Download Heroin-Related Noncardiogenic Pulmonary Edema

Heroin-Related Noncardiogenic
Pulmonary Edema*
A Case Series
Karl A. Sporer, MD; and Elizabeth Dorn, MD
Study objectives: To examine the current clinical spectrum of noncardiogenic pulmonary edema
(NCPE) related to heroin overdose.
Design: Retrospective chart review of all identified patients from August 1994 through December
1998.
Setting: Urban academic hospital.
Patients or participants: Heroin-related NCPE was defined as the syndrome in which a patient
develops significant hypoxia (room air saturation < 90% with a respiratory rate > 12/min) within
24 h of a clinically apparent heroin overdose. This should be accompanied by radiographic
evidence of diffuse pulmonary infiltrates not attributable to other causes, such as cardiac
dysfunction, pneumonia, pulmonary embolism, or bronchospasm, and which resolve clinically
and radiographically within 48 h.
Interventions: None.
Measurements and results: Twenty-seven patients were identified during this 53-month period,
with a majority being male patients (85%; average age, 34 years). Twenty patients (74%) were
hypoxic on emergency department arrival, and 6 patients (22%) had symptoms develop within the
first hour. One patient had significant hypoxia develop within 4 h. Nine patients (33%) required
mechanical ventilation, and all intubated patients but one were extubated within 24 h. Eighteen
patients (66%) were treated with supplemental oxygen alone. Hypoxia resolved spontaneously
within 24 h in 74% of patients, with the rest (22%) resolving within 48 h. Twenty patients (74%)
had classical radiograph findings of bilateral fluffy infiltrates, but unilateral pulmonary edema
occurred in four patients (15%) and more localized disease occurred in two patients (7%).
Conclusion: NCPE is an infrequent complication of a heroin overdose. The clinical symptoms of
NCPE are clinically apparent either immediately or within 4 h of the overdose. Mechanical
ventilation is necessary in only 39% of patients. The incidence of NCPE related to heroin
overdose has decreased substantially in the last few decades.
(CHEST 2001; 120:1628 –1632)
Key words: heroin; noncardiogenic; poisoning; pulmonary edema
Abbreviations: ED ⫽ emergency department; NCPE ⫽ noncardiogenic pulmonary edema; PA ⫽ pulmonary artery
use has increased dramatically in the
H eroin
United States during the past decade, and heroin-related emergency department (ED) visits have
increased 110% between 1990 and 1995.1 In 1996,
heroin overdose was responsible for 14,300 ED visits
as well as 4,178 deaths.2
In certain western cities, the death toll from
*From the Department of Medicine (Dr. Sporer), University of
California, San Francisco, San Francisco General Hospital, San
Francisco; and Department of Emergency Medicine (Dr. Dorn),
Alameda County Medical Center/Highland Campus, University
of California, San Francisco, San Francisco, CA.
Manuscript received August 3, 2000; revision accepted May 1,
2001.
Correspondence to: Karl A. Sporer, MD, Emergency Services,
Room 1E21, San Francisco General Hospital, 1001 Potrero Ave,
San Francisco, CA 94110; e-mail: [email protected]
1628
heroin overdoses has become the largest category of
preventable deaths. In San Francisco, more people
die of a heroin overdose than all causes of trauma
combined.3 The actual mechanism of death caused
by a heroin overdose has never been adequately
explained.4 The one striking finding is the almost
universal occurrence of noncardiogenic pulmonary
edema (NCPE) in the fatal overdose.4 – 6 It is hoped
that a better understanding of NCPE in patients with
nonfatal heroin overdoses may lead to a partial
explanation of the mode of death in a heroin overdose and can lead us in directions that may decrease
the death rate from this street drug. An appreciation
of the incidence of NCPE in heroin-overdose patients may also give us insight into proper treatment
of patients with nonfatal heroin overdoses.
