The trusted source for
healthcare information and
Authors: Brad Frazee, MD, FACEP, Department of Emergency Medicine, Alameda County Medical Center, Assistant Clinical Professor, Department of Medicine, University of California, San Francisco; and Ralph Wang, MD, Department of Emergency Medicine, Alameda County Medical Center, Oakland, CA.
Peer Reviewers: Robert Powers, MD, MPH, FACP, FACEP, Chief and Professor, Emergency Medicine, University of Connecticut School of Medicine, Farmington; and Steven M. Winograd, MD, FACEP, Attending Physician, Emergency Department, The Uniontown Hospital, Uniontown, PA.
Users of intravenous drugs are at risk for complications related to the agents they are injecting as well as other conditions related primarily to transmission of infectious agents. At some urban institutions, especially in the New York and San Francisco metropolitan areas, complications associated with intravenous drug injection are among the most common entities managed in the emergency department (ED) setting.
As experienced emergency physicians (EPs) know, infections associated with intravenous injection encompass a wide range of conditions, from viral hepatitis and bacterial endocarditis to wound botulism and spinal epidural abscess. Frequently, benign-appearing conditions such as soft-tissue cellulitis may be a manifestation of more serious underlying conditions that are systemic in nature. Compromised host response observed in patients with HIV infection may predispose them to opportunistic infections. Many of these conditions not only are life-threatening to the individual patient, but if carriers are not recognized, transmission to other individuals may be accelerated.
The evaluation of fever is especially problematic, since it may be difficult to determine whether the fever is a manifestation of a circumscribed local infection, or whether the patient has a serious underlying infection such as pneumonia or endocarditis. In the final analysis, optimal treatment of this patient population requires a systematic evaluation based on physical, historical, laboratory, radiologic, and bacteriological findings to confirm a suspected diagnosis or exclude it. The authors of this review, who have experience managing this patient population on a day-to-day basis, present an evidence-based strategy for assessment and management of individuals who are intravenous drug users.—The Editor
It is estimated that 3 million Americans have used heroin in their lifetimes, and that there were 400,000 active users in the United States in the year 2000. There are about half as many cocaine and methamphetamine injection drug users.1 Because most heroin users inject the drug, and because EDs serve as a regular source of medical care for this patient population, many EPs will encounter patients who engage in injection drug use (IDU). At some urban hospitals, as many as 10% of admissions are related to IDU.2
Injection drug users are at risk for a set of unique and well-defined problems, both social and medical in nature: addiction and related crime; overdose; withdrawal; and infections directly resulting from drug injection. Among these problems, infections account for the most ED visits and hospitalizations.1,3,4 In fact, infections related to IDU were reported to be the most frequent admitting diagnosis from the San Francisco General Hospital ED during the late 1990s.5
The list of infections resulting from IDU reported in the literature virtually spans the entire spectrum of infectious disease— from viral infections like hepatitis C that are ubiquitous in this population, to bizarre bacterial infections such as pyo-pneumothorax,6 to malaria. Diagnosis of the infections commonly encountered in the ED often is challenging. Soft-tissue infections encompass not only the obvious subcutaneous abscess and simple cellulitis, but also occult necrotizing fasciitis, which may be fatal if not diagnosed and treated rapidly. Wound botulism related to IDU easily is misdiagnosed, and failure to initiate specific therapy can lead to respiratory failure. Similarly, delayed diagnosis of spinal epidural abscess —a notorious complication of IDU that may present simply as back pain—can result in irreversible paralysis. Underlying HIV and associated opportunistic infections must be considered. The issue of appropriate disposition of injection drug users who are febrile often is surprisingly difficult because even an exhaustive ED workup cannot exclude the possibility of bacteremia and endocarditis.
The following article discusses the major bacterial infections associated with IDU. The contents can be divided roughly into three sections: 1) endocarditis and the approach to fever without an obvious source; 2) soft-tissue and skeletal infections; and 3) infections due to neurotoxins such as botulism and tetanus. The emphasis is on occult presentation of serious disease, strategies for accurate and timely diagnosis, and important treatments that should be initiated in the ED.
Injection Drug Use with Fever
During the past three decades, the terminology used to indicate injection drug use (IDU) has changed. "Intravenous drug abuse" (IVDA) and "intravenous drug use" (IVDU) largely have been discarded for the more accurate, inclusive, and less pejorative IDU.
When faced with a febrile IDU patient, the EP should proceed with a strong suspicion that an identifiable, serious infection exists. Approximately 60% of injection drug users who present to the ED with a fever have an apparent serious infection that requires admission.7 In studies that examined source of fever in injection drug users, soft-tissue infections and bacterial pneumonia were the leading etiologies.7,8 Infectious endocarditis (IE) was found in only 5-10% of patients, and in about half of the cases, it was evident upon initial ED evaluation that a serious infection requiring admission existed. Other common infections included pyelonephritis, pelvic inflammatory disease, HIV-related opportunistic infections, and skeletal infections, all of which can be diagnosed in the ED with a careful workup. Hence, the source of fever should be sought aggressively during the history and physical exam. The exam should not focus simply on whether there is a murmur or other stigmata of endocarditis. Patients should be undressed and the skin carefully examined for signs of a soft-tissue infection. Any musculoskeletal or neurologic complaint should be viewed as a possible sign of infection. Likewise, findings such as decreased range of motion, weakness, or numbness may point to the source of infection. The chest x-ray has a high yield in this patient population because pneumonia and HIV-related diseases like tuberculosis are common, and because endocarditis frequently produces chest x-ray abnormalities. Occasionally, more aggressive diagnostic testing, such as magnetic resonance imaging (MRI) for spinal epidural abscess, is required to make the correct diagnosis in the ED.
Roughly 40% of febrile IDU patients who present to the ED have no apparent source of fever and seem well enough to be discharged.7 Unfortunately, a significant proportion of these patients harbor an occult serious infection, particularly IE. The issue of proper disposition of febrile but well-appearing IDU patients with no apparent source of fever has been examined in three prospective studies involving a total of approximately 500 patients. Six to 13% of such patients eventually were diagnosed with IE. Furthermore, these studies consistently found that the subgroup with occult bacteremia could not be separated initially from those with, for example, a viral illness or a pyrogen reaction to injected material.7-9 This finding, together with the notion that injection drug users, as a group, are prone to becoming lost to follow-up, forms the basis for the recommendation that all febrile injection drug users without an obvious source of fever, even if they appear well, should have two blood cultures drawn and be admitted to the hospital until cultures remain negative for 48 hours.7-9 This approach remains the standard of practice at urban teaching hospitals that serve large IDU populations. Some authors have allowed that febrile but well-appearing injection drug users who can be reached by phone and seem reliable could be discharged after blood cultures are drawn.8 A system would need to be in place to follow the blood culture results and contact the patient immediately if they become positive.
