The emergence of human immunodeficiency virus (HIV) infection and the increase in cocaine abuse have added new dimensions to this problem. Despite this, few general reviews have been written [4]. Many questions from the original reviews remain unanswered (for example, the endocrine, hematologic, and immunologic sequelae of drug abuse). Our review is intended to provide a historical context to this vast subject.

Methods

Top Methods Author & Article Info References

Study Selection

We performed a manual search of the internal medicine and subspeciality literature of the past three decades. Selected studies were of three types—baseline studies for the period ending in 1968, subsequent studies (after 1968) that emphasized changes from baseline, and studies after 1968 that emphasized change (or the absence of change) and the manner in which clinicians conceptualized problems.

Definition of Terms

"Intravenous drug user" is now the preferred term instead of the previous, broad term of "drug abuser". This term also places the focus on the major common cause of these medical complications—needle use. Although the term intravenous drug user excludes subcutaneous usage (see tetanus) and the inhalation route used for "crack" cocaine intoxication, we use it because of its connection with the major infectious disease complications.

Certain drugs prevalent in the 1960s, such as intravenous methamphetamine (methedrine or "speed"), have returned. Others, such as cocaine in its various forms—inhaled (powder and free-base), smoked ("crack"), and injected—have expanded to become major parts of the current market. Heroin, which penetrated the inner city areas in the 1950s, remains the major drug abused by addicts currently admitted to drug-abuse treatment programs [5]. However, this may represent selection bias and may systematically underestimate cocaine and amphetamine use [6-8]. However, intravenous drug users are seldom drug purists. In addition to heroin or cocaine, they may inject other opiates available (methadone, hydromorphone [Dilaudid], morphine, or meperidine [Demerol]) and "nonopiate" drugs of other classes, with their unique side effects [9, 10]. A partial list of drugs of abuse and their street names is given in Table 1. Many new slang names and many new drugs that are now abused have appeared during the past 25 years.

The Apparent "Natural History" of Intravenous Drug Use

Experience suggests variability in the natural history of intravenous drug use, but as a disease archetype it is a useful pattern for predicting behavior. The following "natural history" of intravenous drug use emerges from multiple patient interviews and is consistent with what is reported by drug abuse counselors and in the literature [11, 12]. Intravenous administration is usually first tried in adolescence. At this time, the person, who may not yet have his or her own injection paraphernalia ("works"), may use a communal needle and syringe. He may also patronize "shooting galleries" where a common "works" is rented. Most transmission of hepatitis occurs within the first few years; we know less about when HIV transmission occurs, although it may be at the same time. As drug use increases and continues (late teens and early 20s), the person has his own works and shares it less and less, until a steady needle partner may be the only other person using the same hypodermic equipment.

During their late 20s and early 30s, drug users often try to quit or to decrease drug use. They begin to enter rehabilitation programs voluntarily or are forced to do so by the court. Finally, in their late 30s and 40s, they may stop intravenous drug use. Some turn to other drugs, especially alcohol, or other oral agents (prescription or illegal). Some are abstinent for long periods only to relapse repeatedly, and some are weekend recreation intravenous drug users (one form of "weekend chipper"). Fifty- and 60-year-old intravenous drug users are not rare [8, 13].

The relapsing nature of intravenous drug abuse belies the notion of cure in this group of persons. "Stars fall" was the title used by a prominent English therapist writing about relapses in his patients who had previous stellar cures [14, 15]. By this reasoning, the abstinent are only in remission, a profoundly pessimistic view.

Changes in the Injection Apparatus

In the past two decades, the "works" [16] have been modernized. What began as a reusable, metal needle wedged onto a medicine dropper, with rolled paper as an adaptor, is now a disposable syringe with a disposable needle, both being reused. Proponents of the distribution of needles and syringes are not providing anything that is not already available [17]. However, they argue for providing them in new condition to decrease needle sharing, which might help decrease transmission of viral agents, although it will not decrease bacterial or fungal contamination of the skin or injected material. Counseling of intravenous drug users on improving their self-injection sterile technique has also been considered [18]. "Cooking" the material in a spoon or bottle cap to achieve solubility, using contaminated water, and improvising filters (cotton or cigarette filters) are practices that continue, along with the resulting infectious complications.

