Evaluation of a Paradigm Shift From Intravenous Antibiotics to Oral Step-Down Therapy for the Treatment of Infective Endocarditis: A Narrative Review

Abstract

IMPORTANCE

The requirement of prolonged intravenous antibiotic courses to treat infective endocarditis (IE) is a time-honored dogma of medicine. However, numerous antibiotics are now available that achieve adequate levels in the blood after oral administration to kill bacteria. Moreover, prolonged intravenous antibiotic regimens are associated with high rates of adverse events. Accordingly, recent studies of oral step-down antibiotic treatment have stimulated a reevaluation of the need for intravenous-only therapy for IE.

OBSERVATIONS

PubMed was reviewed in October 2019, with an update in February 2020, to determine whether evidence supports the notion that oral step-down antibiotic therapy for IE is associated with inferior outcomes compared with intravenous-only therapy. The search identified 21 observational studies evaluating the effectiveness of oral antibiotics for treating IE, typically after an initial course of intravenous therapy; none found such oral step-down therapy to be inferior to intravenous-only therapy. Multiple studies described an improved clinical cure rate and an improved mortality rate among patients treated with oral step-down vs intravenous-only antibiotic therapy. Three randomized clinical trials also demonstrated that oral step-down antibiotic therapy is at least as effective as intravenous-only therapy in right-sided, left-sided, or prosthetic valve IE. In the largest trial, at 3.5 years of follow-up, patients randomized to receive oral step-down antibiotic therapy had a significantly improved cure rate and mortality rate compared with those who received intravenous-only therapy.

CONCLUSIONS AND RELEVANCE

This review found ample data demonstrating the therapeutic effectiveness of oral step-down vs intravenous-only antibiotic therapy for IE, and no contrary data were identified. The use of highly orally bioavailable antibiotics as step-down therapy for IE, after clearing bacteremia and achieving clinical stability with intravenous regimens, should be incorporated into clinical practice.

Adogma in medicine has long been that effective treatment of infective endocarditis (IE) requires prolonged intravenous antibiotic therapy. This dogma developed during an era when penicillin was the most effective antibiotic to treat IE, for which it was only administered intravenously because its oral absorption was regarded as unreliable. The underpinning of this dogma was the rationalization that a high number of bacteria are “buried” deep within cardiac vegetations, limiting both antimicrobial access and an effective host inflammatory response.

The first truly effective antibacterial agents were the sulfonamides, which became available in the mid-1930s.¹ Early experiences with oral sulfonamides in the treatment of IE were disappointing. A 1943 pre-post study reported a 99% mortality rate among 2596 patients with IE prior to sulfa drugs compared with a 96% mortality rate among subsequent patients with IE treated with oral sulfa drugs.² Other case series reported similar mortality rates.³ When parenteral penicillin G became available in the 1940s and achieved better cure rates among patients with IE (from 1% before parenteral penicillin G to 85% after its introduction), it quickly became the standard of care for this infection.³

Nevertheless, interest in the possibility of using oral therapy for IE continued owing to the difficulties inherent in prolonged intravenous therapy. When oral tetracyclines and macrolides became available in the late 1940s to early 1950s, they too were administered to patients with IE, also with unfavorable outcomes (eg, cure rates of <30% in case series).⁴

Oral formulations of penicillin became available in the mid-1950s but were not viewed as reliable options for IE therapy, given concerns about bioavailability and the preceding high rates of failure of other oral agents. In 1954, a preeminent authority of infectious diseases, Maxwell Finland, published a 2-part review article on the treatment of IE.^3,4 He wrote, “in this disease, oral administration…has generally been discarded as inadequate. Presumably, the oral route is at times successful…it is more likely, however, that such usage is responsible for many therapeutic failures…However, little of this type of experience is recorded, and therefore this assumption cannot be authenticated.”^3(p373) Thus, the general sense that oral antibiotic therapy was unlikely to be effective for IE dates back to the 1950s.