IE, defined as fungal or bacterial infection of the heart valves and perivalvular tissue, is a well-recognized and notorious complication of IDU. In large case series of IE, the proportion of patients who injected drugs ranged from 4% to as high as 67%.10 Yet even among active injection drug users, IE is fairly uncommon, with an incidence of 1-20 per 10,000 users per year,11 accounting for 5-15% of hospitalizations for an IDU-related infection.2 Patients who are HIV positive or who inject cocaine are at an increased risk for IE.12,13 In non-injection drug users, IE occurs almost exclusively in patients with underlying valvular pathology, such as congenital heart malformations, rheumatic heart disease, and prosthetic valves. In contrast, such underlying valvular pathology is present in only 10-26% of injection drug users who develop IE.2,14 The pathophysiology leading to the majority of IE cases in injection drug users is not well understood. Injected material may cause subtle endocardial damage and turbulent flow in the absence of frank valvular pathology. The bacteremia associated with IDU appears to result largely from introduction of skin flora more often than from contaminated drugs or syringes.15
Staphylococcus aureus causes 51-82% of cases of IE in injection drug users in contrast to non-IDU-related cases, where Streptococcus viridans is the predominant pathogen.2,11,16-19 The rate of methicillin-resistant S. aureus varies from approximately 10% to 26%, depending on locale, with the highest rates reported in the upper Midwest and northeastern United States.20 Other organisms found in IDU-related IE include Streptococcal species, enterococcus, and enteric Gram-negatives.17 Pseudomonas, a frequent pathogen in some early case series, was found in only 5-10% of cases in contemporary studies.11 Culture-negative IE, usually caused by the Haemophilus species, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corro-dens, and Kingella kingae (HACEK) organisms, typically accounts for 5-10% of cases.21 Use of antibiotics prior to presentation also is associated with culture-negative endocarditis.7
Endocarditis is classified on the basis of location of the vegetation, route of infection, etiologic organism, and whether the valve is native or prosthetic. (See Table 1.) The pattern of valvular involvement in IDU-related IE differs from that in non-injection drug users in that vegetations more commonly are located on the tricuspid valve. (See Table 1 and Table 2.) Right-sided endocarditis has a distinctive pathophysiology and clinical presentation. Whereas valvular regurgitation from left-sided endocarditis usually causes a significant murmur and may lead to pump failure, tricuspid regurgitation is of lesser hemodynamic consequence and may be silent both clinically and on auscultation. Left-sided IE classically is associated with prominent vascular and immune phenomena, such as splinter hemorrhages. Such findings usually are absent in isolated, right-sided disease. S. aureus endocarditis of the aortic or mitral valve is another fairly common pattern among injection drug users. In contrast to tricuspid disease, aortic valve vegetations are associated with a high morbidity and mortality, with up to 40% of cases resulting in death or heart failure that requires valve replacement. Septic embolization to the brain, coronary arteries, or kidneys can occur. IE presenting with altered mental status or heart failure is considered an ominous sign, indicating likely infection of the aortic valve. A study that underscores the differences between S. aureus IE in injection drug users and non-injection drug users found that the former group on average was younger, had a majority (76%) of right-sided heart involvement, and had a case mortality about one-tenth that of non-injection drug users.16
Endocarditis typically presents with nonspecific symptoms such as arthralgias, malaise, back pain, and weight loss. Pulmonary symptoms such as cough, dyspnea, and chest pain are more common in tricuspid disease. Fever is a cardinal sign of IE, present in 97-100% of cases.2,10,19,22 Classically, IE causes a new or changing murmur; however, a murmur is heard at the time of presentation in only 30-50% of IDU-related IE patients, and in even fewer of those with isolated right-sided disease.
Various diagnostic criteria have been developed to assist in arriving objectively at this difficult diagnosis. In the early 1990s, the Duke criteria were introduced, which emphasized echocardiography in addition to blood culture results and the presence of risk factors.23,24 The Duke criteria are summarized in Table 3.
Blood cultures are a cornerstone of diagnosis in IE. Endocarditis produces a continuous bacteremia, and blood cultures are positive in more than 95% of cases. In patients presenting to the ED with potential IE, it is the responsibility of the EP to ensure proper blood culture collection prior to the administration of antibiotics. Ideally, 2 or 3 separate cultures should be obtained, each containing 10 mL of blood. The recommendation that they be separated in time by as much as one hour may be difficult to adhere to in the ED setting. Factors associated with reduced blood culture yield include sample volume of less than 10 mL, recent antibiotic therapy, and prosthetic valve disease.15 In the setting of a prosthetic valve, it is recommended that three cultures be obtained to improve sensitivity.15
The chest x-ray is an essential diagnostic study in all febrile IDU patients for two reasons: 1) community-acquired pneumonia is a far more common cause of fever than IE; and 2) the chest x-ray is abnormal in up to 72% of injection drug users with IE, probably because of the high proportion of right-sided disease. Septic pulmonary emboli classically appear on chest x-ray as multiple round infiltrates that may show evidence of cavitation. Other findings on chest x-ray include nonspecific infiltrates, pleural effusions, and pulmonary edema. Laboratory abnormalities that support the diagnosis of IE include hematuria, proteinuria, and anemia.7,8 Conduction abnormalities, particularly AV block, may be seen on electrocardiogram (ECG) due to an associated valve ring abscess that erodes into the His-Purkinji system.