Complications of Illicit Drug Use

Skin and Superficial Structures

Skin and soft tissue lesions, once the most common reason for emergency room visits by intravenous drug users, result from the nonsterile injections, sharing of equipment, poor personal hygiene, subcutaneous injection into deltoid muscles and thighs in the absence of an available vein, or injection into the veins of the neck or groin. Some active intravenous drug users have stigmata (scars due to old abscesses or "track marks," a darkening of the skin over the antecubital veins, which is literally tattooing with carbon particles and other materials pushed under the skin). It is easy to understand how these repeated injections result in cellulitis, skin abscess, septic thrombophlebitis, necrotizing fasciitis, gas gangrene, pyomyositis [1, 2, 19-21], localized Fournier gangrene [22], lymphedema, or even infected, pulsatile pseudoaneurysms of the neck or groin [23-25].

The microorganisms involved in the skin and local vascular infections vary, but Staphylococcus aureus is the main agent, followed by various streptococci, aerobic gram-negative rods, anaerobic cocci, and bacilli [19-23]. The same range of organisms has been found in the metastatic bone and joint infections and the endocarditis of intravenous drug users.

However, new and unique symptoms continue to be reported; a syndrome of severe dermatitis, eosinophilia, and dermatopathic lymphadenopathy has recently been described in two intravenous drug users with dual infection with HIV-I and human T-lymphotropic virus II [26].

Musculoskeletal Involvement

To the rheumatologist, the complications of intravenous drug use cover a broad area [27], ranging from the transient rheumatologic prodrome of hepatitis B antigenemia [28] to chronic amyloidosis [29, 30]. One salient change in the past decades has been in the agents causing bone and joint infections in intravenous drug users [31-33]. Staphylococcus aureus has apparently replaced Candida species and gram-negative bacilli (in particular, Pseudomonas) as the primary cause of these infections—not unlike what occurred with endocarditis (see below). In the current setting of extensive cocaine abuse, there are reports of muscle and skin infarction [34] and of rhabdomyolysis after "free-basing" cocaine, sometimes with renal failure and shock after intravenous injection or inhalation of free-base cocaine (crack) or both [35-38]. A similar syndrome of rhabdomyolysis and shock has been reported due to intravenous methamphetamine or phenmetrazine use [39]. Rhabdomyolysis and myoglobinuria have also been reported with heroin and various other drugs [40-44]. Some cases were clearly due to pressure-induced crush injuries during a prolonged period of unconsciousness.

Two unique reports, frequently cited in the literature, must be discussed. A report from Boston City Hospital, in the late 70s, of a musculoskeletal syndrome was associated with "Brown Mexican heroin" [45]. This syndrome was never reported again nor confirmed by clinical experience elsewhere. It was either a truly local phenomenon or an example of the hazard of asking leading questions of intravenous drug users.

The other report [46] presented a syndrome of small middle vessel angiitis (which was disseminated and necrotizing) as a consequence of intravenous drug use. What was insufficiently emphasized was that most patients with this syndrome were intravenous users of methamphetamine. This drug abuse from the 1960s to early 1980s produced a considerable volume of literature [47-53] covering both cerebral and disseminated vasculitis as well as intracerebral hemorrhage and cerebral infarction. Some of these same findings were noted later when cocaine abuse spread, not surprisingly, because amphetamines and cocaine are sympathomimetic agents.

Pulmonary Complications

The nature of pulmonary complications has been greatly altered by the arrival of HIV (Table 2). Aside from Pneumocystis pneumonia [54], the opportunistic infections of HIV have had a smaller effect on hospital admissions of intravenous drug users than have the more frequent bacterial pneumonias. It had long been suspected, although never proven [1, 20, 44], that intravenous drug users were at greater risk for bacterial pneumonia. Although no difference appears to occur in progression to the acquired immunodeficiency syndrome (AIDS) between intravenous drug users and nonintravenous drug users, substantive evidence exists in drug users of increased incidence of bacterial pneumonia, especially that due to Haemophilus influenzae [56-59]. Bacterial pneumonia in immunocompromised persons is also described in the list of complications occurring in intravenous drug users. Unusual organisms include Corynebacterium equi [60] (now Rhodococcus equi) and Nocardia [61], and the list will surely expand.

