Smalley W, Shatin D, Wysowski DK, Gurwitz J, Andrade SE, Goodman M, Chan KA, Platt R, Schech SD, Ray WA. Contraindicated Use of CisaprideImpact of Food and Drug Administration Regulatory Action. JAMA. 2000;284(23):3036-3039. doi:10.1001/jama.284.23.3036
Author Affiliations: Departments of Medicine and Preventive Medicine, Vanderbilt University School of Medicine and Geriatric Research, Education and Clinical Center, and VA Medical Center, Nashville, Tenn (Drs Smalley and Ray); Center for Health Care Policy and Evaluation, UnitedHealth Group, Minnetonka, Minn (Dr Shatin and Ms Schech); Office of Postmarketing Drug Risk Assessment, Food and Drug Administration, Rockville, Md (Dr Wysowski); Fallon Healthcare System and Meyers Primary Care Institute, University of Massachusetts Medical School, Worcester (Dr Gurwitz); College of Pharmacy, University of Rhode Island, Kingston (Dr Andrade); HealthPartners Research Foundation, Bloomington, Minn (Dr Goodman); Channing Laboratory, Brigham and Women's Hospital (Drs Chan and Platt), Department of Ambulatory Care and Prevention, Harvard Pilgrim Health Care (Dr Platt), and Harvard Medical School, (Drs Chan and Platt) Boston, Mass.
Context Cisapride, a gastrointestinal tract promotility agent, can cause life-threatening
cardiac arrhythmias in patients susceptible either because of concurrent use
of medications that interfere with cisapride metabolism or prolong the QT
interval or because of the presence of other diseases that predispose to such
arrhythmias. In June 1998, the US Food and Drug Administration (FDA) determined
that use of cisapride was contraindicated in such patients and informed practitioners
through additions to the boxed warning in the label and a "Dear Health Care
Professional" letter sent by the drug's manufacturer.
Objective To evaluate the impact of the FDA's 1998 regulatory action regarding
contraindicated use of cisapride.
Design and Setting Analysis of data for the 1-year periods before (July 1997-June 1998)
and after (July 1998-June 1999) the regulatory action from the population-based,
pharmacoepidemiology research databases of 2 managed care organizations (sites
A and B) and a state Medicaid program (site C).
Participants Patients with at least 180 days of prior enrollment in 1 of the 3 sites
who were prescribed cisapride at least once in the period before (n = 24 840)
or after (n = 22 459) regulatory action. Patients could be included in
Main Outcome Measures Proportion of cisapride users in each period for whom cisapride use
was contraindicated by the product label, based on computerized patient medical
Results In the year prior to regulatory action, cisapride use was contraindicated
for 26%, 30%, and 60% of users in study sites A, B, and C, respectively. In
the year after regulatory action, use was contraindicated for 24%, 28%, and
58% of users, a reduction in contraindicated use of approximately 2 per 100
cisapride users at each site. When the analysis was restricted to new users
of cisapride after regulatory action, only minor reductions in contraindicated
use were found.
Conclusion The FDA's 1998 regulatory action regarding cisapride use had no material
effect on contraindicated cisapride use. More effective ways to communicate
new information about drug safety are needed.
Cisapride is a gastrointestinal tract promotility agent that was first
marketed in the United States in August 1993 with a label indication for nocturnal
heartburn.1 Use grew rapidly so that in 1995
there were approximately 5 million outpatient cisapride prescriptions filled
in the United States.2 However, by this time,
the Food and Drug Administration (FDA) had received 34 cases of torsade de
pointes and 23 of prolonged QT interval in cisapride users, including 4 deaths.3 Since many of these cases were in patients taking
drugs that inhibited the cytochrome P450-3A4 enzymes that metabolize cisapride,1 a "black-box" warning was added to the cisapride label
in 1995 contraindicating use in patients taking drugs that affected cisapride
metabolism and the manufacturer sent a "Dear Health Care Professional" letter.
In April 1996, more information on the cases was published.3
The second report also suggested that use of cisapride could be hazardous
in patients taking medications that prolonged the QT interval or with chronic
diseases that predisposed to cardiac arrhythmias.
