Journal Logo

Axioms in EM

Admit All Febrile IV Drug Users

Kim, Kyu MD

Author Information

    The febrile patient who is a current IV drug user presents to the ED with a diagnostic dilemma. His fever may be the result of apparent or occult bacterial, viral, or fungal infection, febrile reactions to injected materials, or even fever of unknown origin.

    IV drug users are at high risk for numerous serious infections because of their often unpredictable immune status, poor psychosocial conditions, and often less than optimal hygiene and nutritional status. One of the most life-threatening causes of infection is infective endocarditis.

    Several studies demonstrate the physician's inability to predict final diagnosis, especially endocarditis, in febrile intravenous drug users based on history, physical exam, and initially available laboratory studies. It is axiomatic to admit all febrile IV drug users because serious infection, especially endocarditis, cannot be excluded in the initial ED evaluation.

    In fact, the provisional admission diagnosis in febrile IV drug users is often incorrect. For example, a prospective evaluation of clinical and laboratory data in 75 febrile intravenous drug users determined the final diagnosis was endocarditis in 12.6 percent of patients. The sensitivity and specificity of and ED physician's prediction that a patient had endocarditis were only 57% and 76%, respectively.

    This may be because the classic signs and symptoms of endocarditis (chills, pleuritic chest pain, back pain, heart murmur, splenomagaly, and peripheral embolic phenomena) were seen less often than expected. Although patients with endocarditis had a slightly higher median temperature and lower median serum levels of sodium and potassium, there were no significant differences in leukocyte counts, hematocrit, erythrocyte sedimentation rate, and chest x-ray between patients with and without endocarditis.

    Two of seven patients diagnosed in the emergency department as having trivial illness had very serious illnesses, endocarditis in one and septic staphylococcal arthritis in the other. The study concluded that discharging febrile IV drug users with suspected viral syndrome or other trivial illness causes potential risk to individual patients.1

    Scoring System

    Even a predictive scoring system developed retrospectively for endocarditis could not be validated in symptomatic IV drug users presenting to the ED.2 Of 125 hospitalized febrile IV drug users in the development data set during the study period, 57 (46.3%) patients had a final diagnosis of endocarditis. Significant differences between the patients with and without endocarditis in the development set were noted — past history of endocarditis, temperature, total white blood cell count, percentage of neutrophils and bands on the WBC differential count (“left shift), infiltrate on chest x-ray, and arterial hypoxemia.

    Multivariable logistic regression analysis revealed a past history of endocarditis (negative correlation), total WBC count, and the presence of an infiltrate on the chest x-ray, and arterial oxygenation (negative correlation) as having independent correlation with a current episode of endocarditis. The endocarditis score was developed based on likelihood ratios derived from a development set using these variables.

    The investigators were unable to transport the score to a similar group of admitted IV drug users who subsequently came to the ED. Using presenting clinical features, they were unable to differentiate between symptomatic IV drug users with and without endocarditis. The authors conclude that presenting clinical, radiologic, and laboratory features are not helpful in estimating the clinical likelihood of endocarditis in symptomatic IV drug users.2

    To determine the frequency and identify predictive factors of occult major illness in febrile intravenous drug users presenting to the ED, a prospective follow up study examined 296 IV drug users with a history of IV drug use within the past six months. Of the 286 presentations evaluated, 180 (64%) had a major illness at presentation, (defined as a disease requiring hospital admission), and 103 (36%) had no obvious major illness at presentation.

    Of the 103 without apparent major illness at presentation, the final diagnosis in 11 were infective endocarditis (7), pneumonia (2), disseminated intravascular coagulation (1), and DVT (1). The study identified three univariate predictors of occult major illness: last use of intravenous drugs less than five days, fever greater than 102°F, and proteinuria greater than trace. The combination of univariate predictors that yielded the highest sensitivity and specificity was last drug use less than five days and temperature. However, the sensitivity and specificity were a mere 64% and 77%, respectively. Physicians' estimates of the probability of bacteremia and endocarditis obtained in 204 (70%) of the patients showed a substantial inaccuracy and was too insensitive to base hospitalization decisions.

    Bacteremia was present on seven of 11 (64%) patients who presented with apparent trivial illness but were later diagnosed with major illness. The authors suggest that because outpatient follow-up is unreliable in this population and waiting 24 to 48 hours for culture results before instituting therapy poses significant risk to this population in which the presence of occult bacteremia indicates presence of serious illness, hospitalization of all febrile IV drug users is recommended.3

    Bacteremia Found

    Another prospective study of 121 febrile IV drug users found that 51 (42%) had bacteremia. Of these, 16 were found to have infective endocarditis, 21 had other apparent sites of infection, and 14 had no specific site of infection. This study showed that even febrile IV drug users with an apparent and identifiable source of fever have high incidence of concurrent bacteremia. Even though the obvious infection was being treated, further complications may ensue that can not be identified on an outpatient basis. In infective endocarditis, murmurs due to valve destruction are often not appreciated until late in the disease process.

