Postmarketing drug dosage changes of 499 FDA-approved new molecular entities, 1980–1999

pharmacoepidemiology and drug safety 2002; 11: 439–446Published online 12 August 2002 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/pds.744

ORIGINAL REPORT

Postmarketing drug dosage changes of 499 FDA-approvednew molecular entities, 1980–1999y

James Cross MS*1, Howard Lee MD, PhD2, Agnes Westelinck PharmD2, Julie Nelson MBA2,Charles Grudzinskas PhD2 and Carl Peck MD2

1Division of Metabolic and Endocrine Drug Products, Center for Drug Evaluation and Research, FDA, Rockville, MD, USA2Center for Drug Development Science, Georgetown University, Washington DC, USA

SUMMARY

Purpose Risks and benefits of marketed drugs can be improved by changing their labels to optimize dosage regimens forindicated populations. Such postmarketing label changes may reflect the quality of pre-marketing development, regulatoryreview, and postmarketing surveillance. We documented dosage changes of FDA-approved new molecular entities (NMEs),and investigated trends over time and across therapeutic groups, on the premise that improved drug development methodshave yielded fewer postmarketing label changes over time.Methods We compiled a list of NMEs approved by FDA from 1 January 1980 to 31 December 1999 using FDA’s website,Freedom of Information Act request, and PhRMA (Pharmaceutical Research and Manufacturers of America) database. Ori-ginal labeled dosages and indicated patient populations were tracked in labels in the Physician’s Desk Reference1. Time andcovariate-adjusted risks for dosage changes by 5-year epoch and therapeutic groups were estimated by survival analysis.Results Of 499 NMEs, 354 (71%) were evaluable. Dosage changes in indicated populations occurred in 73 NMEs (21%).A total of 58 (79%) were safety-motivated, net dosage decreases. Percentage of NMEs with changes by therapeutic groupranged from 27.3% for neuropharmacologic drugs to 13.6% for miscellaneous drugs. Median time to change followingapproval fell from 6.5 years (1980–1984) to 2.0 years (1995–1999). Contrary to our premise, 1995–1999 NMEs were3.15 times more likely to change in comparison to 1980–1984 NMEs ( p¼ 0.008, Cox analysis).Conclusions Dosages of one in five NMEs changed, four in five changes were safety reductions. Increasing frequency ofchanges, independent of therapeutic group, may reflect intensified postmarketing surveillance and underscores the need toimprove pre-marketing optimization of dosage and indicated population. Copyright # 2002 John Wiley & Sons, Ltd.

key words— drug dosage; label changes; postmarketing changes; dosage changes; new molecular entity

INTRODUCTION

Having selected a drug for treating a patient with anindicated condition, caregivers need reliable informa-

tion to determine appropriate dosage. Federal regula-tions require drug labels to include recommendeddoses, dose range, dosing intervals, treatment dura-tion, and modifications for special patient popula-tions.1,2 Ideally, the labeled dosage regimen of anewly marketed drug in the indicated patient popula-tion should remain unchanged throughout its lifetime.Such stability of dosage and indicated populationwould reflect an adequate pre-market determinationof dosage to ensure continued safe and effective pro-duct use. This paper studies dosage stability asreflected in the drug label. Changes in labeled dosefor approved indications are to be contrasted withlabel changes for expanded uses (new indications,

Received 2 May 2002Copyright # 2002 John Wiley & Sons, Ltd. Accepted 19 May 2002

* Correspondence to: James Cross, U.S. Food and Drug Adminis-tration, Center for Drug Evaluation and Research, Division ofMetabolic and Endocrine Drug Products, HFD-510, 5600 FishersLane, Rm. 14-B-04, Rockville, MD 20857, USA.E-mail: [email protected] The statements and opinions expressed in this manuscript are thoseof the authors, and not necessarily those of the US Food and DrugAdministration. Data were gathered from publicly availableinformation.

Contract/grant sponsor: DIA.

including populations), which do not reflect inade-quate pre-market evaluation, and are not of concernto this study.

