The postoperative course of patients with infective endocarditis is marked by relatively high mortality, with increased risk of severe complications leading to prolonged intensive care unit (ICU) and hospital lengths of stay. Nevertheless, over 90% of endocarditis patients who undergo surgery survive to hospital discharge, but with higher mortality rates among patients with prosthetic infections compared to native valve infections (13% vs. 5.6%) [ ]. 80% of patients are alive at 1 year, irrespective of valve type used [ ]. The average postoperative length of stay among IE patients is 18 ± 2 days [ ]. Overall outcomes of patients with IE who undergo surgery are dependent on existing preoperative risk factors, such as age, persistent bacteremia, embolic strokes, congestive heart failure, cirrhosis, renal insufficiency, and other major comorbidities [ , ].
In addition to preoperative factors, perioperative factors either specific to or more pronounced with surgical repair of IE such as long cardiopulmonary bypass times, operating on septic patients, debriding infected tissue, and major valvular and tissue reconstruction increase the risk and severity of complications such as severe vasoplegia and hypotension, mixed shock, coagulopathy requiring treatment, excessive postoperative bleeding that may require reexploration of the chest for hemostasis, or relief of tamponade [ , ].
Additional complications that carry significant morbidity and mortality in this patient population include cardiac arrhythmias, cerebrovascular accidents (CVAs), acute renal failure, and pneumonia [ ]. Finally, health-care providers need to consider special patient populations, in particular, pain management and relapse prevention, in the intravenous drug user (IVDU) to improve outcomes and reduce recurrence in this vulnerable group.
Cardiac surgery patients are prone to profound vasoplegia postoperatively, occurring in 5%–25% of patients [ , ]. Postoperative vasoplegic syndrome occurs when there is markedly low systemic vascular resistance, frequently associated with lack of volume responsiveness and increased cardiac demand to maintain oxygen delivery [ ]. Patients with infective endocarditis who undergo surgery are at increased risk for vasoplegic syndrome [ ]. This is likely multifactorial, due to long cardiopulmonary bypass runs, preoperative use of vasodilators, endothelial dysregulation, preoperative cardiac dysfunction, in addition to the profound inflammatory response that characterizes IE [ ]. Initial treatments for vasoplegia include the need for high-dose alpha adrenergic agents such as epinephrine, norepinephrine, phenylephrine, and vasopressin. When these first-line measures fail resulting in refractory vasoplegia, methylene blue can be added [ ].
In addition to vasoplegia, marked by distributive shock due to a profound inflammatory state and sepsis, IE patients are at risk of hypotension postoperatively owing to a mixed picture which includes low cardiac output as myocardium recovers. Further, IE patients who undergo surgery are prone to pulmonary edema, especially following operations for tricuspid valve endocarditis. As a result, careful and frequent attention needs to be paid to the patient’s fluid and hemodynamic status in the early postoperative period, and intraoperative transesophageal echocardiography in addition to frequent point-of-care ultrasound assessments of fluid status postoperatively can help guide resuscitation [ ].
Although pulmonary artery catheters (PACs) have been associated with right-sided endocarditis, the mixed cardiogenic, distributive, and vasoplegic shock state seen in endocarditis patients often argues in favor of PACs [ , ].
The increased risk of heart block postoperatively [ ] argues for the intraoperative placement of both atrial as well as ventricular epicardial leads. When pacing is required for bradycardia, depending on the AV interval which can be significantly prolonged, A pacing or AV pacing with an appropriate AV interval is required. Clearly, for third-degree heart block, AV pacing is preferred over V pacing alone in the face of any LV dysfunction [ , ].
Postoperatively, the decision to start or resume anticoagulation should be made on a case-by-case basis depending on the patient’s preoperative need for anticoagulation, valve(s) implanted, and additional competing risk factors for both bleeding and thrombosis. CNS hemorrhage is a major cause of mortality among patients with prosthetic valve endocarditis who are anticoagulated [ ]. Further, anticoagulation does not effectively prevent septic emboli but does significantly increase the risk of hemorrhagic conversion of ischemic stroke or intracranial hemorrhage. In postoperative patients, anticoagulation should only be initiated or resumed if there is a strong indication to do so, such as mechanical prosthetic valves, deep vein thrombosis, or pulmonary embolisms [ ]. Even in these scenarios, the benefits of preventing thrombus or embolization should be considered in the context of an increased risk of postoperative intracranial hemorrhage, especially among patients diagnosed with ischemic stroke preoperatively.
