Opioid use disorder and endocarditis

How does endocarditis occur in PWID?

Endocarditis has long been known to be a consequence of IDU with many studies dating back at least to the 1950s describing the increasing rates of infections due to IDU, with a focus typically on right-sided endocarditis given the frequency of these cardiac valves being involved [ ]. The risk of infection in these individuals is multifactorial and relates to the drug product, the practice of injection itself, and the susceptibility of the host ( Table 8.1 ).

The pathology of endocarditis results in the establishment of bacteria on the cardiac valves but the initial structural abnormality of the cardiac valve is typically a sterile vegetation composed of fibrin and platelets to which bacteria adhere [ ]. The proposed mechanism of disease begins with impurities in the heroin or drug product that may damage the cardiac valves (or the valves already have preexisting abnormalities) through repetitive injury to the tricuspid valve causing endothelial damage as this is the first valve to encounter the foreign material when injecting into the venous system [ ]. Aside from causing direct mechanical damage, injected diluents can cause vasospasm and damage to the intima layer of the endothelium, leading to thrombus formation and platelet aggregation [ ]. Different organisms have been reported to affect valves with differing frequency [ ] due to the propensity of the individual bacteria to adhere to damaged valves and subsequently colonize this inflamed endothelium producing cytokines and tissue factor which serves to recruit monocytes and platelets [ , ].

The process itself of injecting heroin, for example, is complex and comes with attendant risks at different stages of the injecting process. There are many different factors that are at play both with respect to the effects of heroin and other impurities on the ability of the individual to clear bacteria as well as the actual process of injecting opioids can result in the introduction of bacterial or fungal pathogens leading to disease.

Opioids are often “cut” with impurities which refers to substances added to increase the weight of the drug product being sold but often does not necessarily contain an active drug (although, the addition of synthetic fentanyl to heroin is likely responsible for many overdose deaths in this current epidemic [ ]). These impurities such as talc or baking soda, when injected, may cause local vascular injury which, in conjunction with poor skin antisepsis, can lead to localized or systemic infection. Heroin that is adulterated in this fashion may be limited in its potency requiring the person to inject with greater frequency, enhancing the potential of infectious complications. In addition, the opioid may not have been produced in a sterile fashion and there have been reports of heroin contaminated with spore-producing bacteria [ , ], leading to well-described outbreaks of severe infectious complications of IDU [ ].

The typically solid preparation of the heroin product is subsequently dissolved often with the use of a spoon or an aluminum can in the presence of water and/or heat. The source of the water may range from sterile water to water from a toilet and with it the organisms that may become intermixed with the opioid preparation will then be injected [ ]. Infections that result from gram-negative bacteria that live in water sources have been well described [ ], and the possibility that these organisms are present needs to be considered when treating individuals empirically for life-threatening bacterial infections.

Candida species and other fungi are a relatively rare cause of IE but well known to occur in PWID [ , ] due to environmental contamination. Recent reports from both Denver, CO, and Massachusetts [ , ] point to Candida species increasing in frequency in this population as a cause of both fungemia and endocarditis. Also, syringe tips often will be put into the mouth of the individual prior to injecting for multiple reasons such as habit, to test the drug prior to injecting, or to ensure that the bevel of the needle remains sharp [ ]. This has the potential to introduce mouth pathogens into the bloodstream once the syringe is introduced [ ] into the vein or skin. In addition, a sterile site is often not prepared on the skin with the use of an alcohol pad, and with any penetration of the skin by use of an injectable device can lead to infection typically with skin pathogens, mostly prominently Staphylococcus aureus and Streptococcal species of bacteria.

In addition to infectious agents being introduced through injecting drugs, opioids have been demonstrated to affect the immune system broadly as organs and tissue throughout the body contain opioid receptors, not solely limited to the central nervous system, suggesting a greater overall role in human health. In vitro data suggest opioids have broad effects on all aspects of the immune system including both adaptive and innate immunity [ ], but the research is far from conclusive to suggest either a clear detrimental or beneficial effect on overall immune health [ , ]. While case series and cohort studies have suggested an increased risk for certain infections [ ], more rigorous studies will need to be undertaken to determine the interplay between opioids and the immune system, the impact of length of opioid treatment on these immune effects, and whether all opioid analgesics have the same properties with respect to immune activity.

