Antimicrobial Chemotherapy, 5th Edition
Part 4 - Therapeutic Use of Antimicrobial Agents
Bacteraemia and Endocarditis
The blood is normally sterile in health. However, in disease, micro-organisms invade the bloodstream as either part of the disease process (e.g. malaria), or in the case of blood-borne viruses (e.g. HIV, hepatitis B and C) as a result of the intimate association between an infected target tissue and the blood.
In the case of bacterial infections, the blood may be infected as a consequence of a primary endovascular infection (e.g. infective endocarditis) or more commonly as a result of infection affecting a major organ or other body tissues.
Bacteraemia simply refers to the presence of viable bacteria in the blood. The patient may be completely asymptomatic or present with fever, rigors, tachycardia, shock, and multi-organ failure, sometimes leading to death. When bacteraemia is associated with clinical signs and symptoms it has often been referred to as septicaemia; however, this term is imprecise and should not be used. Bacteraemia is a laboratory finding, the clinical features of systemic response to infection should be described using objectively defined criteria that define sepsis, severe sepsis and septic shock (Chapter 13). The sepsis syndrome is a consequence of the release of cytokines (inflammatory mediators such as tumour necrosis factor, interleukin-1, and interleukin-6) stimulated by microbial products (structural component of organisms, such as lipopolysaccharide, teichoic acid, or exotoxins). The sepsis syndrome can occur in association with any infection and is not confined to patients with bacteraemia.
The distinction between sepsis, sepsis syndrome, and septic shock, including refractory shock, is clinically useful (see Chapter 13) and of prognostic importance. The judicious use of fluid replacement, vasopressors and inotropes, mechanical ventilation and dialysis support in response to target organ failure can be life-saving. However, in addition, it is essential that there be a prompt assessment of the likely source and nature of the triggering infection. Relevant cultures should be collected, including blood cultures.
Any traumatic procedure that facilitates entry of organisms from an infected cutaneous lesion or bacteria-laden mucosal surface may cause bacteraemia (Table 22.1). In addition, invasive infections such as pneumococcal pneumonia, meningitis, or osteomyelitis may be associated with bacteraemia.
In the last 50 years, changes have taken place in the type of organism most frequently encountered. In the pre-antibiotic era Streptococcus pyogenes and Str. pneumoniae accounted for most positive blood cultures and fatalities, but by 1960 Staphylococcus aureus had become dominant. Today, staphylococci and pneumococci are still important, but are outnumbered by Gram-negative bacilli as causes of sepsis and death (Table 22.2). Anaerobes such as Bacteroides fragilis are also encountered more frequently, perhaps because of improved anaerobic techniques.
The increase in bacteraemic infections due to Gram-negative bacilli follows the success of antibiotics in controlling many Gram-positive infections, but advances in medical and surgical expertise have also played an important part: Gram-negative sepsis is common in patients undergoing intra-abdominal surgery or aggressive immunosuppressive therapy and invasive procedures, and in those whose normal defences are already compromised by underlying disease. These infections are mostly hospital acquired, and because of the widespread use of antibiotics the infecting organisms are often multiresistant.
Vascular catheters are widely used in medical management and have resulted in an increase in bacteraemia caused by Gram-positive cocci, notably Staph. aureus and Staph. epidermidis. Polymicrobial bacteraemia and recurrent bacteraemia have also become more common in recent years.
Table 22.1 Procedures that may produce transient bacteraemia
Table 22.2 Distribution of organisms in 4856 episodes of bacteraemia (1988-1998) at University Hospital, Nottingham
Bacteraemia may be transient (lasting for several minutes) intermittent or continuous (lasting for several hours to days). The danger of transient bacteraemia depends on the host and the organism. Thus, transient bacteraemia due to viridans streptococci after dental extraction is of no consequence in an otherwise healthy individual, but in those patients with abnormal heart valves it may produce endocarditis (see below). Staph. aureus may localize in the metaphyses of long bones in children or the vertebrae in adults and lead to osteomyelitis. Transient bacteraemia with Gram-negative bacilli following instrumentation of an infected urinary tract may produce rigor and fever.
Continuous bacteraemia is the hallmark of intravascular infection and also occurs in infections in patients with neutropenia, overwhelming sepsis, acute haematogenous osteomyelitis, and infections with intracellular organisms such as Salmonella enterica serotype Typhi andBrucella spp. during the early stage of the illness.