Before the arrival of HIV, tuberculosis was considered to have a higher prevalence among intravenous drug users [2, 20, 44, 62]. Increased reactivation, dissemination, high rates of transmission, and re-infection with multiresistant strains are important clinical features of tuberculosis in the HIV-infected person [63-66]. Parenteral drug use is apparently not necessary, because an increased incidence of tuberculosis with HIV infection is noted in crack users [67, 68]. Although the number of cases of tuberculosis occurring in drug abusers since 1984 is not directly related to HIV infection, the resurgence of tuberculosis appears closely related to the HIV epidemic. The largest increases have certainly occurred in areas with the highest HIV prevalence [64-66].

The recent outbreaks of multidrug-resistant tuberculosis among HIV-infected substance abusers in different geographic areas [69, 70] make drug users [66, 71] a difficult group to treat; many more of these patients are now being cared for by public institutions, compared with 20 years ago. Prison and hospital outbreaks are frequently caused by intravenous drug users; two newly devised programs are aimed at this subset of drug users. The first program, which is directly observed therapy [72], is indeed desirable and feasible for compliant, intravenous drug users with a known permanent address and is not for those users who do not possess these characteristics. Because multidrug-resistant tuberculosis has spread among intravenous drug users and effective therapy is not yet available, directly observed therapy is not a viable solution for these patients. With the addition of the immunocompromised state, consideration must be given to a second program, long-time tuberculosis detention [73] (or custodial care) as a relevant but politically sensitive alternative.

The use of cocaine, particularly in its alkaloidal (free-base) form, crack, has extended the range of the noninfectious pulmonary disease in illicit drug users. Pulmonary findings attributed to the inhalation of cocaine include focal air-space disease, atelectasis [74, 75], alveolar hemorrhage [76], pneumothorax, pneumomediastinum [77-79], bronchiolitis obliterans, organizing pneumonia [80], pulmonary infarction [81], and pulmonary edema [82, 83]. It should be recalled that pulmonary edema was an important feature of heroin abuse and a major cause of death [1, 2, 84].

The focal complications of intravenous drug use previously mentioned are the same whether heroin or cocaine, or a cocaine and heroin mixture ("speedball"), is used. Complications occur with the "pocket shot," an injection into the supraclavicular fossa in an attempt to hit the jugular, subclavian, or brachiocephalic veins. These complications include pneumothorax, [85, 86], hemothorax and pyopneumothorax [87], cellulitis, abscess, or even a large hematoma forming a pseudoaneurysm [25, 86].

Septic thrombophlebitis producing pulmonary emboli [20, 88, 89] or infective endocarditis from the same source [90] (discussed more fully in the next section) or mycotic aneurysms of the subclavian, carotid [86, 9193], and pulmonary arteries [94] have persisted as complications during the past 25 years.

The intravenous drug user is susceptible to the hazards of injecting insoluble additives or fillers such as talc and starch that can produce chronic granulomatous inflammation and angiothrombosis and even delayed emphysema [95-99]. Indeed, intravenous drug users were reported to have a decreased diffusing capacity [100, 101]. However, in a recent study of former intravenous drug users, a normal diffusing capacity was found once adjustment was made for the effect of cigarette smoking [101]. Because 95% of older intravenous drug users are also heavy cigarette smokers, this factor probably confounded the results of some previous studies.

Cardiovascular Complications

Cocaine is also associated with the noninfectious complications of myocardial infarction [103], coronary artery constriction with angina [104-106], and cardiomyopathy [107]. Rhabdomyolysis appears to produce pectoral muscle pain [37, 38] mimicking the pain associated with myocardial infarction; rhabdomyolysis also causes an increase in the MB isoenzyme bands of creatine phosphokinase (see Table 2).

Before the arrival of HIV and the resurgence of cocaine use, endocarditis was the second major acute medical problem of intravenous drug users requiring hospital admission, exceeded only by the number of users who had overdosed. The first reported cases of infectious endocarditis in intravenous drug users appeared in the 1930s [108, 109].

In a previous review of this long history of endocarditis from 1936 to 1974 [44], we noted three obvious overlapping eras of reporting: Candida, 1940 to 1962; Staphylococcus aureus, 1951 to 1972; and gram-negative rods, especially Pseudomonas, 1972 to 1974. These overlapping eras were probably due more to publication biases than to changes in the disease itself.

