Abstract
Objectives
The incidence of Serratia endocarditis is increasing, yet optimal treatment has not been defined. Our objective was to investigate the outcomes of patients with Serratia endocarditis by treatment strategy.
Methods
We reviewed adult patients with definitive Serratia endocarditis at two independent health systems between July 2001 and April 2023. Combination therapy was defined as receipt of ≥2 in vitro active agents for ≥72 h.
Results
Seventy-five patients were included; 64% (48/75) were male and 85% (64/75) were people who inject drugs. Compared with monotherapy, receipt of combination therapy was associated with lower rates of microbiological failure (0% versus 15%, P = 0.026) and 90 day all-cause mortality (11% versus 31%, P = 0.049). Antimicrobial discontinuation due to an adverse event was more common among patients receiving combination therapy compared with monotherapy (36% versus 8%, P = 0.058).
Conclusions
In the largest series of Serratia endocarditis to date, combination antibiotic treatment was associated with improved outcomes. However, larger, prospective studies are warranted.
Introduction
The first case of endocarditis presumed to be secondary to Serratia species was reported on 11 October 1950 in San Francisco due to a genitourinary source.1 From 1969 to 1978, 36 cases of Serratia endocarditis were described among people who inject drugs (PWID).2,3 Since then, there have been no studies comparing antibiotic management strategies for Serratia endocarditis. We previously reported that Serratia species have emerged as the most common cause of non-HACEK (species other than Haemophilus spp., Aggregatibacter actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens or Kingella kingae) Gram-negative endocarditis in our region.4 Across Gram-negative pathogens we did not demonstrate a clinical benefit of combination therapy; however, clinical outcomes varied significantly by species. The decision to prescribe monotherapy or combination therapy for Serratia endocarditis remains uncertain.2–6 At the same time, the utility of β-lactam agents that induce chromosomal AmpC β-lactamases is under debate given that Serratia species do not demonstrate clinically significant AmpC production in vitro or in clinical studies.7 The objective of this study was to compare the clinical outcomes of patients with definitive Serratia species endocarditis by treatment strategy with a focus on monotherapy versus combination therapy regimens.
Methods
Adult patients with definitive Serratia endocarditis were identified from 12 hospitals from two independent health systems between July 2001 and April 2023. Patients from the University of Pittsburgh Medical Center (UPMC) were identified by searching electronic medical records using the Boolean search terms ‘endocarditis’ and ‘Serratia’ as previously reported.4 At Yale New Haven Health System (YNHHS) all patients with positive blood cultures growing Serratia spp. were manually reviewed. All patients had positive blood cultures and met definitive Duke criteria for infective endocarditis as previously reported.4,5 Unless the cultured valve yielded Serratia, patients with blood cultures growing Gram-positive bacteria or yeast for >24 h were excluded.
Combination therapy was defined as receipt of ≥2 in vitro active agents for ≥72 h. Clinical failure was defined as death or microbiological failure by Day 42. Microbiological failures included escalation of antimicrobial therapy following treatment-emergent resistance, increased vegetation size, or persistently positive blood cultures for ≥14 days. Intracardiac complications were defined as the presence of an intracardiac abscess or valve perforation. Cefepime, carbapenems, fluoroquinolones and trimethoprim/sulfamethoxazole were classified as treatment regimens stable against AmpC derepression.7 Relapse was defined as recurrence of endocarditis secondary to the same Serratia spp. within 9 months. Patients with an indication for cardiac surgery were identified as previously described.4
Categorical variables were compared by a chi-squared or Fisher’s exact test. Continuous variables were compared using a Wilcoxon rank-sum test. A stepwise multivariate logistic regression analysis was performed to assess for predictors of clinical failure. Statistical significance in multivariable analyses was defined as a P value of <0.05 (two-tailed). All analyses were performed using Stata, version 15.1 (College Station, TX, USA).
Ethics
The study was approved by the University of Pittsburgh Institutional Review Board (STUDY21110056) with a waiver of written informed consent.
Results
Seventy-five patients were included; 64% (48/75) were male and 85% (64/75) were PWID. Among those who were not PWID (n = 11), sources of infection included skin and soft tissue (n = 2), line-related (n = 2), urine (n = 1), pulmonary (n = 1), left ventricular assist device driveline (n = 1) or unknown (n = 4). The causative Serratia species were Serratia marcescens (n = 71), Serratia liquefaciens (n = 3) and an unidentified Serratia spp. (n = 1).