New information warranting dosage changes in thefirst indicated population often becomes apparentafter marketing approval. For example, the FDA’sCenter for Drug Evaluation and Research reportedthat over 500 drug products underwent safety-relatedlabel changes during 1999.3 Optimization of the first-labeled dosage regimen for a new molecular entity(NME) in the indicated population represents acritical challenge for pharmaceutical developers andregulators. Thus, dose-related label changes for initi-ally approved indications may be considered a signalof poor quality of a drug development program. Littledata have been published on dosage changes of FDA-approved drugs,4,5 despite general recognition of theimportance of defining clinically useful regimens.6–9

In this study, we: (1) document the changes in drugdosage regimens in indicated populations of newdrugs approved by FDA during 1980–1999 and (2)investigate trends in approved dosage regimens overtime and across theraepeutic groups. We hypothesizedthat advances in drug development and regulation inrecent years would be reflected in a decrease in therate of postmarketing changes relative to that of theearly 1980s. We also hypothesized that fewer labeleddosing changes would occur with oncology and HIVdrugs, relative to other drug groups because ofincreasing use of modern approaches to determiningdose–response in HIV drug development programs,and the routine use of the ‘maximum tolerated dose’strategy in development of cancer drugs.

METHODS

Data sources

Approval dates of all NMEs (n¼ 499) approved dur-ing 1980–1999 were compiled from databasesobtained from the Pharmaceutical Research andManufacturers Association and FDA (direct requestand www.fda.gov/cder). The Physicians’ Desk Refer-ence1 (PDR1, Medical Economics Company, Mon-tvale, NJ), 1981–2000, provided the only availablecomprehensive database of approved labels. Dosagechanges were documented by comparing the first pub-lished label and all successive labels until a changecould be confirmed. To be evaluable, a drug’s labelhad to appear in at least two PDR1 editions. Tworeviewers independently examined the labels for allNMEs, followed by resolution of differences byconsensus.

Study selection

Only dosage regimens of NMEs, first-approvedindications, and formulations were considered. Adosage regimen consisted of dose strength, dosinginterval, treatment duration, and contraindications,each keyed to original indicated patient populations.A decrease in dosage was defined by one or more ofthe following: (1) decreased dose strength, (2)decreased treatment duration, (3) increased dosinginterval (i.e. decreased frequency), (4) new restric-tions of the original indicated population includingcontraindications and (5) removal of an originalindication (including market withdrawal).

Increases in drug dosage included any changesopposite of those defining a dosage decrease. Whensimultaneous changes occurred to a regimen, the totaldaily or weekly dosage was assessed for net increaseor decrease. Postmarketing contraindications for newlydiscovered idiopathic hypersensitivity reactions wereexcluded.

Time to (first dosage) change was defined as the dif-ference between product approval date and year ofPDR1 publishing the revised label. Although requiredby regulation since 1979,10 revision dates were oftenmissing, requiring uniform use of the PDR1 year forconsistency. The lag between change date and publi-cation in the PDR1 was assumed to be similar (4–15months) for all drugs since drug approval could haveoccurred after the publication deadline for entry intothe following year’s PDR1. Market time was recordedas the time from the FDA-approval date to the label’slast year in the PDR1.

Data extraction

Descriptive statistics were used to evaluate raw data.Approval year, therapeutic group, and market timewere prospectively identified as candidate covariatesin the survival analysis. Drug approval years and theirmarket times were grouped into four 5-year epochs.Prospectively designated therapeutic groups wereCardiovascular/Pulmonary, Antiinfective (non-HIV),Oncology/HIV, Neuropharmacology, and Other/Miscellaneous products. Therapeutic groups wereprospectively devised to reflect factors postulated toinfluence dosage derivation during drug development:(a) Disease severity (severe, cardiovascular/pulmonary, oncology/HIV, antiinfective; mild, neuro-pharmacology) (b) Availability of successful treat-ments (high, antiinfective; low, oncology/HIV;variable, cardiovascular/pulmonary and neurophar-macology) (c) Objectivity of study endpoints

440 j. cross et al.

Copyright # 2002 John Wiley & Sons, Ltd. Pharmacoepidemiology and Drug Safety, 2002; 11: 439–446

(objective, antiinfective, oncology/HIV, cardiovascu-lar/pulmonary; subjective, neuropharmacology).