Most other bacteraemias are intermittent and are characteristic of abscesses and certain types of chronic infection such as meningococcal or gonococcal sepsis or brucellosis.
Laboratory investigation and antibiotic therapy
There are no specific clinical findings that are diagnostic of bacteraemia or fungaemia or, for that matter, that differentiate between Gram-negative and Gram-positive sepsis. Hence the importance of blood cultures so that the pathogen is identified and specific therapy instituted as soon as possible. Mortality in patients with shock is over 50%. If death is to be prevented and shock avoided, the clinician must react promptly to the early signs of sepsis with appropriate ‘best-guess’ parenteral therapy.
Most episodes of bacteraemia are intermittent, hence the importance of more than one set of blood cultures before starting antibiotics. Ideally, at least two sets should be taken from separate venepunctures at intervals of a few minutes to a few hours (depending on the clinical urgency). Since bacteraemias are usually low-grade, the volume of blood drawn at each venepuncture is important: in adults at least 10 ml should be taken; in infants and young children 1-3 ml. Specimens from other likely foci of infection must also be sent to the laboratory; these may include urine, sputum, cerebrospinal fluid, pus, pleural or joint fluids, etc. Often a Gram-stained smear of a specimen from the presumed site of infection provides an early clue on which the choice of best-guess therapy can be based.
When bacteraemia is secondary to a focus of infection that is readily apparent or clinically suspected it is usually possible to predict the most likely organisms from the clinical presentation. The selection of antibiotics is guided by the suspicion of the focus of infection and by whether the infection was community or hospital acquired. The most appropriate antibiotic or combination can then be chosen in the light of local knowledge of resistance patterns. The antibiotic regimen can be changed later on the basis of bacteriological findings.
Certain groups of patients give no clue as to the primary focus of infection or its likely source. Such episodes of ‘primary’ bacteraemia of unknown source may occur in neonates, immunocompromised patients, and rarely, normal individuals.
The newborn baby is comparatively more susceptible to bacterial invasion of the bloodstream, but the recognition and localization of infection may be difficult because the manifestations are frequently non-specific. However, it is imperative that the diagnosis is made early, specimens collected, and antibiotic treatment started at once.
The initial empirical choice in a neonate with ‘early-onset’ (< 7 days) sepsis is usually a combination of benzylpenicillin and gentamicin, which covers the two most common organisms, Escherichia coli and group B haemolytic streptococci (Str. agalactiae) as well as other streptococci, many other Gram-negative bacteria and Listeria monocytogenes. If, however, there is an obvious staphylococcal skin infection, flucloxacillin should be substituted for benzylpenicillin.
Empirical treatment for ‘late-onset’ (> 7 days) sepsis varies according to the clinical setting. The combination of cefotaxime and gentamicin should be considered in neonates who are ventilated, known to be colonized with enterobacteria, or have had previous exposure to antibiotics. Staph. epidermidis is the commonest isolate in patients with infection associated with intravenous catheters and vancomycin is the only reliable agent against these organisms, so that a combination of vancomycin with gentamicin or cefotaxime would be a reasonable choice in neonates in whom catheterassociated sepsis is strongly suspected.
Bacteraemia is common in immunocompromised patients, especially those with haematological malignant disease complicated by profound neutropenia, mucosal ulcerations, and administration of corticosteroid, cytotoxic, and immunosuppressive drugs. Despite the fact that blood cultures from these patients yield Gram-positive cocci more frequently than Gram-negative bacilli, it is the Gram-negatives, in particularPseudomonas aeruginosa, that are feared, for if left untreated most patients will die within 48 h.
It is customary to provide a broad-spectrum synergistic combination of antibiotics as initial empirical therapy. Widely used regimens include an aminoglycoside (e.g. gentamicin) together with either an antipseudomonal penicillin (e.g. piperacillin/tazobactam) or an expanded spectrum cephalosporin with antipseudomonal activity (e.g. ceftazidime). If blood cultures are positive modifications to the regimen are made.
If fever persists for more than 72 h despite broad-spectrum antibiotics and blood cultures remain negative, the patient should be reassessed and repeat blood cultures obtained. Vancomycin or teicoplanin may be added to the regimen of those who have vascular catheters in situ. Empirical antifungal therapy with amphotericin B should be considered in selected patients who remain febrile and neutropenic for 7 days despite broad-spectrum antibiotics.