Conceptual Framework for Evaluating Reasonableness of Oral Antibiotics for IE

The conceptual framework for considering oral therapy for IE rests on several points. First, bacteria are unable to recognize by which route antibiotics are administered to the patient they are infecting. Bacteria are affected by the concentration of the antibiotic that arrives at the site of infection, and whether that concentration is sufficiently high to kill them. Thus, the real question is whether antibiotics administered by the oral route are absorbed at levels sufficient to kill bacteria at the site of infection.

Second, older oral tetracyclines, macrolides, and sulfonamides achieve levels in the blood and tissue that are likely below those necessary to kill many bacteria that cause IE (Table 1).^5–15 In contrast, some modern oral antimicrobial agents achieve levels above those necessary to kill IE pathogens in the blood (Table 1).^5–15 Consistent with these pharmacokinetic observations, numerous randomized clinical trials (RCTs) have found that these agents can sterilize the blood in patients with bacteremia caused by gram-positive cocci, affirming that, with proper dosing, the levels achieved are adequate to kill bacteria in the blood.^16–22

Table 1.

Peak Blood Levels vs MICs Achieved by Antibiotics Used to Treat Infective Endocarditis in Published Studies

Oral drug	Peak blood level, μg/mLa	MIC90, μg/mLb	Peak blood level to MIC90 ratio
Antibiotics with peak blood level to MIC90 ratio ≤1
Tetracycline, 250 mg	1	≥4	0.125
Erythromycin, 500 mg	0.5	≥4	0.125
Sulfanilamide, 4000 mg5–9	50	50–70	0.8
Antibiotics with peak blood level to MIC90 ratio >1
Penicillin V, 500 mg, for Streptococcus spp	5	1	5
Amoxicillin, 1000 mg, for Streptococcus spp	10	1	10
Levofloxacin, 750 mg, for Staphylococcus spp	9	4	2.25
Moxifloxacin, 400 mg, for Staphylococcus or Streptococcus spp
Staphylococcus aureus	4	4	1
Streptococcus spp	4	0.25	16
Rifampin, 600 mg, for gram-positive cocci	7	1	7
TMP-SMX, 320 mg/1600 mg, for Staphylococcus spp	100	4.75	22
Linezolid, 600 mg, for gram-positive cocci	15	2	7.5
Clindamycin, 600 mg, for Staphylococcus spp	10	2	5

Abbreviations: MIC, minimum inhibitory concentration; MIC₉₀, minimum inhibitory concentration of the antibiotic needed to kill 90% of clinical isolates encountered; spp, species; TMP-SMX, trimethoprim-sulfamethoxazole.

^aData on peak blood levels are from Table 9A in The Sanford Guide to Antimicrobial Therapy 2019¹⁰; data on TMP-SMX are from the sulfamethoxazole component.

^bData from the JMI Laboratories SENTRY program and other studies of both staphylococci and streptococci.^11–15 Data on MIC₉₀ of penicillin and amoxicillin are shown only for susceptible or intermediate streptococci, not including resistant strains or staphylococci. These results are shown for comparative purposes only. Pharmacodynamic considerations for antibiotic effectivenness are based on more than just peak levels (eg, time above MIC is associated with the effectiveness of β-lactams, and total exposure over a 24-hour period divided by MIC is associated with the effectiveness of most of the other agents). However, if the peak level never exceeds the MIC, then the drug cannot achieve any time above the MIC.

Third, oral therapy avoids prolonged intravenous catheterization, which is associated with a 10% to 60% rate of dangerous, potentially severe adverse events, including deep venous thrombosis, central line infection, catheter fracture and/or migration, and central venous stenosis.^23–27 Intravenous therapy is also more expensive and labor intensive and usually mandates a longer inpatient stay, with additional risks. Therefore, if data support the relative clinical effectiveness of oral vs intravenous therapy for IE, oral therapy would be reasonable to consider. With this framework in mind, we evaluated the medical literature to determine whether there was compelling evidence for the superiority of prolonged intravenous therapy vs oral step-down therapy for IE.

