вход по аккаунту


FDA arthritis advisory committee meetingrisks of agranulocytosisaplastic anemia flank pain and adverse gastrointestinal effects with the use of nonsteroidal antiinflammatory drugs.

код для вставкиСкачать
Agranulocytosislaplastic anemia with the use of
nonsteroidal antiinflammatory drugs. A discussion of
this subject was prompted by the recently published
report of the International Agranulocytosis and
Aplastic Anemia (IAAA) Study Group on the risks of
those events in relation to use of analgesics (1).
Scattered reports that associate agranulocytosis and
aplastic anemia with various nonsteroidal antiinflammatory drugs (NSAIDs) have been received over the
years by the Voluntary Adverse Reaction Reporting
System, and some case reports have been published.
Based on Inman’s estimate of 22 deaths from blood
dyscrasia per million patients (2), and FDA estimates
of 16 deaths per million (3), it has been recommended
that phenylbutazone is “not for initial use” for any
condition. Its label contains strong warning statements, and oxyphenbutazone has been voluntarily
withdrawn from the market by its major manufacturer.
To investigate the problem, the IAAA study
used a different approach. For 4 years, 40 investigators
in 7 countries monitored 300 hospitals that served a
total population of 22.3 million, and this provided
almost 90 million person-years of experience. Unexplained severe selective neutropenia (granulocytes <500/mm3) was identified in 422 cases, and
aplastic anemia (leukocytes s3,500/mm3, platelets
s50,000/mm3, hemoglobin 5 10 gm/dl, and reticulocytes s30,000/mm3) in 168 cases at the various hospitals. Aplastic anemia was confirmed by bone marrow
examination and evaluation by a committee of hematologists. The overall annual incidence of agranHarold E. Paulus, MD, American Rheumatism Association
Liaison to the FDA Advisory Committee, UCLA School of Medicine, Department of Medicine, Division of Rheumatology, Los
Angeles, CA 90024.
Arthritis and Rheumatism, Vol. 30, No. 5 (May 1987)
ulocytosis was 6.2 cases/million/year, with a range of
1.7-9.0 in the various regions studied, compared with
2.2/million/year (range 0.6-3.1) for aplastic anemia.
Nine percent of the patients with agranulocytosis
(0.5/million/year) died, and 49% of the aplastic anemia
patients died during 2 years of followup. The highest
incidence rates were among females older than age 60.
It was not determined whether this was dqe to greater
drug use or greater susceptibility in this group. Fewer
than 20 of the agranulocytosis patients had rheumatoid
arthritis, and exclusion of these patients did not
change the results.
After various exclusions and losses to followup,
a case control study was done. Investigators carefully
interviewed 221 patients with agranulocytosis and
1,425 appropriate control subjects, and 113 patients
with aplastic anemia and 1,224 controls. Specific efforts were made to record all drugs used by the
subjects in the agranulocytosis study during the 7 days
prior to the first indication or symptom of agranulocytosis (or a predefined index day for the controls). For
aplastic anemia patients and controls, all drugs used
during days 29-1 80 before hospital admission were
Among the 221 patients with agranulocytosis,
43 reported taking at least 1 dose of a salicylate during
the 7-day index period, 11 reported taking phenylbutazone or oxyphenbutazone, 10 reported taking
indomethacin, and 5 reported taking other NSAIDs.
Because the usage rates were similar among the controls, there was no evidence that salicylates or other
NSAIDs increased the risk of agranulocytosis. Controls, however, used butazones and indomethacin at
lower rates than did the agranulocytosis patients;
thus, these drugs were significantly associated with
agranulocytosis, and were used, respectively, 4 and 9
times more frequently by patients than by controls.
The excess risk of developing agranulocytosis was
calculated to be 0.2/million among persons who took
butazones for up to 7 days, as 0.6imillion among those
who took indomethacin for up to 7 days.
During days 29-180 before admission, among
the 113 patients with aplastic anemia, 59 reported
salicylate use at least once; 9 reported use of
butazones; 7, indomethacin; and 9, other NSAIDs.
Again, no increased risks of developing aplastic
anemia were associated with salicylates or other
NSd41Ds, but the risk was increased 8.7 times with
butazones and 12.7 times with indomethacin. Although only 5 patients who took butazones and 4 who
took indomethacin received the drug 4 or more days
per week for 2 4 weeks, there were so few controls
who took these drugs in that way, that the multivariate
rate ratio risk estimates for users increased to 43.4 and
49.9, respectively. The estimated excess risk of developing aplastic anemia for persons taking any
butazones during a 5-month period was calculated
to be 6.6/million, and for indomethacin, it was
By way of comparison, it was noted that the
risk of death from anaphylaxis after a penicillin injection is 20/million injections, the risk of death from
thro mboembolic events associated with oral contraceptive use is 18/million/year, and the risk of death
from lightning strike is l/million in the US.
The indomethacin risk rates reported in the
IAAA study are consistent with the evidence in the
earlier study by Inman (2), but the phenylbutazone
risks are lower than those previously reported (4). The
consensus of the discussants and the Committee members was that the risk of developing agranulocytosis or
aplastic anemia from NSAID use is exceedingly small,
although some measurable risk is probably present
with the butazones, and may be present with
indornethacin. Pending additional evidence to support
these conclusions, no changes in label warnings were
recommended. In view of the rarity of these problems
and their unpredictable nature, monitoring for their
occurrence with frequent routine blood cell counts
was not deemed necessary. Some of the many problems and pitfalls inherent in case control studies such
as the IAAA study were discussed, and its accuracy in
differentiation of events that occur with a frequency
5 lO/million was questioned.