Primary bacteraemia in previously healthy individuals is rare. The most common type seen the UK is that due to Neisseria meningitidis. In infants, children or young adults it can produce a fulminating sepsis with petechial rash progressing to shock and death in a matter of hours. Mortality can be as high as 30%. General practitioners suspecting this condition should give benzylpenicillin immediately before transferring the patient to hospital. Since it is unlikely that subsequent blood or cerebrospinal fluid cultures from these patients would be positive, a throat or pernasal swab should be obtained and a special request made for the isolation of N. meningitidis.
Str. pneumoniae occasionally causes bacteraemia in an otherwise healthy, but febrile child. Those aged 6 months-2 years are most at risk. Although such bacteraemias may resolve spontaneously, a few patients remain ill and some develop severe disease, including meningitis. Antibiotic therapy with amoxicillin is warranted. Other important causes of primary bacteraemia include Salmonella enterica serotypes Typhi and Paratyphi (see p. 300), and, rarely, Brucella spp.
Management of septic shock
Septic shock is characterized by hypotension, decreased systemic vascular resistance, myocardial depression, maldistribution of blood flow, and multi-organ system failure. The management of septic shock and severe sepsis is based on three basic principles:
- to eradicate the source of infection with appropriate antibiotics and drainage or debridement of the septic focus, whenever possible;
- to stabilize the patient haemodynamically with the administration of intravenous fluids, oxygen, inotropic agents, and vasopressors;
- to provide organ support using mechanical ventilation and dialysis as necessary.
Even when managed aggressively in intensive care units, the mortality in those with multi-organ failure stage is depressingly high (about 50%). Since the manifestations of sepsis are associated with the release of cytokines by macrophages in response to microbial products, there has been much interest in the possibility that interference with bacterial or host mediators of inflammation might act as a beneficial adjunct to antimicrobial therapy in septic patients.
Monoclonal antibodies against lipid A of Gram-negative bacteria and against various cytokines, such as interleukin-1 and tumour necrosis factor, have been subjected to therapeutic trial. To date, none has shown clinical benefit. Most recently, recombinant activated Protein C (drotrecogin alfa) has shown some benefit in reducing mortality in selected patients with septic shock and organ dysfunction. The pathophysiology of sepsis is extremely complex, suggesting that no single strategy is likely to prove effective in all patients.
Endocarditis is inflammation of the endocardial surface of the heart. When caused by micro-organisms it is known as ‘infective’ endocarditis, and may be caused by bacteria including, rickettsiae, chlamydiae, and also fungi. Endocarditis usually affects the heart valves but may involve the adjacent endocardium. The terms acute and subacute endocarditis originated in the pre-antibiotic era when all patients with endocarditis died. Those who died in less than 6 weeks due to infection of normal valves by virulent organisms such as Staph. aureus, Str. pneumoniae, or N. gonorrhoeae were said to have acute bacterial endocarditis. In contrast, those who suffered a more indolent course due to infection of abnormal valves by relatively avirulent organisms (e.g. viridans streptococci), died much later and were said to have subacute bacterial endocarditis.
Nowadays, most patients with infective endocarditis are cured provided the diagnosis is made, and treatment with appropriate antibiotics begun sufficiently early. It is also more useful to classify endocarditis according to the infecting organism and the underlying site of infection (e.g. Staph. aureus tricuspid endocarditis). In addition, distinguishing native from prosthetic valve infection is also important. These definitions are of value in predicting the probable course of the disease and also have therapeutic implications with regard to the antibiotic regimen to be used.
Infective endocarditis affects about 2000 people per year in England and Wales and has a mortality of 15-30%. In recent years the spectrum of recognized underlying cardiac lesions in endocarditis in adults has changed as a result of a decline in rheumatic heart disease (in the developed world), the increase of endocarditis complicating degenerative heart disease and improvement in diagnostic techniques (echocardiography). The following trends have been noted:
- The mean age of the patient has increased; it is now over 50 years of age for the following reasons:
- people with congenital heart disease or rheumatic heart disease survive longer because of advances in medical and surgical expertise to correct valve dysfunction;
- increased life expectancy is associated with a raised incidence of degenerative valve disease; almost 30% of elderly patients who develop endocarditis do not have a pre-existing cardiac condition. Minor degenerative changes produce valvular lesions that serve as a nidus for infection;
- infectious complications of genitourinary and gastrointestinal disease predispose to bacteraemia in elderly people.