Acknowledged and unacknowledged biases of patient selection may affect the relative frequency of certain organisms. For example, a series from Cook County Hospital in 1970 reported Streptococcus viridans as the most frequently cultured organism, which was distinctly unusual for this disease in the intravenous drug user [110]. However, 3 years later, a series from the same hospital reported a 70% frequency of Staphylococcus aureus of the tricuspid valve [111]. The staff at Cook County recalled that the cases in the first study were identified from the files of the cardiac catheterization laboratory. Therefore, these cases were selected for valvular deformity and also for a less acute clinical course, a virtual description of Streptococcus viridans subacute endocarditis. The second paper [111] was based on a clinical series of patients seen by the infectious disease consultant and thus, not unexpectedly, patients with acute staphylococcal endocarditis predominated. A comparison in New York City of intravenous drug users with endocarditis identified by autopsy, general medical rounds, or infectious disease consultation showed clearly that the means of identification markedly influenced the patient composition [112].

In Detroit, patients with left-sided Pseudomonas aeruginosa endocarditis [113], based on confirmation by surgical or autopsy evidence, had both high mortality and more frequent valvulotomy. Similarly, in Charity Hospital in New Orleans [114], a report of a series of patients who were intravenous drug users, which was based on the records of the cardiac echocardiographic laboratory and required a visible vegetation for inclusion, naturally stated that left-sided endocarditis was more frequent than right-sided endocarditis in these drug users.

Despite inevitable biases, most contemporary series find Staphylococcus aureus infections in more than 50% of clinical series, and most infections are presumed to be of the tricuspid valve [44, 90, 111, 112, 115]. The remaining organisms are various staphylococci and streptococci [112, 116-123], gram-negative bacilli (notably Pseudomonas aeruginosa, Pseudomonas species and Serratia, [112, 124-128]), respiratory flora, Candida, anaerobes [129-131], polymicrobial infections [129, 131, 132], and others not easily categorized [133, 134].

Local and temporal variations occur. Serratia endocarditis was a notable cause of endocarditis in the San Francisco area for more than a decade [127]. Pseudomonas cepacia, by contrast, was seen for only a few years in the New York City area [128]. Pseudomonas aeruginosa remains as a persistent but minor cause of infectious endocarditis in the Detroit area, where it is reported in about 3.5% of cases [115, 124, 126]. The long-term use of pentazocine and tripelennamine ("T's and blues") by addicts in Detroit has been held responsible for the persistence of Pseudomonas endocarditis among intravenous drug users [115]. About 10 years ago, a cluster of patients with Pseudomonas endocarditis was noted at Cook County Hospital in Chicago and was associated with the use of these agents by intravenous drug users [125]. Neither this endocarditis nor this drug use is seen in Chicago at this time.

Because intravenous injection of heroin and cocaine is a risk factor for Staphylococcus aureus endocarditis [135], it has been suggested that this organism probably originates from the drug user's own skin [136, 137]. This probably includes methicillin-resistant Staphylococcus aureus [138, 139]. A high proportion of intravenous drug users are skin carriers of Staphylococcus aureus [137, 140], as are other types of patients who receive frequent injections [141-144].

In contrast, some evidence exists that other bacteria and fungi may be derived from the injected material or the diluent. This is thought to cause the association between "T's and blues" and Pseudomonas endocarditis [115, 125]. This may also be true of Candida species. Tuazon and Sheagren [136] found Candida in street heroin, and Candida parapsilosis has been isolated several times from street heroin containers and injection apparatus [145, 146]. Finally, the syndrome of disseminated Candida albicans infection described in Europe during the 1980s [147] (see below) undoubtedly had its source in the contaminated lemon juice used as a diluent [148, 149].

By the 1980s, the number of intravenous drug users with endocarditis and, as a consequence, staphylococcal endocarditis, had become great enough that comparative therapeutic trials could be done. Initial reports determined that monotherapy with a penicillinase-resistant penicillin (oxacillin or nafcillin) or a selected cephalosporin for 4 weeks yielded high cure rates [151, 152]. The addition of an aminoglycoside produced only minor clinical improvement but substantially increased the occurrence of nephrotoxicity if the aminoglycoside was maintained for the duration of therapy [151, 153].

Further trials attempted to further shorten therapy for uncomplicated, right-sided Staphylococcus aureus endocarditis [154, 155]. However, uncomplicated, right-sided staphylococcal endocarditis occurred in only a few patients with staphylococcal bacteremia who were not identified on initial evaluation. Short-course vancomycin showed a high rate of failure [154] (indeed, doubts have subsequently been raised about the adequacy of vancomycin even when given for a longer therapeutic duration [156]).