Survival analysis (‘survival’ of initial dosage inindicated population) using Kaplan–Meier curvesand the Cox proportional hazards model were per-formed to determine if time and candidate covariateswere significantly related to the risk of dosage change.Initially, survival curves for different therapeuticclasses and epochs were compared and differencesassessed using the log-rank test. For the Cox modelanalysis,11 reflecting our prior hypotheses, ‘Oncol-ogy/HIV’ drug and ‘1980–1984’ were selected asbaseline covariates for therapeutic class and epoch,respectively. All covariates are categorical; indicatorvariables were used to model their effects. For thefinal Cox model, Martingale and Schonfeld residualswere plotted to evaluate model assumptions.12 Theseanalyses were repeated using dosage decrease orincrease, each treated separately as failure endpoint.SAS PROC LIFETEST and PROG PHREG (version6.12) were used, while plots were created using theS-PLUS coxph function.13,14 A two-tailed p-value of0.05 or less was regarded as statistically significant;due to the few hypotheses tested, no correction forthe multiple comparisons was made.

RESULTS

FDA approved 499 NMEs during 1980–1999. Labelsfor 354 products (71%) were available in at least twoseparate PDRs1. NMEs appearing once (n¼ 35) ornever (n¼ 110) in the PDR1 were excluded. Of theevaluable NMEs 21% (n¼ 73 of 354) underwent adrug dosage change: 79% (n¼ 58) a reduction and21% (n¼ 15) an increase. Restrictions of the originalindicated population or contraindications comprised45% (n¼ 26 of 58) of the reductions. New dosingrestrictions associated with the cytochrome P4503A4 metabolic pathway accounted for 12% (n¼ 9)of all observed dosage changes (16% of dosage reduc-tions). New restrictions associated with renal–hepaticimpairment accounted for another 12% (n¼ 9) ofchanges (16% of dosage reductions). Appendix 1 liststhe drugs with dosage changes, and the type of change.

Therapeutic group

In the evaluable database, 79, 68, 42, 77, and 88products were in the Cardiovascular/Pulmonary,Antiinfective (non-HIV), Oncology/HIV, Neurophar-macology, and Other/Miscellaneous groups, respec-tively. Table 1 illustrates the raw percentage ofdrugs by therapeutic group that had dosage changes.

The miscellaneous group, followed by the oncology/HIV group had the smallest percentage of changes(13.6 and 19.0%), whereas 28.6% of neuropharmaco-logic products had changes. No group-related differ-ences were statistically significant (Figure 1).

Approval year epoch

A total of 71, 78, 95, and 110 NMEs were approved inthe evaluable database in epochs 1980–1984, 1985–1989, 1990–1994, and 1995–1999 respectively, repre-senting a 54% increase of NME approvals from first tolast epoch. Table 2 shows the raw percentage (unad-justed for market time) of drugs approved in each5-year epoch that underwent a dosage change. Unad-justed for market time, the 1985–1989 epoch had thegreatest percentage of drugs with changes (25.6%)whereas the 1995–1999 epoch had the smallest(11.8%). Market times for the evaluable NMEs are

Table 1. Raw percentage of NMEs with change in dose-relatedlabel information, by therapeutic class

Therapeutic class No. of No. of PercentageNMEs changed NMEs of class (%)

Neuropharmacology 77 21 27.3Antiviral/Antiinfective 68 17 25.0Cardiovascular/Pulmonary 79 15 19.0Oncology/HIV 42 8 19.0Miscellaneous 88 12 13.6

Figure 1. Kaplan–Meier survival curves of dosage change bytherapeutic class. The oncology/HIV category maintained aconsistently greater proportion of drugs without changes. Thiscontinued until late in the observation period when the smallpopulation size of this group magnified the decrease resulting fromonly one or two dosage changes of the remaining marketed drugswithout changes

postmarketing drug dosage changes 441


shown in Table 3. Predictably, drugs in the oldestepoch had the greatest market times (median¼ 16.2years) while recently approved products had the least(3.1 years); thus, epoch served as a surrogate measureof market time.