- Acute Staph. aureusendocarditis is an important complication of intravenous drug-abusers.
- Mitral valve prolapse with regurgitation predisposes to endocarditis and is now recognized more frequently.
- Prosthetic valve endocarditis has increased in proportion to cardiac valve surgery.
- The ‘classic’ physical signs of subacute bacterial endocarditis are seen in fewer patients as a result of earlier diagnosis.
Infective endocarditis is the consequence of several events (Fig. 22.1):
- haemodynamic or disease associated damage to the endothelial surface of the valve;
- deposition of platelets and fibrin on the edges of the valve, initially resulting in the formation of sterile or non-bacterial thrombotic vegetation;
- colonization by micro-organisms transiently circulating in the blood to produce an infected vegetation.
Transient bacteraemia is common. A wide variety of trivial events (e.g. chewing and tooth-brushing) can induce bacteraemia with oral streptococci. Some 85% of cases of streptococcal endocarditis cannot be related to any medical or dental procedure. To cause endocarditis, organisms must also be able to survive natural complement-mediated serum bactericidal activity and adhere to thrombotic vegetations. Certain streptococci produce extracellular dextran, which promotes adherence to fibrin-platelet vegetations. These strains cause endocarditis more frequently than non-dextran producing streptococci. However, organisms such as enterococci and Staph. aureus that do not produce dextran are also important causes of endocarditis. In these cases host proteins such as fibronectin and fibrinogen may mediate adherence.
Fig. 22.1 Pathogenesis of infective endocarditis.
Once colonization occurs there is rapid deposition of additional layers of platelets and fibrin over and around the growing colonies, causing the vegetation to enlarge. Within 24-48 h marked proliferation of bacteria occurs, leading to dense populations of organisms (109-1010bacteria/g tissue). Micro-organisms deep within the vegetations are often metabolically inactive, whereas the more superficial ones proliferate and are shed continuously into the bloodstream. Fresh vegetations are composed of colonies of micro-organisms in a fibrin-platelet matrix with very few leucocytes.
Any organism can cause infective endocarditis, but streptococci and staphylococci account for more than 90% of culture-positive cases. However, the frequency with which various organisms are involved differs not only for the type of valve that is infected (native or prosthetic) but also with the causative event (e.g. dental manipulation, intravenous drug abuse, or a hospital-acquired infection) (Table 22.3).
Streptococci account for about 65% of all cases of native valve endocarditis. Most common of all are the ‘viridans streptococci’, which include Streptococcus mitis, Str. sanguis, Str. mutans, the Str. milleri group, and Str. salivarius, which are mouth commensals; most are highly sensitive to penicillin and cause infections primarily on abnormal heart valves. Str. bovis is an important cause of endocarditis in elderly people and is frequently associated with bowel pathology, notably colonic polyps and carcinoma. Recovery of this organism should prompt investigation for colonic disease.
Enterococci are gut streptococci and cause 10% of cases. Haemolytic streptococci of Lancefield groups B and G are occasionally incriminated in endocarditis. Diabetic patients are particularly at risk of group B infections.
Staphylococci account for 25% of cases of native valve endocarditis, but over 90% is due to Staph. aureus, which is the leading cause of acute endocarditis. The course is frequently fulminant with widespread metastatic abscesses and death in about 40% of cases. The organism can attack normal or damaged valves and cause rapid destruction of the affected valves. Surgery is often required. Staph. epidermidis, in contrast, causes an indolent infection on previously damaged valves.
Table 22.3 Aetiological agents in infective endocarditis and their approximate frequency
Most commonly encountered are the fastidious, slow-growing Gram-negative bacilli of the ‘HACEK’ group (Haemophilus spp., Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, and Kingella spp.)
Gram-negative enteric bacteria rarely cause endocarditis, except in intravenous drug abusers and patients with prosthetic valves. However, salmonellae have an affinity for abnormal cardiac valves and aneurysms of major vessels.