Three patients had pacemaker endocarditis; the remaining 72 patients had endocarditis involving the mitral valve (n = 28), tricuspid valve (n = 18), aortic valve (n = 13), pulmonic valve (n = 3), multiple valves (n = 9) or right atrium (n = 1). Thirty percent (22/75) of patients had intracardiac complications, including valve perforation (n = 13), intracardiac abscess (n = 6) or both (n = 3). Cardiac surgery was indicated in 77% (58/75); indications included embolic prevention (n = 12), heart and/or valve failure (n = 10), vegetation on implantable cardiac device (n = 3), an intracardiac complication (n = 2) or multiple indications (n = 31).
Monotherapy and combination therapy were prescribed in 52% (39/75) and 48% (36/75) of patients, respectively (Table S1, available as Supplementary data at JAC Online). The median (IQR) duration of combination therapy was 31 (10–44) days. The second agents of combination regimens were started at a median (IQR) of 3 (2–7) days from the time of positive blood cultures. Combination regimens consisted of β-lactam + aminoglycoside (n = 12), β-lactam + fluoroquinolone (n = 20) and other combinations (n = 4). Of the 12 patients who received a β-lactam + aminoglycoside, 5 patients had a goal peak of 5–7 μg/mL and trough of ≤2 μg/mL, 3 patients were dosed using the Hartford Nomogram, 3 patients had variable target peaks but goal troughs of <1 μg/mL and the final patient had a goal peak of 6–10 μg/mL and trough of <2 μg/mL.8 Patients who received combination therapy were more likely to be female; otherwise, patients who received monotherapy or combination therapy demonstrated similar median Charlson comorbidity index scores and Pitt bacteraemia scores, as well as rates of surgical intervention. The overall median (IQR) time to surgery was 12 (8–15) days; a higher proportion of patients who received monotherapy had Serratia-positive valve cultures (P = 0.013). All patients had an Infectious Diseases specialist formally consulted who documented treatment recommendations in the electronic medical record. A cardiologist was consulted for 49% (37/75) of patients and a cardiothoracic surgeon was consulted for 76% (57/75) of patients. Eleven percent (8/75) of patients had neither a cardiology nor a cardiothoracic surgery consult.
Thirty-six percent of patients (14/39) treated with monotherapy experienced clinical failure compared with 11% (4/36) who were treated with combination regimens (P = 0.015). Other factors associated with clinical failure on univariate analysis included older age (P = 0.043), Pitt bacteraemia score (P = 0.004), left-sided or multiple valve involvement (P = 0.024), and central nervous system (CNS) septic emboli (P = 0.033). Patients who experienced clinical failure were also less likely to receive surgical intervention despite an indication for surgery (Table 1) and less likely to receive valve surgical management overall (17% versus 54%, P < 0.001). After adjusting for age, severity of illness, valve involved and surgical intervention, receipt of combination therapy was independently associated with a lower odds of clinical failure (OR = 0.17, 95% CI: 0.03–0.86, P = 0.032). Not receiving surgical management despite an indication for surgery was the only significant predictor of clinical failure (OR = 3.84, 95% CI: 4.5–105, P < 0.001).
Table 1.
Clinical and microbiological factors associated with 42 day clinical outcomes
| 42 day cure (n = 57) |
42 day failure (n = 18) |
Univariate P value |
Logistic regressiona,b (OR, 95% CI; P value) |
|
|---|---|---|---|---|
| Age, years, median (IQR) | 36.8 (30.7–45.5) | 41.6 (37.3–54.3) | 0.043 | 1.06, 1.01–1.12; P = 0.038 |
| Female gender, n (%) | 20 (35.1) | 7 (38.9) | 0.769 | |
| White ethnicity, n (%) | 50 (87.7) | 14 (77.8) | 0.444 | |
| Pitt bacteraemia score, median (IQR) | 1 (0–3) | 4 (2.3–7) | 0.004 | 1.15, 0.88–1.51; P = 0.3 |
