Joranson DE, Ryan KM, Gilson AM, Dahl JL. Trends in Medical Use and Abuse of Opioid Analgesics. JAMA. 2000;283(13):1710-1714. doi:10.1001/jama.283.13.1710
Author Affiliations: Department of Pharmacology (Dr Dahl) and the Pain and Policy Studies Group, Comprehensive Cancer Center (Mr Joranson, Ms Ryan, and Dr Gilson), University of Wisconsin Medical School, Madison.
Context Pain often is inadequately treated due in part to reluctance about using
opioid analgesics and fear that they will be abused. Although international
and national expert groups have determined that opioid analgesics are essential
for the relief of pain, little information has been available about the health
consequences of the abuse of these drugs.
Objective To evaluate the proportion of drug abuse related to opioid analgesics
and the trends in medical use and abuse of 5 opioid analgesics used to treat
severe pain: fentanyl, hydromorphone, meperidine, morphine, and oxycodone.
Design and Setting Retrospective survey of medical records from 1990 to 1996 stored in
the databases of the Drug Abuse Warning Network (source of abuse data) and
the Automation of Reports and Consolidated Orders System (source of medical
Patients Nationally representative sample of hospital emergency department admissions
resulting from drug abuse.
Main Outcome Measures Medical use in grams and grams per 100,000 population and mentions of
drug abuse by number and percentage of the population.
Results From 1990 to 1996, there were increases in medical use of morphine (59%;
2.2 to 3.5 million g), fentanyl (1168%; 3263 to 41,371 g), oxycodone (23%;
1.6 to 2.0 million g), and hydromorphone (19%; 118,455 to 141,325 g), and
a decrease in the medical use of meperidine (35%; 5.2 to 3.4 million g). During
the same period, the total number of drug abuse mentions per year due to opioid
analgesics increased from 32,430 to 34,563 (6.6%), although the proportion
of mentions for opioid abuse relative to total drug abuse mentions decreased
from 5.1% to 3.8%. Reports of abuse decreased for meperidine (39%; 1335 to
806), oxycodone (29%; 4526 to 3190), fentanyl (59%; 59 to 24), and hydromorphone
(15%; 718 to 609), and increased for morphine (3%; 838 to 865).
Conclusions The trend of increasing medical use of opioid analgesics to treat pain
does not appear to contribute to increases in the health consequences of opioid
Unrelieved pain, whether due to trauma, surgery, cancer or noncancer
conditions, and including pain occurring at the end of life, continues to
be a major public health concern.1- 4
Although numerous nonpharmacologic treatments can be used to relieve pain,
the use of opioids in the class of morphine is the cornerstone of pain management.5- 8
However, because opioids have the potential to be abused, they are regulated
under international and national narcotics and controlled substances laws.9,10 International and US federal drug
laws embody a dual imperative to ensure the availability of controlled substances
for medical and scientific purposes, while at the same time to prevent their
diversion and abuse.11
Concerns related to drug abuse permeate efforts to treat pain with opioids.
Patients are concerned about becoming addicted to opioids.6,12,13
Health care professionals may be reluctant to prescribe, administer, dispense,
or stock controlled substances for fear of causing addiction or contributing
to the drug abuse problem.14- 17
There are few studies of the extent to which prescription opioid analgesics
contribute to the national drug abuse problem.18
In this descriptive study, we examine the abuse of opioid analgesics in relationship
to their medical use.