The early development of resistance to rifampin and ciprofloxacin was reported [157] in an intravenous drug user who was HIV positive and was treated with a similar short regimen. First, ciprofloxacin resistance develops rapidly in Staphylococcus aureus [158]. Moreover, there are indications that advanced HIV infection makes endocarditis more difficult to treat [159]. The question of shortened therapy is now moot, because each case of endocarditis in an HIV-infected intravenous drug user is automatically a complicated one.

The full consequences of HIV infection in endocarditis in the intravenous drug user are not yet fully understood. Although the number of patients with endocarditis in New York City in the 1970s appeared to be stable [160], an increase occurred in the 1980s. Further evidence of the large number of intravenous drug users with endocarditis is their inclusion in pharmacokinetic and therapeutic studies of teicoplanin, daptomycin, and vancomycin in patients with gram-positive infections [161-163].

Septicemia and Disseminated Infections (Other Than Endocarditis)

Group A, ß-hemolytic, streptococcal septicemia with high fever, toxicity, and renal failure from either septic thrombophlebitis or cellulitis in intravenous drug users was noted in the early 1970s [20]. It is again reported in intravenous drug users [115, 162-168], probably due to the resurgence of group A streptococcal infections in the population [169-171]. The source of the septicemia with group A streptococci is probably skin colonization. Groups B, C, and G streptococci, also capable of skin colonization, have been reported to cause bacteremia and endocarditis in intravenous drug users [119-121, 172-174].

A most unusual, disseminated infectious syndrome of the past two decades was that of Candida albicans with scalp, ocular, and osteoarticular involvement [147-150, 175-185], which occurred in intravenous drug users only in Europe. It is not the sole evidence for the spread of intravenous drug use in Europe during the past decade [186]. First reported from France [147, 175-177] and then from Italy [178], Spain [147, 183], Switzerland [179], and the United Kingdom [140, 141, 150, 180], it was originally attributed directly to "Iranian Brown Heroin". The syndrome began with postinjection fever, chills, and myalgias followed by the formation of painful small nodules ( 1 cm) on the scalp and other hairy parts of the body. These usually resolved (leaving alopecia) but sometimes became suppurated and drained. One to two weeks after the onset, Candida ophthalmitis might have occurred [147, 177-183]; osteoarticular involvement was seen at a lesser frequency [140, 177-183] 2 weeks to 5 months after the initial episode. Costochondral disease was most frequent, whereas involvement of the vertebral or sacroiliac areas or peripheral joints was uncommon [140, 177]. Pleuropulmonary disease was also noted in one study [177].

The source of the Candida albicans was probably contaminated lemon juice [148-150] used to dissolve the brown heroin [148, 183]. Lemon juice is an excellent growth media for Candida albicans [149, 150, 181], and both fresh lemons and bottled juice and injection paraphernalia can harbor the yeast [150]. The use of lemon juice as a vehicle for dissolving drugs is prevalent in Europe and especially Spain, but the syndrome has not been reported in this country. However, two patients were recently reported who had disseminated candidiasis after intravenous use of oral methadone for which the source was shown to be methadone diluted with another citrus juice, orange juice [186]. This syndrome should not be confused, as it has been by some authors [187], with fungal endocarditis reported from this country. These are two different candidal syndromes separated by decades in time, an ocean, and species of Candida.

In contrast, sporadic cases of candidal and fungal (again, Candida albicans, Aspergillus, and Penicillium) endophthalmitis in the intravenous drug user have been seen both here and in Europe [188-193]. Clinical findings are similar to the reported cases in the nonintravenous drug user patient [194]. Because most of the cases of Candida ophthalmitis in the nonintravenous drug user have a history of prolonged intravenous access or therapy [194], the occurrence of the same entity in intravenous drug users is understandable. The source is again probably in the injection apparatus or injected mixture. Similarly, fungal (Aspergillus and mucormycotic [zygomycotic]) brain abscesses have been noted many times [195-201]. Aspergillus has been detected in contaminated street heroin [3, 136].