Clinical Investigations in Critical Care
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21969/ on 05/15/2017
NCPE is a well-known severe complication of
heroin overdose that was first described by William
Osler during an autopsy in 1880.7 The presentation
and clinical course of NCPE patients with nonfatal
heroin overdoses were first described8 in 1953, and
later in a variety of case series and reports.9 –28
A heroin overdose is routinely clinically diagnosed
by noting the combination of altered mental status,
severely decreased respiratory drive, pinpoint pupils,
and circumstantial evidence of drug use.29 NCPE
related to a heroin overdose usually presents as the
combination of persistent hypoxia after resolution of
opiate respiratory depression along with frothy, pinktinged pulmonary secretions, and a characteristic
radiograph pattern of fluffy diffuse pulmonary infiltrates. Some of these patients require mechanical
ventilation to treat their hypoxia. The symptoms
usually resolve rapidly with supportive care alone
within hours to 1 or 2 days.
NCPE has also been reported with overdoses of
methadone,30 propoxyphene,31 codeine,32 buprenorphine,33,34 and nalbuphine.35 There is also an unrelated syndrome of cardiogenic pulmonary edema
that has been rarely reported with the use of naloxone36 or nalmefene37 to reverse therapeutic doses of
opiates after anesthesia.
Materials and Methods
This study took place at San Francisco General Hospital, an
urban hospital with ⬎ 75,000 ED visits per year. The University
of California, San Francisco Committee of Human Research
approved this study.
Most heroin overdoses are diagnosed and initially treated by
paramedics. After making the presumptive diagnosis of a heroin
overdose (a patient with an abnormal level of consciousness,
pinpoint pupils, and decreased or absent respirations), the patient
is treated with a bag-valve mask and supplemental oxygen along
with an IM dose of naloxone, 2 mg. Approximately 70% of the
heroin-overdose patients in San Francisco are transported to San
Francisco General Hospital.38 In the ED, heroin-overdose patients are observed for 2 h for signs of significant resedation,
significant hypoxia, fever, or other complications. IV fluids are
rarely administered.
We defined opiate-related NCPE as the syndrome in which a
patient develops significant hypoxia (room air oxygen saturation
⬍ 90% with a respiratory rate ⬎ 12/min) within 24 h of a
clinically apparent heroin overdose. This should be accompanied
by radiographic evidence of diffuse pulmonary infiltrates not
attributable to other causes, such as cardiac dysfunction, pneumonia, pulmonary embolism, or bronchospasm, and which resolve clinically and radiographically within 48 h.
A database of all ED patients seen during the period of August
1994 through December 1998 was queried for this study. The
absolute number of patients who presented to our ED with
the diagnosis of a heroin overdose was determined by searching
the diagnosis field with a combination of the terms, “heroin
overdose” or “HOD.” All of the admissions to the medicine
service at our hospital with a test diagnosis containing the words
“heroin overdose,” “HOD,” “heroin,” or “opiate” were reviewed.
The clinical charts and radiographs of 162 patients were reviewed
by one of the authors (E.D.) for appropriateness of the diagnosis
of NCPE. Most patients were admitted to the hospital for a
variety of other medical reasons, such as pneumonia, possible
endocarditis, persistent confusion, cellulitis, etc. Borderline cases
(eg, atypical pneumonia) were reviewed by the first author
(K.A.S.) for consistency with the diagnosis.
Results
There were 1,278 patients who presented with the
diagnosis of heroin overdose in our hospital during
this 53-month period; 162 of these patients were
admitted to the medical service, and a total of 27
patients met the criteria for the diagnosis of NCPE.
There were 23 male patients in this NCPE cohort,
and the average age was 34 years. There were no
repeat cases among the NCPE group. The repeat
rate for all heroin-overdose patients has been
shown38 to be extremely small in our city.
All patients but one received prehospital naloxone.
Twenty patients (74%) were hypoxic on presentation
to the ED, and 6 patients (22%) had severe hypoxia
develop during the first hour. One patient (4%)
presented with a room air oxygen saturation of 93%
and a respiratory rate of 18/min that worsened within
4 h to a room air oxygen saturation of 81%. Characteristic radiographic findings of pulmonary edema
developed in this patient.