Methods

In October 2019, we conducted a review of the literature for studies of oral antibiotic therapy for IE (with an update of the search in February of 2020). We searched PubMed for title words “endocarditis” and “oral” (“endocarditis [ti] AND oral [ti]”). A total of 169 articles were identified. References within these articles, including in 5 review articles,^3,4,28–30 were also evaluated to identify other relevant literature. We excluded studies of the pathogenesis or epidemiology of IE, reports of single patient outcomes, studies of prophylaxis for dental or other procedures, editorials or letters to the editor commenting on other studies, review articles, and studies of oral agents other than antibiotics (eg, opiates and anticoagulation). We also excluded 2 studies in which linezolid was associated with high cure rates among 84 patients with IE, but it was not clear what proportion of patients received oral vs intravenous linezolid.^31,32 We ultimately included 21 observational or quasi-experimental studies and 3 RCTs, all of which focused on the effectiveness of antibiotics administered orally for part of the therapeutic course for patients with IE (for a total of 24 publications).

Results

Observational Data

Consistent with the pharmacokinetic profiles summarized in Table 1,^5–15 6 small, uncontrolled case series (total, 66 patients) published in the 1950s to 1960s reported that, in contrast to older sulfa, tetracycline, or macrolide agents, oral penicillins were associated with high cure rates for streptococcal, enterococcal, and gonococcal IE (eTable in the Supplement).^33–38 Cure rates ranged from 78% to 100%. For some of the patients in these studies, parenteral amino-glycosides were added for partial courses of therapy, and for some patients, adjunctive probenecid was used (which prolongs the serum half-life of penicillin).

We also identified 15 uncontrolled, observational studies of oral therapy for the treatment of IE published since 1975, using more modern therapeutic regimens (for total of >1000 patients, of whom >500 were treated orally).^39–53 In 12 uncontrolled case series of fewer than 50 patients each, high-dose oral β-lactams, oral ciprofloxacin plus rifampin, or oral linezolid monotherapy regimens achieved cure rates of 77% to 100% among patients of various ages, who had right-sided and/or left-sided native or prosthetic valve IE, mainly caused by traditional IE pathogens (ie, streptococci, staphylococci, and enterococci; eTable in the Supplement).^41–52

Three controlled, observational studies compared the out-comes of patients whose treatment began with intravenous therapy and then was switched to oral “step-down” therapy with the outcomes of patients who were treated with intravenous antibiotics for their entire course (Table 2).^39,40,53 Demonchy et al⁵³ reviewed the experience of 66 French patients with IE, of whom 19 with left-sided IE were switched to oral antibiotic therapy after a mean of 18 days of intravenous therapy. The mean (SD) age of these patients was 62 (17) years, and they had a large variety of comorbidities. Fifty-one patients (77.3%) had left-sided IE, and 9 (13.6%) had intracardiac devices. All 19 patients who were switched to oral therapy had experienced complications from the prior intravenous therapy. The infecting pathogens included methicillin-susceptible Staphylococcus aureus (MSSA; n = 8), methicillin-resistant S aureus (MRSA; n = 4), and oral streptococci (n = 7). Treatment regimens included linezolid (n = 3), fluoroquinolones (n = 3), amoxicillin (n = 4), or combinations of fluoroquinolones, amoxicillin, or clindamycin plus rifampin (n = 9). Patients who were switched to oral therapy had numerically, but not statistically significantly, lower mortality rates (0% vs 21%; P = .05).

Table 2.

