Q J Med 2010; 103:327–335

doi:10.1093/qjmed/hcq019 Advance Access Publication 7 March 2010

Inappropriate utilization of intravenous proton pump

inhibitors in hospital practice—a prospective study
of the extent of the problem and predictive factors
D.G.N. CRAIG1, R. THIMAPPA1, V. ANAND2 and S. SEBASTIAN1
From the 1Department of Gastroenterology and 2Department of Acute Medicine, Hull Royal
Infirmary, Hull, HU3 2JZ, UK

Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024

Address correspondence to Dr S. Sebastian, Department of Gastroenterology, Hull Royal Infirmary, Anlaby
Road, Hull, HU3 2JZ, UK. email: darrencraig@doctors.org.uk

Received 5 July 2009 and in revised form 31 January 2010

Summary
Background: Intravenous (IV) proton pump inhibi-
tors (PPI) reduce rebleeding from high-risk peptic
ulcers following endoscopic therapy. The majority
of IV PPI prescriptions in US hospital practice are
inappropriate, leading to unnecessary drug costs,
drug shortages and potential adverse events. To
date, little is known about UK hospital IV PPI pre-
scribing practice.
Aims: To examine IV PPI use in a large university
teaching hospital to determine factors predicting
inappropriate prescribing practices.
Methods: Prospective study of 276 recently hospita-
lized patients initiated on IV PPI over a 6–month
period. IV PPI use was deemed appropriate for the
following indications: endoscopic evidence of
recent upper gastrointestinal (UGI) haemorrhage,
patient nil by mouth with a valid indication for
oral PPI therapy and stress ulcer prophylaxis in a
critical care setting.
Results: The majority (208/276, 75.4%) of IV PPI
prescriptions were deemed inappropriate in terms
of either indication for use, dose or duration of ther-
apy. The majority (168/276, 60.9%) of prescriptions
were initiated on non-medical wards. Inappropriate
prescribing was more common amongst female
patients, surgical admissions, non-UGI haemorrhage
cases and when initiated by junior hospital doctors.
Surgical admission [odds ratio (OR) 2.88, 95% con-
fidence interval (CI) 1.12–7.42] and female gender
[OR 3.92 (95% CI 1.84–8.34)] were independently
predictive of inappropriate use.
Conclusions: This study suggests that the majority of
IV PPI prescriptions in hospital are inappropriate,
particularly when initiated for non-UGI bleeding
indications. Improving prescribing awareness
through education of junior medical staff on
non-medical wards could reduce inappropriate IV
PPI use.

Introduction £425-million per annum.1 In the inpatient setting,

Proton pump inhibitors (PPI) have emerged as the
leading treatment for gastro-oesophageal reflux dis-
ease and peptic ulcer disease due to their efficacy
and low toxicity in treating these conditions. The
NHS expenditure on these treatments exceeds
adjuvant use of high-dose intravenous (IV) PPI has
been demonstrated to reduce rebleeding from
high-risk peptic ulcers following endoscopic hae-
mostasis and has been endorsed for this indication
by both British and North American consensus

