors outside pharmaceutical industry stan-
`dards in 23% of methotrexate infusions administered to
`
`children with leukemia and among 31% of the methotnex—
`ate admixtures prepared for treatment. The impact of these
`errors on measured drug concentrations was small, and
`most of the variability (58%) in steady-state methotrexate
`concentrations remains unexplained. In this model of drug
`error, patient-related differences in renal drug elimination
`and the dose administered in the final hours of infusion
`
`were more important predictors of the serum drug conCen-
`tration than variations in the total dose administered.
`
`The results of our study are consistent with previous ob-
`servations of methotrexate pharmacology.“ The finding
`that serum creatinine level was more strongly associated
`with a serum methotrexate concentration than a calculated
`
`The total doac administered per meter squared was not
`significantly associated with any of the clinical toxicities
`
`creatinine clearance may be related to previous descrip-
`tions of the limitations of calculated creatinine clearance in
`
`Table 1. Prediction ol the 24 Hour Serum Methotrexate Concentration-
`
`Perameter
`
`Mean Min
`
`Max Estimate
`—0.02
`
`193
`
`p Value
`0.34
`
`1 41
`
`92
`
`the serum methotrexate concentration by 0.5—4.9 pmol/L.
`When all variables were used in multiple regression, 43%
`of the variability in methotrexate concentrations could be
`explained. The serum creatinine level at the time of diag-
`nosis explained more than half (24%) of this variability
`(Table 1).
`Repeated measures regression found that the dose ad-
`ministered over the last hours of infusion (p = 0.006) and
`the serum creatinine level at diagnosis (p = 0.05) were the
`only variables significantly associated with the serum
`methotrexate concentration.
`
`awrheannais .com
`
`I"
`0.02
`0.09
`
`0.32
`
`0.92
`
`6.10
`
`0.04
`
`65
`
`.
`
`221
`
`—0.003
`
`0.85
`
`0.0008
`
`0.36
`
`4.5
`
`140
`
`4.00009
`
`0.19
`
`0.73
`
`0.04
`
`0.002
`
`1233
`1233
`94
`
`0.0022
`-0.0030
`0.0356
`
`175
`87
`2.0
`
`0.04
`0.03
`1.43
`
`000 at diagnosis
`(mUrninli .73 m2}
`Creatinino at diagnosis
`{mgldt}
`CCC at start of infusion
`(tnUminl‘t 1’3 m')
`Creetlnine at start oi
`infusion (mgl‘dLl
`Time between obtaining
`serum concentration
`and documented
`completion of inlusion
`(min)
`Protocol
`
`Total dose tmgfm’)
`Last bag dose (mglm‘)
`0055 administered in
`the last 1-2 h of
`Inlusion (rngim’rl'h)
`Height {cm}
`Weight (kg)
`BSA (m2)
`All oi the above
`Creatinine at diagnosis
`and last hours dose
`only
`
`0.4?
`
`171
`
`0.39
`
`16
`
`.
`
`BSA = body surface area; CCC 2 calculated creatinine clearance; Max = maximum:
`Min = minimum.
`'Single variable linear tegrossion comparing the 24 hour serum methotrexate con-
`centration with potentially predictive factors.
`
`808 e The Annals ofPhnrmacotherapy I 2006 May, Volume 40
`
`children with malignancy?"m One previous
`study found that a calculated creatinine clear-
`ance could explain only 10% of the variability
`in methotrexate clearance in children." The
`
`serum creatinine level at diagnosis was more
`strongly associated with the serum mediotrcx~
`ate concentration than the serum creatinine at
`the time of infusion. The serum creatinine at di-
`
`agnosis was 0.08 mgde. higher than at the time
`of infusion. This most likely reflects reduced
`muscle mass during treatment rather than altered
`renal Motion?”
`
`The variable most significantly associated
`with a 24 hour serum methotrexate concentra-
`
`tion was the dose administered during the last
`hours of infusion. This finding raises 3 issues.
`First, it highlights the importance of consider-
`in g drug administration as an explanation for
`apparent variability in therapeutic drug con-
`centrations such as peak antibiotic concentra-
`tions and other clinical outcomes 3"" Second,
`
`it questions the value of 24 hour serum
`methotrexate concentrations as a measurement
`
`of steady-state. The wide variability of the
`dose per meter squared per hour administered
`during the last 1A2 hours of infusion is appar-
`ent from Figure 2. Because the half-life of
`methotrexate is 648 hours in this population. it
`should be expected that the serum methotrex-
`
`Page 00004
`
`
`
`ate concentration would change in this 1- to 2-hour period,
`although a new steady-state would not be reached for at
`least i8 hours.“ Our data suggest that the clinical impact of
`the change in serum methotrexate concentration is small.