Flank pain with the use of NSAIDs. Shortly after
the NSAID suprofen was placed on the market, in
January 1986, for pain and dysmenorrhea, spontaneous reports were received that described the abrupt
onset of severe bilateral flank pain within 2 or 3 hours
after the first or second dose of the drug. The pain
often radiated to the groin or lower abdomen, lasted 2
or 3 days, and in approximately 75% of patients, was
followed by a rise in serum creatinine, microscopic
hematuria, mild proteinuria, and polyuria. By November 1, 1986, 301 cases had been reported to the FDA;
the reporting was prompted in part by several warning
letters sent to all US physicians. Males <40 years old
were most commonly affected. The estimated incidence was 1.5/1,000 suprofen users. Intravenous
pyelograms were performed on some patients and
showed faint or absent visualization of the collecting
system, with a good nephrogram. In all patients, the
pain resolved, and renal function returned to baseline.
A careful study, which included frequent analysis of urine and plasma from well-hydrated normal
volunteers, demonstrated that suprofen increases uric
acid excretion approximately 50% and doubles the
undissociated uric acid concentration in urine, but
only during the 2 hours immediately following administration of the drug. Fractional excretion of sodium
decreased, but there was no evidence of renal tubular
or glomerular toxic injury. There was also a transient
decrease in creatinine clearance and renal plasma
flow, but all of these changes reversed rapidly within a
few hours.
It was hypothesized that suprofen-induced
acute renal failure is caused by intratubular precipitation of uric acid, with temporary tubular obstruction
causing both the flank pain and the rise in creatinine.
Although suprofen is a potent uricosuric, its half-life is
so short that this effect lasts only a few hours, and
serum uric acid levels fall only modestly (from 5.8
mg/dl to 4.7 mg/dl) during continued treatment with
the drug. Uric acid precipitation is more likely to occur
in acid urine with low urine flow rates and high uric
acid concentrations; thus, dehydration and hyperuricemia predispose to this condition.
Because of the fairly frequent occurrence
of this syndrome, suprofen is not recommended for
initial use as an analgesic, and the manufacturer has
stopped promoting it. Patients are advised to drink 1-2
glasses of water with the medication, and thus be well
This syndrome of flank pain and renal dysfunction may not be unique to suprofen, although it appears to occur more frequently with use of this drug. A
search of the FDA’s Voluntary Adverse Reaction
Report files revealed 55 reports of a flank pain syndrome with similar characteristics, associated with a
variety of other NSAIDs, although only a few of the
patients had elevations of serum creatinine levels.
The Advisory Committee recommended that
the FDA emphasize the current warning statement on
the label by highlighting it in a printed box, and that a
patient package insert be developed for suprofen. In
addition, more detailed information is to be obtained
regarding the 301 reported cases, and these subjects
are to be followed for 2 years to document the lack of
permanent sequelae and to better establish predisposing factors and mechanisms in this syndrome.
Based on current knowledge, physicians should
caution patients to be well hydrated before taking the
first dose of an NSAID. Weekend athletes, who become acidotic and dehydrated from exercise and then
take an NSAID to relieve their aches and pains, may
be particularly susceptible. If the syndrome is observed, it should be reported to the FDA, either
directly or through the pharmaceutical company representative.
Gastrointestinal adverse reactions with longterm use of NSAIDs. A general paragraph regarding the
risk of gastrointestinal ulcers, bleeding, and perforation with long-term NSAID treatment is being developed for inclusion on the labels of all NSAIDs. By life
table analysis of prospectively collected data from
multiple NSAID submissions, the FDA estimates that
these serious events occur in approximately 1-2% of
patients using NSAIDs for 3 months, and in approxi-
mately 2-5% using them for 1 year. The cumulative
risk appears to increase with the duration of therapy
and to be greater in patients with previous peptic ulcer
disease. Fatal outcomes are more likely in elderly or
debilitated patients. Higher dosages of NSAID probably entail greater risk than lower dosages. The patient’s disease, age, and degree of inflammation need
to be considered in the choice of the optimal dosages
for individual patients.
Regarding piroxicam, dosages >20 mg/day are
not recommended because of a greater risk of the more
serious gastrointestinal reactions. Data justifying the
efficacy of 10 mg/day will be submitted by the manufacturer. Meanwhile, physicians should consider starting elderly or other high-risk patients with a dosage of
10 mg/day, and should attempt to reduce maintenance
dosages to 10 mg/day if initial control of symptoms has
required 20 mg/day.
1. The International Agranulocytosis and Aplastic Anemia
Study: Risks of agranulocytosis and aplastic anemia: a
first report of their relation to drug use with special
reference to analgesics. JAMA 256: 1749-1757, 1986
2. Inman WHW: Study of fatal bone marrow depression
with special reference to phenylbutazone and oxyphenbutazone. Br Med J 1:15W1505, 1977
3 . Paulus HE: Government affairs: FDA Arthritis Advisory
Committee meeting. Arthritis Rheum 28:450451, 1985
4. Faich GA: Analgesic risks and pharmacoepidemiology
(editorial). JAMA 256: 1788, 1986
Без категории
Размер файла
290 Кб
adverse, flank, meetingrisks, agranulocytosisaplastic, anemia, pain, effect, drug, gastrointestinal, advisory, arthritis, nonsteroidal, antiinflammatory, fda, committee, use
Пожаловаться на содержимое документа