Survival analysis

Survival curves of dosage change by epoch, unad-justed for other covariates, are plotted in Figure 2,which shows that the dosage change patterns among

epochs are marginally different ( p¼ 0.072). The lat-est epoch (1995–1999) shows the steepest decline.

Addition of therapeutic group into the Cox model,either as main effect or interaction with epoch, did notimprove the fit as judged by �2*log likelihood differ-ences or global p-values. Therefore, a main effectepoch model with therapeutic group treated as stratawas selected (Table 4). Compared to the 1980–1984epoch, drugs approved in the 1995–1999 wereapproximately 3.15 times more likely to incur adosage change ( p¼ 0.008), despite the greater markettime exposure of drugs in the earliest epoch. Theepochs 1985–1989 and 1990–1994 showed a similartrend of increased risk for dosage change( p¼ 0.012, ordered trend analysis). This analysiswas reconfirmed by plotting the cumulative hazardfunction or logarithmically transformed survival func-tion of each epoch over time. Figure 3 shows that theincreased risk of dosage change for drugs approved inthe 1995–1999 epoch was evident relatively soon afterapproval, and remained steady over time.

The epoch effect was analysed separately by dosageincreases and decreases (Table 5). Although statisticalsignificance was not attained and the precision of esti-mated relative risk was low because of the small num-ber of NMEs in each category, drugs approved during1995–1999 had greater risk for both dosage increaseand decrease.

CONCLUSIONS

Our hypothesis that the rate of post-approval dosagechanges would be less for more recently approveddrugs was not confirmed by this study. In fact, twoopposite trends emerged. During 1980–1999: (a)dosage change rates of newly approved NMEs,increased in each 5-year epoch, the 1995–1999 epochhaving a rate 3.15 times greater than 1980–1984, and(b) dosage changes have been occurring sooner—median times dropped by 69%.

Analysed by 5-year epoch, an observed decrease inthe raw percentage of NMEs with dosage changesappeared to support our a priori hypothesis thatimproved drug development and regulatory practicescontributed to pre-market dosage optimization. How-ever, analysing time to dosage change by epoch(implicitly adjusting the rates for market time) usingsurvival analyses (Figures 2 and 3) revealed a statisti-cally significant trend with the most recent epochshowing the steepest rise in dosage change rate.

With regard to therapeutic group, we postulatedthat fewer dosage changes would occur among oncol-ogy and HIV drugs. Despite a twofold range of label

Figure 2. Kaplan–Meier survival curves of dosage change by5-year epoch of drug approval. At the end of the observation period,the oldest and earliest epochs showed similar proportions ofunchanged drugs. However, newer drugs attained that level duringan observation period that was equal in duration to one-quarter ofthat for the earliest cohort

Table 2. Percentage of NMEs with change in dose-related labelinformation, by approval epoch

Approval epoch No. of No. of changed Percentage ofNMEs NMEs epoch (%)

1980–1984 71 18 25.41985–1989 78 20 25.61990–1994 95 22 23.21995–1999 110 13 11.8

Table 3. Range of market time by epoch for all evaluable products(two or more labels found in PDR1)

Approval epoch Market time Minimum Maximum(median in years)

1980–1984 16.2 3.0 19.71985–1989 11.6 2.0 14.71990–1994 7.1 2.0 9.91995–1999 3.1 1.0 4.8

442 j. cross et al.


change rates, significant differences in the observedrates of change between therapeutic groups were notsupported even after adjustment for market experience(Figure 1).

Ironically, the steep rate of early dosage changesseen recently may reflect post-approval market condi-tions accelerating the discovery of suboptimal regi-mens. Thus, we urge caution in interpreting increaseddosage change rates since the risk-conscious marketfor new drugs of the late 1990s differs significantly

from that of the early 1980s; greater consumer aware-ness and advocacy for drug safety and regulatorytracking systems have intensified pharmacovigilanceand data-mining of postmarketing information.15,16

More immediate regulatory action (e.g. label revi-sions) in response to signals in reported adversereactions data may be contributing to increaseddosage change rates.