    Peripheral vascular changes such as Janeway lesions, petechial and splinter hemorrhages, and Osler nodes occur too infrequently and too late to be useful in the majority of cases especially in acute right side endocarditis as demonstrated in this and other studies. The authors of this study also recommend admission of all febrile IV drug users.4

    In injection drug users, infective endocarditis of the right side heart valves affecting the tricuspid valve occurs commonly as compared with endocarditis of the left side of heart affecting mitral or aortic valves in general population. Up to 76 percent of cases of endocarditis among IV drug users occur on the right side, compared with nine percent in nonusers. Although the mortality rate associated with right side endocarditis is lower than that associated with left side endocarditis, cardiopulmonary, neurologic, renal, ophthalmologic, and abdominal and extremity vascular complications can cause significant morbidity.

    The higher incidence of left side endocarditis in general population is related to the three factors: The higher pressure on the left side of the heart creating turbulent flow leading to endothelial damage; the higher oxygen content of the left side circulation supportive of bacterial growth; and congenital and acquired lesions of the heart valves epidemiologically more common on the left side.

    Although there is no single hypothesis that adequately explains the increased frequency of right side endocarditis among IV drug users, proposed explanations include damage to right side cardiac valves as a result of repeated bombardment by particulate matter; right side vasospasm, intimal damage, and thrombus formation induced by injected drug; drug-induced pulmonary hypertension and increased right side cardiac turbulence; increased right side expression of matrix molecules capable of binding microorganisms in intravenous drug users; directly injected large bacterial loads associated with increased likelihood of right side infection, and intravenous drug-related immune dysregulation.

    A dynamic combination of host, environmental, immunologic, and microbial factors cause predominantly right side endocarditis in this population. Again, because of the high incidence of right side endocarditis in injection drug users, the classic signs of left side endocarditis more common in the general population may not be present in the febrile IV drug users.5

    Illicit Drug Use

    How common is the problem? In 1999, approximately 14.8 million Americans were current users of illicit drugs; they reported the use of an illicit drug at least once during the month prior to being interviewed for the latest National Household Survey on Drug Abuse, an annual nationwide survey among Americans age 12 and older. About 3.5 million were dependent on illicit drugs. In 2000, there were an estimated 243 drug abuse-related ED visits per 100,000 population.

    Cocaine, heroin/morphine, marijuana, and methamphetamine were the most frequently used illicit drugs reported. Among those four, the use of cocaine and heroin constituted the major portion of illicit drug use. In 1995, the total economic cost was $97.7 billion. The White House Office of National Drug Control Policy found that between 1988 and 1995, Americans spent $57.3 billion on drugs: $38 billion on cocaine, $9.6 billion on heroin, $7 billion on marijuana, and $2.7 billion on other illegal drugs and the misuse of legal drugs. With the high prevalence of drug use, the emergency physician must always ask about past and present drug use in all patients presenting to emergency department.6

    Infective endocarditis is one of the most serious infectious illnesses in IV drug users. However, fever may represent a variety of other infectious processes that cause significant morbidity and mortality in this population. Any history of intravenous drug use should prompt the consideration of the following infectious entities. Hepatitis, specifically hepatitis B, is the most common. It is estimated that more than 80 percent of IV drug abusers have had hepatitis and that 80 percent of these cases are due to hepatitis B.

    Clinicians also should consider skin and soft tissue infections due to frequent injection, nonsterile technique, sharing equipment, and skin-popping. Cellulitis, skin abscesses, thrombophlebitis, necrotizing fasciitis, gas gangrene, and pyomyositis have been reported. Tetanus caused by exotoxin produced by Clostridium tetani is likewise a possibility. Increased incidence of tetanus among IV drug users has been reported likely from cutting heroin with quinine, which causes a favorable condition for growth of C. tetani.

    Skeletal infections including septic arthritis and osteomyelitis are the third most common complication of IV drug use. These infections may occur as a result of hematogenous spread from a distant site, most commonly endocarditis, or an overlying skin or soft tissue infection. Pulmonary infections occurs with increased incidence of pneumonia in intravenous drug users due to decreased immune function, increased aspiration, and septic pulmonary emboli as a result of endocarditis or septic thrombophlebitis.