`Finally, if the 24 hour serum methotrexate concentration
`is not a true steady-state level. then the assumptions about
`the nature of relationships between potential explanatory
`variables and this concentration may be attenuated. This in
`turn may have impacted the ability of this study to show a
`significant effect of dosing error in the relatively small
`sample studied.
`
`LlMlTATlONS
`
`There are 5 limitations to this study. First, the method
`used to determine the methotrexate dose may be less accu-
`rate than suggested. We excluded one under-mixed sam-
`ple. However, the pharmacy technicians making the infu-
`sions were aware of the purpose of the study and the im-
`
`portance of thorough mixing, and our previous work
`suggests that incomplete mixing is uncommon.‘ At most.
`the measurements varied by 5% in the sampling study. The
`other sources of measurement error were the weight and
`methotrexate assay. We estimate maximum errors of 1%
`from weighing measurements (precision 0.1 g) and 5%
`from the methotrexate assay from the measuring instru-
`ments used (CV ~2%). These errors are multiplicative and
`could suggest a measured dose between 89% (0.95 x 0.95
`
`Umndcipatea‘ Variation in Intravenous Method-exam Dosing
`
`x 0.99) and 111% (1.05 x 1.05 x 1.01) ofthe actual dose.
`Thus. in the worst-case scenario, the process of measure-
`ment could have added up to l 1% error. This is consider-
`ably less than the majority of the errors found (Figure 1).
`Second, the study may have been underpowered to de«
`tect important differences in each of the predictor vari-
`ables.“ There were 11 infusions with 10% or more error.
`
`Aggressive preemptive toxicity management may have
`prevented the development of the historically well-recog-
`nized complications. However, variations in folinic acid
`rescue, postchemotherapy hydration, urinary alkaliniza—
`tion, and the duration of methotrexate exposure may be
`important determinants of later toxicities?!“ Further ob-
`servational studies are needed to evaluate the effect of vari-
`
`ations from protocol on methotrexate-associated toxicity.
`Third, as the largest prospective study reported that 4% of
`medication errors are associated with adverse events, a
`
`sample of 1] errors may not be large enough to demon-
`strate adverse drug events .29 Dosing errors may also con-
`tribute to subclinical adverse drug events.
`Fourth, our results are from a single institution and a
`single drug and may not be generalizable to other centers
`or medications. However, results from studies in other in-
`
`stitutions and with other drugs suggest that this is not the
`case.” We did not evaluate the origins of variability and.
`consequently. were not able to exclude the stock solution
`(from the vial) as a major source of error.
`Finally, the inaccuracies of the serum methotrexate con-
`centration assay may have contributed to the limited corne-
`
`plhtltte'I
`
`59
`
`809
`
`Table 2. Mean Laboratory-Based Toxicity Measurement?
`Ouartllos
`Methotraxete
`Parameters Lowest
`
`Middle Upper
`Total dose in”
`AST (max; UtLt
`ALT (max: Ur'L)
`creatinine (max: mgldL]
`neutrophits (min; 10'tLl
`platelets (min: 10le
`hemoglobin (rnin: gtdL]
`24 h concentration
`AST (max; Ul’L]
`ALT (max: UJ'L]
`creatinine (max: mgr'dLl
`neutrophils (min : 10M.)
`platelets (min: 109.1)
`hemoglobin (min: gtdL)
`
`
`
`No.infusions
`
`43
`90
`
`43
`73
`
`Methotrexate administration tmgrm’th]
`Flgura 2. Histogram ot the rate at which methotrexate was administered dur-
`ing the test documented period of the infusion. The infusion rate in mLi'h was
`determined from the volumes infused divided by the time over which the vol-
`ume was infused. The dose per hour was calculated from the measured
`methotreirate concentration. The dashed line represents the anticipated infu-
`stat! rate if the 300 mglrn" dose was administered censistently throughout the
`23 hour and 40 minute infusion. Greater doses were associated with greater
`serum concentrations at the completion of the infusion (p = 0.01).
`
`ALT = alanine aminotransferase: AST = aspartete aminotrensterase:
`max = maximum; min = minimum.
`'Presented for the 4 weeks beginning at the completion ot each of 4?
`methotrexate infusions administered to 19 patients. Mean values are
`presented by quartiles of total dose per meter squared and the 24
`hour serum methotrexate concentration.
`“From linear regression of continuous toxicity values versus the total
`dose or 24 hour concentration.
