Jones JK, Fife D, Curkendall S, Goehring, Jr E, Guo JJ, Shannon M. Coprescribing and Codispensing of Cisapride and
Contraindicated Drugs. JAMA. 2001;286(13):1607-1609. doi:10.1001/jama.286.13.1607
Author Affiliations: The Degge Group, Ltd, Arlington, Va (Drs Jones and Curkendall and Mr Goehring); Department of Epidemiology, Drug Safety and Surveillance, Janssen Research Institute, Titusville, NJ (Dr Fife); College of Pharmacy, University of Cincinnati Medical Center, Cincinnati, Ohio (Dr Guo); Jorie Kent Associates, Inc, Milton, Del (Ms Shannon). Corresponding Author and Reprints: Suellen Curkendall, PhD, Director of Research, The Degge Group, Ltd, 1616 N Fort Myer Dr, No. 1430, Arlington, VA 22209 (e-mail: firstname.lastname@example.org).
Context Cisapride, an oral prokinetic drug indicated for the
symptomatic treatment of nocturnal heartburn due to gastroesophageal
reflux disease, was approved by the US Food and Drug Administration in
July 1993. After reports of serious cardiac arrhythmias and deaths
during administration of cisapride, most involving concomitant exposure
to another drug, a series of label changes and warnings were issued in
February 1995, October 1995, June 1998, and June 1999. Cisapride was
removed from general distribution in July 2000.
Objective To determine the frequency of contraindicted
coprescribing and codispensings, in which cisapride and a
contraindicated drug were prescribed or dispensed to the same patient
for overlapping periods, and the proportion of contraindicated
coprescribing by the same physicians and codispensing by the same
Design and Setting Retrospective study of prescription claims from
a managed care organization database for all patients with cisapride
prescriptions between July 1993 and December 1998.
Participants A total of 38 757 adult and pediatric patients
who had a cisapride prescription immediately preceded by at least 60
days of insurance eligibility.
Main Outcome Measure Proportion of cisapride prescriptions or
dispensing occurring during the same treatment period as a drug
contraindicated at that time prescribed by the same physicians or
dispensed by the same pharmacies.
Results Of 131 485 cisapride prescriptions dispensed after
the warnings began, 4414 (3.4%) overlapped with at least 1 drug
contraindicated in the labeling at the time of the prescription. Of all
overlapping prescription pairs, 2190 (50%) were by the same
physicians, 3908 (89%) were by the same pharmacies, and 765 (17%)
were dispensed on the same day.
Conclusion Prescriptions dispensed by the same pharmacies
accounted for a far higher proportion of contraindicated medication
pairs than prescriptions from the same physicians. The pharmacy may be
an important and underutilized intervention point to prevent
contraindicated drugs from being used together.
Use of contraindicated drug combinations can result in serious adverse
reactions. Cisapride, an oral prokinetic agent indicated for the symptomatic
treatment of nocturnal heartburn due to gastroesophageal reflux disease,1 was approved for use in the United States in July
1993 and voluntarily removed from the market in July 20002
after more than 270 serious cardiac arrhythmias, including 70 fatalities,
were reported among patients in the United States.3- 5
Most case reports indicated that cisapride was taken concurrently with one
or more drugs with which it interacts. Four label changes and physician notifications
("dear doctor" letters) were issued in February 1995, October 1995, June 1998,
and June 1999 (1999 is outside of our study period).
Traditionally, physician notification letters and drug label changes
are designed for and sent to physicians to alert them to potential problems
with a drug, although the information is also provided to pharmacies. The
letters often include a list of drugs that, taken concomitantly, could predispose
the patient to an adverse reaction. Label changes have not been entirely successful
and makers of health policy continue to search for more effective methods
of preventing inappropriate drug use.6- 9
Since a previous study of cisapride prescribing indicated that there was little
change in prescribing cisapride to patients with medical or drug contraindications
after the June 1998 label change,10 we
studied the extent of coprescribing by the same physicians and codispensing
by the same pharmacies.
Protocare Sciences' managed care claims database contains private health
care claims and enrollment data representing health care services provided
to approximately 3 million members of several preferred provider organizations
and health maintenance organizations. The database covers a wide geographic
area, with members residing predominantly in 22 states. Medical claims and
patient eligibility data are linked with a unique, encrypted patient identification
number (ID). During the period that a patient is eligible for drug benefits
coverage, the database contains all outpatient prescription claims filed with
the insurer, regardless of where the prescription is filled. Each prescription
claim record includes the patient ID, the dispensing date, a national drug
code (NDC), and an encrypted Drug Enforcement Administration (DEA) number
identifying a unique physician. An encrypted provider ID that identifies a
unique pharmacy was also provided by the data source for the subset of 10 491
study prescriptions that were overlapping contraindicated prescriptions. This
provider ID represents an individual pharmacy location, regardless of whether
the pharmacy belongs to a chain.