While dosage changes may represent a small frac-tion of the overall number of label changes occurringannually,3 recent findings have suggested that thequantity and quality of clinical pharmacology infor-mation in drug labels could be improved.17 FDA hasrecently proposed improvements in the prescriptiondrug label. Among the proposed improvements aretwo new label sections entitled ‘Use in Specific Popu-lations’ and ‘Drug Interactions’.18 Steps have alreadybeen taken to improve quality of labels by promoting‘label-driven’ drug development (e.g. the ‘TargetedPackage Insert’ program developed jointly by FDAand PhRMA).19 However, it is not yet evident thatthese initiatives will lead to fewer postmarketingdosage changes. To promote derivation of dose–response information, FDA and ICH (InternationalConference on Harmonization) have publishednumerous related initiatives and guidances.5,8,20–22

Our inclusion of restrictions on originally indicatedpopulations and market withdrawals may appearto overstate the rate of dosage changes. However,

Table 5. Relative risk and 95% confidence interval of dosage increase and decrease by epoch

Epoch Dosage increase (n¼ 15) Dosage decrease (n¼ 58)

Relative risk* (95%CI) p-value Relative risk* (95%CI) p-value

1980–1984 1.00 1.001985–1989 3.55 (0.40, 31.9) 0.260 1.27 (0.61, 2.61) 0.5201990–1994 4.96 (0.59, 41.7) 0.140 1.39 (0.65, 2.96) 0.3901995–1999 13.64 (1.28, 146.0) 0.031 2.35 (0.91, 6.04) 0.076

p-value for trendy 0.024 0.109

*1980–1984 as reference.yEach epoch was treated as a continuous variable.

Figure 3. Cumulative hazard function for dosage change overtime by epoch. The drugs of the most recent epoch were exposed toa 3.15 times greater risk of undergoing a dosage change ( p¼ 0.003)

Table 4. Relative risk and its 95% confidence interval of dosage change by epoch

Epoch No. approved No. changed (%) Relative risk* 95%CI p-value p-value for trendy

1980–1984 71 18 (25.4) 1.001985–1989 78 20 (25.6) 1.42 (0.72, 2.80) 0.3101990–1994 95 22 (23.2) 1.68 (0.84, 3.34) 0.1401995–1999 110 13 (11.8) 3.15 (1.36, 7.33) 0.008 0.012

*1980–1984 as reference.yEach epoch was treated as a continuous variable.



these changes were included because adequate pre-marketing studies might have provided informationsufficient to avert such postmarketing label modifica-tions. The importance of describing a new drug’smetabolism and pharmacokinetics–dynamics in spe-cial populations, encouraged by FDA,23 is confirmedby the observation that nearly one-quarter of allchanges could be attributed to renal–hepatic impair-ment and CYP3A4-mediated drug interactions (one-third of dosage reductions). Such dosage regimenchanges are less likely in the future due to routineattention to these recognizable causes of postmarketingsafety problems.

Certain dosage changes and factors were unac-counted for in this study. (1) Indiscernible dosagechanges: some labels were ‘modified’ to includenew dosage information without ‘changing’ netdosages. For example, some anti-cancer agentsretained the original dosage regimen but added anindividualized regimen using creatinine clearanceinformation, not effecting a net dosage change forthe original indicated population. (2) Off-label dosagechanges occurring in practice. (3) The potentialimpact of decreased foreign marketing experience ofnewly approved drugs: the percentage of NMEs withforeign marketing experience prior to US approval hasdropped from 72 to 26% from 1987–1989 to 1996–1998.24 Therefore, foreign-use outcomes of new drugsin clinical practice may be less commonly availablenow in FDA’s pre-market review to corroborate phaseIII study conclusions.

This study was constrained by availability of druginformation. The PDR1 appears to be the most com-prehensive collection of FDA-approved labels.Nevertheless, we found that 30% of NMEs in thestudy could not be included in our database becauseless than two labels were available. Many of the 145‘drop-outs’ were approved too recently to appear innewer volumes of the PDR1. Other missing drugsincluded diagnostic products or drugs to treat rarediseases, whose manufacturers chose not to publishin the PDR1. Biases due to selection or detection pro-cedures were minimized since the short-fall becameapparent prior to application of standardized dosagechange criteria.

