Prevalence of drug–drug interactions in geriatric patients at an ambulatory care pharmacy in a tertiary care teaching hospital
Aljadani and Aseeri BMC Res Notes
Prevalence of drug-drug interactions in geriatric patients at an ambulatory care pharmacy in a tertiary care teaching hospital
Rawabi Aljadani 0
Mohammed Aseeri 1
0 Pharmacy Department, King Saud bin AbdulAziz University for Health Sciences/King AbdulAziz Medical City , Riyadh , Saudi Arabia
1 Clinical Pharmacy Services, Pharmacy Department, King Saud bin AbdulAziz University for Health Sciences/King AbdulAziz Medical City , PO Box 9515, Jeddah 21423 , Saudi Arabia
Objective: A cross-sectional study was performed from February to May 2015, to estimate the prevalence of drugdrug interactions in geriatric patients at the ambulatory care pharmacy at King Abdul-Aziz Medical City in Jeddah, Saudi Arabia. Results: A total of 310 patients were included, with a mean age (± SD) of 73.78 ± 6.96, and 48.70% were female. The overall prevalence of DDIs of all categories was 90.64%. Category B DDIs was 55.80%, category C DDIs 87.74%, category D DDIs 51.93%, and category X DDIs 16.45%. Atorvastatine plus omeprazole was identified as the most common interacting pair, with a prevalence of 25.26%. Multivariate logistic regression analysis showed that category D or X DDIs are more likely to occur in the female patient (OR = 1.79; 95% CI 1.07, 2.97), the patient taking more than three medications (OR = 22.62; 95% CI 2.93, 174.83), and the patient with more than two conditions (OR = 3.09; 95% CI 1.81, 5.27).
Prevalence; Drug-drug interactions; Geriatric; Ambulatory care; Pharmacy
The issue of poly-pharmacy in geriatric patients has
led to concern over an increase in drug–drug
interactions (DDIs) prevalence. The precise minimum number
of medications used to define “poly-pharmacy” varies
but generally ranges from 5 to 10 [
]. Some combined
medications may result in undesired pharmacodynamic
or pharmacokinetic interactions, leading to
under-treatment or harmful effects [
]. DDIs are an important cause
of adverse drug reactions. Approximately 3–26% of all
adverse drug reactions that require hospital admission
are caused by drug–drug interactions [
involving community-dwelling individuals report prevalence
rates ranging from 4.6 to 17.6% [
]. Studies conducted
in hospital settings report prevalence rates ranging
from 0.6 to 18.3%, and a study carried out in a primary
health care unit reports a prevalence rate of 6.8% [
Janchawee studied the pharmacoepidemiologic of
potential drug–drug interactions in the outpatient pharmacy of
a university hospital, where the overall rate of potential
DDIs was 27.9%, with a maximal value of 57.8% at the
Department of Psychiatry. The rate of the most
potentially significant interactions was 2.6%, being the
highest in the Department of Medicine (6%), with isoniazid
vs. rifampin as the most common interacting
combination . Another study of drug–drug interactions
potential pattern in a tertiary care teaching hospital diabetic
outpatient pharmacy, including a total of 182 patients,
found a higher risk of DDIs among patients 51–60 years
of age (43 patients, or 23.6%), It was found that 10 (5.3%)
of the potential drug–drug interactions were major, 5
(2.7%) minor, and 174 (92.1%) moderate, the most
common potential drug–drug interaction observed was
between metformin and enalapril (n = 64) [
though nature and prevalence of potential DDIs has been
widely studied in different settings, DDIs nature in the
ambulatory setting still a gray area with many
contradictions, ambulatory care pharmacy serves a huge number
of patients on a daily basis, should provide the highest
possible level of quality, with the understanding that
safety assurance is a cornerstone for the quality of any
type of service. We realized the harmful effect of drug–
drug interactions on patients’ conditions and progress,
as well as the effect on the pharmacoeconomics of the
hospital itself. Based on that, this study was conducted to
estimate the prevalence of drug–drug interactions in
geriatric patients at the ambulatory care pharmacy at King
Abdul-Aziz Medical City in Jeddah, Saudi Arabia.
This cross-sectional study was approved by King
Abdullah International Medical Research Center (KAIMRC)
and was conducted to estimate the prevalence of drug–
drug interactions in geriatric patients at the ambulatory
care pharmacy at King Abdul-Aziz Medical City in
Jeddah, Saudi Arabia, from February to May 2015.