We evaluated abuse trends for opioid analgesics as a class, as well
as the medical use and abuse of 5 specific Schedule II opioids: fentanyl,
hydromorphone, meperidine, morphine, and oxycodone. We chose these 5 drugs
because they are effective in treating severe pain and are marketed as analgesics.5- 7,19 We excluded
opioid analgesics classified in lower schedules (ie, Schedules III and IV),
such as hydrocodone and codeine combinations because they are not indicated
for severe pain. We also excluded opioid analgesics for which consumption
data would include amounts used for other major indications (such as codeine
for cough and diarrhea, and drugs used for treatment of opioid addiction such
We used the Drug Abuse Warning Network (DAWN) as the source for data
on opioid abuse. DAWN, sponsored by the Substance Abuse and Mental Health
Services Administration in the US Department of Health and Human Services,
provides estimates of the health consequences of the nonmedical use of individual
drugs. It is a large-scale, ongoing retrospective survey of medical records
that is used to monitor national drug abuse trends. The system collects information
from DAWN-affiliated hospital emergency departments (EDs) to identify substances
that are abused; monitor drug abuse patterns and trends and detect new drug
entities and combinations; assess health hazards associated with drug abuse;
and provide data for national, state, and local drug abuse policy and program
Data are collected on patients 6 years and older from the EDs of approximately
500 US hospitals in 21 metropolitan and other nonmetropolitan areas. Hospitals
eligible to participate in DAWN are nonfederal, short-stay general hospitals
with 24-hour EDs that are located in the coterminous United States. DAWN has
been in existence since 1972, and began collecting data from a nationally
representative sample of EDs in 1990.20,21
For our analysis, we used data from the 7-year period from 1990 to 1996 (the
most recent year for which data were available at the time this study was
Each hospital has a designated reporter, usually a member of the ED
or medical records department, who is responsible for obtaining information
from medical records each time a patient visits the ED with a presenting problem
related to drug use; each visit is defined as an episode. Reported episodes
typically involve drug overdoses, but also may be the result of long-term
drug use and adverse reactions. The reporter collects information for each
episode, including gender, ethnicity, age, concomitant use of other drugs,
motive for use, reason for ED contact, source of substance, dosage form, and
route of administration. The national estimates of abuse are derived from
these data. Standardized procedures are used to collect DAWN data; however,
there may be some variability from facility to facility.20
Drug abuse in the DAWN system is defined as the nonmedical use of a
substance for psychic effect, dependence, or suicide attempt or gesture. Drug
abuse can involve the use of illicit drugs or any other substance (eg, heroin,
marijuana, peyote, glue, aerosols); prescription drugs in a manner inconsistent
with accepted medical practice; and over-the-counter drugs contrary to approved
For each episode of drug abuse, as many as 4 different substances, in
addition to alcohol, can be recorded. Each is referred to as a drug mention.
More than half of DAWN episodes involve multiple drug mentions.20
If DAWN reporters are not able to classify a drug, these mentions are classified
as other/unspecified. DAWN reports do not include information about drugs
for which the frequency of annual mentions is less than 200. However, for
our study, the Substance Abuse and Mental Health Services Administration performed
a special data run to extract annual abuse mentions for one of the study drugs,
fentanyl, for which the number of mentions is consistently less than 200 per
We renamed several drug categories and reclassified a number of drugs
to reflect current medical terminology and pharmacology. The categories designated
as narcotic analgesics and nonnarcotic analgesics were renamed as opioid analgesics
and nonopioid analgesics. Some drugs were reclassified: codeine combinations
were recategorized from other drugs to opioid analgesics. Hydrocodone was
reclassified from other/unspecified narcotic analgesics to opioid analgesics.
Ibuprofen and naproxen were transferred from other drugs to nonopioid analgesics.
Methamphetamine and methaqualone, originally classified as amphetamines and
nonbarbiturate sedatives, respectively, were moved to the category of illicit
drugs. Other/unspecified drugs listed within each subcategory of DAWN reports
were aggregated into 1 "other drugs" category.
DAWN reports routinely combine heroin and morphine mentions into a single
category, making it impossible to distinguish between the abuse of an illicit
drug and an essential pain medication. We requested that the Substance Abuse
and Mental Health Services Administration separate morphine from the heroin-morphine
category for the period from 1990 to 1996. Morphine mentions accounted for
an annual average of 1.9%, and never exceeded 2.5%, of the combined heroin-morphine
We obtained data on medical use of opioids from the US Drug Enforcement
Administration's Automation of Reports and Consolidated Orders System (ARCOS)
for the years 1990 to 1996. ARCOS is a federal, computerized data system,
required by the 1970 Controlled Substances Act.10
ARCOS monitors the lawful distribution of controlled substances in Schedules
I and II and narcotic substances in Schedule III from manufacturers to the
retail level of consumption, including hospitals, pharmacies, and licensed
practitioners. The Drug Enforcement Administration makes reports on ARCOS
data, providing information on individual states and national totals.22 Information is provided for each drug in total grams
and grams per 100,000 population.23,24
ARCOS is the only nonproprietary source of information on medical use of opioids.
The percentage of 1996 total DAWN ED mentions represented by each of
the 5 drug categories is shown in Table
1. Mentions for opioid analgesics account for less than 4% of total
DAWN mentions, mentions for nonopioid analgesics account for 8.6%, and mentions
for illicit drugs account for 33.2%. Table
2 presents the abuse levels for the same drug categories as number
of mentions and as a percentage of total DAWN mentions for the period 1990
to 1996. From 1990 to 1996, the number of mentions for drug abuse in DAWN
increased from 635,460 to 907,561 (42.8% increase), with an increase in total
mentions for all drug categories. For opioid analgesics, the total number
of mentions increased from 32,430 in 1990 to 34,563 in 1996 (6.6% increase),
but declined as a percentage of total mentions from 5.1% in 1990 to 3.8% in
1996. Illicit drugs is the only category of drug abuse that exhibited a continual
increase in both number of mentions and percentage of total mentions over
the study period.