These organisms seldom cause native valve endocarditis. Risk factors include major underlying illnesses, prolonged courses of broad-spectrum antibiotics, corticosteroids or cytotoxic agents, and a central venous line in situ for a considerable length of time. Fungal endocarditis is more often seen in intravenous drug abusers or after reconstructive cardiovascular surgery. Candida and Aspergillus species are usually implicated. The course is indolent but grave. Large vegetations frequently embolize, occluding major vessels in the lower extremities. Culture of material obtained at embolectomy may yield the offending organism when blood cultures are negative.
Prosthetic valve endocarditis
Endocarditis complicating prosthetic valve or other devices is divided into ‘early’ and ‘late’ onset disease. Early-onset disease usually reflects contamination during the peri-operative period. Despite prophylactic antibiotics, staphylococci account for 50% of all cases; Staph. epidermidis is more common than Staph. aureus. Early-onset infection is a serious complication and is often associated with valve dehiscence, a fulminant course, and a high mortality. Late-onset prosthetic valve endocarditis occurs after the valve has become endothelialized. The source of infection, as in native valve endocarditis, is seeding of the valves following transient bacteraemia and viridans streptococci again become the commonest organism. Late-onset disease caused by Staph. epidermidis, diphtheroids, or other organisms of the early-onset type may just reflect a delayed manifestation of infection acquired in the peri-operative period.
Infective endocarditis in intravenous drug abusers
The skin is the commonest source of micro-organisms responsible for infective endocarditis in intravenous drug abusers, although contaminated drug and syringes are other possibilities. Staph. aureus is the predominant cause, but other organisms including Pseudomonasspp., group A haemolytic streptococci, other streptococci, and fungi are also important. Endocarditis often involves the tricuspid valve, especially when Staph. aureus is the causative agent.
Blood cultures must be obtained from all patients with fever and heart murmur before contemplating antibiotic therapy, irrespective of the initial diagnosis. Bacteraemia is usually low grade, but continuous, so it does not really matter when blood cultures are obtained. In the absence of previous antimicrobial therapy blood cultures are positive in more than 90% of cases.
At least 10 ml of blood should be withdrawn at each venepuncture and divided equally into two blood culture bottles (one set). Strict attention should be paid to skin preparation and aseptic technique. At least three sets of blood cultures are obtained from three separate venepunctures over a period of time. Staph. epidermidis and diphtheroids are important causes of endocarditis, as well as common blood culture contaminants from the skin, and isolation of the same relatively ‘avirulent’ organism repeatedly in the absence of an intravascular catheter is highly suggestive of endocarditis.
The interval between each venepuncture depends on the clinical urgency. In acute cases, when antimicrobial therapy should be commenced promptly, three sets from separate venepunctures should be taken at intervals at least 30 min apart before the start of therapy. If there is no urgency, then the three sets can be taken over a 24-h period. All blood culture bottles must be taken to the laboratory and incubated for up to 3 weeks to cater for fastidious organisms.
Blood culture may be persistently negative in some patients with suspected endocarditis. Likely explanations are:
- prior administration of antibiotics (the most common cause);
- infection with fastidious organisms including those of the HACEK group, Coxiella burnetii, Chlamydia psittaci,or Brucella spp.;
- infection with Candidaspp. in which only 50% of blood cultures may be positive;
- infection of the right side of the heart, which is occasionally accompanied by negative blood cultures;
- cardiac disease other than infective endocarditis (e.g. left atrial myxoma).
When blood cultures are negative, paired samples of sera (one taken on admission and another 10-14 days later) should be examined for antibodies against other infective agents of endocarditis.
Echocardiography (transthoracic and transoesophageal) is key to the diagnosis, assessment, and management of patients with suspected infective endocarditis, but negative results do not exclude the diagnosis, especially in those with prosthetic valves.
General principles of therapy
The chief aims of management are to sterilize the vegetation and to ensure that relapse will not occur.
In endocarditis, organisms reach extremely high densities within the vegetation and are encased in layers of fibrin, where they are free to divide without interference from phagocytic cells or humoral defences. Hence, bactericidal antibiotics are essential to sterilize the vegetation. Bacteristatic agents such as tetracycline or chloramphenicol may produce a symptomatic response, but once discontinued, relapse is common and they should not be used. The most commonly used bactericidal agents are the penicillins, in particular benzylpenicillin. In penicillin-hypersensitive patients vancomycin or a cephalosporin are suitable alternatives.