Inclusion and exclusion criteria
We included patients 65 years and older of both
genders who visited the ambulatory care pharmacy at King
Abdul-Aziz Medical City in Jeddah and who had a drug
profile containing two or more medications, drug
profile containing herbal products or topical products only
including: creams, ointments, gels, patches, drops, sprays
and inhalers were excluded.
In this study, we did a retrospective review of 310
patients’ medication profiles of ambulatory care
pharmacy dispensed prescriptions. The assigned pharmacist
retrieved the dispensed prescriptions randomly. Then
the baseline characteristic and patient demographics
data (including patient age, gender, number of drugs, and
chronic illness at the time of dispensing) and the patient
drug profile data (including drug, dose, and route) were
collected. The drug profile for each patient was
analyzed by Lexi-Interact (a comprehensive drug-to-drug,
drug-to-herb, and herb-to-herb interaction analysis
program); Lexi-Interact categorized DDIs into five
categories according to its risk rating, category A: data have
not demonstrated either pharmacodynamic or
pharmacokinetic interactions between the specified
medications, B: the specified medications may interact with each
other, but there is little to no evidence of clinical concern
resulting from their concomitant use, C: the medications
agents may interact with each other in a clinically
significant manner, but the benefits of concomitant use of these
two medications usually outweigh the risks, D: the two
medications may interact with each other in a clinically
significant manner, a patient-specific assessment must
be conducted to determine whether the benefits of
concomitant therapy outweigh the risks, and X: the specified
medications may interact with each other in a clinically
significant manner, but the risks associated with
concomitant use of these medications usually outweigh the
Descriptive statistics were used to describe continuous
(Mean ± SD) and categorical variables (frequency and
percent). We used Chi square test to assess the
association between D or X DDIs and categorical variables,
while Independent-t test was used to assess the
association between D or X DDIs and continuous variables.
Prevalence was used to determine the proportion of
geriatric patients visiting the ambulatory care pharmacy
who have a medication profile containing at least one
interacting pair with no regard for DDI risk rating.
Prevalence was also used to determine the proportion of
geriatric patients with at least one interacting pair for each
drug–drug interaction risk rating (B, C, D, and X) alone.
It was also used to estimate the proportion of patients for
each one of the most common interacting pairs.
Multivariate logistic regression analysis was used to determine
if gender, age, number of drugs, and chronic conditions
can predict D and X DDIs risk rating prevalence. The
odds ratio (OR) and respective confidence interval (CI)
was calculated in this analysis for each variable. Wald Chi
square test P values < 0.05 were considered statistically
significant. Using IBM SPSS Statistics for Microsoft,
Version 22.0, we preformed all the statistical analysis [
A total of 310 patients were included in this study, with
a mean age of 73.8 (± SD 7) years. Approximately half
of them (48.70%) were female. The majority (91.3%) had
a prescription with more than three medications, and
62.3% had more than two chronic conditions. Baseline
characteristics and patient demographics are listed in
Drug–drug interactions prevalence
The overall prevalence of DDIs of all categories was
90.64% (95% CI 86.8%, 93.6%). Category B DDIs
prevalence was 55.80% (95% CI 50.1%, 61.4%), category C DDIs
87.74% (95% CI 83.6%, 91.2%), category D DDIs 51.93%
(95% CI 46.2%, 57.6%), and category X interacting pairs
16.45% (95% CI 12.5%, 21.1%).
Most common interacting pairs
Atorvastatin + omeprazole was identified as the most
common interacting pair, with a prevalence of 25.26%.
Next was atorvastatin + calcium, with a prevalence of
Interactions D or X DDIs
22.90%. The prevalence of all the identified most
common interacting pairs, their risk rating, severity,
reliability rating, predicted impact on the clinical outcome, and
suggested patient management plan are all mentioned in
Drug–drug interactions predictors
Univariate logistic regression analysis including all 310
patients showed that the odds of D or X DDIs were 44
times higher when more than three medications were
prescribed (OR = 44.75, 95% CI 5.99, 334.61). The
percentage of D or X DDIs was higher in patients taking
more than three medications (3.7% were using ≤ 3
medications vs. 63.3% using > 3 medications, P = 0.001). Also,
univariate analysis revealed that the odds of D or X DDIs
were four times higher when the patient had more than
two chronic conditions (OR = 4.21, 95% CI 2.59, 6.86).