In this study, one-fifth of evaluable drugs approvedby FDA during 1980–1999 underwent a change indosage and/or indicated population for the firstapproved indication(s). The majority (79%) of thesechanges were safety-motivated decreases. This patternmay represent a systematic flaw in pre-marketingdosage evaluation; it has been common practice inthe pharmaceutical industry to undertake phase III

trials evaluating drug effectiveness at or near maximum-tolerated dosages (MTD) established in phase Inormal volunteer studies or small phase II proof-of-concept studies. Dose or concentration responsestudies in phase II-generated data are often not fullyanalysed.7,25 Phase III trials are frequently initiatedprior to completion of a full analysis of phase II trials,precluding the opportunity to influence phase III doseselection prior to phase III.

Although the MTD approach may enable efficientdemonstration of effectiveness, it has led, as our datashow, to frequent postmarketing dosage reductionsdue to inadequate pre-marketing dose optimization.To this end, we recommend intensified efforts to deriveadequate understanding of dose–response, dose–concen-tration–response, and underlying pharmacokinetic–pharmacodynamic relationships prior to finalselection, testing, and regulatory approval of theto-be-marketed dosage and target population.

ACKNOWLEDGEMENTS

This study was supported in part by the 2000 DIA Research

Grant Award. We thank James Knoben, at the National

Library of Medicine, for his insights and assistance.

REFERENCES

1. General Requirements on Content and of Labeling for HumanPrescription Drugs. 21 Code of Federal Regulations Sect.201.56 (2000). Available at: http://www.access.gpo.gov/nara/cfr/waisidx_00/21cfr201_00.html. [11 November 2000].

2. Specific Requirements on Content and Format of Labeling forHuman Prescription Drugs. 21 Code of Federal RegulationsSect. 201.57(j) (2000) Available at: http://www.access.gpo.-gov/nara/cfr/waisidx_00/21cfr201_00.html. [11 November2000].

3. US Food and Drug Administration. Index of Safety-relatedDrug Labeling Change Summaries Approved by FDA.Available at: http://www.fda.gov/../medwatch/safety/1999/index99.htm. [19 January 2001].

4. World Health Organization Collaborating Centre forDrug Statistics Methodology. Available at: http://www.whocc.nmd.no/DDs_alterations.htm. [17 January 2001].

5. Cohen JS. Dose discrepancies between the Physician’s DeskReference and the medical literature, and their possible role

KEY POINTS

* Postmarketing changes to labeled dosage regi-mens may reflect suboptimal drug development

* Dosage changes occur frequently and appearoverwhelmingly to be safety motivated

* The rate of these changes is greater for newerdrugs than older drugs

444 j. cross et al.


in the high incidence of dose-related adverse drug events. ArchIntern Med 2001; 161: 957–964.

6. International Conference on Harmonisation. Guidance forIndustry: Dose–response Information to Support Drug Regis-tration. Available at: http://www.fda.gov/cder/guidance/iche4.pdf. [16 February 2001].

7. Slone D, Shapiro S, Miettinen OS, Finkle WD, Stolley PD.Drug evaluation after marketing. Ann Intern Med 1979; 90:257–261.

8. Temple R. Dose–response and registration of new drugs. InDose–Response Relationships in Clinical Pharmacology.Elsevier Science Publishers B.V.: Amsterdam, 1989.

9. Temple R. Government viewpoint of clinical trials. DrugInf J 1982; 1: 10–17.

10. General Requirements on Content and Format of Labeling forHuman Prescription Drugs. 21 Code of Federal RegulationsSect. 201.56(e) (2000). Available at: http://www.access.gpo.gov/nara/cfr/waisidx_00/21cfr201_00.html. [16 January 2001].

11. Cox DR. Regression models and life-tables. J Royal Stat Soc1972; 34: 187–202.

12. Schonfeld D. Partial residuals for the proportional hazardregression model. Biometrika 1982; 69: 239–241.

13. The SAS System1. SAS Institute Inc.: Cary, NC, USA, 1996.14. S-PLUS1. MathSoft Inc.: Cambridge, MA, USA, 1998.15. US Food and Drug Administration. Update from the Office of

Postmarketing Drug Risk Assessment. 36th Annual Meeting ofthe Drug Information Association. 2000 Jun 12; San Diego,USA. Available at: http://www.fda.gov/cder/present/dia-62000/opdra/index.htm. [19 January 2001].