There were no episodes of hypoventilation in this
group of patients. The respiratory rate ranged from
16 to 44/min (average, 24/min). The initial oxygen
saturation ranged from 47 to 89% on room air, with
an average oxygen saturation of 76%. Thirty-three
percent of these patients required mechanical ventilation to achieve adequate oxygenation, while most
of the patients (66%) were treated with oxygen via
nonrebreather face mask and observation only.
There were three patients who presented with hypotension (systolic BP ⬍ 90 mm Hg). These episodes resolved spontaneously with ⬍ 500 mL of
crystalloid fluid resuscitation.
The hypoxic symptoms of a majority of patients
(74%) resolved within 24 h, several within 8 h. The
other six patients (22%) had resolution of symptoms
within 24 to 48 h. All intubated patients but one were
extubated within 24 h. This patient required a
prolonged ICU stay partially because of an iatrogenic
pneumothorax.
Classic chest radiographic findings of bilateral
pulmonary edema were noted in 20 of the 27
patients (74%). Interestingly, four patients presented
with chest radiographic findings of unilateral pulmonary edema, a finding that has been previously
reported.14,16,39 Three other patients had chest radiographic findings of “patchy atelectasis” (n ⫽ 2)
and “nodular interstitial pattern” (n ⫽ 1). These
CHEST / 120 / 5 / NOVEMBER, 2001
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21969/ on 05/15/2017
1629
patients did not undergo bronchoscopy to definitively rule out aspiration pneumonia, but the absence
of an increased WBC count and the very rapid
resolution of the symptoms and the radiographic
findings make this less likely. Rhabdomyolysis was
diagnosed in two of these patients, and they were
treated with fluid replacement and sodium bicarbonate. Neither patient developed renal insufficiency.
Discussion
We have presented the clinical characteristics of
the largest case series of heroin-related NCPE since
the 1960s. NCPE is an infrequent complication of
patients who present with an acute heroin overdose.
The symptoms of significant hypoxia with an adequate respiratory rate usually occur rapidly after
treatment. The radiographic findings are usually
bilateral fluffy infiltrates, but a percentage of patients can have unilateral or lobar disease. Only 33%
of these patients require intubation. Most patients
required only supplemental oxygen, and most of
these patients’ symptoms resolve within 24 h.
The epidemiology of NCPE has changed over the
last 40 years. Earlier retrospective case series14,17,18
of heroin-overdose patients admitted to the hospital
have described a NCPE rate of 48 to 80%. The full
denominator, including heroin-overdose patients admitted to and discharged from the hospital, was not
available in these studies. More recent case series38,40 – 43 of patients presenting with heroin overdoses have a NCPE rate of 0.8% to 2.4%, similar to
our rate of 2.1%.
The decreased rate of NCPE in patients with
nonfatal heroin overdoses can be partially explained
by including the full denominator of heroin-overdose
patients admitted to and discharged from the hospital. But this alone cannot explain all of the difference. These earlier case series14,17,18 occurred before
naloxone was readily available. This specific narcotic
antagonist may be instrumental in decreasing the
number of patients with nonfatal heroin overdoses
with NCPE. The improvement of emergency medical services with the earlier prehospital administration of naloxone and oxygen may also add to this
improvement.41
This early experience with high rates of NCPE is
the likely reason for the extended recommended
observation period of 12 to 24 h that is commonly
quoted.14,44 Because NCPE is a relatively uncommon event in heroin-overdose patients and usually
presents early, a 1-h to 2-h observation period should
be adequate for these patients. One prospective
study43 demonstrated that these patients rarely return to the hospital for subsequent treatment.
1630
The pathophysiology of NCPE related to heroin
overdose has never been adequately explained. This
is of interest not because of its relative infrequency
in nonfatal heroin overdose, but because of its almost
universal occurrence in fatal heroin overdose.4 – 6
This may suggest that this entity may be a partial
explanation for the mode of death in a heroin
overdose.