When we compare the number of chronic conditions
among patients with D or X DDIs, we can see that the
percentage of patients with more than two conditions
who have D or X DDIs almost doubles the percentage
of patients with two or fewer conditions (36.8% with ≤ 2
conditions vs. 71.0% > 2 conditions, P = 0.001). The
combined effects of all the factors were further investigated
by conducting a multivariate logistic regression
analysis, which indicated that D or X DDIs were more likely
to occur in the female patient (OR = 1.79; 95% CI 1.07,
2.97), the patient taking more than three medications
(OR = 22.62; 95% CI 2.93, 174.83), and the patient with
more than two conditions (OR = 3.09; 95% CI 1.81, 5.27).
Regression analysis results are stated in Table 3.
The overall DDIs prevalence estimate (90.64%) could be
misleading if not linked to the expected clinical impact
of each category, to correctly judge these results, we must
consider that category C, which has the highest influence
on this estimate, has less of a clinical impact compared to
category D and X DDIs. D and X DDIs’ estimated
prevalence of 51.93 and 16.45% respectively reflects the need
for in-depth investigations. Björkman et al. studied drug–
drug interactions in elderly outpatients in six European
countries. They found that 46% of patients had at least
one potential clinically significant DDI, and 10% of these
DDIs were classified as (to be avoided) according to the
Swedish interaction classification system [
et al. reported a moderate DDIs prevalence of 69.9%, and
a 21.2% prevalence of major DDIs. The difference in the
results may be attributed to the use of different DDIs
analysis database; DDIs categories differ according to the
database used, the difference in the population
characteristics, and the sample size of these studies [
]. A direct
comparison can’t be done without a further adjustment
for these factors. Other studies reported the prevalence
of DDIs either in different settings, such as the
hospital, or in a specific population, such as cancer patients,
so also can’t be directly compared with the results of this
Several studies investigated the most common
interacting pairs, and different results were reported. This is to
be expected due to the variations in the medication
availability and medical practice of each institute. Of the top
20 identified most common interacting pairs we
identified, atorvastatine was responsible for four, as was
aspirin. Hypoglycemic agents, PPIs, and calcium were each
responsible for three interacting pairs. Dinesh reported
aspirin as the fourth-ranked of the top 10 medications
with a high risk for DDIs [
Our multivariate logistic regression analysis results
agree with the results of many previous studies. Neto
et al. studied the prevalence and predictors of potential
drug–drug interactions in the elderly in the Brazilian
iiltccenh iltccaooum rtscaeeeh r-ceuodCA r-cceuonnm iiirtsseohnb -cceuonnm
on iln iyn GM se tca rse
t c a H e u e
cap teh rsm fon teh red teh
itedm tcaonp iiitohnb irttaeon rso rscaeed -oCAG rscaeed illtsycae
ic im p cn it ya HM ya sa
,regdp itrceedd rtuoonpm rsceuom iitsaenhb itscandm ifrttaoon itscandm ifrttaoon
i P P A A
a 4 id is ) if ib
tc t o t p r h
s r g o in
u a yh en P
d le ( d p
primary public health system. Their multivariate
regression analysis results showed that being female and having
three or more diagnosed diseases were positively
associated with exposure to DDIs (OR = 2.49; 95% CI 2.29, 2.75
and OR = 6.43; 95% CI 3.25, 12.44, respectively) [
Teixeira et al. performed a multivariate regression that
showed that patients with three to five medications have
high odds of DDIs (OR = 4.84; 95% CI 2.85, 7.91), while
patients with six or more medications have higher odds
of DDIs (OR = 25.11; 95% CI 9.98, 48.63) [
The prevalence of DDIs among geriatric patients
attending the ambulatory care pharmacy was high. This
indicates the need for special care while handling their
prescriptions. Even though the majority of the most
common interacting pairs (top 20) were category C, we
still need to look for D and X interacting pairs carefully.
Finally, the identified DDIs’ main predictors of number of
prescribed medications (more than four) and number of
chronic conditions (more than three) could guide us to
provide better care and to develop a safer practice.
DDIs analysis was based only on one database
(LexiInteract). Our study was conducted in a specific type of
setting and population, which is why our findings are
not generalizable to other settings or different
populations, especially populations such as cancer or transplant
patients. The study may underestimate the prevalence of
DDIs in these populations. Additionally, non-prescribed
medications such as over-the-counter and herbal
products were excluded, which may lead to an
underestimate of the prevalence of DDIs in our sample. Despite
these limitations, the results of this study represent a
step forward in exploring the occurrence and nature of
DDIs in geriatric patients attending the ambulatory care
We thank Dr. Anwar Ahmed and Dr. Fayssal Farahat for their useful criticism
and their constructive recommendations.