Nonrandomized, Controlled Studies of Oral Step-Down Therapy for the Treatment of IE

Source	Design	Findings	Mortality, No./total No. (%)
			Oral step-down therapy	Intravenous therapy	P value
Demonchy et al,53 2011	Retrospective case-control study	66 Patients total: 51 with left-sided IE, 3 with right-sided IE, 9 with devices, and 3 with unspecified IE Mean age 62 y, extensive comorbidities 11 Patients infected with Staphylococcus aureus, 9 with coagulase-negative staphylococci, 25 with streptococci, 2 with enterococci, and 7 with other organisms 19 Patients were switched to oral therapy, all had had complications on intravenous therapy, 12 had left-sided IE Switch to oral therapy occurred after a mean (SD) of 18 (9) d 3 Patients treated with oral linezolid; 3 with oral fluoroquinolones; and 6 with combinations of oral fluoroquinolones, rifampin, or clindamycin; all patients infected with streptococci were treated with amoxicillin	0/19 (0)	10/47 (21)	.05
Mzabi et al,40 2016	Retrospective case-control study	426 Patients with IE, 214 treated orally vs 212 treated intravenously 79% Had left-sided IE, 6% had right-sided IE, 46% had a prosthetic, and 15% had device IE 40% Of patients infected with streptococci, 19% infected with S aureus, 11% infected with coagulase-negative staphylococci, 12% infected with enterococci, 18% infected with other organisms Variety of oral regimens used (β-lactams with or without rifampin, clindamycin with or without rifampin or fluoroquinolones, fluoroquinolones with or without rifampin, fluoroquinolones alone, clindamycin alone, or TMP-SMX with or without rifampin) Oral switch occurred at a median of 21 d of treatment Relapse rate: 2 of 214 (0.9%) receiving oral step-down therapy vs 9 of 212 (4%) receiving intravenous therapy (P = .05)	17/214 (8)	76/212 (36)	<.001
Tissot-Dupont et al,39 2019	Quasi-experimental, pre-post study	Initial 170 consecutive control patients with S aureus IE treated intravenously vs subsequent 171 consecutive patients treated with a new oral antibiotic protocol 23% Of patients had prosthetic valve IE and 28% had cardiac devices Intravenous control regimen was oxacillin or vancomycin plus gentamicin for 5 d TMP-SMX dosed at 960 mg/4800 mg per day plus clindamycin, 1800 mg/d, intravenously for 7 d, followed by 5 wk of oral TMP-SMX, 160 mg/800 mg, 6 tablets/d without clindamycin The TMP-SM and clindamycin group was older (64 vs 59 y; P = .007) and had higher blood pressures Median follow-up of 166 d Relapses less common with oral therapy than intravenous therapy (4% vs 6%; P = .046) No differences in recurrence between oral therapy and intravenous therapy (3.5% vs 7%; P = .15)	30-d Follow-up: 12/171 (7) Final follow-up: 32/171 (19)	30-d Follow-up: 24/170 (14) Final follow-up: 51/170 (30)	30-d Follow-up: 0.05 Final follow-up: 0.02

Abbreviations: IE, infective endocarditis; TMP-SMX, trimethoprim-sulfamethoxazole.

In a large, nonrandomized study, Mzabi et al⁴⁰ studied 426 cases of IE, including 214 patients who received oral step-down therapy and 212 who received intravenous therapy only. The patients predominantly had left-sided IE (nearly half of which involved prosthetic valves), while another 15% had IE associated with an intracardiac device. The change from intravenous to oral therapy occurred at a median of 21 days (range, 0–70 days). Several oral treatment regimens were used, including β-lactams, clindamycin, fluoroquinolones, and trimethoprim-sulfamethoxazole (TMP-SMX), each with or without rifampin. There was no difference in clinical cure rates, but patients who received oral step-down therapy had a significantly lower mortality rate than patients who received intravenous therapy only (8% vs 36%; P < .001). Furthermore, patients who received oral step-down therapy had a numerically lower rate of microbiologic relapses (0.9% vs 4%; P = .05).