Echocardiography, now considered nearly as important as blood cultures in the workup of IE, is used both for definitive diagnosis and to assess complications and prognosis. Although by no means always present, the finding of an oscillating intra-cardiac mass attached to a valve is considered definitive. Other possible echocardiographic findings include valvular incompetence, reduced ventricular function, perivalvular abscess, or new dehiscence of a prosthetic valve. For the diagnosis of IE, transthoracic echo is 60-80% sensitive and approximately 90% specific.26,27 Transesophageal echocardiogram is better overall (sensitivity 85-95%, specificity ~95%), and it better visualizes the aortic valve and prosthetic valves. The tricuspid valve is not well visualized by either modality.26,27 Echocardiogram findings that place patients with IE into a higher risk category include: vegetations greater than 1 cm in size, left-sided valvular involvement, severe regurgitation, and ring abscess. Vegetations greater than 1 cm may indicate the need for surgery. In the setting of isolated tricuspid valve involvement, a vegetation greater than 1 cm represents a contraindication to short-course antibiotics.11,26
Appropriate antibiotic choice for suspected IE depends on whether it is classified as native valve non-IDU, IDU-associated endocarditis, or prosthetic valve endocarditis. Empiric treatment of suspected IE related to IDU should cover S. aureus and Streptococcal species. The regimen of choice is nafcillin plus genta-micin.28 (See Table 4.) Vancomycin is used in patients with proven penicillin allergy; however, vancomycin has been shown to be less effective than nafcillin in eradicating S. aureus. The addition of gentamicin shortens the time to negative blood cultures.18 In well-appearing patients, treatment can be delayed until culture results are known; however, empiric treatment should be begun in the ED for all ill-appearing patients and those in heart failure. A two-week course of nafcillin and gentamicin has been shown to effectively treat isolated tricuspid valve IE in IDU, provided the vegetation is small and there is no evidence of pulmonary emboli.29
Intracardiac complications include valvular destruction and incompetency, resultant heart failure, intracardiac abscess, purulent pericarditis, and septic embolization to the coronary arteries.11 Embolic complications also can be devastating, and are more frequent with left-sided S. aureus IE. Septic cerebral emboli may present with delirium or focal neurological findings. Meningitis may occur in association with IE and is thought to result from microemboli to the meningeal arteries. Some form of neurologic complication is found in up to 50% of left-sided IE cases.11 Spinal epidural abscess is a feared, albeit rare, complication of IE. Renal sequelae include glomerulonephritis as well as renal infarction. Right-sided IE causes downstream pulmonary complications such as septic pulmonary embolism, pneumonia, empyema, and pyopneumothorax.6,11
Mortality due to IDU-related IE ranges from 2% to 39%.10,16 Isolated right-sided disease has a mortality of less than 10%.11,18,19,28 The effect of HIV on the course of IDU-related IE has been examined in several studies. Overall, AIDS confers a 2.5-fold increase in case mortality, and mortality is inversely related to CD4 count.13,30-32
Although the actual incidence is unknown, subcutaneous abscess is clearly the most common bacterial complication of IDU.33-35 A recent study conducted in San Francisco reported a 32% prevalence of abscess among active injection drug users.36 At San Francisco General Hospital in 1998, IDU-related soft-tissue infection (abscess or cellulitis) was the most common admitting diagnosis from the ED. The practice of subcutaneous injection, or "skin popping," has been shown to increase the risk of abscess approximately five-fold.36,37 Skin popping also is associated with necrotizing soft-tissue infections, tetanus, and wound botulism. It is believed that injection of bacteria-laden material beneath the skin, a relatively anaerobic environment, creates a nidus where both aerobic and anaerobic organisms may proliferate. Skin popping is more common in female injection drug users and those who have used up superficial veins. Needle licking, another common practice, probably accounts for the high percentage of oral flora found in IDU-related abscesses. Other risk factors for abscess formation include the following: cocaine injection, lack of skin preparation, and use of dirty needles.33,37
Despite the fact that surgical drainage is all that usually is required for treatment, the bacteriology of IDU abscesses has been well elucidated in several studies,34,35,38 as outlined in Table 5. Skin flora and oral flora account for the majority of pathogens. S. aureus and Streptococcal species are the predominant aerobes. Eikenella, a gram-negative aerobe and notorious pathogen in human fight bites, was cultured from 25% of abscesses in one series.34 Peptostreptococcus and fusobacterium are the predominant anaerobes. Bacteroides fragilis, an antibiotic-resistant gut anaerobe, rarely was isolated from upper extremity abscesses.34
The diagnosis of IDU-related abscess usually is straightforward, with most patients complaining of a painful mass at a previous injection site. Erythema, fluctuance, and drainage are confirmatory physical findings, present in a majority of cases. However, abscesses that present without specific physical findings, particularly without fluctuance, easily are misdiagnosed as cellulitis. Often these are deep subcutaneous or intramuscular collections. In such cases, ED ultrasound can be indispensable in revealing the correct diagnosis. Using a high frequency transducer, abscess cavities are visualized as a hypoechoic mass with acoustic enhancement in the far field.39 An alternative to ultrasound is blind needle aspiration. Plain x-rays are recommended because they may reveal a subcutaneous needle, tissue gas suggestive of a necrotizing infection, or bony involvement.
Most of the case series describing IDU-related abscesses consist of patients requiring admission, and therefore the literature on this topic is skewed toward severe disease. Forty to 60% of these patients were febrile, and the incidence of bacteremia among febrile patients was 17-24%.33,34,40 Despite the selection bias, these data underscore the fact that abscesses alone can cause fever, and suggest that such cases are associated with an approximately 20% rate of bacteremia. In contrast, a study of 50 abscess cases (13 IDU-related) that excluded febrile patients drew blood cultures before and after incision and drainage, and in every case found no cases of bacteremia.41
Incision and drainage is the primary treatment for IDU-related abscess. In most cases, this can be performed in the ED. Achieving adequate anesthesia often is a challenging issue for the EP. Options include local anesthesia, a regional block, a parenteral narcotic combined with a short-acting sedative (conscious sedation), or deep procedural sedation with an agent such as methohexital or propofol. (See Table 6.) The best choice depends on the size, number, and location of abscesses, availability of intravenous (IV) access, and patient preference. For large abscesses, deep procedural sedation usually is the best solution. In patients who otherwise have no peripheral IV access, ultrasound-guided cannulation of the deep brachial or basilic vein offers a rapid and less invasive alternative to a central line.42
A generous incision should be made and the abscess cavity explored with a hemostat or other instrument. Exploration with a gloved finger is to be avoided, especially without a preoperative x-ray proving there is no subcutaneous needle. In the case of deep procedural sedation, preoperative injection of local anesthetic is unnecessary, but 10-20 cc of bupivacaine is injected into the surrounding tissue after drainage to provide local anesthesia when the patient awakens. To achieve hemostasis, the cavity is packed tightly for the first 24 hours. A four-inch gauze roll can be used to pack large abscesses. Patients are instructed to return at least once, at 24 hours, for wound check, dressing change, and to review wound care procedures.
The need for prophylactic antibiotics prior to incision and drainage to suppress bacteremia and prevent endocarditis is controversial. American Heart Association guidelines list abscess incision and drainage among the procedures for which preoperative antibiotics should be considered in patients with high-risk cardiac abnormalities. However, there is very little evidence to suggest that abscess incision and drainage causes bacteremia in the absence of fever.41 Pre-incision and drainage antibiotics can be limited to patients at highest risk for endocarditis: those with prosthetic valves, congenital cardiac abnormalities, or a prior history of endocarditis.
The decision about whether to treat patients with a course of antibiotics following incision and drainage of an uncomplicated abscess also deserves mention. Although conventional wisdom states that uncomplicated abscesses are cured with drainage alone, there are almost no research data to support this. In a case series of 133 patients with abscesses treated in an outpatient setting (a small proportion of whom were injection drug users), antibiotics were withheld in 74%, and all patients did well.38 Antibiotics should be given when there appears to be significant surrounding cellulitis and should be considered in patients with diabetes or HIV infection. If the abscess is accompanied by fever or if the patient appears ill, parenteral antibiotics and admission to the surgical service is indicated. Adequate staphyloccocal and streptococcal coverage can be achieved with a first-generation cephalosporin, dicloxacillin, or clindamycin. Gram-negative and anaerobic coverage, such as with amoxicillin/clavulanate or a quinolone, usually is unnecessary.33,43
When faced with an IDU-related abscess, the EP must be vigilant for associated complications and less obvious coexisting infections such as necrotizing soft-tissue infection, septic arthritis, osteomyelitis, and epidural abscess. Such complications were discovered in 15-19% of patients admitted for IDU-related abscess.34,44 Endocarditis also is a consideration. However, we do not routinely obtain blood cultures or otherwise pursue this diagnosis, provided the soft-tissue infection appears sufficient to explain the fever.