Despite the black-box warning, use of cisapride continued to increase
in the United States so that in 1998 there were 7 million prescriptions dispensed.2 Further information accumulated on the potential adverse
effects of cisapride. Clinical studies demonstrated that cisapride prolonged
the QT interval in healthy volunteers,1 which
suggested a mechanism underlying the reported cases of torsade de pointes.3 Further cases of serious arrhythmias were reported.4- 6 The FDA thus expanded
the black-box warning in June 1998 to note that cisapride use also was contraindicated
in patients taking medications that could prolong the QT interval or in patients
with baseline heart disease or other conditions that could predipose to cardiac
arrhythmias. The FDA also drew attention to this change through circulation
of a press release and the manufacturer of cisapride distributed a Dear Health
Care Professional letter informing practitioners of the revised label.7 This letter was distributed to 800 000 health
care professionals in the United States including physicians (both primary
care and specialists), pharmacists (retail and hospital), and vendors of drug-alert
databases, such as Medispan and First Data Bank.
However, the effect of the 1998 and previous FDA regulatory actions
is unknown. This study was conducted to quantify the prevalence of contraindicated
use of cisapride before and after the June 1998 warning, using data from 3
large populations with computerized records of medical encounters.
The amended cisapride labeling and Dear Health Care Professional letter
were dated June 26, 1998. Thus, the study periods included the years before
(July 1997-June 1998) and after (July 1998-June 1999) this regulatory action.
The study was conducted at 3 pharmacoepidemiology research sites, each
of which included a defined population enrolled in a health plan for which
there are automated records of medical care encounters, including prescriptions
dispensed, hospital admissions, and outpatient visits.8,9
Each of the 3 study sites conducts postmarketing surveillance studies in collaboration
with the FDA, but this activity is not linked to clinical practice or guidelines
at any of the sites. The 3 sites included a large managed care organization
(site A) with geographically dispersed health plans (primarily independent
practice associations); a consortium (site B) of 3 health maintenance organizations
(2 on the Eastern seaboard and 1 in the Midwest) that included independent
practice associations and staff and group practice models; and a state Medicaid
program (site C). During the study period, there were no changes in membership
or coverage in any of the 3 sites that were material to this study.
Each of the sites used computerized plan files to provide study data.
Enrollment files identified persons who qualified for the study and provided
demographic information. Pharmacy files identified prescription drug use and
included information on the date a drug was dispensed and the prescribed days
of supply. The latter defined medication use days, or those days on which
a medication was likely to be taken. Inpatient and outpatient files were used
to identify previous encounters for medical conditions identified by International Classification of Diseases, Ninth Revision, Clinical
Modification (ICD-9-CM)10 or Current Procedural Terminology, Fourth Edition (CPT-4)11
codes. Confidentiality was preserved because the files used for the study
analysis did not have linkable individual identifiers. The study was approved
by each site's committee for the protection of human subjects.
The cohorts for the primary analysis consisted of persons who had at
least 1 prescription for cisapride dispensed during either year and, prior
to the dispensing of the first such prescription, had at least 180 days of
continuous enrollment (required to identify contraindicated medical conditions).
Two cisapride cohorts thus were identified, 1 for each of the years before
and after the regulatory intervention. A person was included in both cohorts
if cisapride was dispensed in both years and enrollment criteria were met.
In a secondary analysis, we identified the patients in each cohort with a
new episode of cisapride use, as these patients should provide a better test
of the effect of the regulatory intervention. New use was defined as dispensing
of a cisapride prescription with no cisapride use in the preceding 180 days;
if in a given year there was more than 1 episode of new use, only the first
was included in the study.
Cisapride use was defined as contraindicated if there was at least 1
day that dispensed prescriptions indicated was both a day of cisapride use
and a concurrent day of use for a medication listed in the label as contraindicated.
These were drugs that interfere with P450-3A4 metabolism (clarithromycin,
erythromycin, troleandomycin, nefazodone hydrochloride, fluconazole, itraconazole,
ketoconazole, indinavir, ritonavir) or those that can prolong QT intervals
(class IA or III antiarrhythmics, cyclic antidepressants, and antipsychotics).
Use was also defined as contraindicated if either an inpatient or outpatient
encounter record indicated a medical condition listed as contraindicated in
the cisapride label in the 180 days preceding dispensing of any cisapride
prescription in that year. The listed conditions were ventricular arrhythmia,
heart failure, other ischemic heart disease, electrolyte disorders, renal
failure, or respiratory failure. If the use days of a cisapride prescription
encompassed both study years, we classified all days for the prescription
according to the year in which it was dispensed. Detailed lists of medications
and other study codes are available on request.
For each of the study years, we calculated the proportion of cisapride
users for whom such use was contraindicated according to the revised label.
We calculated the difference between the proportions for the before and after
years to estimate the reduction in contraindicated use. We calculated 95%
confidence intervals based on normal approximation to the binomial. To compensate
for the fact that some cisapride users would be in both the before and after
cohorts, the width of the confidence intervals was adjusted to reflect the
numbers of unique individuals in each cohort.
