The Epidemiology of GI Drugs Use Among Persons with RA, OA and Fibromyalgia Frederick Wolfe1, Kaleb

1
The Epidemiology of GI Drugs Use Among Persons
with RA, OA and FibromyalgiaFrederick Wolfe1,
Kaleb Michaud1, Jodi Messer21National Data Bank
for Rheumatic Diseases, Wichita, KS2Wichita
State University, Wichita, KS
Abstract
Results
Table 1. Prevalence of GI drug use
Table 4. Comparison of Clinical and Demographic
Data for Users and Non-users of GI Drugs in
11,442 Rheumatic Disease Patients.
Background. It is widely accepted that
non-steroidal anti-inflammatory drug (NSAIDs) and
rheumatic disease therapies, such as prednisone,
are associated with dyspeptic symptoms and
increased risk of GI ulceration, and such
symptoms and ulcers may also occur in the absence
of NSAIDs. A wide variety of drugs are available
to treat GI symptoms and for ulcer prophylaxis in
rheumatic disease patients. However little is
known about the rates of GI drug use, predictors
of use, and costs associated with GI drug
Therapy. Methods. 11,442 patients with rheumatoid
arthritis, osteoarthritis reported comprehensive
details of their medical history, drug use,
illness severity, and service utilization. Results
. 40 of all patients used some GI drug, and 30
used an H2 blocker, a PPI, sucralfate, or
misoprostol. Compared to patients not on NSAIDs,
the odds ratio (OR) for use in COX-2 and
non-specific NSAID users (NS NSAID) was 1.8 (1.6
to 2.0) and 1.2 (1.3 to 1.2), respectively. Total
direct medical costs were 3,337 per 6 month
period for non GI drug users compared to 3,765,
4,187, and 5,178 for H2, sucralfate and
misoprostol, and PPI users, respectively. After
adjustment for age and sex, persons on Medicaid
and Medicare had rates of PPI use of 27 to 22
compared with PPO and HMO use of 16. GI drug
use was strongly associated with COX-2 therapy,
current and past GI symptoms, and higher rates of
illness severity, including measure of function,
inflammation, quality of life, and work
disability. Simulation studies, however,
suggested that disease severity contributed less
than 2 to the overall prevalence of GI drug
use. Conclusions. GI drug use is common in
rheumatic disease (40). COX-2 users are 1.6 (1.4
to 1.7) times more likely to use GI drugs than NS
NSAID users, and NS NSAID users 1.2 2 (1.1 to
1.3) times more likely to use GI drugs than
non-NSAID users. Regardless of measure studied,
GI drug users have more severe illness across all
rheumatic conditions, and have greater costs and
more severe illness outcomes. Simulation studies
suggest that GI symptoms and NSAID. Prednisone,
and analgesic use rather than disease severity
underlie GI drug prescription. This observation
should also clarify the relationship of
functional disability to GI ulceration that has
been observed previously, suggesting possibly a
more limited role for functional disability than
has previously been noted.
Prevalence of GI drugs in rheumatic in RA, OA,
and fibromyalgia (Table 1). 40 of all patients
report the use of a GI drug in the preceding 6
months. About 36 are directly used for reasons
related to stomach and esophageal use, including
gastric and duodenal ulcers and their prevention,
and heartburn and/or acid reflux conditions.
Antacids use is about 8. The major classes of
Gastro-esophageal drugs are used by 31 of
rheumatic disease patients, including 21 for
PPIs, 11 for H2 drugs, and 2 for the remainder
(sucralfate or misoprostol). Not included in this
list is the misoprostol/diclofenac combination
that was used by 2 of patients.
When total direct medical costs are considered,
the difference between PPI use and no GI Drug use
for all patients is 2333 and is 2453 for those
with RA. Adjusted for severity, these differences
decrease to 1667 and 1727. When the data of
Tables 2 is considered, they show that PPI users
have greater drug costs and greater total costs
than non-PPI users. For drug costs, severity, as
measured by the HAQ, RADAI, and comorbidity,
accounts for a 233 difference between severity
adjusted and not severity adjusted costs, and for
total costs the difference accounted for by
severity is 669.
Table 2. Total direct medical costs according to
category of GI medication.