16. US Food and Drug Administration. Adverse Event ReportingSystem (AERS). Available at: http://www.fda.gov/cder/aers/index.htm. [18 December 2000].

17. Spyker DA, Harvey ED, Harvey BE, et al. Assessment andreporting of clinical pharmacology information in drug label-ing. Clin Pharmacol Ther 2000; 67: 196–200.

18. Requirements on Content and Format of Labeling for HumanPrescription Drugs and Biologics: requirements for Prescrip-tion Drug Product Labels. 65 Fed Reg 81081–81131 (2000).Available at: http://www.accessdata.fda.gov/scripts/oc/ohrms/dailylist.cfm?yr¼ 2000&mn¼ 128dy=22. [29 April 2001].

19. US Food and Drug Administration. ODE IV Pilot TargetedProduct Information. Available at: http://www.fda.gov/cder/tpi/default.htm. [2 December 2000].

20. US Food and Drug Administration. Guidance for Industry: Invivo Drug Metabolism/Drug Interaction Studies—StudyDesign, Data Analysis, and Recommendations for Dosing andLabeling. Available at: http://www.fda.gov/cder/guidance/2635fnl.pdf. [17 December 2000].

21. International Conference on Harmonisation: guidance ongeneral considerations for clinical trials. 62 Fed Reg (1997Dec 17). Available at: http://www.fda.gov/cder/guidance/1857fnl.pdf. [17 December 2000].

22. US Food and Drug Administration. Comprehensive List of Gui-dance Documents. Available at: http://www.fda.gov/cder/gui-dance/index.htm. [1 May 2001].

23. US Food and Drug Administration. Guidance for Industry:Population Pharmacokinetics. Available at: http://www.fda.gov/cder/guidance/index.htm. [15 October 2001].

24. Kaitin KI, Healy EM. The new drug approvals of 1996, 1997,and 1998: drug development trends in the user fee era.Drug Inf J 2000; 34: 1–14.

25. Lesko LJ, Wiliams RL. The question-based review: a concep-tual framework for good review practices. Appl Clin Trials1999; 8(10): 56–62.

APPENDIX 1: LISTING OF NMES APPROVED 1980–1999 EXPERIENCING A DOSAGE CHANGE INTHE DRUG LABEL

Type of change, 1¼ decrease; 2¼ increase. Time to change (years) was calculated as the period between regulatory approval

and the date of the PDR1 containing the modified drug label. For drugs withdrawn from the market before publication in the

PDR1, market withdrawal date was used as the date of change.

Generic Approval Time to Type of Approved indicationname date change change

Amoxapine 09/22/80 4.3 1 DepressionMaprotiline 12/01/80 2.1 1 Depression; manic-depressionTemazepam 02/27/81 10.8 1 InsomniaCefotaxime 03/11/81 6.8 1 Anti-infectiveCaptopril 04/06/81 3.7 1 HypertensionKetoconazole 06/12/81 12.6 1 Anti-infective (anti-fungal)Verapamil 08/12/81 4.4 1 Supraventricular tachyarrhythmiaAlprazolam 10/16/81 15.2 1 Anxiety disordersAlprostadil 10/16/81 6.2 1 Ductus arteriosus (neonatal)Isotretinoin 05/07/82 3.7 1 Severe acnePindolol 09/03/82 1.3 1 HypertensionDiltiazem 11/05/82 5.2 2 Angina (chronic stable and pectoris)Triazolam 11/15/82 15.1 1 InsomniaEconazole 12/23/82 8.0 1 Anti-fungal (tinea pedis/cruris/corporis)Guanadrel 12/29/82 9.0 1 HypertensionIndapamide 07/06/83 10.5 1 Hypertension; sodium fluid retention associated with congestive