All claims for patients taking cisapride between July 1993 and December
1998 were collected (49 485 patients). Patients with unknown insurance
eligibility, age, or sex, or with incomplete pharmacy claims, were excluded
(6754 patients). Patients younger than 15 years with prescriptions for the
tablet (vs liquid) form of cisapride were excluded because days of drug supply
estimates are unreliable in this population (2023 patients). Patients without
at least 1 cisapride prescription immediately preceded by 60 days of insurance
eligibility were excluded (1951 patients). The final study population was
38 757 patients.
Prescriptions for cisapride and contraindicated drugs were identified
and the number of cisapride prescriptions that were overlapped with a contraindicated
drug was computed. Refills were counted as separate prescriptions. A drug
was only counted as contraindicated beginning in the month that the physician
notification letter was sent to inform professionals that the cisapride label
was changed to include that drug. The overlap of cisapride and a contraindicated
drug was defined to exist when cisapride and the contraindicated drug were
dispensed either on the same day or on overlapping therapy days. If a single
prescription for a contraindicated drug overlapped during more than 1 cisapride
prescription, each overlapping prescription was counted as a separate overlap.
Beginning in February 1995, the month of the first warning, we examined
the proportion of the overlapping contraindicated prescriptions prescribed
by the same physician and the proportion dispensed by the same pharmacy.
Sixty-one percent of the study population were women, 3% were younger
than 15 years (received the liquid form), 17% were aged 15-39 years, 41% aged
40-64 years, and 39% 65 years or older.
There were 141 119 cisapride prescriptions dispensed during the
study period. Of these, 131 485 were dispensed from February 1995 through
December 1998, after the warnings began, and 4414 (3.4%) overlapped with at
least 1 contraindicated drug. The proportion overlapped was 1.9% in 1995,
4.3% in 1996, 3.5% in 1997, and 3.2% in 1998. The overlaps involved 2657 patients,
3301 physicians, and 1715 pharmacies. Of the patients, 1716 (65%) were women,
58 (2%) were younger than 15 years old, 379 (14%) were aged 15-39 years, 1339
(50%) were aged 40-64 years, and 881 (33%) were 65 years or older.
The most frequent overlapping contraindicated drugs were clarithromycin
(43% of all overlapping prescriptions), erythromycin (26%), and fluconazole
(22%). The average days of supply were 26 for cisapride, 10 for clarithromycin,
10 for erythromycin, 9 for fluconazole, 18 for ketoconazole, 19 for itraconazole,
29 for nefazodone, 30 for troleandomycin, 21 for indinavir, and 27 for ritonavir.
Thus, concurrent use was mainly associated with prescription pairs that were
dispensed within 60 days.
Of all 4414 overlapped prescription pairs, 2190 (50%) were from the
same physicians for the same patients, while 3908 (89%) were from the same
pharmacies for the same patients, and 765 (17%) were dispensed on the same
day for the same patient. Figure 1
shows the trend in the proportion of the contraindicated prescriptions that
were associated with the same pharmacy or the same prescriber each month.
In nearly every month, more than 85% of the contraindicated drugs were dispensed
by the same pharmacy that had dispensed the prescriptions for cisapride.
Of the overlapped prescriptions, 2125 involved a refill prescription
for one or both of the medications. When refills were involved, 955 (45%)
of the prescriptions were from the same physician for the same patient, while
1870 (88%) were dispensed by the same pharmacies for the same patient. For
the 2289 combinations of new prescriptions, 1235 (54%) were from the same
physicians and 2038 (89%) were dispensed by the same pharmacy that had dispensed
the prescription for cisapride.
Throughout the study period, most patients with overlapping contraindicated
dispensings had obtained both drugs from the same pharmacy. This suggests
that a pharmacy-based intervention to prevent codispensing of contraindicated
medication pairs could be accomplished without involving complicated communications
among different pharmacies.
Because the present study was based on dispensings, we were unable to
determine how many contraindicated medication pairs were successfully identified
by pharmacists and not dispensed or how often the pharmacists called physicians
to question the prescriptions. We also could not determine how often pharmacists
dispensed overlapping prescriptions for cisapride and a contraindicated medication
but instructed the patient to discontinue one of the medications while taking
the other. Other evidence suggests that while most pharmacies have computer-based
warning systems, these systems do not consistently prevent the dispensing
of contraindicated drugs.11 Possible reasons
for such inconsistent prevention include label design that embeds contraindications
in a large volume of other material, warning systems that do not present current
information in a rationally prioritized layout, and pharmacists' concerns
about questioning the prescribing physician's decision.
The 3% of cisapride prescriptions that overlapped with contraindicated
medications represent a substantial and preventable risk to patients. If pharmacists
and physicians had better methods to use the information available to them,
they might be able to reduce this risk for other contraindicted drug combinations.
We found that 89% of the contraindicated codispensings were associated with
the same pharmacies that also dispensed the prescriptions for cisapride, vs
50% that occurred when the same physician wrote both prescriptions. This suggests
that the pharmacist, who is focused solely on the therapy, more often has
the opportunity to apply critical information necessary to prevent contraindicated