The investigations into the cause of heroin-related
NCPE have been limited. Hemodynamic evaluation
of a total of 11 patients with heroin-related NCPE
demonstrated normal or increased cardiac output,
normal pulmonary capillary pressure, and only moderately elevated pulmonary artery (PA) pressure.24,28,45,46 Pulmonary fluid analysis of patients
with this syndrome confirmed a much higher protein
concentration, as compared to edema fluid from
patients with congestive heart failure, consistent with
the hypothesis that leaky pulmonary capillaries and
not pump failure are the likely cause of heroinrelated NCPE.45,47
The only animal model48 of opiate-induced pulmonary edema examined the effect of morphine on
the pulmonary vasculature in perfused, normally
oxygenated cats and dogs. A large dose of morphine
caused a highly variable increase in pulmonary vascular resistance via venous vasoconstriction. Animals
with a mild-to-moderate increase in PA pressure
recovered spontaneously and returned to baseline.
The animals with a marked increase in PA pressure
had pulmonary edema develop within 20 to 30 min.
The increase in PA pressure was reversed by chlorpheniramine, a histamine type 1-receptor antagonist.
Pretreatment with naloxone blunted this response
but had no effect in posttreatment.
Histamine has been implicated as a possible mediator of NCPE related to heroin overdose. Human
lungs are rich in histamine,49 and opiates have been
shown to cause its systemic and local release.50 –57
Histamine has been shown49,58,59 to increase pulmonary lymph flow and capillary permeability. One
study of patients with fatal heroin overdoses demonstrated a modest correlation with elevated tryptase
levels (a postmortem histamine surrogate), as compared to control subjects.60
Conclusion
To our knowledge, this is the largest case series of
NCPE reported in several decades. NCPE is an
infrequent complication of nonfatal heroin overdoses, with an incidence of 2.1% in our hospital. The
symptoms of significant hypoxia with an adequate
respiratory rate usually occur rapidly after treatment.
The radiographic findings are usually bilateral fluffy
Clinical Investigations in Critical Care
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21969/ on 05/15/2017
infiltrates, but a significant percentage of patients
can have unilateral or lobar disease. Only 33% of
these patients require mechanical ventilation. A majority of patients only required supplemental oxygen,
and symptoms in most of these patients resolved
within 24 h.
References
1 Department of Health and Human Services. Year-end preliminary estimates from the 1996 Drug Abuse Warning
Network, Rockville, MD: Substance Abuse and Mental
Health Services Administration, November 1997
2 Department of Health and Human Services. Drug Abuse
Warning Network: annual medical examiners data, 1995,
Rockville, MD: Substance Abuse and Mental Health Services
Administration, May 1997
3 McLoughlin E, Weitzel M, Skaj P, et al. Profile of injury in
San Francisco. San Francisco, CA: San Francisco Department of Public Health, 1998