Necrotizing Soft-Tissue Infections
Since Joseph Jones first coined the term "hospital gangrene" after the Civil War, a number of terms have been proposed to describe necrotizing soft-tissue infections (NSTI): necrotizing fasciitis, pyomyositis, "gas gangrene," and "flesh-eating bacterial infection." These terms describe a life-threatening soft-tissue infection, often involving muscle and fascial planes, associated with systemic toxicity.46 A definitive diagnosis is made when friable, necrotic fascia associated with vascular thrombosis is found intra-operatively.
Well-known risk factors for NSTI include IDU, diabetes mellitus, and peripheral vascular disease.47-50 In most cases, an inciting infection or breach of the skin barrier can be identified, such as drug injection, trauma, perineal infection, or post-operative wound infection. In community-acquired cases of NSTI (as opposed to hospital-acquired or postoperative cases), by far the most prevalent risk factor and inciting cause is IDU. In two recent case series of predominantly community-acquired NSTI from northern California, 55-64% of patients with NSTI were active injection drug users.49,51,52 These studies describe a sharp increase in the incidence of NSTI between 1994 and 1997. The pathophysiology of NSTI classically involves compromised or frankly devitalized tissue from trauma, surgery, diabetes, or longstanding IDU that becomes infected by a synergistic combination of aerobes and anaerobes, particularly Clostridium perfringens. Virulent strains of group A strep, however, appear to be capable of causing NSTIs in normal hosts in the absence of an obvious inciting cause or compromised tissue. Bacteriology studies reveal that 60-85% of NSTIs are polymicrobial. Aerobes include S. aureus and group A beta-hemolytic streptococcus (GABHS); anaerobes include gas forming Clostridium species, peptostreptococcus, and other so-called oral anaerobes.51,53,54 Although arising from various etiologies, NSTIs can be viewed as a common final pathway in terms of pathophysiology, rapid progression, and need for rapid recognition and surgical treatment.
The timely diagnosis of NSTIs is critical, yet often difficult. Patients may present with a variety of nonspecific symptoms and signs that may lead to initial misdiagnosis. Pain, warmth, edema, and fever commonly are present but are obviously nonspecific. Findings more specific for NSTI include: pain out of proportion to skin findings; signs of skin necrosis such as bullae or blisters; or crepitus representing subcutaneous gas. In IDU-related NSTIs, tense circumferential edema of an extremity, often spreading onto the trunk, is highly characteristic. Such findings should be sought in any soft-tissue infection presenting to the ED, and their significance needs to be recognized. A recent prospective trial identified several objective criteria useful in differentiating necrotizing from non-necrotizing skin infections: white blood cell (WBC) count greater than 14,000; gas on plain soft-tissue x-ray; elevated blood urea nitrogen (BUN); and sodium greater than 135.55 Similarly, previous case series have reported that a WBC count greater than 20,000 is seen in about 50% of NSTIs. While a marked leukocytosis should raise concern for NSTI, its absence by no means excludes the diagnosis. Signs of shock or organ dysfunction are found at the time of presentation in only 0-40% of cases.49,51,55
Various forms of imaging have been evaluated for their ability to differentiate NSTI from less severe skin infections. Plain films demonstrate subcutaneous gas in a stippled pattern in 20-60% of cases.47,49,51,55 Computed tomography (CT) scan is likely somewhat more sensitive than plain film in demonstrating gas in the tissue, and may reveal unsuspected pockets of pus. The typical CT finding in NSTI is asymmetric thickening of deep fascia associated with gas. Of note, IV contrast should be avoided in patients with signs of shock or incipient renal dysfunction.38,56 MRI, although costly and more difficult to obtain, has proven useful in one study in differentiating NSTI from non-necrotizing infections.38,56 MRI should not delay definitive diagnosis and debridement in the operating room.
Various quasi-invasive bedside tests to identify NSTIs have been described. Rapid antigen testing for group A streptococcus has been used to identify NSTIs due to this organism.53 In a single report, bedside frozen section was touted to prevent delay in diagnosis, but this approach has not been widely adopted.57 Finally, bedside exploration and fascial inspection can be performed. The ability to pass a blunt instrument along fascial planes without resistance is considered diagnostic of necrotizing fasciitis. This is a realistic option only in intubated, heavily anaesthetized patients, such as in a surgical intensive care unit.
While a positive result from a confirmatory test is extremely useful in establishing the diagnosis of NSTI and need for immediate surgery, a negative result cannot be relied upon to exclude the diagnosis. Indeed, a false negative result tends to delay definitive treatment. Rather, the proper approach to diagnosis combines the following: an extremely high index of suspicion on the part of the EP, prompt surgical consultation when any suspicion of NSTI arises, and a low threshold for operative exploration on the part of the surgeon—analogous to the traditional approach to appendicitis. Studies looking at risk factors for mortality in NSTI consistently find that delay to operation is the lone modifiable risk factor.51,54,58 In one study, the mortality in the group of patients brought to the operating room 24 hours after presentation was quadruple that of patients undergoing early operation. Of particular importance to EPs was the finding that delay in surgery was associated with admission to a medical service and negative bedside aspiration.58
Initial management of NSTI includes aggressive fluid resuscitation and early broad-spectrum antibiotics. Laboratory and imaging studies should be performed promptly and early surgical consultation expedited. Systemically ill patients will not improve without debridement. If surgical consultants are unfamiliar with NSTI, transfer of the patient to another facility experienced in its management should be considered. Hyperbaric oxygen has been suggested as an adjunct to the treatment of NSTI and has been shown to offer the advantage of early wound closure.59 However, other studies have shown no benefit of hyperbaric oxygen in terms of mortality or number of debridements.46
Reports of mortality in NSTIs range widely from 6% to greater than 70%. In a recent review summarizing 660 cases from several case series, overall mortality was 26%.52 Comparison of IDU-related NSTI vs. non-IDU NSTI at one institution revealed a lower mortality among injection drug users (10% vs 21%), which may be explained by the younger age and lack of co-morbidities in injection drug users.49
Septic Arthritis and Osteomyelitis
Both septic arthritis (SA) and osteomyelitis (OM) may complicate IDU. The incidence of these infections has not been well established, but in one report they jointly accounted for 4% of IDU-related hospital admissions. SA appears to be more common, occurring five times as often as OM in one case series.60
Both SA and IDU-related OM result from hematogenous seeding during bacteremia. Bacteremia may occur transiently after drug injection or may be due to IE. Of 180 IDU-related bacteremias described in one report, 6% were associated with SA.2,61 In another series of 104 cases of IE, 15% of cases were complicated by SA or OM.61 In a series of 36 cases of SA, four were associated with IE.62 These findings underscore the need to consider IE in all cases of IDU-related SA or OM, and vice versa. Joint infections only rarely result from a contiguous spread of a cellulitis.