! The Author 2010. Published by Oxford University Press on behalf of the Association of Physicians.
All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
328 D.G.N. Craig et al.
groups.2–4 High-dose pre-endoscopic IV PPI use in
patients with upper gastrointestinal bleeding (UGIB)
appears to significantly reduce both the need for
endoscopic haemostasis at the index endoscopy
and the length of hospital stay.5 A retrospective
study of tertiary care patients suggested that
pre-endoscopic PPI treatment of non-variceal
UGIB may reduce adverse outcomes such as
rebleeding, need for surgery and mortality, although
this has yet to be demonstrated in a prospective
trial.6,7 However, empirical parenteral treatment of
this group may be cost-effective and this, combined
with biological plausibility, has led one consensus
group to recommend its use.4,8,9 Patients with gas-
tric hyper secretion syndromes and patients who are
nil-by-mouth (NBM) with valid indications for oral
PPI use may also benefit from IV PPI.10,11 There is
some evidence to support oral PPI use for stress
ulcer prophylaxis within a critical care setting
although few trials have focused specifically on IV
PPI use for this indication.12,13
Several studies have demonstrated that inappro-
priate prescribing of IV PPI is widespread, particu-
larly amongst patients with non-UGIB
indications.14–16 One retrospective study identified
predictive factors for inappropriate prescribing in a
Canadian centre, but little is known regarding IV PPI
prescribing patterns within the UK.14
We aimed to prospectively study the incidence
and appropriateness of IV PPI prescriptions in a
UK teaching hospital setting and to identify possible
predictive factors for inappropriate prescribing
practices.
Patients and methods
This study was undertaken at two sites (Hull Royal
Infirmary and Castle Hill Hospital) in the Hull and
East Yorkshire NHS Trust. This is a large university
hospital trust serving a catchment area of 585 000
people and treating approximately 180 000 patients
per annum. Approximately 400 patients per annum
are admitted with acute UGIB, with all UGIB cases
admitted initially under the auspices of general med-
icine prior to review by a consultant-led gastroen-
terology service providing out-of-hours emergency
endoscopy where required.
Two hundred and seventy-six consecutive
patients receiving IV pantoprazole (the only IV PPI
available in our institution) identified during a
6-month period were included. Demographic, clin-
ical and endoscopic data were collected, and
patients were classified into one of two groups:
UGIB and non-UGIB on the basis of the proposed
rationale for IV PPI use by the prescribing physician.
Both the drug charts and clinical notes were exam-
ined to determine the rationale for prescription and
whether the prescribing doctor had initiated the pre-
scription of their own volition or on the behest of
another colleague. IV PPI use in UGIB patients was
only considered appropriate in the presence of stig-
mata of recent haemorrhage (SRH) at endoscopy,
defined as adherent clot or a visible or spurting
blood vessel. In these patients, appropriate IV PPI
dosing involved an 80 mg bolus of pantoprazole
post-endoscopy and a subsequent pantoprazole
infusion at 8 mg/h for 72 h. Other potentially subop-
timal dosing regimens in the presence of SRH, such
Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024
as twice daily bolus IV pantoprazole, were consid-
ered inappropriate. Use of IV PPI was also deemed
inappropriate if used: pre-endoscopy; in patients
without SRH; in patients with isolated variceal
bleeding; in patients deemed too unwell to undergo
endoscopy due to other co-morbid factors or where
inpatient endoscopy was deemed unnecessary. In
keeping with the divided opinion over
pre-endoscopic IV PPI use in actively bleeding
patients, post hoc subgroup analysis was performed
where pre-endoscopic IV PPI use in patients with
evidence of haemodynamic instability and wit-
nessed haematemesis, melaena or haematochezia
was considered appropriate.4,7
In the non-UGIB group, IV PPI use was consid-
ered appropriate for stress ulcer prophylaxis in crit-
ical care patients provided that the patient was NBM
or unable to tolerate any oral medications. The
appropriate dose in this scenario was pantoprazole
40 mg once daily. Continuation of IV PPI treatment
following discharge from the critical care setting, or
when other oral medications were tolerated, was
deemed inappropriate. Likewise, pantoprazole
40 mg IV once daily was considered appropriate in
patients who were NBM with a valid indication for
oral PPI use. Use of IV PPI in patients with abdomi-
nal pain or vomiting was considered inappropriate
unless the patient had another valid reason for oral
PPI use and could not tolerate oral medications.
Following inclusion, patients were followed until
hospital discharge or death and the following data
were collected: oral PPI use at admission, haemo-
dynamic status, transfusion requirements, time to
initial endoscopy, requirement for repeat endo-
scopy, operative record, duration of IV PPI use and
discharge oral PPI use. Predictive factors for inap-
propriate IV PPI use examined included: patient age,
gender, admitting speciality (gastroenterology, gen-
eral medicine or surgery), prescribing doctor status
(consultant, specialist registrar or junior doctor),
timing of initial prescription [working hours (0800–
1700), working hours at weekends or nights], prior
oral PPI use, proposed rationale for use and, in the
 Factors predicting inappropriate utilization of intravenous proton pump inhibitors 329

UGIB group, the presence of active bleeding with medical or gastroenterology prescriptions. The
pre-endoscopy and transfusion requirements. majority (183/276, 66.3%) of prescriptions were
made by junior clinical staff and 124/183 (67.8%)
Statistical analysis of these were inappropriate. Specialist registrars
were significantly more accurate in their prescribing
All statistical tests were performed using SPSS ver-
habits for IV PPIs compared with junior doctors. The
sion 16.0 (Chicago, IL, USA). Data values are pre-
majority (146/276, 52.9%) of IV PPIs were initiated
sented as median and interquartile range (IQR) or
outside normal daylight working hours, but there
percentages unless otherwise stated. Continuous
were no significant differences in the accuracy of
unpaired data were compared using either analysis
these weekend or night-shift prescriptions when
of variance or the Kruskal–Wallis test if inter-group
compared with weekday prescriptions.
variances were unequal. Categorical data were ana-
The proposed rationale for IV PPI use, where
lysed using chi-squared tests or Fishers exact test.
stated in the clinical notes or on the drug prescrip-
Adjustments for multiple comparisons were made

Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024

using Bonferroni’s correction. Univariate and multi-
variate analysis was performed to examine the asso-
ciation between individual predictive factors and
the appropriateness of IV PPI use. A P < 0.05 was
considered significant using two–tailed tests.
Results
A total of 276 patients [153 (55.4%) male] received
IV pantoprazole within the time period of the study.
Baseline demographic and clinical details are
shown in Table 1. The majority (168/276, 60.9%)
of patients had their IV PPI initiated on a surgical
ward, while only 19/276 (6.9%) of prescriptions
were initiated by a gastroenterology specialist
team. Surgical prescriptions were significantly
more likely to be inappropriate when compared
tion chart, is shown in Table 2. A total of 90 (32.6%)
prescriptions were initiated for UGIB and 186
(67.4%) prescriptions were for non-UGIB indica-
tions. In total, only 68 patients (24.6%) received IV
PPI for a valid indication with an appropriate dosing
and duration of therapy. Within the non-UGIB
group, abdominal pain (n = 77, 41.4% of
non-UGIB cases), NBM with previous oral PPI use
(n = 49, 26.3%) and stress ulcer prophylaxis (n = 21,
11.3%) were the most frequent indications. Of the
186 prescriptions initiated for non-UGIB, only 53
(28.5%) were appropriate (Table 2). Patients treated
with IV PPI for abdominal pain were significantly
more likely to have their prescription initiated
within 24 h of admission compared with other indi-
cations for treatment. The median duration of inap-
propriate treatment within this group was 3 (IQR 2–
5) days, with a total of 375 doses of IV pantoprazole

Table 1 Baseline demographic and clinical details of patients initiated on IV PPI

Patient characteristic N (%)/median (IQR) Number of IV PPI

prescriptions (% appropriate)

Age (years) 68 (50–78) —

Length of stay (days) 11 (5–20)
Male 153 (55.4) 76 (49.7)
Female 123 (44.6) 38 (30.9)*
Speciality where IV PPI first prescribed
Gastroenterology 19 (6.9) 14 (73.7)
Medicine 89 (32.2) 57 (64.0)
Surgical 168 (60.9) 44 (26.2)y
Grade of prescriber
Consultant 25 (9.1) 12 (48.0)
Specialist registrar 68 (24.6) 41 (60.3)
Junior doctor 183 (66.3) 59 (32.2)z
Timing of prescription
Working hours (0800–1700) 130 (47.2) 55 (42.3)
Working hours (weekends) 109 (39.5) 43 (39.4)
Nights 1700–0800 hours 37 (13.3) 14 (37.8)

*P = 0.002 vs. male patient prescriptions.

y
P < 0.001 vs. other specialities.
z
P < 0.001 vs. specialist registrars.
330 D.G.N. Craig et al.

Table 2 Proposed indications, duration, and appropriateness of IV PPI use

Proposed N N (%) N (%) N (%) Median (IQR)

indication (% of total Prescriptions Prescriptions Prescriptions duration
for IV PPI prescriptions) appropriate initiated within initiated by of therapy
24 h of admission surgical team (days)

Presumed UGI bleeding 90 (32.6) 15 (16.7)/53 (58.9)a 63 (70.0) 17 (18.9)z 3 (2–4)

Stress ulcer prophylaxis 21 (7.6) 18 (85.7) 16 (76.2) 10 (47.6) 3 (2–5)
Abdominal pain 77 (27.9) 12 (15.6) 63 (81.8)y 74 (96.1)
Vomiting 14 (5.1) 2 (14.3) 7 (50.0) 9 (64.3)
Combination of above 6 (2.2) 3 (50.0) 6 (100.0) 5 (83.3)
Previously on oral PPI 49 (17.8) 18 (36.7) 26 (53.1) 40 (81.6)
now NBM

Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024

Unknown reason/none 19 (6.9) 0 (0.0) 10 (52.6) 12 (63.2)
Total 276 68 (24.6) [106 (38.4)a] 191 (69.2) 167 (60.5)

a
When pre-endoscopy use deemed appropriate in presence of active bleeding.
y
P < 0.05 compared with other prescribing indications.
z
P < 0.05 compared with non-UGIB group.