Tissot-Dupont et al³⁹ more recently described a quasi-experimental, pre-post study in which an initial 170 consecutive patients with IE caused by S aureus were treated with an intravenous-only regimen, after which a new oral step-down protocol was put in place. The intravenous regimen consisted of 6 weeks of oxacillin or vancomycin, with an initial 5 days of synergy-dose gentamicin. After the first 170 patients, the next 171 consecutive patients with IE were treated with intravenous TMP-SMX (960 mg/4800 mg; equivalent to 6 double-strength tablets per day) plus clindamycin (600 mg 3 times daily) for 7 days, followed by 5 weeks of oral TMP-SMX (six 160-mg/800-mg tablets per day, adjusted for weight and kidney function). The enrolled patients had substantial comorbidities, and a high proportion of patients had left-sided disease, with 25% having prosthetic valve IE and 25% having cardiac device infections. Patients treated with oral step-down therapy had a significantly lower 30-day mortality rate than the patients who received intravenous therapy only (7% vs 14%; P = .05), as well as a lower mortality rate at final follow-up (median, 166 days; 19% vs 30%; P = .02). The oral step-down treatment of IE for all patients after the intervention had begun is important because it excludes the possibility that patients were selected for step-down therapy because they were healthier or more likely to succeed with oral therapy than patients previously treated with an intravenous regimen only.

RCTs of Oral Step-Down Therapy for IE

We identified 3 RCTs comparing oral therapy with intravenous therapy for IE (Table 3).^54–58

Table 3.

Randomized Clinical Trials Comparing Oral Step-Down vs Intravenous Antibiotics for the Treatment of IE

Source	Design and findings	Mortality rate, %
		Oral	Intravenous
		Stamboulian et al,54 1991	30 Patients with left-sided streptococcal IE randomized to 4 wk of intravenous ceftriaxone (2 g/d) vs 2 wk of intravenous ceftriaxone +2 wk of oral amoxillin (1 g 4 times/d) Half of patients infected with viridans streptococci, half infected with Streptococcus gallolyticus (formerly known as Streptococcus bovis) 3–6 mo of follow-up, all patients cured	0	0
Heldman et al,55 1996	500 Patients with history of intravenous drug use and suspicion of right-sided IE randomized to oral ciprofloxacin (750 mg twice daily) plus rifampin (300 mg twice daily) vs intravenous oxacillin (2 g every 4 h, or vancomycin 1 g every 12 h for MRSA) plus5d of intravenous gentamicin Ultimately 93 patients received a diagnosis of IE (45 oral, 48 intravenous) Median age, 35 y; two-thirds of patients were HIV positive 1 Treatment failure in the oral group (patient was holding pills in her cheek and spitting them out) vs 3 in the intravenous group (osteomyelitis, left-sided IE, and cerebral hemorrhage, all developing after 14 d of therapy) Only 1 (3%) patient in the oral group developed an adverse event vs 24 (62%) in the intravenous group	0	2
Iversen et al,57 2019; Bundgaard et al,58 2019a	400 Patients with left-sided IE randomized to receive oral step-down therapy (n = 199) vs intravenous therapy (n = 201) 108 Patients had prosthetic valves and 35 had implanted cardiac devices Mean age was 67 y; patients had extensive comorbidities 196 Patients infected with streptococci, 87 infected with Staphylococcus aureus, 23 infected with coagulase-negative staphylococci Patients switched to oral therapy when clinically stable, with no need for surgical intervention: median 17 d of intravenous therapy Pragmatic study with a variety of oral regimens, all of which were dual regimens, and based on high-dose β-lactams (amoxicillin or dicloxacillin 1 g 4 times daily), linezolid (600 mg twice daily), or moxifloxacin (400 mg once daily), plus rifampin (600 mg twice daily), clindamycin (600 mg 3 times daily), or fusidic acid (750 mg twice daily) At 6 mo of follow-up, 9% treatment failure with oral therapy vs 12% with intravenous therapy At 3.5 y of follow-up, 26% treatment failure with oral therapy vs 38% with intravenous therapy (HR, 0.64; 95% CI, 0.45–0.91)	6-mo Follow-up: 4 3.5-y Follow-up: 16b	6-mo Follow-up: 7 3.5-y Follow-up: 27bb

Abbreviations: HR, hazard ratio; IE, infective endocarditis; MRSA, methicillin-resistant Staphylococcus aureus.