| CCI score, median (IQR) | 0 (0–1) | 1 (0–2) | 0.185 | |
| PWID, n (%) | 50 (87.7) | 14 (77.8) | 0.444 | |
| Vegetation > 1 cm, n (%) | 43 (75.4) | 10 (55.6) | 0.106 | |
| Patient-directed discharge, n (%) | 9 (15.8) | 1 (5.6) | 0.435 | |
| Prior endocarditis secondary to a different pathogen, n (%) | 22 (38.6) | 4 (22.2) | 0.263 | |
| Vertebral osteomyelitis, n (%) | 8 (14) | 0 | 0.186 | |
| Prosthetic valve IE, n (%) | 14 (24.6) | 2 (11.1) | 0.328 | |
| Left-sided or multiple valve involvement, n (%) | 34 (59.6) | 16 (88.9) | 0.024 | 3.48, 0.45–27.1; P = 0.234 |
| CNS septic emboli, n (%) | 16 (28.1) | 10 (55.6) | 0.033 | 1.38, 0.27–7.2; P = 0.701 |
| Septic emboli, n (%) | 42 (73.7) | 14 (77.8) | >0.999 | |
| Secondary pathogen, n (%) | 0.847 | |||
| None | 49 (86) | 15 (83.3) | ||
| Gram-negative | 2 (3.5) | 1 (5.6) | ||
| Staphylococcus aureus, VGS, Enterococcus spp. | 4 (7) | 1 (5.6) | ||
| Mixed/other | 2 (3.5) | 1 (5.6) | ||
| Index isolate ceftriaxone resistant, n (%) | 8 (14) | 2 (11.1) | >0.999 | |
| Initial therapy dose optimizedc, n (%) | 48 (84.2) | 14 (77.8) | 0.498 | |
| AmpC-stable drug for therapy initiationd, n (%) | 41 (71.9) | 14 (77.8) | 0.765 | |
| AmpC-stable drug for therapy completione, n (%) | 45 (78.9) | 14 (77.8) | 0.916 | |
| No surgical intervention despite an indication, n (%) | 10 (17.5) | 14 (77.8) | <0.001 | 12.7, 2.15–75.7; P = 0.005 |
| Positive valve culture, n/N (%) | 3/26 (50) | 3/4 (75) | 0.602 | |
| Combination treatmentf, n (%) | 32 (56.1) | 4 (22.2) | 0.015 | 0.16, 0.03–0.84; P = 0.031 |
CCI, Charlson comorbidity index; IE, infective endocarditis; VGS, viridans group streptococci. Bold type indicates statistical significance.
Age (1.05, 0.99–1.1; P = 0.064), left-sided or multiple valve endocarditis (5.36, 0.81–35.4; P = 0.081), combination treatment (0.17,0.03–0.86; P = 0.032) and no surgery despite surgical indication (3.84, 4.5–105; P < 0.001) were retained in the stepwise multivariate model.
Hosmer–Lemeshow goodness-of-fit test: χ2 = 46.5, P = 0.97.
Dose-optimized cefepime, meropenem, ceftazidime was defined as receipt of at least 6 g/day for patients with a CLCR over 50 mL/min, 3 g/day for patients with a CLCR of 30–49 mL/min or on CRRT, 2 g/day for patients with a CLCR of 10–29 mL/min, and 6 g/week for patients on haemodialysis. Dose-optimized ceftriaxone was defined as at least 2 g/day. Dose-optimized ertapenem was defined as at least 1 g/daily for patients with a CLCR above 30 mL/min and 500 mg daily for those with a CLCR less than 30 mL/min. Dose-optimized imipenem/cilastatin was defined as 2 g/daily of imipenem for patients with a CLCR over 50 mL/min. Dose-optimized piperacillin/tazobactam was defined as 18 g/day in patients with a CLCR over 40 mL/min or 9 g/day for patients on haemodialysis. Dose-optimized ciprofloxacin was defined as 750 mg twice daily for patients with a CLCR over 50 mL/min while dose-optimized levofloxacin was defined as 750 mg daily for patients with a CLCR over 50 mL/min.
AmpC-stable regimens for initiation (>7 days of treatment) consisted of cefepime (n = 39), carbapenems (n = 14) or fluoroquinolones (n = 2). Other regimens consisted of ceftriaxone (n = 14), piperacillin/tazobactam (n = 4) or ceftazidime (n = 2).
AmpC-stable regimens for completion consisted of cefepime (n = 34), carbapenems (n = 16), fluoroquinolones (n = 6) or trimethoprim/sulfamethoxazole (n = 3). Other regimens consisted of ceftriaxone (n = 9), piperacillin/tazobactam (n = 4) or ceftazidime (n = 3).
Combination regimens consisted of β-lactam + aminoglycoside (n = 12), β-lactam + fluoroquinolone (n = 20) and other combinations (n = 4) and initial monotherapy regimens consisted of cefepime (n = 19), ceftazidime/ceftriaxone (n = 7), fluoroquinolones (n = 2), carbapenems (n = 8) and piperacillin/tazobactam (n = 3).