All patients who undergo surgery for active infective endocarditis should receive postoperative antibiotics. The standard duration of antibiotics for a full course of treatment of infective endocarditis is 6 weeks from the time of source control, though the duration in part depends on the overall clinical scenario, invasiveness of infection, type of organism, and presence of hardware [ ]. In general, in cases in which blood cultures were initially positive, the 6 weeks duration should begin on the first day in which blood cultures are negative or the day of surgery, whichever is later (Class IIa; Level of Evidence C) [ ]. Tissue cultures should be taken at the time of surgery, and antibiotics should be narrowed and both choice and duration tailored toward the culture results. If operative tissue cultures are positive, then the patient should proceed with the entire antimicrobial course following surgery (American Heart Association Class IIa; Level of Evidence B) [ ]. However, in instances where the operative tissue cultures are negative, depending on the overall clinical scenario it may be reasonable to count the number of days of antimicrobial therapy administered before surgery in the overall duration of therapy, or to shorten the duration of therapy (if blood cultures are also negative) (Class IIb; Level of Evidence C) [ , ]. Additional sources of bacteremia and satellite infections should also be investigated both before and following surgery, such as the teeth, mouth, chronic skin lesions or bone infections, colon (in cases of Streptococcus gallolyticus ), and implanted medical devices. In patients with fungal endocarditis, lifelong oral antifungal therapy is recommended as suppressive therapy due to its predilection for recurrence. A formal infectious disease consultation at time of antibiotic therapy initiation is recommended for assistance with antibiotic management and duration (Class I; Level of Evidence B) [ ].
Postoperative pain management
Postoperative pain management is critical in endocarditis patients, just as it is in all cardiac surgical patients. One special population to consider is opioid IVDUs. Indeed, an estimated 90% of isolated right-sided endocarditis occurs in IVDU patients, and 20% of left-sided endocarditis [ ]. Patients with opioid use disorder develop opioid tolerance and therefore should be expected to require higher doses of opioids postoperatively compared to opioid-naive patients. Furthermore, there is evidence to suggest that sustained opioid exposure leads to hyperalgesia, with lower pain threshold and tolerance, making postoperative pain control in this population even more challenging [ ]. When available, involvement of an acute pain service to assist in determining the optimal analgesic regimen can be helpful.
Generally speaking, postoperative pain control should be multimodal with the use of nonnarcotic adjuncts to decrease the dose of narcotics which minimizes their accompanying side effects. Recommended multimodal agents include acetaminophen, gabapentin, ketamine, and dexmedetomidine. Nonsteroidal antiinflammatory drugs (NSAIDs) are typically contraindicated in the cardiac surgical population given their potential nephrotoxicity and platelet inhibition [ , ]. The potential nephrotoxicity of NSAIDs is particularly relevant in the endocarditis population, as acute kidney injury (AKI) complicates nearly one-third of endocarditis cases [ ].
Acetaminophen should be utilized in the absence of hepatic dysfunction. A randomized clinical trial (RCT) found that six 1000 mg doses of intravenous (IV) acetaminophen administered over the first 24 h postoperatively to cardiac surgical patients reduced opioid consumption and improved patient satisfaction with their overall pain experience [ ]. Another trial of 24 h of IV acetaminophen versus placebo in poststernotomy patients demonstrated significantly reduced pain intensity scores, but no significant reduction in opioid consumption [ ]. There is also some evidence to support the use of gabapentin or pregabalin to address neuropathic pain. A meta-analysis of RCTs found that three of four gabapentin trials and two of four pregabalin trials resulted in decreased opioid consumption, with three RCTs each for gabapentin and pregabalin resulting in lower reported patient pain scores [ ].
Use of ketamine, an N-methyl- d -aspartate (NMDA) antagonist, has also been shown to reduce opioid consumption after cardiac surgery [ ]. One RCT demonstrated significantly less oxycodone consumption in the 48 h postoperatively and superior patient satisfaction in coronary artery bypass grafting (CABG) patients treated with a bolus dose of ketamine followed by a 48-hour infusion, as compared to saline placebo [ ]. If sedation is needed, dexmedetomidine has been demonstrated to reduce the incidence, delay the onset, and shorten the duration of postoperative delirium in elderly patients after cardiac surgery as compared to propofol [ ]. A multicenter trial in the United States and Canada compared a dexmedetomidine-based sedation regimen to a propofol-based regimen in the ICU after CABG and found fewer dexmedetomidine-treated patients remained on the ventilator beyond 8 hours, and morphine use was significantly reduced [ ].