In early reports from Los Angeles totaling 180 IDU patients with bone and joint infections, 66-78% of cases were due to Pseudomonas aeruginosa.51,63,64 In more recent studies, S. aureus has been the predominant pathogen, isolated in 53-75% of cases whereas only 11% involved Pseudomonas.60,62 Eikenella corrodens also has been isolated in cases of OM.65
Bone and joint infections related to IDU typically affect the axial skeleton. In two case series of IDU-related SA, the sacro-iliac, costochondral, hip, and sternoclavicular joints were involved in 61-80% of cases. The remainder involved extremity joints, of which the knee was the most common.62,63 OM in IDU usually occurs in the spine. The lumbar region most commonly is affected, followed by the cervical spine.61 In one case series involving patients who frequently injected into the groin, lower extremity infections predominated, suggesting that both SA and OM may occur distal to the site of injection.60
SA usually presents as an acute infection, with rapid onset of pain, tenderness, and decreased range of motion. Fever occurs in 67-73% of patients; leukocytosis occurs in 50-61%.60,62 By contrast, OM often is indolent, with patients presenting as late as three months after the onset of infection.64 Back pain is the most common chief complaint in OM. In one case series of 67 patients, fever was seen in 42% of patients, transient neurological deficits in 15%, and leukocytosis in 35%.64
The workup of both SA and OM includes an erythrocyte sedimentation rate (ESR), blood cultures, and x-rays of the symptomatic area. The ESR, while non-specific, is elevated above 20 mm in greater than 90% of SA and OM cases.62,64 Blood cultures, although positive in only 20-30% of cases, are particularly important in OM because they may establish the etiology and obviate bone biopsy. As mentioned above, a positive blood culture in the setting of SA or OM generally mandates a search for IE.
Plain x-rays are insensitive for acute OM. There is a lag period of 10 days to three weeks or more before the lytic and demineralizing effects of infection become apparent radiographically.60,64 In one study, fewer than 5% of plain films were positive upon presentation, whereas radiographic signs of OM were present in 90% of cases after 3-4 weeks.66 Radionuclide scintigraphy, using technecium-, gallium-, or indium-tagged WBCs to visualize areas of infection, has a sensitivity of 50% for chronic infections and as high as 90% in acute infection.67 Compared with scintigraphy, MRI can better differentiate soft-tissue infection from OM. The sensitivity and specificity of MRI for OM range from 60-100% and 50-90%, respectively.67
Arthrocentesis remains the main diagnostic test for suspected SA. Synovial fluid WBC greater than 50,000 or positive Gram stain generally indicates SA. Synovial fluid culture reveals the etiology in approximately 75% of cases, provided that arthrocentesis is performed prior to antibiotics.62 In the case of a suspected septic hip, ultrasound guidance may permit successful arthrocentesis in the ED.
Treatment of both OM and SA involves IV antibiotics and immobilization. Suspected OM, and to a lesser degree SA, represents deep-seated infection requiring prolonged antibiotics, in which the results of culture may be critical. For this reason, antibiotics should be administered only after appropriate cultures are obtained, unless the patient is clinically unstable. Recommended empiric therapy for OM and SA is the combination of nafcillin and ciprofloxacin, which covers both Staphylococcus and Pseudomonas.61 In addition to antibiotic therapy, SA generally requires therapeutic arthrocentesis. The benefit of irrigating the joint with an antibiotic solution is unproven.63 The role of surgery in the treatment of OM remains unclear.
Spinal Epidural Abscess
Spinal epidural abscess (SEA) is a rare but feared complication of IDU. Difficulty arises from the facts that SEA may present simply as back pain, MRI generally is required for diagnosis, and diagnostic delay can be associated with sudden and irreversible neurological damage. The incidence of SEA in the general population is estimated at 0.2-1.0/10000 hospital admissions and is thought to be rising.69,70 While IDU is considered a major risk factor for SEA, no study specifically has addressed the incidence in this patient population.
Most SEAs are due to hematogenous seeding of the vertebrae or epidural space. Common sources of bacteremia include skin and soft-tissue infections, dental infections, and urinary tract infections, as well as IDU.68,69 Interestingly, in a series of 18 cases of IDU-related SEA, no co-existing endocarditis was found.69 S. aureus is the predominant pathogen in SEA, and is implicated in 50-65% of SEA cases overall and an even higher percentage of IDU-related cases.68,69,71 Pseudomonas and tuberculosis also have been isolated in IDU-related SEA. Epidural abscesses were associated with adjacent OM, diskitis, or psoas abscess in 14 of 18 cases, all involving IDU.69
There are two proposed mechanisms of cord injury in SEA, which may explain the wide variation in the time of onset of neurological deficits: direct compression and vascular ischemia. Compression due to mass effect is postulated to produce a subacute course, whereas thrombosis and vascular injury may result in a rapid paralysis.
While the classic presentation of SEA includes back pain, fever, and neurological symptoms, the presence of this triad is far from universal. Back pain is present in greater than 90% of patients, whereas fever is present in 60-76%, and neurological deficits are found at presentation in 57-70%.68,71 Neurologic findings include radicular pain, urinary incontinence, leg weakness, and paraplegia or quadraplegia. Patients occasionally may present with sepsis or encephalitis, further confounding and delaying the diagnosis.68
When faced with an IDU patient whose presentation is at all concerning for SEA, emergent imaging is required to exclude or establish the diagnosis. MRI now is the test of choice. With a sensitivity of 90%, it has come to replace the equally sensitive but more invasive CT-myelogram.68 In addition to revealing an abscess, MRI provides images of the cord itself, and may uncover an alternative diagnosis accounting for the patient’s symptoms.72 Plain x-rays cannot be relied upon to exclude the diagnosis of SAE, although they may be abnormal if there is associated vertebral OM. In such cases, end plate destruction or a narrowed disc space may be evident.
An ESR greater than 30 mm, while nonspecific, almost always is found in SEA. ESR also may be used to follow treatment efficacy. Some authors advocate ESR as a screening test for SEA in at-risk patients, such as injection drug users, who present with unexplained back pain.68,71 No prospective data exist to support this approach. Blood cultures are positive in greater than 60% of cases of SEA, and may provide an etiologic diagnosis if surgical cultures are unrevealing.68,69 Therefore, it is critical that blood cultures be obtained prior to beginning empiric antibiotics whenever the diagnosis of SEA is being entertained.