^aThe study by Bundgaard et al⁵⁸ is an update of the results from the study by Iversen et al,⁵⁷ describing longer follow-up from the same patients.

^bHazard ratio, 0.57 (95% CI, 0.37–0.87).

Streptococcal IE

In 1991, Stamboulian et al⁵⁴ randomized 30 patients with left-sided streptococcal IE to receive 4 weeks of intravenous ceftriaxone or 2 weeks of intravenous ceftriaxone followed by 2 weeks of oral amoxicillin (1 g orally 4 times per day). At 3 to 6 months of follow-up, all patients were cured.

Staphylococcal IE

In 1996, Heldman et al⁵⁵ randomized 93 patients who injected drugs and had right-sided staphylococcal IE to treatment with oral ciprofloxacin (750 mg) plus oral rifampin (300 mg) twice daily (45 patients) or to standard intravenous therapy (oxacillin, or vancomycin for MRSA plus gentamicin for the first 5 days) (48 patients). Patients received oral therapy immediately after randomization, with no intravenous lead-in as part of the trial. However, patients were randomized on admission to the hospital and thus likely received a short duration (eg, <1–2 days) of intravenous therapy in the emergency department before admission. Only 5 patients were infected with MRSA, while 6 were infected with coagulase-negative staphylococci. The remainder were infected with MSSA. There was no difference in either clinical or microbiological cure rates or mortality between the oral and intravenous groups. Only 1 antibiotic-associated adverse event (3%) developed among 36 evaluable patients in the oral therapy group vs 24 (62%) among 39 evaluable patients in the intravenous group (P < .001). Many adverse events in the intravenous group were severe, including hepatotoxic effects and nephrotoxic effects.

The limitations of this study were its relatively small sample size, a probably suboptimal rifampin dose⁵⁵ (600 mg once per day has substantially more favorable pharmacokinetics than 300mg twice daily⁵⁶), predominantly young patients, the low rate of MRSA infections, and a focus on right-sided IE only (a form of IE with distinctly more favorable outcomes than left-sided IE).⁵⁵ This rifampin dosing regimen could have disadvantaged the patients receiving oral therapy, who nevertheless recovered equally well as the patients receiving intravenoustherapy. Moreover, while the study was relatively small, enriched for younger patients, and focused on right-sided IE, these limitations should be assessed in the context of the overall data available from the pharmacokinetic and clinical studies already discussed and a large RCT.⁵⁷

We recognize the concern about the relative paucity of cases of MRSA in published literature of oral therapy for IE. However, methicillin resistance does not intrinsically alter the activity of orally bioavailable, non–β-lactam antimicrobial agents (eg, linezolid, fluoroquinolones, TMP-SMX, clindamycin, or rifampin). Thus, as long as the infecting MRSA strain remains susceptible to an oral agent of interest, there is no reason to believe its effectiveness would be lower relative to intravenous therapy for infections caused by MRSA vs MSSA.

All Patients With Left-Sided IE

In the largest RCT conducted to date, to our knowledge, Iversen et al⁵⁷ randomized 400 patients with left-sided IE to receive either oral step-down(n = 199) or intravenous (n = 201) therapy. Intravenous nous therapy was given to all patients for at least the first 10 treatment days. After this intravenous lead-in, those randomized to oral step-down therapy were switched to oral regimens when they were clinically stable and had no echocardiographic indications for immediate surgical intervention. This was a pragmatic trial, in that it used a menu of dual, orally bioavailable antibiotic regimens selected based on the infecting pathogen rather than focusing on one global regimen. The use of the combination oral regimens was not based on prior data showing that 2 agents were needed or synergistic, but rather it was designed to ensure that at least 1 the oral agents would achieve therapeutic concentrations in vivo.