4 Darke S, Zador D. Fatal heroin “overdose”: a review. Addiction 1996; 91:1765–1772
5 Levine SB, Grimes ET. Pulmonary edema and heroin overdose in Vietnam. Arch Pathol 1973; 95:330 –332
6 Karch SB. Narcotics. In: The pathology of drug abuse. 2nd
ed. Boca Raton, FL: CRC Press; 1997; 281– 407
7 Sternbach G. William Osler: narcotic-induced pulmonary
edema. J Emerg Med 1983; 1:165–167
8 Troen P. Pulmonary edema in acute opium intoxication.
N Engl J Med 1953; 248:364 –366
9 Silber R, Clerkin EP. Pulmonary edema in acute heroin
poisoning. Am J Med 1959; 27:187–192
10 Selzman HM, Kahil ME, Fred HL. Pulmonary edema accompanying heroin intoxication. Cardiovasc Res Center Bull
1967; 6:77–78
11 Louria DB, Hensle T, Rose J. The major medical complications of heroin addiction. Ann Intern Med 1967; 67:1–22
12 Cherubin CE. The medical sequelae of narcotic addiction.
Ann Intern Med 1067; 67:23–33
13 Stern WZ, Spear PW, Jacobson HG. The roentgen findings in
acute heroin intoxication. AJR Am J Roentgenol 1968; 103:
522–532
14 Steinberg AD, Karliner JS. The clinical spectrum of heroin
pulmonary edema. Arch Intern Med 1968; 122:122–127
15 Karliner JS, Steinberg AD, Williams MH Jr. Lung function
after pulmonary edema associated with heroin overdose. Arch
Intern Med 1969; 124:350 –353
16 Jaffe RB, Koschmann EB. Intravenous drug abuse: pulmonary, cardiac, and vascular complications. Am J Roentgenol
Radium Ther Nucl Med 1970; 109:107–120
17 Morrison WJ, Wetherill S, Zyroff J. The acute pulmonary
edema of heroin intoxication. Radiology 1970; 97:347–351
18 Duberstein JL, Kaufman DM. A clinical study of an epidemic
of heroin intoxication and heroin-induced pulmonary edema.
Am J Med 1971; 51:704 –714
19 Frand UI, Shim CS, Williams MH Jr. Heroin-induced pulmonary edema: sequential studies of pulmonary function.
Ann Intern Med 1972; 77:29 –35
20 Addington WW, Cugell DW, Basely ES. The pulmonary
edema of heroin toxicity: an example of the stiff lung syndrome. Chest 1972; 62:199 –205
21 Light RW, Dunham TR. Severe slowly resolving heroininduced pulmonary edema. Chest 1975; 67:61– 64
22 Paranthaman SK, Khan F. Acute cardiomyopathy with recurrent pulmonary edema and hypotension following heroin
overdosage. Chest 1976; 69:117–119
23 Banner AS, Muthuswamy P, Shah RS, et al. Bronchiectasis
following heroin-induced pulmonary edema: rapid clearing of
pulmonary infiltrates. Chest 1976; 69:552–555
24 Smith WR, Glauser FL, Dearden LC, et al. Deposits of
immunoglobulin and complement in the pulmonary tissue of
patients with “heroin lung.” Chest 1978; 73:471– 476
25 Leechawengwong M, Berger HW, Jayamanne DS. Longterm serial follow-up after two episodes of heroin-induced
adult respiratory distress syndrome. Mt Sinai J Med 1979;
46:119 –121
26 Wang M, Liaw S, Bullard MJ. Heroin lung: report of two
cases. J Formosa Med Assoc 1994; 93:170 –172
27 Raijmakers PG, Groeneveld AB, de Groot MC, et al. Delayed
resolution of pulmonary oedema after cocaine/heroin abuse.