Trends in the medical use of the 5 selected opioid analgesics from 1990
to 1996 are shown in Table 3.
Substantial increases were observed in use of fentanyl and morphine, which
occurred in both total use and use adjusted for population.
Table 4 presents trends
for the abuse of the 5 selected opioid analgesics for the same 7-year trend.
The number of abuse mentions, as measured by DAWN, for fentanyl, hydromorphone,
meperidine, and oxycodone declined during the study period, whereas abuse
mentions for morphine increased by 3.2%. The abuse levels for each of the
5 opioid analgesics, as a percentage of total DAWN mentions, were less than
1% and declined during the study period despite substantial increases in medical
Official Drug Enforcement Administration data indicate that the amounts
of fentanyl, oxycodone, hydromorphone, and morphine distributed to the retail
level have increased substantially.25 According
to the World Health Organization, increasing medical use of opioids is one
indication that progress is being made to improve pain management.19 Despite these increases, pain is still inadequately
treated due to numerous barriers to pain management.6,26
In the future, as these barriers are addressed, the medical use of opioid
analgesics may be expected to increase further.
In this study, meperidine was the only opioid that decreased in medical
use over the study period. This decrease may reflect increasing awareness
of the shortcomings of the use of meperidine for chronic pain, which includes
short duration of action and accumulation of a long-lived toxic metabolite.6,27
These data suggest that opioid analgesics, including the 5 study drugs,
are a relatively small part of drug abuse as measured by the DAWN system.
Although there are year-to-year variations, the abuse levels have remained
low and relatively stable for the past 7 years despite substantial increases
in the medical use of opioids. Although abuse of most opioids decreased during
the study period, several caveats are needed to place these results in context.
First, these data also indicate that there is some abuse of opioid analgesics.
However, compared with the abuse of other drugs, illicit drugs in particular,
the abuse of opioid analgesics appears to be relatively low, accounting for
3.8% of total DAWN mentions in 1996. Moreover, even though there were increases
in the total number of mentions of abuse for opioid analgesics during the
study period, the proportion of mentions for opioid abuse relative to total
reports of drug abuse decreased by 25% (from 5.1% to 3.8%).
Second, the DAWN system may underestimate the extent of the drug abuse
problem. The DAWN system measures only those episodes of drug abuse that result
in an admission to an ED, and thus underreports the true extent of all drug
abuse, such as drug-related overdoses and deaths occurring out of the hospital.
We believe that this limitation, which is common to all drug abuse monitoring
systems, should not affect the proportional representation of drug categories
because it should affect all drug categories equally. Furthermore, our study
primarily addresses changes in abuse trends rather than abuse levels at a
single point in time. As such, it is important to evaluate these data longitudinally
to examine the stability of the incidence frequency, and to detect relative
changes in abuse levels. However, the apparent stability of abuse levels over
the study period suggests that actual abuse of opioid analgesics, regardless
of total prevalence, is also relatively stable.
Third, although the DAWN system collects some information about the
severity of adverse events resulting from drug ingestion, such as mortality
rates and hospital admissions, we were unable to use the data to detect any
trends in the severity of adverse events related to the abuse of opioid analgesics.
The mortality data in the DAWN ED system are sparse because most deaths do
not occur in the ED. Instead, these deaths appear in the DAWN medical examiner
data. However, the DAWN medical examiner reports were not used because drug-related
mortality data do not differentiate among deaths related to morphine, heroin,
and codeine. Moreover, data on hospital admissions related to abuse of opioid
analgesics are not available for most of the drugs examined in this study.
Fourth, we used data on abuse of opioids from the DAWN system. Two other
sources of drug abuse information were considered but not used. The National
Household Survey on Drug Abuse,28 a widely
known survey measuring drug abuse prevalence, was not used because data are
not available for specific opioid analgesics. Also, the Toxic Exposure Surveillance
System,29 which tracks exposures to toxic substances
from a large sample of regional poison control centers, was not used because
it is not nationally representative and because the cost of acquiring the
data was prohibitive.
Conventional wisdom suggests that the abuse potential of opioid analgesics
is such that increases in medical use of these drugs will lead inevitably
to increases in their abuse. The data from this study with respect to the
opioids in the class of morphine provide no support for this hypothesis. The
present trend of increasing medical use of opioid analgesics to treat pain
does not appear to be contributing to increases in the health consequences
of opioid analgesic abuse. To maintain this trend, manufacturers, pharmacies,
clinicians, and patients should continue their efforts to improve pain management
while exercising care so that the diversion of opioid medications for nonmedical
use is minimized. If the abuse of opioid analgesics should increase, the sources
of diversion should be addressed directly without interfering with medical
availability of opioid analgesics, legitimate medical practices, or patient