The host may be uniquely susceptible to bacterial infections as hepatitis C virus (HCV) antibody positivity is seen in roughly 70% of PWID, and the majority of these individuals likely have active disease without having cleared the infection [ ]. While it is unclear to what degree these individuals with HCV are at increased risk for serious bloodstream infections or endocarditis, HCV does appear to represent an additional risk for S. aureus infection [ ]. In those select individuals with HCV and resulting cirrhosis, their immune function will be compromised in various regards, but characteristically they will have more complications from infection than healthy individuals given defects in innate immunity, phagocytosis, and impaired humoral immunity [ ]. Increasing numbers of HIV infections have been seen in different areas of the country [ , ] with the outbreak of HIV infections in Scott County, IN [ , ], being the most dramatic demonstration of an HIV infection cluster related to IDU. Obviously, HIV infection has long been seen in individuals with a history of IDU, and serious bacterial infections which occur in individuals with HIV which can result in endocarditis [ ]; however, the risk of serious bacterial infections can be mitigated with early antiretroviral treatment [ , ].

Increasing rates and epidemiology of infective endocarditis secondary to injection drug use

Several recent studies [ , ] have clearly demonstrated that the rates of endocarditis secondary to IDU have increased over the past 15 years, all of them showing dramatic increases that coincide with the national opioid epidemic [ , ]. In fact, the sharp increase in cases of endocarditis could serve as a method to mark this new epidemic of IDU as an increase in endocarditis seen in the early 2000s was associated was associated increase in positive opioid toxicology and new HCV infections [ ]. There remains, however, a limited ability to capture exact number of cases of endocarditis related to IDU given the dependence on discharge data or the use of surrogate markers for IDU such as hepatitis C status [ , , ] given the lack of an International Classification of Disease (ICD)-9 or ICD-10 code for IDU and no reporting requirements at either the state or national level which could provide actual numbers of patients with this condition. In addition, gaps in recording data fail to account for long-term morbidity, for example, when a patient suffers a devastating embolic stroke from IDU. Regardless of the failure to systematically capture the number of cases of endocarditis, the number of cases certainly is increasing anecdotally and in the available medical literature, and the total numbers of cases that are being discussed likely represent an underestimation of the real number of infections.

An increase in hospitalizations for IE in PWID was seen by Wurcel et al. with 12% of admissions for endocarditis at community hospitals were secondary to IDU in 2013 compared with 7% in 2000 [ ]. Similarly, in North Carolina, annual hospitalizations for IE associated with IDU increased from 0.92 to 10.95 per 100,000 persons between 2007 and 2017 using data gathered from hospital discharges [ ] and in a national survey based on hospital discharge data, endocarditis secondary to IDU increased from 15.3% to 29.1% of IE cases between 2010 and 2015 [ ].

The incidence of S. aureus infections has increased over the last half-century worldwide and this pathogen is the most common cause of IE in developed countries, with the incidence of S. aureus infection being even greater in those with IE from IDU, with well over half of the cases in PWID being due to S. aureus [ , ], far greater than in those who do not inject drugs. Infection with S. aureus puts patients with IDU-IE at increased risk of severe sepsis, multiorgan dysfunction, major neurological events, and death [ ] given the higher rates of vegetations > 1 cm and extracardiac emboli compared with non-IDU patients. Other gram-positive organisms besides S. aureus , gram-negative bacteria, fungal organisms, and polymicrobial infections all have been shown to cause endocarditis in this population[ , , , ].

The International Collaboration on Endocarditis-Prospective Cohort Study (ICE-PCS) investigators found most cases of IE in a pooled international population involve vegetations on either the aortic (38%) or mitral valves (41%) compared with the tricuspid valve (12%) [ ]. However, in those individuals with IDU-related endocarditis, a far higher percentage of them have right-sided disease [ , ]. The mechanism behind this phenomenon remains elusive [ , ]. But, more recently, right-sided endocarditis is by no means the exclusive presentation for those with IDU-associated endocarditis as several studies have demonstrated a relatively high rate of left-sided involvement as well in individuals who inject drugs [ , , , , ]. Historically, there were more men than women with IE secondary to IDU [ , ], and while this seems to continue to be the case, the gap is narrowing as there is an increase in the number of women with this condition [ , ], a fact that is likely trending with the increase in women who use opioids [ ].

Treatment of IDU infective endocarditis

Empiric treatment of IE in patients who have endocarditis, or bacteremia with a concern for endocarditis, needs to account for the likely pathogens. Therefore, it is important that broad antibiotic therapy to cover gram-positive pathogens (particularly methicillin-resistant and methicillin-sensitive S. aureus given the highest likelihood that S. aureus is the causative agent), but also empiric treatment for gram-negative bacteria, including Pseudomonas [ , ], is initiated in the population of patients who inject drugs. This would be a necessary expansion of typical empiric coverage for those who do not inject drugs which may not be as broad with respect to the gram-negative coverage. While empiric treatment is determined on a case-by-case basis, regimens such as vancomycin and either meropenem, a third or fourth generation cephalosporin (ceftazidime or cefepime), or an extended spectrum beta-lactamase inhibitor with activity against Pseudomonas provide the requisite broad antimicrobial therapy in this setting.