The patients had a mean age of 67 years and had substantial comorbidities, with more than 25% having prosthetic valve(s) and 9% having an intracardiac device. The causative bacteria included the viridans group and other streptococci, staphylococci, and enterococci. The median time to switching to oral therapy was 17 days (interquartile range, 12–24 days) of intravenous therapy. The patients who were switched to oral therapy spent significantly less time in the hospital than the patient who received intravenous therapy only (3 additional days vs 19 additional days; P < .001). At 6 months of follow-up, overall treatment failure rates (10.5% vs 12.1%; hazard ration [HR], 0.72; 95% CI, 0.37–1.36) and mortality rates (3.5% vs 6.5%; HR, 0.53; 95% CI, 0.21–1.32) were numerically, but not statistically significantly, better for patients who received oral step-down therapy than for patients who received intravenous therapy only; the 10% noninferiority margin was met. A subsequent follow-up publication describing the long-term outcomes of the patients in this trial (median follow-up, 3.5 years) reported that the rates of treatment failure (26% vs 38%; HR, 0.64; 95% CI, 0.45–0.91) and of mortality (16%vs 27%; HR, 0.57; 95%CI, 0.37–0.87) were both significantly better for patients who had received oral step-down therapy than for patients who had received intravenous therapy only.⁵⁸

The pragmatic trial design, with a menu of oral options, does not allow for the selection of a single preferred regimen. However, these outcome differences do allow for a degree of clinical confidence that a rather broad array of dual-oral therapeutic options can work in step-down outpatient therapy of IE.

Discussion

Our literature search found no published data demonstrating that intravenous-only antibiotic therapy is more effective than modern, oral antibiotic step-down options as an alternative approach for the treatment of IE. In all studies that we evaluated, oral antibiotic step-down therapy was at least as effective as intravenous-only therapy. In the largest RCT conducted, oral step-down therapy strategies resulted in better outcomes, including reduced long-term mortality, compared with prolonged intravenous regimens.⁵⁸ These results are concordant with multiple prior observational studies, including a quasi-experimental pre-post study.³⁹ Oral antibiotic step-down therapy is intrinsically safer than prolonged intravenous therapy owing to the elimination of the risk of complications from the indwelling line. Thus, absent any data suggesting the superiority of intravenous therapy, any meaningful data supporting the efficacy of modern oral antibiotic therapy for IE creates equipoise, which allows for a reasonable consideration of oral therapy for this disease in well-selected patients.

Limitations

The data do not allow for a conclusion to be drawn regarding the need for an initial course of intravenous antibiotics before changing to oral therapy. Durations of intravenous lead-in therapy in the published, controlled studies ranged from only that given in the emergency department prior to admission,⁵⁵ to 7 days,³⁹ to 18 to 24 days.^{40,53,54,57,58} As with any other serious infection, it is reasonable to begin therapy with intravenous drugs just to be certain that the patients are stable and have cleared their bacteremia, after which a transition to oral antibiotics can be made.

Additional psychosocial considerations regarding oral vs intravenous therapy may come into play. Such concerns include the likelihood of outpatient adherence (particularly among patients who acquired their IE in the setting of intravenous drug use) or justifying a stay in an extended-care facility. Thus, the effectivenss of oral step-down therapy does not necessarily mean that oral therapy is always the right choice for all patients. However, the data support the option to use oral step-down therapy for appropriate patients.

Specifically, based on aggregate pharmacokinetic, observational, and RCT data and on an assessment of clinical appropriateness, it is reasonable to consider oral step-down therapy for patients with IE once each of the following is true (Figure):

Figure. Considerations for Oral Step-Down Antibiotic Therapy for Infective Endocarditis.

Q indicates question.

1. They are clinically stable with no immediate indication for cardiac surgical intervention;

2. The initial course of intravenous therapy has cleared their bacteremia;

3. There are no concerns regarding absorption of oral therapy from the gastrointestinal tract;

4. There are no psychosocial concerns that would cause intravenous therapy to be preferred for adherence or level-of-care concerns; and

5. An oral antibiotic regimen is available to which the causative organism is susceptible in vitro and that is supported by published clinical data.