Compared with monotherapy, receipt of combination therapy was also associated with lower rates of microbiological failure (0% versus 15%, P = 0.026) and 90 day all-cause mortality (11% versus 31%, P = 0.049). Among the six patients treated with monotherapy who experienced microbiological failure, four developed resistance to their primary treatment, one had persistently positive blood cultures for 20 days, and one was found to have increased vegetation size on repeat imaging. Two patients each developed resistance to cefepime or carbapenems as their primary monotherapy. No patient treated with combination therapy developed resistance to treatment. There were 20 patients treated with an initial AmpC-inducing regimen that contained ceftriaxone (n = 14), piperacillin/tazobactam (n = 4) or ceftazidime (n = 2). The corresponding microbiological cure rates for patients initially treated with ceftriaxone, piperacillin/tazobactam or ceftazidime were 93% (13/14), 75% (3/4) and 100% (2/2), respectively. Clinical outcomes did not vary between patients who did or did not receive an AmpC-inducing agent as part of their treatment regimen (data not shown). Nine patients experienced relapse of endocarditis secondary to Serratia spp. at a median (IQR) of 134 (88–159) days from the time of initial diagnosis; 56% (5/9) of patients were documented as reporting injection drug use upon relapse.
Among patients surviving to hospital discharge, the total median length of stay was longer for patients treated with combination therapy compared with monotherapy (34 versus 20 days, P = 0.006). Antimicrobial discontinuation due to an adverse event was more common among patients receiving combination therapy compared with monotherapy (36% versus 8%, P = 0.058). Adverse events included aminoglycoside-induced nephrotoxicity (n = 5) or ototoxicity (n = 1), cefepime-induced CNS toxicity (n = 4), QTc prolongation due to ciprofloxacin (n = 2), abnormal liver function tests attributed to ciprofloxacin (n = 1) and an elevation in serum creatinine from trimethoprim/sulfamethoxazole (n = 1). Two patients experienced multiple adverse effects requiring antimicrobial discontinuation. Of note, the patient who developed aminoglycoside-induced ototoxicity was considered to have received monotherapy since treatment was discontinued within 72 h of gentamicin therapy in addition to ceftazidime.
Discussion
To our knowledge, this is the largest study evaluating the clinical treatment outcomes of Serratia endocarditis. The key findings of this investigation were that combination therapy was associated with favourable outcomes, including lower rates of clinical failure, death and microbiological failure, when compared with monotherapy, and rates of treatment-emergent resistance to therapy were generally low. On balance, receipt of combination therapy was more commonly associated with antibiotic adverse events and prolonged hospitalization. Our findings corroborate historical data from the 1970s where 36 patients with Serratia spp. endocarditis were treated with either aminoglycoside monotherapy or in combination with chloramphenicol or carbenicillin.2,3 The cumulative data suggest that treatment with more than one in vitro active agent may have clinical benefit. Interestingly, in our study, patients treated with monotherapy were more likely to have valve cultures positive for Serratia spp., suggesting that monotherapy was ineffective in sterilizing heart valves prior to surgery. The preferred combination regimen, however, is still unknown. Patients at our centres were most commonly treated with a β-lactam plus an aminoglycoside or fluoroquinolone. While guidelines generally recommend combination treatment for all cases of Gram-negative endocarditis, our prior experience would suggest that the benefit of combination therapy may be limited to Serratia spp. and it is unclear if there is benefit in administering combination therapy for Gram-negative endocarditis secondary to other pathogens such as Pseudomonas aeruginosa.4
Combination regimens in our study included β-lactam agents that were stable against AmpC hydrolysis, like cefepime and meropenem, but also included agents like ceftriaxone and piperacillin/tazobactam previously not recommended for AmpC-producing organisms due to the risk of inducible resistance. Interestingly, among the 10 patients receiving monotherapy with these agents, no cases of treatment-emergent resistance were identified. These data add to a growing body of evidence that suggests S. marcescens and S. liquefaciens are unlikely to demonstrate clinically relevant AmpC production based on in vitro analyses and clinical reports.9–11 Although index blood culture isolates of Serratia were non-susceptible to ceftriaxone in 13% of patients in our cohort, we failed to identify a clear advantage in using an AmpC-stable regimen among those with in vitro susceptibility to AmpC-inducing agents.
The strength of the study lies in its real-world, multicentre design, which better elucidates treatment options for a rare, but life-threatening disease. However, several limitations should be acknowledged. First, our sample size precluded us from conducting more robust analyses despite this being the largest study of Serratia endocarditis reported to date. Second, this study was performed over a prolonged time period across multiple centres. It is possible that changes to the diagnosis and treatment of Serratia endocarditis over time may have impacted our findings. Notably, 81% (61/75) of patients in the study were identified between 2015 and 2023, which is consistent with the increasing prevalence of Serratia endocarditis we have previously reported.4 Third, we identified adverse events at the time of treatment discontinuation, and thus, it is possible that other events may have occurred during treatment, as a result of the patients’ underlying medical conditions, or as a result of concomitant medications. Fourth, patient adherence to outpatient antimicrobials was not assessed. Finally, our observational design may present indication biases for patients who received monotherapy or combination therapy. We have attempted to control for these confounders through multivariable analyses; however, other biases may influence our results. Our findings should ideally be confirmed by larger, multicentre studies.