delivered within this group for inappropriate indica-
tions. One hundred and twenty-five (94.0%) inap-
propriate prescriptions for non-UGIB patients were
given pantoprazole 40 mg once daily with the
remaining eight patients receiving 40 mg twice
daily. No patients within the non-UGIB group
received continuous IV PPI infusions. Twenty-one
(7.6%) of all prescriptions were initiated within a
critical care setting expressly for stress ulcer prophy-
laxis, but 3 out of these 21 prescriptions (14.3%)
were deemed inappropriate due to prolonged use
in patients taking other oral medications (Table 3).
Only 35/165 (21.2%) prescriptions in the non-UGIB
group were appropriate when stress ulcer prophy-
laxis was excluded from analysis. The most
common reason for inappropriate prescribing
within the non-UGIB group was an inappropriate
indication [126/186 (67.7%) non-UGIB prescrip-
tions], with 42/186 (22.6%) of non-UGIB prescrip-
tions being continued for an inappropriately
prolonged period of time.
Upper GI bleeding group
A total of 90 cases were treated for presumed UGI
bleeding, with the majority (63/90, 70.0%) of UGIB
prescriptions initiated within 24 h of hospital admis-
sion. UGIB IV PPI prescriptions were significantly
more likely to be initiated by medical than surgical
teams (Table 2). A total of 22 patients were not
endoscoped, as 14 patients were later deemed to
have minor UGIB not requiring inpatient endo-
scopy, and eight were deemed too unwell due to
co-morbid illness to safely undergo endoscopy.
This latter group consisted mainly of patients who
were terminally ill and in whom endoscopy was
deemed inappropriate. Of the 68/90 (75.6%) who
underwent an endoscopy, 53 (77.9%) received IV
PPI prior to endoscopy, of whom 38 (71.7%) had
evidence of haemodynamic instability and wit-
nessed haematemesis, melaena or haematochezia.
In total, only 16.7% (15/90) IV PPI prescriptions
were considered appropriate in the UGIB group,
but this rose to 58.9% (53/90) when considering
pre-endoscopic use appropriate in the presence of
active haemorrhage. At endoscopy, 26 patients
(38.2%) had evidence of peptic ulcer bleeding
with SRH (corresponding to Forrest class 1a-IIc
ulcers). Eighteen patients (26.5%) had Forrest class
III peptic ulcers, nine patients (13.2%) oesophagitis,
eight (11.8%) variceal bleeding, six (8.8%) normal
endoscopies and one (1.5%) patient had a
post-ERCP sphincterotomy bleed. Use of IV PPI
prior to endoscopy in the 44 patients with peptic
ulcer disease was associated with a significant
reduction in the presence of SRH [15/32 (46.9%)
received PPI, 11/12 (91.7%) no PPI, P = 0.007].
Twenty (76.9%) of 26 patients with SRH received
post-endoscopy pantoprazole infusions appropri-
ately, but six (23.1%) patients with SRH requiring
haemostatic endoscopic therapy received poten-
tially suboptimal dosing with twice daily pantopra-
zole 40 mg injections. Eleven patients with
non-peptic ulcer bleeding or low-risk lesions had
IV PPI continued post-endoscopy, including four
patients who received continuous IV PPI infusions
inappropriately (Table 3). IV PPI was continued for a
median duration of 3 (IQR 1.75–4) days
post-endoscopy in those patients with SRH, but
11 patients (16.2%) were considered to have had
inappropriately prolonged use of IV PPI
post-endoscopy (range 4–11 days). Five patients
 Factors predicting inappropriate utilization of intravenous proton pump inhibitors 331

Table 3 Reasons for inappropriate IV PPI prescribing according to proposed indication for use

Proposed N N (%) N (%) N (%) N (%)

indication (% of total Prescriptions Inappropriate Inappropriate Inappropriate
for IV PPI prescriptions) appropriate indication dosing regimen duration of therapy

Presumed UGI bleeding 90 (32.6) 15 (16.7)/53 (58.9)* 75 (83.3) 26 (28.9) 11 (12.2)

IV PPI used
pre-endoscopy
22 (24.4) not
endoscoped
Stress ulcer prophylaxis 21 (7.6) 18 (85.7) 0 (0.0) 1 (4.8) 3 (14.3)
Abdominal pain 77 (27.9) 12 (15.6) 64 (83.1) 4 (5.2) 15 (19.5)
Vomiting 14 (5.1) 2 (14.3) 11 (78.6) 1 (7.1) 7 (50.0)

Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024

Combination of above 6 (2.2) 3 (50.0) 2 (33.3) 0 (0.0) 1 (16.7)
Previously on oral 49 (17.8) 18 (36.7) 30 (61.2) 2 (4.1) 16 (3.3)
PPI now NBM
Unknown reason/none 19 (6.9) 0 (0.0) 19 (100.0) 0 (0.0) 0 (0.0)
Total 276 68 (24.6) [106 (38.4)a] 164 (59.4) 34 (12.3) 53 (19.2)

a
IV PPI use in UGIB patients was only deemed appropriate in the presence of SRH at endoscopy, and in subgroup analysis, in
patients with haemodynamic instability and active UGIB pre-endoscopy.
In the non-UGIB group, IV PPI use was considered appropriate for stress ulcer prophylaxis in critical care patients unable to
take oral medications or in patients who were NBM with a valid indication for oral PPI use.