During the study period, the 3 study sites included approximately 6.8
million insured lives. During the year preceding the regulatory intervention,
there were 24 840 qualifying cisapride users at sites A (n = 13 613),
B (n = 6848), and C (n = 4379) (Table 1). The proportions of cisapride users that were female were 60%,
61%, and 68%, respectively. The respective proportions for ages 0 to 2 years
were 12%, 13%, and 8% and the respective proportions for age 60 years and
older were 14%, 23%, and 45%. In the year following the regulatory intervention,
there were 22 459 qualifying cisapride users, reflecting a slight drop
in numbers from each of the sites. There were only minor changes in the demographic
characteristics of cisapride users in the postintervention year.
In each of the sites in the year before the regulatory intervention,
the proportion of cisapride users whose use was contraindicated because of
concomitant medication was high, with respective proportions of 14% (1946/13 613),
19% (1280/6848), and 34% (1481/4379) for sites A, B, and C (Table 1). The most common contraindicated medications were amitriptyline,
erythromycin, and clarithromycin. Similarly, the proportion of users with
use contraindicated because of concomitant conditions was also high, with
respective proportions of 15% (2029/13 613), 15% (1047/6848), and 41%
(1809/4379) for sites A, B, and C, respectively. The most common conditions
were heart failure and other ischemic heart disease. In the year before the
regulatory action, the proportions of cisapride users with use contraindicated
because of either a concomitant medication or condition were 26% (3506/13 613),
30% (2025/6848), and 60% (2614/4379), respectively, for the 3 study sites.
There was little change in contraindicated use of cisapride in the year
after the FDA regulatory action (Table 1). During this year, the proportions of study cisapride users with
use contraindicated because of either a concomitant medication or condition
were 24% (2963/12 418), 28% (1598/5812), and 58% (2432/4229), respectively,
for sites A, B, and C; a reduction in contraindicated use of approximately
2 patients per 100 cisapride users at each site.
We conducted a separate analysis of patients with a new episode of cisapride
use because they had the greatest likelihood of benefitting from the regulatory
intervention (Figure 1). At each
of the 3 sites, there were only minor reductions in the proportions of new
users of cisapride for whom such use was contraindicated.
The findings of this population-based study show that a substantial
proportion of patients dispensed cisapride at each of the study sites also
had concurrent medications or concomitant medical conditions thought to increase
the risk of potentially lethal cardiac arrhythmias in these patients. There
was no material reduction in such use at any of the study sites following
FDA regulatory action, which included a black-box warning in the cisapride
label indicating such use was contraindicated and a Dear Health Care Professional
letter. Thus, despite this and previous modifications to the cisapride label,
in the year ending in June 1999, cisapride use was contraindicated for between
24% and 58% of patients dispensed this drug. Even the group most likely to
be affected by the regulatory action—patients beginning a new episode
of cisapride use—had no material change in contraindicated use.
Some of the variation among the 3 sites with regard to absolute proportion
of cisapride use that was contraindicated is probably explained by differences
in the populations served at each site. The study included a Medicaid population,
known to overrepresent patients of advanced age and with chronic illnesses.8 This population thus was likely to include more persons
with medication use or chronic illnesses that would contraindicate cisapride
use. Conversely, the study included plans with employment-based insurance,
in which the prevalence of contraindications to cisapride use may be lower.
Despite these baseline differences, data from the 3 sites were consistent
in that the proportion of cisapride use that was contraindicated was high
in absolute terms and did not change materially following the regulatory intervention.
In March 2000, prior to when an FDA advisory committee was scheduled
to review cisapride's benefits and risks for the approved indication, the
manufacturer terminated marketing of cisapride in the United States effective
as of July 2000.12 Thus, the public health
problem of contraindicated use of cisapride has been resolved in the United
States, although cisapride continues to be marketed in other countries. However,
4 years prior to this action, case reports received by the FDA and published
in the scientific literature3 had identified
the vulnerability of cisapride users taking contraindicated drugs or having
contraindicated conditions to also be at an increased risk of serious arrhythmias.
During this period, millions of patients received cisapride; our data suggest
these would have included hundreds of thousands in whom such use was deemed
contraindicated as of June 1998. Our data also indicate that in the 12 months
following this regulatory action, hundreds of thousands of patients in whom
cisapride use was contraindicated were likely to have received this drug.
The exposure of these patients to inappropriate cisapride use, despite the
prominent publication of case reports, label changes, and Dear Health Care
Professional letters, highlights the need to develop more effective methods
for modifying practice to reflect new information about a drug's risks and