The authors declare that they have no competing interests.
Availability of data and materials
The datasets generated and analyzed during the current study are not publicly
available to protect participant information confidentiality but are available
from the corresponding author on reasonable request.
Consent to publish
Ethical approval and consent to participate
This study was conducted with ethical approval from the Ethical Committee
of King Abdullah International Medical Research Center (KAIMRC). The need
for consent was waived by the IRB, the decision number is (RJ15/003/J), the
research didn’t involve any procedure or represents any harm to participants,
and to avoid any potential breach of confidentiality necessary participant
identifiers were removed from data files, Identifiers were stored in a physically
separated and secured location of the data files.
The authors have no funding to report.
Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
1. Astrand B , Astrand E , Antonov K , Petersson G . Detection of potential drug interactions-a model for a national pharmacy register . Eur J Clin Pharmacol . 2006 ; 62 : 749 - 56 .
2. Peyriere H , Cassan S , Floutard E , et al. Adverse drug events associated with hospital admission . Ann Pharmocother . 2003 ; 37 : 5 - 11 .
3. Ferner RE , Aronson JK . Communicating information about drug safety . BMJ . 2006 ; 333 : 143 .
4. Secoli SR , Figueras A , Lebrao ML , de Lima FD , Santos JL . Risk of potential drug-drug interactions among Brazilian elderly: a population-based, cross-sectional study . Drugs Aging . 2010 ; 27 : 759 - 70 .
5. Bleich GW , Bleich A , Chiamulera P , et al. Frequency of potential interactions between drugs in medical prescriptions in a city in southern Brazil . Sao Paulo Med J. 2009 ; 127 : 206 - 10 .
6. Cruciol-Souza JM , Thomson JC . A pharmacoepidemiologic study of drug interactions in a Brazilian teaching hospital . Clinics (Sao Paulo) . 2006 ; 61 : 515 - 20 .
7. Silva NMO , Carvalho RP , Bernardes ACA , et al. Potential drug interaction evaluation in medical prescriptions in a public hospital specialized in women care , in Campinas, SP. Rev Ciênc Farm Básica Apl . 2010 ; 31 : 171 - 6 .
8. Codagnone Neto V , Garcia VP , Santa Helena ET . Possible pharmacological interactions in hypertensive and/or diabetic elderly in family health units at Blumenau (SC) . Braz J Pharm Sci . 2010 ; 46 : 795 - 804 .
9. Janchawee B , Wongpoowarak W , Owatranporn T , et al. Pharmacoepidemiologic study of potential drug interactions in outpatients of a university hospital in Thailand . J Clin Pharm Ther . 2005 ; 30 : 13 - 20 .
10. Dinesh KU , Subish P , Pranaya M , et al. Pattern of potential drug-drug interactions in diabetic out-patients in a tertiary care teaching hospital in Nepal . Med J Malaysia. 2007 ; 62 : 294 - 8 .
11. Lexicomp Online , Lexi-Interact™ Online, Hudson, Ohio: Lexi-Comp, Inc.; 2015 . http://www.uptodate.com/crlsql/interact/frameset.jsp. Accessed Feb 2018
12. IBM Corp . Released 2013 . IBM SPSS Statistics for Macintosh, Version 22 .0. Armonk: IBM Corp.
13. Bjorkman IK , Fastbom J , Schmidt IK , Bernsten CB , The Pharmaceutical Care Of The Elderly In Europe Research (PEER) Group . Drug-drug interactions in the elderly . Ann Pharmacother . 2002 ; 36 : 1675 - 81 .
14. Venturini C , Engroff P , Ely L , et al. Gender differences, polypharmacy, and potential pharmacological interactions in the elderly . Clinics . 2011 ; 66 ( 11 ): 1867 - 72 .
15. Neto P , Nobili A , Marusic S , et al. Prevalence and predictors of potential drug-drug interactions in the elderly: a cross-sectional study in the Brazilian primary public health system . J Pharm Pharm Sci . 2012 ; 15 ( 2 ): 344 - 54 .
16. Teixeira JJV , Crozatti MTL , dos Santos CA, Romano-Lieber NS . Potential drug-drug interactions in prescriptions to patients over 45 years of age in primary care, Southern Brazil . PLoS ONE . 2012 ; 7 ( 10 ): 47062 .