The intrusion of HIV into the intravenous drug use population has introduced to this group infections characteristic of the immunocompromised host. These include renal abscess with the previously mentioned Aspergillus [202], disseminated Rhodococcus equi infection [203], listeriosis [204], nocardiosis (pulmonary and brain abscess) [61, 205, 206], relapsing Salmonella septicemia [207], and pyomyositis in intravenous drug users with advanced HIV infection [208-210]. (This last is in addition to the more usual, mixed bacterial muscle abscess that occurs after direct injection [211].)

In the miscellaneous category are four patients from Spain who had tick-borne relapsing fever; Borrelia was probably transmitted by intravenous drug users who shared injection needles [212]. Pericarditis due to Bacillus cereus has been described [213] in an intravenous drug user with renal insufficiency on hemodialysis. Bacillus cereus spores were recovered from street heroin and his injection paraphernalia [213].

Finally, botulism secondary to Clostridium botulinum infection of a sinus [214] has been reported, as has Pott puffy tumor [215], a subperiosteal abscess of the frontal bone associated with complicating frontal osteomyelitis. Because chronic sinusitis and septal perforation are well-known complications of chronic cocaine sniffing, such complications can be expected in the future.

Syphilis and Other Sexually Transmitted Diseases

Twenty-five years ago the most discussed aspect of syphilis in the intravenous drug user was the high rate of biologic false reactivity in serologic tests for syphilis [3, 44, 216]. Now that a substantial proportion of intravenous drug users are part of the HIV-infected population, the problems of diagnosis and treatment of syphilis are more complicated. Seronegative secondary syphilis occurs in HIV-infected persons [217, 218]. Lack of reactivity is more of a diagnostic problem than is low-titer biologic false reactivity. The course of the disease has been altered with an increase and persistence in early-onset neurosyphilis (meningitis, meningovasculitis with stroke, and cranial nerve abnormalities) [219-221] and by a questionable response to recommended therapy [222-225]. Finally, the number of persons with syphilis is now increasing rapidly rather than declining. From the viewpoint of the public hospital, the major component of this epidemiologic wave is the drug user [217, 226, 227]; of special concern are the maternal and congenital cases. Other venereal diseases such as gonorrhea, pelvic inflammatory disease, cervical dysplasia, chancroid, and nongonorrheal urethritis are also seen more frequently in the intravenous drug user [227].

Hepatitis Viruses

Multiple episodes of clinical hepatitis (either recurrence or reinfection) affect intravenous drug users [1-3]. We can now identify the individual viruses, and we know that epidemic transmission of multiple, different hepatitis viruses (A, B, C, and delta viruses) occurs in intravenous drug users [228-234]. Hepatitis A is presumably spread by close personal contact or ingestion of contaminated drugs [228], whereas the other viruses require needle sharing.

The interaction of HIV with hepatitis B virus and hepatitis D virus has been described in two studies [235, 236]. Both show that concomitant infection produces an increase in hepatitis B virus and hepatitis D virus replication and a decreased titer of hepatitis D antibodies. No evidence of an increased risk for hepatic injury from the simultaneous infection with hepatitis and HIV has been found.

Human T-Lymphotropic Virus II

Human T-lymphotrophic II is also needle transmitted in this country among intravenous drug users [243-248]. Interestingly, human T-lymphotropic virus II appears to have been an endemic infection in intravenous drug users at least since the early 1970s. Its clinical consequences are not yet understood.

Renal Complications

In the 1970s, the nephrotic syndrome with rapidly progressive renal failure were first reported in intravenous drug users [243-248]. The renal biopsy usually showed focal segmental to diffuse sclerosing glomerulonephritis [249]. Such a large number of cases of renal failure in young addicts was unexpected and led to a supposition that this process was somehow related to heroin abuse. The syndrome was called "heroin nephropathy," despite the lack of clear epidemiologic evidence [247] and true understanding of the cause. Only one study [248] from the Buffalo area offers support for that designation. The incidence of sclerosing glomerulonephritis among intravenous drug users was 30 times greater than in nonaddicted persons in this area. Interestingly, in this study, most new cases of sclerosing glomerulonephritis, whether in intravenous drug users or in nonintravenous drug users, were in young black men.

Nephropathy (HIV associated) may also exist, occurring with similar clinical features of proteinuria or the nephrotic syndrome and progressing to end-stage renal disease with a diffuse sclerosing glomerulonephritis on renal biopsy [250-253]. Again, the predominance of black males in this group is striking. These nephropathies have created a large burden on dialysis programs in public hospitals.