Treatment of SEA involves both surgical decompression and appropriate antibiotic therapy. While conservative, non-operative management has been described, surgical drainage remains the treatment of choice in most cases.68-70 Indications for surgery include neurological deficits, negative blood cultures, and failure of conservative therapy. Antibiotic therapy optimally is guided by the results of operative cultures or blood cultures. The combination of nafcillin and an anti-pseudomonal antibiotic such as ciprofloxacin is the recommended initial empiric regimen.73
Mortality from SEA is approximately 15%.70,74 In one series of IDU-related SEA, permanent neurological damage occurred in 25% of cases.75 The only modifiable determinant of neurological outcome is time from presentation to operative decompression. In one study, all patients with pre-operative neurological deficits for fewer than 36 hours showed some degree of recovery, whereas recovery occurred in only two of 11 in patients with greater than 36 hours of pre-operative neurological deficits.75
While tetanus now is a rare disease in the United States, injection drug users remain at significant risk for developing the disease. Of 124 cases reported in the United States between 1995 and 1997, 11% occurred in injection drug users. In California, 40% of cases were related to IDU.76,77 IDU-related tetanus in California tended to occur in Hispanics in whom under-vaccination against tetanus was documented in every case. The practice of skin popping also appears to be associated with the development of tetanus—of the 14 patients who were questioned about injection technique, all admitted to skin popping.76,77 EPs who regularly see injection drug users must not only remain vigilant for this uncommon disease, but must attend compulsively to the tetanus immunization status in these patients.
The pathophysiology of tetanus begins with the introduction of C. tetani spores into a wound. Similar to botulism, subcutaneous injection of heroin is thought to produce favorable conditions for the growth of C. tetani.78 Spores then germinate and begin to produce the exotoxin tetanospasmin. Tetanospasmin is transported in retrograde fashion to the spinal cord and into inhibitory neurons, where the toxin prevents release of neurotransmitter and ultimately results in muscle spasm.78,79
Presenting symptoms are due to muscle spasm, which may be localized—so-called cephalic tetanus—or generalized. Trismus and neck or back pain are the initial symptoms in 85% of cases.80 Other findings include opisthotonos, risus sardonicus, dysphagia, and drooling. Physical exam reveals palpable muscle rigidity and hyperreflexia.78-80
Complications of tetanus include respiratory failure, severe autonomic dysfunction, and rhabdomyolysis. Muscle spasm may involve the airway and respiratory musculature. Asphyxia, apnea, and pneumonia were common in both case series of IDU-related tetanus.78-80 Autonomic dysfunction, manifesting as tachycardia, blood pressure lability, and hyperpyrexia occurs in the majority of patients. Rhabdomyolysis should be anticipated in cases of severe, generalized muscle spasm.79,80 Mortality ranges from 70% in early reports to 25% in a recent case series of 18 patients.79
The diagnosis of tetanus is made on clinical grounds. Disorders that may present similar to generalized tetanus include the following: seizures, meningitis/encephalitis, drug withdrawal, sepsis, and strychnine poisoning. The differential diagnosis of localized tetanus includes peritonsillar abscess, mandibular disorders, and dystonic reactions.73,80,81 A history and physical consistent with tetanus in an injection drug user, combined with negative results on CT scan and lumbar puncture, is sufficient to establish the diagnosis. Wound cultures for C. tetani are insensitive.81 Antibody titers against tetanus toxoid should be ordered, although results will not be available to assist with ED diagnosis.
The treatment of tetanus primarily involves three strategies: 1) Elimination of all potential sources of toxin production is accomplished by drainage of skin abscesses and administration of antibiotics active against Clostridium. Metronidazole appears to be the antibiotic of choice;73 2) Clearance of extraneuronal tetanospasm toxin is achieved by administration of tetanus immunoglobulin (TIG). Intrathecal plus IV administration of TIG probably offers an advantage over the IV route alone.79 Passive immunization with tetanus toxoid also should be initiated immediately;76,79 3) Aggressive and scrupulous supportive care is critical to prevent associated morbidity and mortality. Patients with muscle spasm that is severe, generalized, or rapidly progressing should be intubated, deeply sedated, and paralyzed if necessary. Long-term ventilatory support and tracheostomy often are required.79,80 Muscle spasm is controlled with benzodiazepines or propofol. Labetolol, a combined beta- and alpha-adrenergic blocker, can be used to manage autononomic instability.79
Viral Seroprevalence and Transmission
IDU has been known for decades to be a major risk factor for a variety of parenterally transmitted viral infections. While a full discussion of IDU-associated viral infections is beyond the scope of this paper, providers who care for injection drug users need to be aware of the high prevalence of viral infections in these patients.
Practices such as needle and paraphernalia sharing readily transmit hepatitis C (HCV), hepatitis B (HBV), as well as HIV. Overall prevalence of HCV and HBV in IDU is 50-90% and 70%, respectively. (See Table 7.) There is a linear relationship between the number of years of IDU and seroprevalence of both HCV and HBV, with a 10-30% incidence of acquiring the infection per year of drug use.82 After 15 years, seroprevalence of HCV reaches 100%. Approximately 10% of HBV positive patients are HB surface antigen positive, and thus considered infectious.83,84
IDU is the second most frequently reported risk behavior leading to HIV infection. Approximately 25% of U.S. AIDS cases reported to the CDC are IDU-related, and as many as 10,000 injection drug users are believed to acquire HIV every year.32 Nevertheless, infection with HIV is far less common than the hepatitis viruses. HIV seroprevalence among injection drug users varies substantially by geographic location, from 1-3% in California to 23-27% in some northeastern cities such as Newark, NJ.84,85
Every EP should be able to recognize and manage the major bacterial infections that complicate IDU.
1. Schoener EP, Hopper JA, Pierre JD. Injection drug use in North America. Infect Dis Clin North Am 2002;16:535-551.
2. Levine DP, Lawrence RC, Zervos MJ. Bacteremia in narcotic addicts at the Detroit Medical Center. Infectious endocarditis: A prospective comparative study. Rev Infect Dis 1986;8:374-395.
3. Cherubin CE, Sapira JD. The medical complications of drug addiction and the medical assessment of the intravenous drug user: 25 years later. Ann Intern Med 1993:119:1017-1028.
4. Stein MD. Injected-drug use: Complications and costs in the care of hospitalized HIV-infected patients. J Acquir Immune Defic Syndr 1994;7:469-473.
5. Harris HW, Young DM. Care of injection drug users with soft tissue infections in San Francisco, California. Arch Surg 2002;137:1217.
6. Zorc TG, O’Donnell AE, Holt RW, et al. Bilateral pyopneumothorax secondary to intravenous drug abuse. Chest 1988;93:645-647.