The use of epidural analgesia has been evaluated in cardiac surgical patients, but epidurals are contraindicated in with the face of ongoing bacteremia [ ]. Single-shot bilateral pectoralis nerve blocks have also been described and have been shown to significantly reduce pain scores [ ]. Evidence on the efficacy of multimodal, nonnarcotic-based pain regimens has led to the creation of cardiac surgery–specific enhanced recovery after surgery (ERAS) protocols that incorporate acetaminophen, gabapentin, ketamine, dexmedetomidine, and regional nerve block [ ]. While these protocols have been designed and tested in a more elective cardiac surgery population rather than the critically ill endocarditis patient, many of these concepts—multimodal therapy, use of nonnarcotic adjuncts, use of nonnarcotic sedatives—generalize to the postoperative endocarditis patient.
Postoperative noncardiac complications
As outlined in Chapter 10 on Preoperative Care of the Critically Ill Endocarditis Patient, complications of endocarditis generally result either from direct damage to the heart, vegetation embolization, or systemic hypoperfusion due to septic or cardiogenic shock. When an endocarditis patient is treated surgically, vegetations should be removed and the infected tissue debrided, decreasing the subsequent risk of vegetation embolization. While the embolic risk is reduced, the hemodynamic changes associated with a major cardiac operation and cardiopulmonary bypass run mean the risk of complications from systemic hypoperfusion persists.
Prior studies have demonstrated that CVAs are some of the most common, and some of the most devastating, complications following surgical repair of infective endocarditis [ ]. Postoperative strokes are more likely in patients who have large vegetations of the aortic and/or mitral valves, and those who experienced embolic stroke preoperatively.
Despite the risk, surgical repair confers survival and morbidity benefits and is associated with lower embolic strokes compared to medical therapy, especially among patients with large vegetations. A randomized study showed that early surgery (within 48 hours of randomization) in patients without preoperative stroke was associated with reduced mortality and embolic strokes (3% of composite outcome of in-hospital death and embolic events within 6 weeks of randomization in the early surgery vs. 28% in the conventional treatment group, P = 0.02) [ ].
Yet despite the benefit of surgical intervention, significant risk remains for postoperative stroke. In a small institutional series, 16% of patients developed a CVA postoperatively, with one patient requiring a craniotomy due to cerebral edema [ ]. Additionally, in a multicenter study of 1345 consecutive episodes of left-side endocarditis, 25% of patients experienced neurologic complications following surgery [ ]. Of these, 14% (188) had ischemic events, 6% (86) had encephalopathy or meningitis, 4% (60) had hemorrhages, and 1% (2) had brain abscesses. Ischemic stroke or intracranial hemorrhage increased mortality from 24% to 45% ( P = 0.01). Regarding perioperative management, early antimicrobial therapy significantly reduced the incidence of neurologic complications, but anticoagulant therapy, which could be viewed as a modifiable factor, increased brain hemorrhage [ ]. Further, among the subset of patients with preoperative intracranial hemorrhage, mortality was higher when surgery was performed within 4 weeks of the hemorrhagic event [ ].
While some studies have demonstrated no survival benefit to delayed versus early surgery among patients with IE complicated by ischemic stroke, current STS guidelines recommend delaying valve replacement for at least 4 weeks in patients who have had either a major ischemic stroke or any intracranial hemorrhage [ , ]. This recommendation is due to the increased risk of spontaneous hemorrhagic transformation while being anticoagulated for cardiopulmonary bypass, or exacerbation or expansion of ischemia due to hypotension during cardiac surgery that is most pronounced within the first few days to weeks following the CVA [ ]. Regardless of timing, in patients with preexisting strokes who undergo surgical repair for IE and then exhibit a change in neurologic status, the possibility of hemorrhagic conversion or worsening cerebral ischemia and/or edema should be ruled out or managed emergently.