Neuropsychiatric Complications

Several hospital-based studies [254, 255] have recorded the effect of cocaine on neurologic and psychiatric complications such as altered mental status, coma, seizures, agitation, disorientation, and psychosis. Thus, young adults seen in emergency rooms with altered consciousness or mentation must now be evaluated for, in addition to other causes, cocaine intoxication.

Many drugs of abuse can produce the organic brain syndrome through hypoxia or cerebral hemorrhage or infarction. The early hope that alterations of the mental examination might be specific for certain classes of drugs has not materialized [44]. After acutely intoxicated persons are excluded, neither the frequency of drug abuse nor HIV seropositivity influences neuropsychologic test performance [256].

One major change in medical training in the past 25 years is that withdrawal protocols are not taught. There has been a concomitant institutionalization of the process as well as sequestration of addiction treatment and "detoxification" programs. The overall effect of institutionalization of such programs has been to remove withdrawal therapy entirely from the general training experience.

Old Diseases

Several classic diseases of the intravenous drug user are now rare. Tetanus is the oldest infectious medical complication of hypodermically administered narcotics, dating back to Victorian times. It was also a prominent feature of the medical complications of illicit narcotic abuse seen in New York City municipal hospitals in the 1950s and 1960s. It was seen primarily in addicts who "skin popped" (used subcutaneous injections). The poorer venous access of women was used to explain the more equal sex ratio than would be predicted given the male predominance within the addict population [257-259]. Patients always had a severe form of the disease, and the high mortality rate (>90%) was not effectively reversed until the creation of respiratory intensive care units. Tetanus began to disappear in the early 1970s, probably because the population at risk was immunized in childhood with diphtheria-pertussis-tetanus. A contributing factor, at least in New York City and some other cities, may also have been the routine tetanus immunizations given to identified drug users beginning in the 1960s. Even so, tetanus still appears in intravenous drug users in some cities [260].

In the 1930s, falciparum malaria was a consequence of intravenous drug use in the Eastern Maritime cities [261-263]. At that time a substantial number of intravenous drug users were sailors, explaining that geographic distribution. After World War II, malaria became rare in the United States as the characteristics of intravenous drug users changed.

A cluster of patients had Plasmodium vivax in California during the Vietnam era, presumably needle-transmitted [264]. The ease of foreign travel to Africa or South America means that future cases of plasmodia transmission might again occur among intravenous drug users.

Hindsight and Foresight

Twenty-five years ago, our most pessimistic vision would have been that intravenous drug use would continue at the same levels, remaining a part of the urban scene. Indeed, in the late 70s, reviewing the indices of addiction-related disease in New York City [160], we thought that they were merely fluctuations around a stable baseline. Unaware of the epidemiology in other cities and in Europe, we could never have predicted the spread of cocaine use. Twenty-five years ago, the major psychoactive sympathomimetic agents available were methamphetamine and amphetamine; cocaine was viewed as a quaint relic. We find that intravenous drug use now has taken a permanent place in our society, along with alcoholism, as addictive disorders associated with constellations of serious medical problems.

Whereas the past 25 years have seen controlled therapeutic trials in infectious endocarditis and other infectious problems in the intravenous drug user, there are few comparable studies of the long-term effectiveness of these therapeutic drug rehabilitation programs. Within the past decade, the legislative justification for these programs has become their putative role in blocking the transmission of HIV. There are other newer policies aimed at the intertwined problems of intravenous drug use and HIV: needle exchange, syringe sterilization, directly observed therapy, and long-term care for multiple-drug-resistant tuberculosis.

Both intravenous drug users and alcoholics depend on public medical facilities for their care. Before 1980, in New York City and elsewhere, admissions due to alcoholism predominated at these institutions. Now in New York City and other areas plagued by HIV infection the reverse is true. The burden on these medical facilities is already enormous and growing. The evidence is that drug users require a substantial amount of medical resources [265], much greater than an age-matched control population, and that the serious clinical disease is predominated by intravenous drug-related infections [266]. This medical need is stimulating some response [267], but the cost to society will be enormous. The task falls not to infectious disease specialists but to primary care physicians. Perhaps medical schools should provide courses in addiction-related studies. For those clinicians who want to know more about the clinical management of the infectious complications of the intravenous drug user, Levine and Sobel's recent book is recommended [268].