(7.4%) underwent repeat diagnostic endoscopy,

seven patients (10.3%) required repeat therapeutic Table 4 Univariate analysis of predictive factors for
endoscopic intervention, while two patients (2.9%) inappropriate IV PPI use
required surgery for protracted UGIB. Patients
receiving IV PPI prior to endoscopy showed a Variable Univariate Multivariate
trend towards requiring fewer repeat diagnostic or analysis analysis*
P-value Odds ratio
therapeutic procedures compared with patients
(95% CI)
where IV PPI was initiated post-endoscopy [8/52
(15.4%) received PPI, 6/16 (37.5%) no PPI, Age (per decade) 0.81 –
P = 0.056]. Female gender 0.001 3.92 (1.84–8.34)a
Surgical admission <0.001 2.88 (1.12–7.42)a
Junior prescribing doctor <0.001 2.06 (0.84–5.0)
Predictors of inappropriate prescribing Night prescription 0.96 –
No prior oral PPI use 0.057 1.99 (0.91–4.34)
Univariate analysis was performed to examine any Reason for use (non-UGIB) <0.001 NS
association between potential factors for inappropri-
ate IV PPI use. IV PPI initiated for reasons other than *Factors with P > 0.1 on univariate analysis excluded.
UGIB was strongly associated with inappropriate a
Both female gender and surgical admission indepen-
use (P < 0.001), Table 4. Both admissions to a sur- dently predict inappropriate IV PPI prescriptions.
gical ward and prescriptions initiated by junior doc-
tors were also strongly associated with inappropriate
IV PPI use (P < 0.001). Inappropriate prescribing
appeared to be more common amongst female Subgroup analysis was performed where IV PPI
patients with 85/123 (69.1%) of all female prescrip- pre-endoscopy in UGIB patients with haemody-
tions deemed inappropriate compared with 77/153 namic instability was deemed appropriate.
(50.3%) of all male prescriptions. There was a trend Predictive factors for inappropriate IV PPI use after
towards more frequent inappropriate prescribing in univariate and multivariate analysis are shown in
patients not previously on oral PPI (P = 0.057). After Table 5. Female gender and admission to a surgical
multivariate logistic regression, admission to a gen- ward were again independently predictive of inap-
eral surgical ward and female gender remained sig- propriate use in this scenario, with the requirement
nificant as predictors for inappropriate IV PPI use for blood transfusion strongly predicting appropriate
(Table 4). use in UGIB patients. Patients not on oral PPI prior
332 D.G.N. Craig et al.

Table 5 Univariate and multivariate logistic regression There are several limitations to the results of our
of predictive factors for inappropriate IV PPI use study. IV PPI prescriptions are not regulated within
(pre-endoscopic IV PPI use deemed appropriate in pres- our institution and do not require specialist gastro-
ence of active bleeding) enterology or pharmacy input prior to dispensing. As
a result, it was difficult to ensure that all prescrip-
Variable Univariate Multivariate tions were identified and we may have underesti-
analysis analysis* Odds
mated the use of IV PPI. Some patients may have
P-value ratio (95% CI)
received IV PPI during short inpatient stays and
Age (per decade) 0.41 – failed to be identified during ward surveys, but it is
Female gender 0.005 3.69 (1.77–7.68)a likely that such patients would have received the
Surgical admission <0.001 2.71 (1.33–5.52)a drug for inappropriate indications or for inappropri-
Junior prescribing doctor <0.001 1.45 (0.83–2.53) ately short durations of therapy. The study was also
Night prescription 0.79 – promoted within acute treatment areas of the hospi-
(1.00–4.15)a

Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024

No prior oral PPI use 0.057 2.04
UGIB cases only
Requirement for 0.001 0.06 (0.01–0.43)a
blood transfusion
*Factors with P > 0.1 on univariate analysis excluded.
a
Female gender, surgical admission and no previous oral
PPI use independently predict inappropriate IV PPI
prescriptions.
to admission were also more likely to be prescribed
IV PPI inappropriately.
Discussion
This prospective observational study confirms previ-
ous studies which suggested that IV PPI use is fre-
quently inappropriate within hospital practice.14–16
In total, only 24.6% of all IV PPI prescriptions writ-
ten within our institution had correct indications,
dosing and duration of therapy, compared with 31
and 6%, respectively, in smaller Canadian and Irish
cohorts.14,15 The majority of prescriptions were
initiated on surgical wards, and these were signifi-
cantly more likely to be for inappropriate reasons
than those initiated on general medical or gastroen-
terology wards. Most IV PPI prescriptions were
ordered by junior hospital staff, and again this
group were significantly less likely to prescribe the
drug for appropriate reasons when compared with
more experienced clinicians, particularly specialist
registrars. In keeping with other studies, the majority
(67.4%) of IV PPI prescriptions were initiated for
non-UGIB indications, in particular abdominal
pain and NBM patients previously prescribed oral
PPIs.14–18 Frequently, the original indications for
oral PPI use were inappropriate, or IV PPI treatment
was continued despite the patient eating or taking
other oral medications. This large non-UGIB cohort
deserves close consideration since targeting inter-
ventions at this group could produce significant
cost savings.
tal to aid case identification, which may have influ-
enced prescribing patterns, but again, it would be
expected that this would reduce, rather than
increase, inappropriate prescribing.
High-dose IV PPI infusions, initiated post-
endoscopy, reduce recurrent gastrointestinal bleed-
ing, need for surgery or repeat endoscopic haemo-
static treatment in patients with endoscopically
treated bleeding lesions.2 In contrast to Afif
et al.,14 we deemed pre-endoscopic IV PPI use
amongst UGIB patients to be inappropriate.
Empirical IV PPI treatment of UGIB has been sug-
gested to reduce adverse outcomes in a retrospec-
tive study but has not been shown in a prospective
randomized trial to reduce clinically important out-
comes such as recurrent bleeding rates, need for
surgery or mortality.6,7 In this study, 35.3% of
UGIB patients had non-peptic ulcer related pathol-
ogy at endoscopy in keeping with the lower rates of
peptic ulcer bleeding amongst UK UGIB cohorts
compared with Asian populations.19,20 This, com-
bined with the higher prevalence of poor metabo-
lism of PPIs in Asian compared with Caucasian
populations, makes empirical IV PPI treatment of
all UGIB cases within the UK difficult to justify at
present.21 High-dose empirical IV PPI treatment of
low-risk UGIB patients is unlikely to be the most
cost-effective approach and IV PPI treatment alone
in high-risk patients is probably inferior to definitive
haemostatic treatment, making effective resuscita-
tion followed by timely endoscopy in these patients
the most important management strategy.22–24
On the other hand, targeted IV PPI initiated prior
to endoscopy downgrades endoscopic stigmata,
reducing the need for endoscopic haemostatic ther-
apy at the index endoscopy, and may reduce lengths
of hospital stay.5,7 Within the UGIB cohort, the
majority (77.9%) of patients undergoing UGI endos-
copy received IV PPI prior to endoscopy and this
was associated with a significant reduction in the
incidence of SRH at endoscopy. This is similar to
the findings of a Canadian cohort where 86% of
patients undergoing endoscopy for suspected
 Factors predicting inappropriate utilization of intravenous proton pump inhibitors
UGIB received empirical IV PPI.14 Although our
study was not designed to examine post-endoscopy
reintervention rates, there was a trend towards fewer
repeat endoscopies in the empirically treated UGIB
cohort. The use of post-endoscopy bolus doses fol-
lowed by continuous infusions of PPI causes intra-
gastric pH to rise more effectively than single doses
of IV PPI,25 which may explain the reduction in
rebleeding following endoscopic haemostasis, due
to the effects of raised pH upon platelet aggregation
and coagulation.26 Within our cohort of UGIB
cases, only 76.9% of cases with SRH at endoscopy
were subsequently fully treated with bolus PPI and
subsequent PPI infusion, suggesting suboptimal
treatment of patients with endoscopically confirmed
active haemorrhage, who might benefit most from
IV PPI use.
Predictors of inappropriate prescribing
Previous studies have attempted to identify predic-
tive factors for inappropriate prescribing of IV PPI in
an attempt to improve future practice and guide
educational interventions. Afif and colleagues14 per-
formed a multivariate analysis of factors predicting
appropriate IV PPI use and identified increasing
patient age and lower mean daily PPI dose as pre-
dictive factors, with a trend towards inappropriate
prescriptions being written more frequently during
evening shifts. There was no association between