Multiple clinical studies describe outcomes of patients with IE treated with specific antimicrobial regimens. Based on these studies, reasonable oral considerations for IE (for organisms shown in vitro to be susceptible) would include the following: amoxicillin for highly penicillin-susceptible streptococci (minimum inhibitory concentration, ≤0.12 μg/mL); TMP-SMX for S aureus IE, only after initial intravenous lead-in period with an alternative agent and/or possibly with an initial lead-in period of intravenous TMP-SMX in combination with intravenous clindamycin; levofloxacin or moxifloxacin (used in lieu of ciprofloxacin owing to enhanced staphylococcal activity) only in combination with a second agent; and linezolid, alone or possibly in combination with a second agent, for susceptible gram-positive organisms. For each of these agents, it is reasonable to use dosing and frequency based on those in published studies (Table 4).^{39,45,49,50,52,54–58}

Table 4.

Summary of Oral Step-Down Antibiotic Dosing Used in Published Clinical Studies^a

Drug	Organism	Dose	Reference No.
Amoxicillin	Sensitive streptococci or enterococci (for streptococci, with or without combination; and for enterococci, only in combination with rifampin, moxifloxacin, linezolid, or clindamycin)	1 g 4 times daily	54, 57, 58
Dicloxacillin	Sensitive staphylococci (only in combination with rifampin or fusidic acid)	1 g 4 times daily	57, 58
Levofloxacinb	Sensitive staphylococci (only in combination with rifampin or fusidic acid)	750 mg once daily	55
Moxifloxacin	Sensitive streptococci, enterococci, or staphylococci (only in combination with amoxicillin, rifampin, clindamycin, or linezolid)	400 mg once daily	57, 58
TMP-SMX	Sensitive staphylococci	960 mg/4800 mg daily	39
Linezolid	For sensitive gram-positive cocci (for most patients in published studies, linezolid was used alone; in some studies,52,57,58 linezolid was given as a combination regimen with rifampin, fusidic acid, moxifloxacin, clindamycin, or amoxicillin)	600 mg twice daily	45, 49, 50, 52, 57, 58
Rifampin	Only as adjunctive agent (see above for other antibiotics rifampin has been combined with) and never as a single agent	600 mg once or twice daily	55–58
Clindamycin	Only as adjunctive agent (see above for other antibiotics clindamycin been combined with) and never as a single agent	600 mg 3 times daily	39, 57, 58
Fusidic acid	Only as adjunctive agent (see above for other antibiotics fusidic acid has been combined with) and never as a single agent	750 mg twice daily	57, 58

Abbreviation: TMP-SMX, trimethoprim-sulfamethoxazole.

^aThis is not a guideline for treatment. Rather, it summarizes the oral step-down regimens in published studies of infective endocarditis in which patients had favorable outcomes. Most published regimens were combinations of antibiotics, and some were used in only a small number of cases. Several monotherapy regimens have been described for oral step-down therapy in published studies.

^bThe study used ciprofloxacin rather than levofloxacin, the latter of which was not yet clinically available. However, in the ensuing years, there was a rapid emergence of staphylococcal resistance to ciprofloxacin. Levofloxacin is preferred to ciprofloxacin owing to enhanced in vitro activity against staphylococci, but only in combination with a second agent.

More robust data on oral therapy for MRSA IE are needed to confirm effectiveness for this infection in particular. Also, whether or not to include rifampin in linezolid-based regimens is unclear. The combination provides insurance that 1 of the 2 agents will achieve therapeutic concentrations; however, linezolid monotherapy may be preferable given that rifampin lowers linezolid blood levels.^59,60

Conclusions

The inherent conservatism of the medical profession delays incorporation of evidence-based changes to treatment regimens into clinical practice.^61–64 The dogma that intravenous antibiotic strategies are required for the treatment of IE appears to reflect an “eminence-based” rather than evidence-based standard practice.⁶¹ Sufficient data now exist to establish that oral step-down antibiotic therapy is as effective as intravenous therapy for select patients with IE after bacteremia has been cleared and clinical stability achieved.