Roughly 7% of IE patients develop postoperative renal failure, a rate significantly higher with prosthetic valve compared to native valve endocarditis (11% vs. 4.4%, P = 0.008) [ ]. The kidneys remain at particular risk postoperatively in endocarditis patients, with potential etiologies of AKI and renal failure including preop glomerulonephritis, infarcts from hypoperfusion, and acute interstitial nephritis from antibiotics [ ]. Antibiotics are critical in endocarditis treatment but can be nephrotoxic and must be dosed appropriately, particularly if glomerular filtration rate acutely decreases in the setting of kidney injury. As with preoperative care, maintenance of cardiac output and an adequate blood pressure both protect the kidney from global ischemic injury and acute tubular necrosis. Careful fluid management similarly plays a role. Nevertheless, hemodialysis should be initiated when indications exist.
Additional early noncardiac complications that occur with high frequency regardless of prosthetic or native tissue valve endocarditis include respiratory failure (17%), sepsis (13%), and reoperation for bleeding (6%–9%) [ ]. It goes without saying that in this patient population so prone to complications, multifactorial in origin, all of which carry significant mortality and morbidity, their prompt recognition and aggressive management is paramount so as to improve outcomes and maximize the overall potential benefit of surgical repair.
IVDU is a known risk factor for endocarditis, particularly right-sided endocarditis. As mentioned above, an estimated 90% of isolated right-sided endocarditis and 20% of left-sided endocarditis occur in IVDU patients [ ]. The incidence of endocarditis in IVDUs is estimated to be 1.5 to 3.3 cases per 1000 person-years [ , ]. Kaiser et al. retrospectively compared the postoperative outcomes of IVDU patients versus non-IVDU patients requiring valve replacements for endocarditis [ ]. Operative mortality did not significantly differ—11% in the IVDU patients and 12% in the non-IVDU patients. The perioperative complication rate and age-adjusted long-term survival were also similar between the two groups, but IVDU patients more often require reoperation (17% vs. 5% of non-IVDU patients). Similarly, in a study investigating risk factors for recurrence of endocarditis after valve replacement for native valve endocarditis, Fedoruk et al. found that IVDU/human immunodeficiency virus (HIV) were associated with a hazard ratio of 12.8 for recurrence requiring surgery [ ].
Specific to HIV, a study using data from the Nationwide Inpatient Sample found that the proportion of HIV-positive patients undergoing surgery for endocarditis decreased from 32% to 8% over the study period of 2000–10 [ ]. Importantly, when making decisions on operative management of endocarditis in HIV-positive patients, this same study found that HIV was not an independent predictor of operative mortality.
One specific concern in treating IVDU endocarditis patients is opioid dependence and tolerance, which can make pain management a particular challenge. As outlined above, the care team must recognize these patients’ higher opioid requirements and potential for hyperalgesia. Multimodal analgesic regimens should be utilized, and an acute pain service consulted for guidance if available.
An additional concern in the IVDU population with endocarditis is the potential need for long-term antibiotic administration with outpatient, durable IV access. One solution is a shorter inpatient antibiotic course. Endocarditis is most commonly right-sided in IVDU, and the most common pathogen is Staphylococcus aureus [ ]. In uncomplicated right-sided S. aureus endocarditis, a 2-week inpatient regimen of a beta-lactam, with or without an aminoglycoside, has been shown to be effective [ ]. IVDU patients, however, are also at an increased risk for unusual pathogens such as gram-negative bacilli and polymicrobial infections [ ]. In this context—or in the case of endocarditis with complications—longer parenteral antimicrobial therapy is required, and outpatient therapy is challenging both due to poor compliance frequently seen in this population as well as the use of the IV line for potential illicit drug use. The American Heart Association (AHA) Scientific Statement on Infective Endocarditis, endorsed by the Infectious Diseases Society of America, suggests the following criteria when considering outpatient parenteral antibiotic therapy in an IVDU endocarditis patient: a reliable home support system, easy access to a hospital, regular visits by a home infusion nurse, and regular visits with an experienced physician [ ].
Finally, addiction medicine should be consulted in the care of IVDU patients. The patient should be provided with resources and a referral to a substance abuse disorder program.
The postoperative management of patients with infective endocarditis should encompass both the routine recommendations of postoperative care following valve surgery as well as considerations specific to IE, which includes profound vasoplegia, infectious and embolic sequelae, acute brain injury, and complicated pain management in those opioid-tolerant patients. Given the high rates of morbidity and mortality in this oftentimes intensely ill population, and the variety of complications that they experience, IE patients are best managed actively in an intensive care setting by an experienced, multidisciplinary team, which continues to be involved in their care even once they are transferred to the floor.