the prescribing doctor, patient gender or the pre-
senting diagnosis and the appropriateness of PPI
therapy. In contrast, we found no relationship
between patient age or timing of prescriptions and
appropriateness of IV PPI therapy, but did identify
female gender and admission to a surgical ward as
independent predictive factors for inappropriate IV
PPI use.
It is unclear why female gender should have been
so strongly predictive of inappropriate PPI prescrib-
ing since there were no significant gender differ-
ences with regard to the indications for IV PPI
prescribing. Female gender has been demonstrated
to be independently associated with inappropriate
community prescribing amongst elderly UK primary
care patients but, to our knowledge, has not pre-
viously been demonstrated to predict inappropriate
IV PPI prescribing.27 The majority of inappropriate
prescriptions were initiated by non-medical teams or
by junior staff and surgical admission independently
predicted inappropriate IV PPI prescribing. This is in
marked contrast to Afif et al.,14 where the disciplines
of the prescribing doctors were not associated with
inappropriate prescribing. Relatively junior surgical
doctors triage the majority of new admissions with
abdominal pain in our institution and we speculate
333
that they may be using IV PPI as an aid to excluding
dyspepsia from their differential diagnosis, since IV
PPI prescriptions were significantly more likely to be
initiated for abdominal pain during the first 24 h of
admission compared with other non-UGIB indica-
tions. IV pantoprazole is well tolerated with low tox-
icity and, since it is readily available in our
institution, may be seen by some prescribers as an
effective treatment that can be rapidly initiated
during the initial evaluation of acute abdominal
pain while awaiting definitive diagnostic tests.28
This may be particularly true amongst non-specialist
or junior prescribers given that the practical PPI pre-
Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024
scribing patterns of general practitioners often con-
flict with their theoretical knowledge of indications
for PPI therapy.29 Lack of theoretical knowledge is
associated with prescribing errors and this may have
contributed towards the poorer prescribing practices
of junior surgical staff since in our institution junior
medical (but not surgical) doctors receive specific
guidance on IV PPI prescribing during job
induction.30
Cost-effectiveness models have suggested that
pre-endoscopy IV PPI use amongst all patients
with UGIB is unlikely to be the most cost-effective
approach.22 This study supports this view, since
many patients treated empirically for UGIB within
our hospital practice were not haemodynamically
unstable. However, the debate regarding relative
cost benefits from use of IV PPI in only selected
UGIB patients is overshadowed by the widespread
inappropriate use of IV PPI amongst non-UGIB
patients. Targeting inappropriate prescribing pat-
terns and educating doctors is an attractive option
to produce cost-savings, although previous attempts
to do so have suggested that this may prove
cost-neutral due to reduction of under-prescribing
habits.31 Such interventions are likely to require a
multi-modality approach involving pharmacists,
gastroenterologists and nurse specialists, and
should be targeted initially towards IV PPI prescrip-
tions initiated in non-medical hospital areas.
Restricting the availability of IV PPI to specialist hos-
pital areas could reduce demand but these drugs
would still need to be available on a 24-h basis,
particularly if pre-endoscopic treatment of active
UGIB cases becomes widely adopted.
Conclusion
This large prospective observational study further
confirms that the majority of IV PPI prescriptions
within hospital practice are inappropriate.
Significant factors influencing prescribing patterns
appear to include female patient gender and
334 D.G.N. Craig et al.
initiation of prescriptions by non-medical teams.
Targeting interventions towards particular groups
of prescribers could produce the greatest
cost-savings, but changing entrenched practice is
likely to require a multifaceted approach.
Conflict of interest: None declared.
References
1. Forgacs I, Loganayagam A. Overprescribing proton pump
inhibitors. Br Med J 2008; 336:2–3.
2. Lau JY, Sung JJ, Lee KK, Yung MY, Wong SK, Wu JC, et al.
Effect of intravenous omeprazole on recurrent bleeding after
endoscopic treatment of bleeding peptic ulcers. N Engl J Med
2000; 343:310–6.
3. British Society of Gastroenterology Endoscopy, Committee
Non-variceal upper gastrointestinal haemorrhage: guidelines.
Gut 2002; 51:1–6.
4. Barkun A, Bardou M, Marshall JK. Nonvariceal Upper GI
Bleeding Consensus Conference,Group. Consensus
recommendations for managing patients with nonvariceal
upper gastrointestinal bleeding. Ann Intern Med 2003;
139:843–57.
5. Lau JY, Leung WK, Wu JC, Chan FK, Wong VW, Chiu PW,
et al. Omeprazole before endoscopy in patients with gastro-
intestinal bleeding. N Engl J Med 2007; 356:1631–40.
6. Keyvani L, Murthy S, Leeson S, Targownik LE. Pre-endo-
scopic proton pump inhibitor therapy reduces recurrent
adverse gastrointestinal outcomes in patients with acute
non-variceal upper gastrointestinal bleeding. Aliment