Each of the IE guidelines has addressed the topic of empiric antimicrobial treatment slightly differently. The AHA approach to empiric treatment has a statement that simply recognizes that therapy initially is usually broad, will likely be narrowed, and infectious diseases consultation should take place at the time of the initiation of empiric antibiotic therapy. The BSAC guideline provides explicit recommendations for native valve, empiric treatment for endocarditis in an individual who injects drugs and is septic which includes vancomycin dosed as per local guidelines and meropenem 2 g intravenously q8h to cover staphylococci (including methicillin-resistant staphylococci), streptococci, enterococci, HACEK, Enterobacteriaceae , and P seudomonas aeruginosa . The ESC recommends that consideration be given to the type of drug and solvent used as well as the infection location, while ensuring that S. aureus is covered. If the patient is injecting pentazocine, Pseudomonas is emphasized as a potential pathogen and must be included as a target of empiric treatment due to outbreaks of this pathogen [ ], while if brown heroin was injected, Candida spp. should be covered [ ].

Once the microbiologic diagnosis of IE has been established in PWID, the treatment is antibiotic therapy typically for 4 to 6 weeks with much of the historical data suggesting this should be parenteral therapy. The guidelines again provide similar detailed guidance that is organism specific with the most common causative agents in this population in both those with native and prosthetic valves.

Surgical interventions

Surgical indications for IE have been developed and assessment by the surgical team should occur early in the course of a hospitalization and with IE, surgery occurs in up to 50% of the patients affected [ ]. The generally defined indications for which prompt evaluation and consideration for cardiac surgery are listed in Table 8.3 . These include large valvular vegetations, valve rupture, paravalvular abscess, difficult to treat organisms, new embolic phenomena despite appropriate treatment, and heart failure that fails to respond to medical therapy.

The decision to intervene surgically in patients with endocarditis is often reached through a multidisciplinary approach involving the cardiothoracic surgical team, internal medicine providers, and cardiologists, taking into account the risk of the surgical procedure and weighing the benefit that the patient is likely to experience from the procedure.

The ESC IE guidelines specifically include recommendations regarding surgery in PWID as it relates to right-sided endocarditis [ ]. They suggest that avoiding surgery for this condition is advised given the risk of relapse related to opioid use and the chances of recurrent infection. These guidelines are consistent with US practice as clinical data support the relative infrequency of the necessity of surgical intervention for right-sided endocarditis as it is successfully medically managed in the vast majority of cases.

Timing of cardiac surgery for endocarditis

There has been a great deal of discussion in the medical literature about the appropriate timing of surgery for endocarditis [ ]. Several studies have demonstrated that early surgery is not associated with worse clinical outcomes and the US and European guidelines suggest there is no need to delay surgery in general once an indication for early surgery has been identified [ , , ]. Even in patients with ischemic stroke, several studies have suggested that there is no apparent survival benefit in delaying surgery [ , ].

However, surgery will often be delayed when there is evidence of a large cerebral infarction or a hemorrhagic stroke given concerns for neurological compromise. Patients can bleed into the infarcted areas of the brain [ ] or further bleed into the hemorrhage with the use of the cardiac bypass machine and the anticoagulation associated with that procedure. Much of the medical consensus currently suggests waiting at least 2 weeks and ideally up to four before considering surgery in patients with large cerebral infarction and for at least 4 weeks in those with intracerebral hemorrhage if possible [ , ].

In the patient with OUD, it would be expected that the timing of surgery would be consistent with those who do not inject drugs given that the recent thoracic surgery guidelines for the treatment of IE that suggest the same criteria are used for determining surgery whether the patient has a history of IDU or not [ ]. While right-sided disease is successfully treated medically most of the time, stable patients with OUD with large emboli or stroke will occasionally be discharged from the hospital with the plan to complete much of their IV antibiotic therapy only to return at a future date for a planned cardiac surgery procedure. This may be a chance to engage the patient with treatment for their OUD prior to their surgery. Delays for other reasons, including not performing surgery until the patient engages in treatment for OUD, are sometimes considered without data supporting that this strategy motivates patients to engage in OUD treatment, and may not benefit either the patient's cardiovascular health or in overcoming their OUD.