Association between GI drug classes and
concomitant clinic variables. Table 3 shows that
PPIs are the class of GI drug that is most
strongly associated with clinical variables, with
few exceptions. For example, the odds ratio (OR)
for PPIs compared to H2s is 4.95 vs. 2.77 for
ulcer diagnosis by a physician. However,
regardless of drug class, worse clinical status
is associated with GI drug use compared with no
GI drug. For HAQ, PPI use is associated with a
0.28 difference in HAQ score and, for
comorbidity, with 1.24 additional comorbid
conditions. GI drug use is also associated with
higher rates of COX-2 drugs, prednisone, and
biologics. COX-2 users are 1.6 (1.4 to 1.7) times
more likely to use GI drugs than NS NSAID users,
and NS NSAID users 1.2 (1.1 to 1.3) times more
likely to use GI drugs than non-NSAID users.


Table3. Association between GI Drug Classes and
Clinical Variables Compared to no GI drug use.
Relationship between clinical variables and GI
drug use. Table 4 shows that patients using GI
drugs, defined as any GI drug except antacids or
miscellaneous drugs, have more abnormal score
compared with patients not using such drugs.
Among clinical severity pain scores are almost a
unit greater and HAQ scores 0.25 units greater.
The VAS QOL scale and both utility measure are
more abnormal in GI drug treated patients. Among
concomitant treatment variables there are also
clinically important differences between GI drug
users and non-users prednisone use (41 vs.
30), TNF use (24 vs, 20), COX-2 use (40 vs.
29), and analgesic use (57 vs.38). These data
indicate that patients with more severe disease
received GI drugs. There is other evidence as
well. Work disability (42 vs. 28), social
security disability (29 vs. 18), and total
joint replacement (16 vs. 13) are additional
markers of increased severity among GI drug
users. Two other makers are of interest. GI drug
users have higher scores for non-articular pain
and have a high rate of survey fibromyalgia
diagnosis (27 vs. 16).
Table 5. Increase in GI Drug use as a Function of
the Presence of Risk Factors Based on Monte Carlo
Simulations.
Introduction
In the current report we examine the prevalence,
costs, and predictors of GI drugs in patients
with rheumatoid arthritis (RA), osteoarthritis of
the hip or knee (OA), and fibromyalgia. In
addition, we report on the factors underlying the
association between functional status and GI
disease and GI drug prescription.
Methods
This study was performed using the National Data
Bank for Rheumatic Diseases (NDB). The NDB is a
rheumatic disease research data bank in which
patients complete detailed self-report
questionnaires at 6 month intervals. The
characteristics of the NDB have been reported
previously. Patients in the NDB are recruited
from two sources 1) non-selected patients from
the practices of US rheumatologists and 2)
patients enrolled as part of pharmaceutical
company sponsored registries. Eligible patients
in this study were those with rheumatoid
arthritis (RA), osteoarthritis of the hip or knee
(OA), or fibromyalgia who had completed at least
2 of 6 possible biannual surveys for events
between January 1 1999 and December 31 2001. In
addition, all patients who were recruited as part
of pharmaceutical company registries were
excluded to avoid possible bias. The resultant
data set contained 11,442 patients who were then
studied at their most recent observation after
the creation of lagged variables for the
immediately preceding 6 month survey. The
diagnosis of RA, OA, or fibromyalgia was by the
patients rheumatologist.
Conclusions
GI drug use is common on rheumatic disease, with
30.1 using H2 drugs, PPIs, sucralfate, or
misoprostol. GI drug us associated with more
severe manifestations of rheumatic disease
illness. Patients receiving GI drugs have more
abnormal clinical severity scores, lower quality
of life, higher rates of work and social security
disability, as well as increased us of
prednisone, TNF agents, analgesics, and COX-2
NSAIDs. In addition, GI drug use is associated
with greater drug and total direct medical costs.
However, Monte Carlo simulations suggest that
disease severity has little independent effect,
but instead works through the drug used to treat
severity.
However, further insight into the partition
between severity and drug therapy can be seen
from the Monte Carlo simulations (Table 5). Based
on no risk factors being present (No risk
factors age 25 years, HAQ, prednisone, COX-2
drugs, NS NSAIDs, analgesics, GI factors
(epigastric pain, heartburn, ulcers diagnosis,
previous GI history) all set to 0), the simulated
GI drug use is 3.8. When the HAQ increases to
its mean value (1.06) the additional contribution
to GI drug use is 1.4. Even when prednisone at
25 (the usage in the data set) is added to the
HAQ of 1.06, the additional GI drug use goes to
1.9. It seems likely, then, that severity or
inflammation does not contribute much to overall
GI drug use.