Thorax 1994; 49:1038 –1040
28 Otto D, Franke M, Madden JF. Hypothermia accompanied
by noncardiogenic pulmonary edema: a case report. Del Med
J 1996; 68:499 –502
29 Sporer KA. Acute heroin overdose. Ann Intern Med 1999;
130:584 –590
30 Kjeldgaard JM, Hahn GW, Heckenlively JR, et al. Methadone-induced pulmonary edema. JAMA 1971; 218:882– 883
31 Bogartz LJ, Miller WC. Pulmonary edema associated with
propoxyphene intoxication. JAMA 1971; 215:259 –262
32 Sklar J, Timms RM. Codeine-induced pulmonary edema.
Chest 1977; 72:230 –231
33 Gould DB. Buprenorphine causes pulmonary edema just like all
other mu-opioid narcotics [letter]. Chest 1995; 107:1478–1479
34 Thammakumpee G, Sumpatanukule P. Noncardiogenic pulmonary edema induced by sublingual buprenorphine. Chest
1994; 106:306 –308
35 Stadnyk A, Grossman RF. Nalbuphine-induced pulmonary
edema. Chest 1986; 90:773–774
36 Schwartz JA, Koenigsberg MD. Naloxone-induced pulmonary edema. Ann Emerg Med 1987; 16:1294 –1296
37 Henderson CA, Reynolds JE. Acute pulmonary edema in a
young male after intravenous nalmefene. Anesth Analg 1997;
84:218 –219
38 Sporer KA, Firestone J, Isaacs SM. Out-of-hospital treatment
of opioid overdoses in an urban setting. Acad Emerg Med
1996; 3:660 – 667
39 Stern WZ, Subbarao K. Pulmonary complications of drug
addiction. Semin Roentgenol 1983; 18:183–197
40 Smith DA, Leake L, Loflin JR, et al. Is admission after
intravenous heroin overdose necessary? Ann Emerg Med
1992; 21:1326 –1330
41 Bertini G, Russo L, Cricelli F, et al. Role of a prehospital
medical system in reducing heroin-related deaths. Crit Care
Med 1992; 20:493– 498
42 Osterwalder JJ. Patients intoxicated with heroin or heroin
mixtures: how long should they be monitored? Eur J Emerg
Med 1995; 2:97–101
43 Christenson J, Etherington J, Grafstein E, et al. Early discharge of patients with presumed opioid overdose: development of a clinical prediction rule. Acad Emerg Med 2000;
7:1110 –1118
44 Goldfrank L, Weisman R. Opioids. In: Goldfrank LR,
Flomenbaum NE, Lewin N, et al, eds. Goldfrank’s toxicologic
emergencies. 6th ed. Norwalk, CT: Appleton & Lange, 1994;
769 –783
45 Karliner J. Noncardiogenic forms of pulmonary edema. Circulation 1972; 156:212–215
46 Gopinathan K, Saroja D, Spears R, et al. Hemodynamic
studies in heroin-induced acute pulmonary edema [abstract].
Circulation 1970; III-44:61– 63
47 Katz S, Aberman A, Frand UI, et al. Heroin pulmonary
edema: evidence for increased pulmonary capillary permeCHEST / 120 / 5 / NOVEMBER, 2001
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21969/ on 05/15/2017
1631
ability. Am Rev Respir Dis 1972; 106:472– 474
48 Hakim TS, Grunstein MM, Michel RP. Opiate action in the
pulmonary circulation. Pulm Pharmacol 1992; 5:159 –165
49 Pietra GG, Szidon JP, Leventhal MM, et al. Histamine and
interstitial pulmonary edema in the dog. Circ Res 1971;
29:323–337
50 Brashear RE, Kelly MT, White AC. Elevated plasma histamine after heroin and morphine. J Lab Clin Med 1974;
83:451– 457
51 Schurig J, Cavanagh R, Buyniski J. Effect of butorphanol and
morphine on pulmonary mechanics, arterial blood pressure
and venous plasma histamine in the anesthetized dog. Arch
Int Pharmacodyn Ther 1978; 233:296
52 Philbin DM, Moss J, Akins CW, et al. The use of H1 and H2
histamine antagonists with morphine anesthesia: a doubleblind study. Anesthesiology 1981; 55:292–296
53 Fahmy NR. Hemodynamics, plasma histamine, and catecholamine concentrations during an anaphylactoid reaction to
morphine. Anesthesiology 1981; 55:329 –331
1632
54 Rosow CE, Moss J, Philbin DM, et al. Histamine release
during morphine and fentanyl anesthesia. Anesthesiology
1982; 56:93–96
55 Moss J, Rosow CE. Histamine release by narcotics and muscle
relaxants in humans. Anesthesiology 1983; 59:330–339
56 Barke KE. Opiates, mast cells, and histamine release. Life Sci
1993; 53:1391–1399
57 Withington DE, Patrick JA, Reynolds F. Histamine release by
morphine and diamorphine in man. Anesthesia 1993; 48:
26 –29
58 Brigham KL, Owen PJ. Increased sheep lung vascular permeability caused by histamine. Circ Res 1975; 37:647– 657
59 Pietra GG, Magno M, Johns L. Morphological and physiological study of the effect of histamine on the isolated perfused
rabbit lung. Lymphology 1979; 12:165–176
60 Edston E, van Hage-Hamsten M. Anaphylactoid shock: a
common cause of death in heroin addicts? Allergy 1997;
52:950 –954
Clinical Investigations in Critical Care
Downloaded From: http://journal.publications.chestnet.org/pdfaccess.ashx?url=/data/journals/chest/21969/ on 05/15/2017