    Ocular infections including fungal and bacterial endophthalmitis can be a primary infection or a result of secondary seeding. Also possible are nervous system infections including meningitis, epidural abscess, and brain abscess as extension of local osteomyelitis or embolic complications. And malaria was transmitted among IV drug users early in this century in the United States. This is now an uncommon infectious complication. The decline may be due to the frequent use of quinine as an adulterant.

    HIV Infection

    Complicating the infectious picture in IV drug users is the prevalence of HIV infection in this population. Injection drug use is the second most frequently reported risk behavior for infection with HIV. As of December 1995, of 513,486 cases of acquired immunodeficiency syndrome were reported to the Centers for Disease Control and Prevention, and 184,359 (36%) were directly or indirectly associated with injection drug use. Of the 184,359, 161,891 were injection drug users. Current estimates for HIV positive status in the IV drug users range from 10 to 70 percent. Selwyn et al demonstrated that the presence of serious bacterial infection in the HIV-positive IV drug users were important as predictors of progression of AIDS and as a source of HIV-related morbidity and mortality. It is important to consider IV drug users to be positive for HIV until proven otherwise.7,8

    Initial diagnostic studies for febrile IV drug users in addition to standard history and physical examination include chest radiography, at least two sets of blood cultures obtained from two different sites for aerobic, anaerobic, and fungal organisms, hemoglobin, hematocrit, RBC count, WBC count, and differential. Urinalysis, urine culture, and sputum culture should be obtained when indications of infection are present. Following blood cultures and other indicated cultures, broad spectrum antibiotic coverage should be initiated because of the importance of early treatment of infective endocarditis.

    During these patients' hospital stays, HIV status should be determined if unknown. Echocardiography should be obtained in patients with positive blood culture for diagnosis of infective endocarditis. If positive, these patients should be treated for endocarditis. However, if initial echocardiography is negative, repeat echocardiogram is recommended in one week. In patients with negative blood cultures, echocardiography should be performed if they demonstrate pulmonary or systemic embolism or if they have cardiac complications. If patients with negative blood cultures show no complications, they can be managed clinically.9

    Management of febrile IV drug users continues to be a challenge to the emergency physician. Infection in this population can be apparent or occult, single or mixed. In fact, patients with multiple sites of infection (such as pneumonia and meningitis) should have infective endocarditis considered in their differential diagnosis. Intrinsic impairment of the immune response limits their ability to respond to infectious agents.

    This impairment also masks the classic symptoms and signs of the disease, making diagnosis difficult. The initial ED impression is notorious inaccurate and unreliable for final diagnosis. IV drug users have limited access to medical care and are unreliable for follow-up. Delayed diagnosis and treatment can lead to significant morbidity and mortality in this population. Therefore, it is axiomatic to admit all febrile IV drug users.

    Possible Causes of Fever in IV Drug Users

    • ▪ Hepatitis, specifically hepatitis B
    • ▪ Skin and soft tissue infections
    • ▪ Tetanus caused by exotoxin produced by Clostridium tetani.
    • ▪ Skeletal infections including septic arthritis and osteomyelitis
    • ▪ Pulmonary infections
    • ▪ Ocular infections including fungal and bacterial endophthalmitis
    • ▪ Nervous system infections including meningitis, epidural abscess, and brain abscess
    • ▪ Malaria (uncommon)


    1. Marantz PR, et al. Inability to predict diagnosis in febrile intravenous drug users. Ann Intern Med 1987;106:823.
    2. Young GP, et al. Inability to validate a predictive score for infective endocarditis in intravenous drug users. J Emerg Med 1993;11:1.
    3. Samet JH, et al. Hospitalization decision in febrile intravenous drug users. Am J Med 1990;89:53.
    4. Weisse AB, et al. The febrile parenteral drug user: A prospective study in 121 patients. Am J Med 1993;94:274.
    5. Fontera JA, et al: Right-side endocarditis in injection drug users: Review of proposed mechanisms of pathogenesis. Clin Infect Dis 2000;30:374.
    6. The Lewin Group Report on Cost of Drug Abuse in the U.S. The National Institute on Drug Abuse and the National Institute on Alcohol Abuse and Alcoholism;.
    7. AIDS Associated with Injecting-Drug Use — United States, 1995. MMWR 1996;45(19):392.
    8. Selwyn PA, et al. Clinical manifestations and predictors of disease progression in drug users with human immunodeficiency virus infection. New Engl J Med 1992;327:1697.
    9. Shepherd SM, et al. Other infectious complications in intravenous drug users. Emerg Med Clinics of NA 1990;8(3)683.
    © 2002 by Lippincott Williams & Wilkins