7. Samet JH, Shevitz A, Fowle J, et al. Hospitalization decisions in febrile intravenous drug users. Am J Med 1990;89:53-57.
8. Marantz PR, Linzer M, Feiner CJ, et al. Inability to predict diagnosis in febrile intravenous drug abusers. Ann Intern Med 1987;106:823-828.
9. Weisse AB, Heller DR, Schimenti RJ, et al. The febrile parenteral drug user: A prospective study in 121 patients. Am J Med 1993;94:274-280.
10. Bouza E, Menasalvas A, Munoz P, et al. Infective endocarditis—a prospective study at the end of the twentieth century: New predisposing conditions, new etiologic agents, and still a high mortality. Medicine 2001;80:298-307.
11. Roberts R, Slovis CM. Endocarditis in intravenous drug abusers. Emerg Med Clin North Am 1990;8:665-681.
12. Chambers HF, Morris DL, Tauber MG, et al. Cocaine use and the risk for endocarditis in intravenous drug users. Ann Int Med 1987;106:833-836.
13. Spijkerman IJ. Human immunodeficiency virus and other risk factors for skin abscesses and endocarditis among injection drug users. J Clin Epidemiol 1996;49:1149-1154.
14. Levitt MA, Snoey ER, Tamkin GW, et al. Prevalence of cardiac valve abnormalities in afebrile injection drug users. Acad Emerg Med 1999;6:911-915.
15. Delaney KA. Endocarditis in the emergency department. Ann Emerg Med 1991;20:405-414.
16. Chambers HF, Korzeniowski OM, Sande MA. Staphylococcus aureus endocarditis: Clinical manifestations in addicts and nonaddicts. Medicine 1983;62:170-177.
17. Mathew J, Addai T, Anand A, et al. Clinical features, site of involvement, bacteriologic findings, and outcome of infective endocarditis in intravenous drug users. Arch Intern Med 1995;155:1641-1648.
18. Chambers HF, Miller RT, Newman MD. Right-sided Staphylococcus aureus endocarditis in intravenous drug abusers: Two-week combination therapy. Ann Intern Med 1988;109:619-624.
19. Hecht SR, Berger M. Right-sided endocarditis in intravenous drug users. Ann Intern Med 1992;117:560-566.
20. Crane LR, Levine DP, Zervos MJ, et al. Bacteremia in narcotic addicts at the Detroit Medical Center. I. Microbiology, epidemiology, risk factors, and empiric therapy. Rev Infect Dis 1986;8:364-373.
21. Brouqui P, Raoult D. Endocarditis due to rare and fastidious bacteria. Clin Microbiol Rev 2001;14:177-207.
22. Mathew J, Addai T, Anand A, et al. Clinical features, site of involvement, bacteriologic findings, and outcome of infective endocarditis in intravenous drug users. Arch Intern Med 1995;155:1641-1648.
23. Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective endocarditis: Utilization of specific echocardiographic findings. Am J Med 1994;96:200-209.
24. Bayer AS, Ward JI, Ginzton LE, et al. Evaluation of new clinical criteria for the diagnosis of infective endocarditis. Am J Med 1994;96:211-219.
25. Cecchi E, Parrini I, Chinaglia A, et al. New diagnostic criteria for infective endocarditis. A study of sensitivity and specificity. Eur Heart J 1997;18:1 149-1156.
26. Mortara LA, Bayer AS. Staphylococcus aureus bacteremia and endocarditis. New diagnostic and therapeutic concepts. Infect Dis Clin North Am 1993; 7:53-68.
27. Mugge A, Daniel WG. Echocardiographic assessment of vegetations in patients with infective endocarditis: prognostic implications. Echocardiography 1995;12:651-661.
28. DiNubile MJ. Short-course antibiotic therapy for right-sided endocarditis caused by Staphylococcus aureus in injection drug users. Ann Intern Med 1994;121:873-876.
29. Dinubile MJ. Surgery in active endocarditis. Ann Intern Med 1982;96: 650-659.
30. Cicalini S, Forcina G, De Rosa FG. Infective endocarditis in patients with human immunodeficiency virus infection. J Infect 2001;42:267-271.
31. Ribera E, Miro JM, Cortes E, et al. Influence of HIV 1 infection and degree of immunosuppression in the clinical characteristics and outcome of infective endocarditis in intravenous drug users. Arch Intern Med 1998;158: 2043-2050.
32. Cohn JA. HIV-1 infection in injection drug users. Infect Dis Clin North Am 2002;16:745-770.
33. Bergstein JM, Baker EJ 4th, Aprahamian C, et al. Soft tissue abscesses associated with parenteral drug abuse: Presentation, microbiology, and treatment. Am Surg 1995;12:1105-1108.
34. Orangio GR, Pitlick SD, Della Latta P, et al. Soft tissue infections in parenteral drug abusers. Ann Surg 1984;199:97-100.
35. Schnall SB, Holtom PD, Lilley JC. Abscesses secondary to parenteral abuse of drugs. A study of demographic and bacteriological characteristics. Am J Bone Joint Surg 1994;76:1526-1530.
36. Binswanger IA, Kral AH, Bluthenthal RN, et al. High prevalence of abscesses and cellulitis among community-recruited injection drug users in San Francisco. Clin Infect Dis 2000;30:579-581.
37. Murphy EL, DeVita D, Liu H, et al. Risk factors for skin and soft-tissue abscesses among injection drug users: A case-control study. Clin Infect Dis 2001;33:35-40.
38. Meislin HW, Lerner SA, Graves MH, et al. Cutaneous abscesses. Anaerobic and aerobic bacteriology and outpatient management. Ann Intern Med 1977;87:145-149.
39. Struk DW, Munk PL, Lee MJ, et al. Imaging of soft tissue infections. Radiologic Clin North Am 2001;39:277-303.
40. Gonzalez MH, Garst J, Nourbash P, et al. Abscesses of the upper extremity from drug abuse by injection. J Hand Surg 1993;18A:868-870.
41. Bobrow BJ, Pollack CV, Gamble S, et al. Incision and drainage of cutaneous abscess is not associated with bacteremia in afebrile adults. Ann Emerg Med 1997;29:404-408.
42. Keyes LE, Frazee BW, Snoey ER, et al. Ultrasound-guided brachial and basilic vein cannulation in emergency department patients with difficult intravenous access. Ann Emerg Med 2000;36:399.
43. Simmen HP, Giovanoli P, Battaglia H, et al. Soft tissue infections of the upper extremities with special consideration of abscesses in parenteral drug abusers. A prospective study. J Hand Surg 1995;20:797-800.
44. Henriksen BM, Albrektsen SB, Simper LB, et al. Soft tissue infections from drug abuse. A clinical and microbiological review of 145 cases. Acta Orthop Scand 1994;65:625-628.