Pharmacol Ther 2006; 24:1247–55.
7. Dorward S, Sreedharan A, Leontiadis GI, Howden CW,
Moayyedi P, Forman D. Proton pump inhibitor treatment
initiated prior to endoscopic diagnosis in upper gastrointest-
inal bleeding. Cochrane Database Syst Rev 2006; 4:005415.
8. Gagnon YM, Levy AR, Eloubeidi MA, Arguedas MR,
Rioux KP, Enns RA. Cost implications of administering intra-
venous proton pump inhibitors to all patients presenting to
the emergency department with peptic ulcer bleeding. Value
Health 2003; 6:457–65.
9. Enns RA, Gagnon YM, Rioux KP, Levy AR. Cost-
effectiveness in Canada of intravenous proton pump
inhibitors for all patients presenting with acute upper
gastrointestinal bleeding. Aliment Pharmacol Ther 2003;
17:225–33.
10. Metz DC, Forsmark C, Lew EA, Starr JA, Soffer EF,
Bochenek W, et al. Replacement of oral proton pump inhi-
bitors with intravenous pantoprazole to effectively control
gastric acid hypersecretion in patients with Zollinger-Ellison
syndrome. Am J Gastroenterol 2001; 96:3274–80.
11. Metz DC. Potential uses of intravenous proton pump
inhibitors to control gastric acid secretion. Digestion 2000;
62:73–81.
12. Levy MJ, Seelig CB, Robinson NJ, Ranney JE. Comparison of
omeprazole and ranitidine for stress ulcer prophylaxis.
Dig Diseases & Sci 1997; 42:1255–9.
13. Azevedo JR, Soares MG, Silva G, Palacio G. Prevention of
stress ulcer bleeding in high risk patients: comparison of
three drugs. Crit Care Med 1999; 27:A41.
14. Afif W, Alsulaiman R, Martel M, Barkun AN. Predictors of
inappropriate utilization of intravenous proton pump inhibi-
tors. Aliment Pharmacol Ther 2007; 25:609–15.
15. Sebastian SS, Kernan N, Qasim A, O’Morain CA, Buckley M.
Appropriateness of gastric antisecretory therapy in hospital
practice. Ir J Med Sci 2003; 172:115–7.
16. Slattery E, Theyventhiran R, Cullen G, Kennedy F, Ridge C,
Nolan K, et al. Intravenous proton pump inhibitor use in
hospital practice. Eur J Gastroenterol Hepatol 2007;
19:461–4.
17. Guda NM, Noonan M, Kreiner MJ, Partington S, Vakil N. Use
Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024
of intravenous proton pump inhibitors in community prac-
tice: an explanation for the shortage? Am J Gastroenterol
2004; 99:1233–7.
18. Nardino RJ, Vender RJ, Herbert PN. Overuse of acid-suppres-
sive therapy in hospitalized patients. Am J Gastroenterol
2000; 95:3118–22.
19. Daneshmend TK, Hawkey CJ, Langman MJ, Logan RF,
Long RG, Walt RP. Omeprazole versus placebo for acute
upper gastrointestinal bleeding: randomised double blind
controlled trial. Br Med J 1992; 304:143–7.
20. Hawkey GM, Cole AT, McIntyre AS, Long RG, Hawkey CJ.
Drug treatments in upper gastrointestinal bleeding: value of
endoscopic findings as surrogate end points. Gut 2001;
49:372–9.
21. Qiao HL, Hu YR, Tian X, Jia LJ, Gao N, Zhang LR, et al.
Pharmacokinetics of three proton pump inhibitors in
Chinese subjects in relation to the CYP2C19 genotype.
Eur J Clin Pharmacol 2006; 62:107–12.
22. Leontiadis GI, Sreedharan A, Dorward S, Barton P,
Delaney B, Howden CW, et al. Systematic reviews of the
clinical effectiveness and cost-effectiveness of proton pump
inhibitors in acute upper gastrointestinal bleeding.
Health Technol Assess 2007; 11:1–164.
23. Spiegel BM, Dulai GS, Lim BS, Mann N, Kanwal F,
Gralnek IM. The cost-effectiveness and budget impact of
intravenous versus oral proton pump inhibitors in peptic
ulcer hemorrhage. Clin Gastroenterol Hepatol 2006;
4:988–97.
24. Barkun AN, Martel M, Toubouti Y, Rahme E, Bardou M.
Endoscopic hemostasis in peptic ulcer bleeding for patients
with high-risk lesions: a series of meta-analyses. Gastrointest
Endosc 2009; 69:786–99.
25. Netzer P, Gaia C, Sandoz M, Huluk T, Gut A, Halter F, et al.
Effect of repeated injection and continuous infusion of ome-
prazole and ranitidine on intragastric pH over 72 hours. Am J
Gastroenterol 1999; 94:351–7.
26. Green FW Jr, Kaplan MM, Curtis LE, Levine PH.
Effect of acid and pepsin on blood coagulation and
platelet aggregation. A possible contributor prolonged
gastroduodenal mucosal hemorrhage. Gastroenterology
1978; 74:38–43.
27. Carey IM, De Wilde S, Harris T, Victor C, Richards N,
Hilton SR, et al. What factors predict potentially inappropri-
ate primary care prescribing in older people? Analysis of UK
primary care patient record database. Drugs Aging 2008;
25:693–706.
 Factors predicting inappropriate utilization of intravenous proton pump inhibitors
28. Cheer SM, Prakash A, Faulds D, Lamb HM. Pantoprazole: an
update of its pharmacological properties and therapeutic use
in the management of acid-related disorders. Drugs 2003;
63:101–33.
29. Raghunath AS, Hungin AP, Cornford CS, Featherstone V.
Use of proton pump inhibitors: an exploration of the atti-
tudes, knowledge and perceptions of general practitioners.
Digestion 2005; 72:212–8.
335
30. Dean B, Schachter M, Vincent C, Barber N. Causes of pre-
scribing errors in hospital inpatients: a prospective study.
Lancet 2002; 359:1373–8.
31. Kaplan GG, Bates D, McDonald D, Panaccione R,
Romagnuolo J. Inappropriate use of intravenous pantopra-
zole: extent of the problem and successful solutions. Clin
Gastroenterol Hepatol 2005; 3:1207–14.
Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024
Downloaded from https://academic.oup.com/qjmed/article/103/5/327/1537505 by guest on 15 April 2024