`
`wwrheanmlsmm
`
`The Annals of Pharmcotherapy I
`
`2006 May, Volume 40 II
`
`Page 00005
`
`
`
`CS Parshumm er a1.
`
`lation between the steady-state concentration and clinical
`outcomes, although the apparatus is regularly calibrated
`and has a quoted precision of 1% and CV of5%.
`
`Conclusions
`
`mrheannalscom
`
`. Buclin T. Perrottet N. Biollaz l. The importance of assessing the dose ac-
`tually administered in phannaookinetic trials. Clin Pbar'rnacol Ther 2005;
`77:235—40.
`. Richards CF. Cannon CP. Reducing medication errors: potential ballet-IE
`of bolus thrombolytic agents. Acad Emerg Med 2000:T:1285-9.
`. Evans WE. Crom WR. Abromowitch M. et al. Clinical pharmacody-
`namics of high—dose methotrexate in acute lymphocytic leukemia. iden—
`tification of a relation between concentration and effect. N Eng] 1 Med
`1986;314:47l-T.
`. Stoller RG. Hand: KR. Jacobs SA. Rosenberg SA.Chabner BA. Use of
`plasma pharmacokinetics to predict and prevent methotrexate toxicity. N
`EnglJ Med l9'l7t297:630-4.
`. Evans WE. Pratt CB.Taylor RH. Barker LF,Crom WR. Phnnnacokinet-
`ic monitoring of high-dose rncthotrcxatc. Early recognition of high-risk
`patients. Cancer Chemother Phannaool 19393: [61-6.
`. Crom WR. Evans WE. Medics-circle. In: Evans WE. lusko Wl.eds. Ap-
`plied clinica] pharmacoltinetics. 3rd ed. Vancouver. WA: Applied Thera—
`peutics Inc.. [992291—2941
`. Evans WE. Rolling MV. Rodman JH. Crorn WR. Boyctt JM. Pui CH.
`Conventional compared with individualized chemotherapy for childhood
`acute lymphoblastic leukemia. N Eng] J Med 19983383399605.
`. Pharmacy Discipline Committee. Pharmacology Manual. 3rd ed. Arcar
`dia. CA: Childrens Oncology Group. 2003” i9-23.
`. Murry Dl. Synold TW. Pui Cl-l. Rodman “1. Renal function and
`methorrcxate clearance in children with newly diagnosed leukemia.
`Pitarmacothempy 1995:15:144-9.
`. Crorn WR. Glynn AM. Abromowitch M. Pui CH. Dodge R. Evans WE.
`Use of the automatic interaction detector method to identify patient char—
`acteristics related to mclhotnertate clearance. Clin Pharmacol Thor 1986'.
`39:592-7.
`. Cash. M. Schafhauser E. Byers J F. Venipuncture versus central venous
`access: a comparison of methotrexate levels in pediatric leukemia pa-
`tientsJ Podiarr Dnool Nurs 1999;16:189-93.
`. Schwartz (3]. Brion LP. Spirzcr A. The use of plasma cmstinine concen-
`tration for estimating glomerular filtration rate in infants. children, and
`adolescents. Pcdiatr Clin North Am 198?:34:571-9U.
`. Skinner R. Cole M. Pearson AD. or al. lnaccuracy of glomerular filtra-
`tion rate estimation from heighb'plasma creatinine ratio. Arch Dis Child
`l994;70:38T-90.
`. Joannon P. Oviedo 1. Campbell M. Tonlecilla J. High-dose merhotmxatc
`therapy of childhood acute lymphoblastic leukemia: lack of relation be-
`tween serum methotrexate concentration and creatinine clearance. Pedi-
`on Blood Cancer ZOO-1:43: 17-22.
`. Koskclo EK.Saarinen UM. Siimes MA. Skeletal muscle wasting and
`protein-energy malnutrition in children with a newly diagnosed acute
`leukemia. Cancer 1990;66:373-6.
`. Sala A. Tamopolsky M. Webber C. Norman G. Barr R. Serum creati-
`nine: :1 sunngate measurement of lean body mass in children with acute
`lymphoblastic leukemia. Pediatr Blood Cancer 2305;45:16-9.
`. Nahata MC. Paediatric drug therapy. II—Drug administration errors. 1
`Clin Pharm Thor 198833399402.
`. Nahata MC. Crist KD. Comparison of two infusion methods for pharma-
`cokinetic monitoring of gentamicin. Ther Drug Monit [9903 2: ISO-3.
`. Errick IK. Torre MS. Coleman 113. Hryciuk FL. Phannaookinetic consid-
`erations of variation in intermittent infusion time duration. Clin Pharm
`1982;l:61-2.