heart failure

Continues



APPENDIX 1: Continued

Generic Approval Time to Type of Approved indicationname date change change

Etoposide 11/10/83 14.2 1 Testicular tumorsPimozide 07/31/84 3.4 1 Tourette’s disorderLeuprolide 04/09/85 6.7 1 Prostate cancerTerfenadine 05/08/85 8.7 1 Seasonal allergic rhinitisBetaxolol 08/30/85 7.3 2 Ocular hypertension; open angle glaucomaFlecainide 10/31/85 2.2 1 Ventricular arrhythmiasMidazolam 12/20/85 2.0 1 Sedation; induction of general anesthesiaNorfloxacin 10/31/86 6.2 1 Anti-infective (bacterial)Zidovudine 03/19/87 3.8 1 HIVLovastatin 08/31/87 4.3 1 Cholesterol reductionFluoxetine 12/29/87 3.0 1 DepressionUrsodiol 12/31/87 4.0 2 GallstonesNaftifine 02/29/88 2.8 1 Anti-fungal (tinea cruris/corporis)Diclofenac 07/28/88 8.4 1 RA, OA, ankylosing spondylitisOctreotide 10/21/88 6.2 2 Carcinoid and vasoactive intestinal peptide tumorsAstemizole 12/29/88 5.0 1 Seasonal allergic rhinitis; chronic urticariaOxistat 12/30/88 5.0 2 Anti-fungal (dermatological infections)Selegiline 06/05/89 2.6 1 Parkinson’s diseaseClozapine 09/26/89 3.3 1 SchizophreniaPropofol 10/02/89 5.3 1 AnesthesiaKetorolac 11/30/89 6.1 1 PainGoserelin 12/29/89 3.0 1 Prostate cancer (carcinoma)Fluconazole 01/29/90 7.9 1 Anti-infective (candidiasis, meningitis)Halobetasol 12/17/90 2.0 1 Corticosteroid (inflammation, pruritis)Ofloxacin 12/28/90 7.0 1 Anti-infective (bacterial)Etodolac 01/31/91 5.9 1 RA, OA, pain managementFelodipine 08/25/91 4.4 1 HypertensionFoscarnet 09/27/91 5.3 1 CMV retinitis (in AIDS patients)Clarithromycin 10/31/91 4.1 1 Anti-infective (bacterial)Quinapril 11/19/91 6.1 2 HypertensionMivacurium 01/22/92 2.9 2 Muscle relaxant (neuromuscular blocker)Cefpodoxime 08/07/92 5.4 1 Anti-infective (bacterial)Itraconazole 09/11/92 3.3 1 Anti-fungalSumatriptan 12/28/92 5.0 1 MigraineCisapride 07/29/93 2.4 1 GI motility stimulatorFelbamate 07/29/93 2.4 1 EpilepsyTrimetrexate 12/17/93 5.0 2 PneumoniaAprotinin 12/29/93 6.0 1 Blood loss associated with coronary artery bypass graft surgeryCalcipotriene 12/29/93 4.0 1 Scabies; pruritic skinGranisetron 12/29/93 4.0 2 Nausea/vomiting associated with cancer treatmentRisperidone 12/29/93 5.0 2 Psychotic disordersFluvastatin 12/31/93 3.0 2 HypercholesterolemiaTacrolimus 04/08/94 3.7 1 Organ rejectionImiglucerase 05/23/94 5.6 1 Gaucher diseaseDirithromycin 06/19/95 3.5 1 Bacterial infectionLamivudine 11/17/95 4.1 1 HIVIndinavir 03/13/96 2.8 2 HIVIrinotecan 06/14/96 3.6 2 AntineoplasticCidofovir 06/26/96 1.5 1 CMV retinitisButenafine 10/18/96 2.2 2 Interdigital tinea pedisTroglitazone 01/29/97 1.9 1 Type II diabetesMibefradil 06/20/97 1.0 1 Hypertension; chronic stable angina pectorisCerivastatin 06/26/97 2.5 2 HypercholesterolemiaBromfenac 07/15/97 0.9 1 PainTrovafloxacin 12/18/97 2.0 1 AntibacterialTolcapone 01/29/98 1.9 1 Parkinson’s diseaseSildenafil 03/27/98 1.8 1 Erectile dysfunction

446 j. cross et al.