45. Summanen PH, Talan DA, Strong C, et al. Bacteriology of skin and soft-tissue infections: Comparison of infections in intravenous drug users and individuals with no history of intravenous drug use. Clin Infect Dis 1995; 20(suppl 2): s279-s282.
46. Quirk W F, Sternbach G. Joseph Jones: Infection with flesh eating bacteria. J Emerg Med 1996;14:747-753.
47. Sudarsky LA, Laschinger JC, Coppa GF, et al. Improved results from a standardized approach in treating patients with necrotizing fasciitis. Ann Surg 1987;206:661-665.
48. Francis KR, Lamaute HR, Davis JM, et al. Implications of risk factors in necrotizing fasciitis. Am Surg 1993;59:304-308.
49. Chen JL, Fullerton KE, Flynn NM. Necrotizing fasciitis associated with injection drug use. Clin Infect Dis 2001;1:6-15.
50. Gonzalez MH, Kay T, Weinzweig N, et al. Necrotizing fasciitis of the upper extremity. J Hand Surg 1996;21:689-692.
51. Bosshardt TL, Henderson VJ, and Organ CH Jr. Necrotizing soft-tissue infections. Arch Surg 1996;131:846-854.
52. Callahan TE, Schecter WP, Horn JK. Necrotizing soft tissue infections masquerading as cutaneous abscess following illicit drug injection. Arch Surg 1998;133:812-818.
53. Ault MJ, Geiderman J, Sokolov R. Rapid identification of group A Streptococcus as the cause of necrotizing fasciitis. Ann Emerg Med 1996;28: 227-230.
54. McHenry CR, Piotrowski JJ, Petrinic D, et al. Determinants of mortality for necrotizing soft-tissue infections. Ann Surg 1995;221:558-565.
55. Wall DB, de Virgilio C, Black S, et al. Objective criteria may assist in distinguishing necrotizing fasciitis from non-necrotizing soft tissue infection. Am J Surg 2000;179:17-21.
56. Rahmouni A, Chosidow O, Mathieu D, et al. MR imaging in acute infectious cellulitis. Radiology 1994;192:493-496.
57. Stamenkovic I, Lew PD. Early recognition of potentially fatal necrotizing fasciitis. The use of frozen-section biopsy. N Engl J Med 1984;310: 1689-1693.
58. Lillie ST, Sato TT, Engrav LH, et al. Necrotizing soft tissue infections: Obstacles in diagnosis. J Am Coll Surg 1996;182 7-11.
59. Elliot DC, Kufera JA, Myers RA. Necrotizing soft tissue infections. Risk factors for mortality and strategies for management. Ann Surg 1996;224: 672-683.
60. Chandrasekar PH, Narula AP. Bone and joint infections in intravenous drug abusers. Rev Infect Dis 1986;8:904-911.
61. Sapico FL, Liquete JA, Sarma RJ. Bone and joint infections in patients with infective endocarditis: Review of a 4-year experience. Clin Infect Dis 1996; 22: 783-787.
62. Brancos MA, Peris P, Miro JM, et al. Septic arthritis in heroin addicts. Semin Arthritis Rheum 1991;21:81-87.
63. Roca RP, Yoshikawa TT. Primary skeletal infections in heroin users: A clinical characterization, diagnosis and therapy. Clin Orthop 1979;144: 238-248.
64. Sapico FL, Montgomerie JZ. Vertebral osteomyelitis in intravenous drug abusers: Report of three cases and review of the literature. Rev Infect Dis 1980;2:196-206.
65. Swisher LA, Roberts JR, Glynn MJ. Needle licker’s osteomyelitis. Am J Emerg Med 1994;12:343-346.
66. Schauwecker DS, Braunstein EM, Wheat LJ. Diagnostic imaging of osteomyelitis. Infect Dis Clin North Am 1990;4:441-463.
67. Tehranzadeh J, Wong E, Wang F, et al. Imaging of osteomyelitis in the mature skeleton. Radiologic Clin North Am 2001;39:223-250.
68. Hlavin ML, Kaminski HJ, Ross JS, et al. Spinal epidural abscess: A ten-year perspective. Neurosurgery 1990;27:177-184.
69. Koppel BS, Tuchman AJ, Mangiardi JR, et al. Epidural spinal infection in intravenous drug abusers. Arch Neurol 1988;45:1331-1337.
70. Reihsaus E, Waldbaur H, Seeling W. Spinal epidural abscess: A meta-analysis of 915 patients. Neurosurg Rev 2000;23:175-204.
71. Del Curling O Jr, Gower DJ, McWhorter JM. Changing concepts in spinal epidural abscess: A report of 29 cases. Neurosurgery 1990;27:185-192.
72. Stabler A, Reiser MF. Imaging of spinal infection. Radiol Clin North Am 2001;39:115-135.
73. Tunkel AR, Pradhan SK. Central nervous system infections in injection drug users. Infect Dis Clin North Am 2002;16:589-605.
74. Martin RJ, Yuan HA. Neurosurgical care of spinal epidural, subdural, and intramedullary abscesses and arachnoiditis. Orthop Clin North Am 1996; 27:125-136.
75. Danner RL, Hartman BJ. Update of epidural abscess: 35 cases and a review of the literature. Rev Infect Dis 1987;9:265.
76. Talan DA, Moran GJ. Tetanus among injecting drug users-California,1997. Ann Emerg Med 1998;32:385-386.
77. Bardenheier B, Prevots DR, Khetsuriani N. Tetanus surveillance—United States, 1995-1997. MMWR CDC Surveill Summ 1998;47:1-13.
78. Francois MP, Roberts JR, Hewlett D. Tetanus in a parenteral drug abuser: Report of a case. J Natl Med Assoc 1994; 86:223-225.
79. Sun KO, Chan YW, Cheung RTF, et al. Management of tetanus: A review of 18 cases. J R Soc Med 1994;87:135-137.
80. Brust JC, Richter RW. Tetanus in the inner city. N Y State J Med 1974;74: 1735-1742.
81. Abrahamian FM, Pollack CV Jr, LoVecchio F, et al. Fatal tetanus in a drug abuser with "protective" antitetanus antibodies. J Emerg Med 2000;18: 189-193.
82. Hagan H, Des Jarlais DC. HIV and HCV infection among injecting drug users. Mt Sinai J Med 2000;67:423-428.
83. Garfein RS, Vlahov D, Galai N, et al. Viral infections in short-term injection drug users: The prevalence of the hepatitis C, hepatitis B, human immunodeficiency, and human T-lymphotropic viruses. Am J Public Health 1996;86: 655-661.
84. Zeldis JB, Jain S, Kuramoto IK, et al. Seroepidemiology of viral infections among intravenous drug users in northern California. West J Med 1992; 156:30-35.
85. [No Author Listed.] AIDS associated with injecting drug use. Morb Mortal Wkly Rep 1996;45:392-398.