`. Sackctt DL. Clinical epidemiology: a basic science for clinical medicine.
`2nd ed. Boston. MA: Little Brown and Company. l991 : 183.
`. Sand TE. Jacobson S. Effect of urine pH and flow on renal clearance of
`methotrexate. Eur J Clin Pharmacol [981:19:453-6.
`. Sasaki K. Tanalca J. Fujimotc T. Theoretically required urinary flow dur-
`ing high-dose methotrexatc infusion. Cancer Chemother Pharmacol
`l984;l3:9-l3.
`. Christensen ML. Rivera GK. Cram WR. Hancock ML. Evans WE. Ef-
`fect of hydration on methonexatc plasma concentrations in children with
`acute lymphocytjc leukemia] Clin Oncol l988:6:797*30|.
`. Kinoshita A. Kurosawa Y. Kondoh K. et al. Efiocts of sodium in hydra-
`tion solution on plasma methotrexate concentrations following high
`dose methotrexare in children with acute lymphoblastic leukemia. Can-
`cer Chemotlter Pharmacol 2033;51:256-60.
`. Kaushal R. Bares DW. Landrigan C. ct a1. Medication errors and adverse
`drug events in pediatric inpatients. .lAMA 2001 :285:2l 14-20.
`
`Dosing errors of 10% or more were found in 23% of
`methotrexate infusions and in 31% of bags making up
`those infusions. The impact of these errors in total dose on
`the serum concentration was small. Variations in adminis-
`
`tration, incorporating dosing errors in the last bag and drug
`elimination by the patient. were major determinants of
`serum drug concentration. Further evaluation of the clini-
`cal impact of unexpected variability in the doses of medi-
`cations administered to children is needed.
`
`Christopher S Parshuram MB ChB D Phil. Stafl Physician. De-
`partment of Critical Care Medicine: Investigator. Population Health
`Sciences. The Research Institute. Hospital for Sick Children; As-
`sistant Prolossor, Faculty of Medicine, University of Toronto; Career
`Scientist. Ontario Ministry ol Health and Long Term Care. Toronto.
`ON. Canada
`L Lee Dupuls B PharmSci, Pharmacist. Department at Pharmacy.
`Hospital tor Sick Children
`Teresa To 550 MSG PhD. Senior Scientist. Program Head. Popu-
`lation Health Sciences. The Ftesearch Institute. Hospital for Sick
`Children: Associate Professor. Department of Public Health Sci-
`ences. University of Toronto: Investigator. Canadian Institutes of
`Health Research and Ministry of Health and Long Term Care: As-
`sociate Professor, Faculty of Medicine. University of Toronto
`Sheila S Weltzman MBChB. Stair Physician. Division of Haema-
`tology Oncology. Department of Pediatrics. Hospital for Sick Children;
`Associate Protessor. Faculty of MediCine. Univer5ity oi Toronto
`Gideon Koren MD FABMT FFlCP MBBS, Stall Physician. Division
`of Clinical Pharmacology and Toxicology. Department ol Pediatrics.
`Hospital lor Sick Children: Professor. Faculty of Medicine. Universi—
`ty of Toronto
`Andreas Laupacls MD MSc. CEO. Institute of Clinical and Evalu-
`ative Sciences: Professor. FaCuIty of Medicine and Health Policy.
`Management and Evaluation. University of Toronto
`Reprints: Dr. Parshuram. Department of Critical Care Medicine,
`Hospital for Sick Children, 555 University Ave. Toronto M5G 1x3.
`ON. Canada. tax 41 6181 3-7299. chn’stopher.parshuram@sickkids.ca
`Dr. Parshurarn is a recipient of support from The Research Institute
`at the Hospital tor Sick Children. This work was supported by the
`Physicians Services Incorporated Foundation. Toronto. ON. Cana-
`da. and the Canadian Institutes of Harm Research.
`
`We mm the Physicians Services Incorporated Foundation at Toronto ON. Canada.
`for “Pinning this wont. We also acknowledge the contributions of Lisa Mak BScN.
`Karen Bundle BNSc MN. Rachael Patel BN5: MN. rid Sergei Gauguin MD for col-
`lection oi the data and Ahmed Naqvl MD for his comments on the design at the study.
`
`References
`
`I. Parshurarn CS.NgG.l-lo T.etal. Dimmcics between ordered and deliv-
`ered concentrations of opiate infusions in critical care. Crit Care Med
`20033 1 :243351'.
`
`. Ferner RE. Lengford NJ . Anton C. l-