Route and duration of therapy
Parenteral high-dose bactericidal therapy is essential to guarantee the penetration of relatively avascular vegetations. The duration of therapy is determined by the nature (native or prosthetic valve) and site of the infection and by the susceptibility of the target pathogen. Most patients with endocarditis are cured by 4 weeks of treatment; some may require treatment for 6 weeks or more, although a selected group may be cured in only 2 weeks. Shorter courses are associated with relapse. Oral treatment is recommended only after initial parenteral therapy, and if patient compliance is assured, for the last 2 weeks of treatment of selected cases of native valve endocarditis.
Synergistic combination therapy
Aminoglycosides, though generally bactericidal, have little activity against streptococci and cannot be used alone. However, the combination of gentamicin with penicillin is synergistic and produces a more rapid and complete bactericidal effect than is obtained with penicillin alone. It is therefore common practice to recommend combination therapy in the initial stages of management of infective endocarditis.
Laboratory control of antibiotic therapy
Determination of minimal inhibitory concentration (MIC) and minimal bactericidal concentration
Standard routine sensitivity testing is not recommended in the laboratory management of endocarditis. Instead, the MIC and minimal bactericidal concentration of the antibiotics to be used, for the organism isolated, should be determined. In difficult cases tests for antibiotic synergy may also be required for optimal combination therapy.
When aminoglycosides (usually gentamicin) are used to treat endocarditis, a serum concentration lower than that considered therapeutic for Gram-negative infections is adequate, thus lessening the potential for toxicity. Patients with normal renal function should receive a loading dose appropriate to the age and body weight, followed by a maintenance dose. The serum concentration should be periodically monitored and the dose adjusted accordingly (a pre-dose concentration <1.0 mg/l; post-dose 4-5 mg/l is adequate).
Specific antimicrobial regimens
Isolates highly sensitive to penicillin (MIC < 0.1 mg/l)
This includes most viridans streptococci, Str. bovis, and other streptococci. Viridans streptococci are highly sensitive to penicillin and a 99% cure rate can be achieved with a 4-week regimen of benzylpenicillin alone. Such a regimen is recommended for the treatment of uncomplicated native valve endocarditis in elderly patients, or those with impaired renal function, in whom aminoglycosides are best avoided.
For uncomplicated native valve endocarditis it is common practice in the UK to give 2 weeks of combination therapy with high-dose benzylpenicillin together with an aminoglycoside and then to consider oral amoxicillin for the last 2 weeks if patient compliance can be guaranteed. Patients with prosthetic valve endocarditis should be treated for longer to ensure cure; 6 weeks is recommended.
Isolates relatively resistant to penicillin (MIC 0.1-0.5 mg/l)
This category includes some viridans streptococci such as the ‘nutritionally variant’ streptococci. Viridans streptococci that are relatively resistant to penicillin are increasingly encountered. The relapse rate in endocarditis caused by ‘nutritionally variant’ streptococci is high, even when 2 weeks of combination therapy is followed by 2 further weeks of benzylpenicillin. Endocarditis caused by such strains or other streptococci that are relatively resistant to penicillin is best treated with high-dose benzylpenicillin and gentamicin for 4 weeks, with appropriate monitoring of serum gentamicin levels.
Isolates resistant to penicillin (MIC > 0.5 mg/l)
Examples are Enterococcus faecalis, Ent. faecium, and other streptococci. Enterococcal endocarditis is the third most common type of endocarditis and is among the most difficult to treat. Mortality is about 20% and relapses are not uncommon. Although penicillin, ampicillin, and vancomycin inhibit the growth of enterococci they are not bactericidal for most strains, and therapy with these agents alone results in a high relapse rate. For a bactericidal effect, it is usually necessary to add an aminoglycoside; this results in marked enhancement of killing. A combination of penicillin or ampicillin with an aminoglycoside is the treatment of choice for enterococcal endocarditis; gentamicin is the preferred aminoglycoside. High-level resistance to gentamicin among enterococci is becoming more common and in-vitro testing (with a disc containing 100 mg of gentamicin) should be a routine procedure in all isolates of enterococci.
Ent. faecium is generally more resistant to β-lactam antibiotics than Ent. faecalis; β-lactamase-producing Ent. faecalis strains have also been reported. Such patients are best treated with vancomycin and gentamicin. Enterococci are uniformly resistant to all cephalosporins. Patients with enterococcal endocarditis should receive at least 4-6 weeks of combination therapy.
In about one-third of patients with Staph. aureus endocarditis there is no evidence of pre-existing valvular heart disease. The infection results in rapid and severe valvular destruction and a mortality of about 40% even with appropriate treatment.
Patients over 50 years of age with Staph. aureus endocarditis secondary to infected intravascular devices have the highest mortality rate. Many require surgery to replace the infected valve, because of valvular dysfunction, dehiscence, and myocardial abscesses. In contrast,Staph. aureus endocarditis involving the tricuspid valve in intravenous drug abusers is much easier to cure and carries a mortality below 10%.
Since the vast majority of Staph. aureus strains produce a β-lactamase that destroys penicillin, the initial choice is a penicillinase-stable penicillin such as flucloxacillin. However, the choice and duration of treatment with a synergistic agent (e.g. gentamicin) remains controversial. Flucloxacillin plus gentamicin is associated with a more rapid clearance of bacteraemia. In the UK, low-dose gentamicin is widely used with flucloxacillin for both native valve and prosthetic valve endocarditis, at least for the first 2 weeks of therapy.
In selected patients, such as an intravenous drug abuser with a right-sided endocarditis, or a patient with uncomplicated native valve endocarditis, who has responded fully to 2 weeks of combination therapy, oral flucloxacillin may be considered for the remaining 2 weeks of therapy. The remaining patients should receive high-dose flucloxacillin intravenously for at least 4 weeks.
Table 22.4 Recommended antibiotic treatment regimens for streptococcal endocarditis
Table 22.5 Recommended antibiotic treatment regimens for staphylococcal endocarditis
If the patient is allergic to penicillin or the Staph. aureus is multiresistant, then vancomycin should be used. Rifampicin, a very potent antistaphylococcal agent, should be added in difficult cases but is never used alone, owing to the emergence of resistance.
This organism rarely infects a native valve, but is a common cause of both early and late onset prosthetic valve endocarditis. It is difficult to cure with antibiotics alone, and surgery is often required, particularly in patients with early-onset endocarditis. Isolates are frequently resistant to flucloxacillin, which must not be used unless the isolate is confirmed to be sensitive. Therapy must therefore be started with vancomycin and rifampicin. Gentamicin may be added for the first 2-3 weeks, if the strain is sensitive. Frequent monitoring of gentamicin levels is mandatory to minimize the risk of toxicity. If the organism is indeed sensitive to flucloxacillin then vancomycin is stopped and flucloxacillin and rifampicin are continued.
Table 22.6 Recommended antibiotic treatment regimens for endocarditis other than that caused by streptococci and staphylococci (doses are for adult with normal renal and hepatic function)
Recommended antibiotic regimens for streptococcal and staphylococcal endocarditis are shown in Tables 22.4 and 22.5 respectively.
Recommended antibiotic regimens for endocarditis caused by Gram-negative bacilli and other organisms are shown in Table 22.6. Culture-negative cases are occasionally caused by Coxiella burnetii or Chlamydia psittaci and may be detected by serology.
In clinically suspected acute endocarditis, it is prudent to start treatment with a combination of benzylpenicillin, flucloxacillin, and gentamicin and to modify the regimen appropriately once the causative organism has been identified.
Emergency valve replacement in patients with infective endocarditis is an important adjunct to medical therapy. In selected patients, it is a life-saving procedure at any stage of the disease. The major indications for surgical intervention include:
- refractory heart failure related to structural valvular damage;
- myocardial or perivalvular abscess;
- untreatable or uncontrolled infection (e.g. infection of a prosthetic valve with fungi or Gram-negative bacilli);
- repeated relapses with a difficult organism (e.g. Ent. faecium);
- multiple embolic episodes.
Infective endocarditis remains a life-threatening infection. The prognosis varies according to the infecting micro-organism, the type of cardiac valve (native versus prosthetic and aortic versus mitral versus tricuspid), the age of the patient, and the presence or absence of complications. Mortality is lowest in viridans streptococcal endocarditis of the native valve and highest in early-onset prosthetic valve endocarditis.
Patients who recover from an episode of infective endocarditis carry a lifelong risk of a further attack. It is important that they maintain high levels of dental hygiene supplemented by regular dental reviews. Antibiotic prophylaxis plays an essential part in the risk of bacteraemia associated with many dental procedures (see p. 240).