Conclusions
In the largest series of Serratia endocarditis to date, predictors of clinical failure included failure to receive surgical intervention despite an indication for surgery, and receipt of antibiotic monotherapy. Combination antibiotic treatment was associated with a decreased risk of clinical failure compared with monotherapy and may be preferred. Larger, prospective studies are warranted.
Supplementary Material
Contributor Information
Sunish Shah, Antibiotic Management Program, University of Pittsburgh Medical Center, Pittsburgh, Falk Medical Building, Suite 3A, Room 317, 3601 Fifth Avenue, Pittsburgh, PA 15213, USA; Department of Pharmacy, University of Pittsburgh Medical Center, Pittsburgh, PA, USA; Department of Pharmacy, Yale New Haven Hospital, New Haven, CT, USA.
Madeline McCrary, Division of Infectious Diseases, University of North Carolina, Chapel Hill, NC, USA.
Asher J Schranz, Division of Infectious Diseases, University of North Carolina, Chapel Hill, NC, USA.
Lloyd Clarke, Antibiotic Management Program, University of Pittsburgh Medical Center, Pittsburgh, Falk Medical Building, Suite 3A, Room 317, 3601 Fifth Avenue, Pittsburgh, PA 15213, USA; Department of Pharmacy, University of Pittsburgh Medical Center, Pittsburgh, PA, USA.
Matthew W Davis, Department of Pharmacy, Yale New Haven Hospital, New Haven, CT, USA.
Ashley Marx, Department of Pharmacy, University of North Carolina Medical Center, Chapel Hill, NC, USA.
Douglas Slain, Department of Clinical Pharmacy and Division of Infectious Diseases, West Virginia University, Morgantown, WV, USA.
Bobbi Jo Stoner, Department of Pharmacy, University of Kentucky Medical Center, Lexington, KY, USA.
Jeffrey Topal, Department of Pharmacy, Yale New Haven Hospital, New Haven, CT, USA.
Ryan K Shields, Antibiotic Management Program, University of Pittsburgh Medical Center, Pittsburgh, Falk Medical Building, Suite 3A, Room 317, 3601 Fifth Avenue, Pittsburgh, PA 15213, USA; Department of Medicine, University of Pittsburgh, Pittsburgh, PA, USA.
Funding
All authors state that this work was not directly supported by any funding agencies. This study was performed as part of our routine work.
Transparency declarations
No funding sources were received for this study. R.K.S. has served as a consultant for Allergan, Cidara, Shionogi, Menarini, Melinta, Merck, Entasis, Utility and Venatorx, and has received investigator-initiated funding from Merck, Melinta, Shionogi and Venatorx. A.J.S. received support from NIDA (K23DA049946). All other authors report no potential conflicts of interest or funding sources.
Supplementary data
Table S1 is available as Supplementary data at JAC Online.
References
- 1. Wheat RP, Zuckerman A, Rantz LA. Infection due to chromobacteria; report of 11 cases. AMA Arch Intern Med 1951; 88: 461–6. 10.1001/archinte.1951.03810100045004 [DOI] [PubMed] [Google Scholar]
- 2. Mills J, Drew D. Serratia marcescens endocarditis: a regional illness associated with intravenous drug abuse. Ann Intern Med 1976; 84: 29–35. 10.7326/0003-4819-84-1-29 [DOI] [PubMed] [Google Scholar]
- 3. Cooper R, Mills J. Serratia endocarditis. A follow-up report. Arch Intern Med 1980; 140: 199–202. 10.1001/archinte.1980.00330140057018 [DOI] [PubMed] [Google Scholar]
- 4. Shah S, Clarke LG, Shields RK. Epidemiology and clinical outcomes of non-HACEK Gram-negative infective endocarditis. Open Forum Infect Dis 2023; 10: ofad052. 10.1093/ofid/ofad052 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Baddour LM, Wilson WR, Bayer AS et al. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association. Circulation 2015; 132: 1435–86. 10.1161/CIR.0000000000000296 [DOI] [PubMed] [Google Scholar]
- 6. Calderón Parra J, De Castro-Campos D, Muñoz García P et al. Non-HACEK Gram negative bacilli endocarditis: analysis of a national prospective cohort. Eur J Intern Med 2021; 92: 71–8. 10.1016/j.ejim.2021.04.021 [DOI] [PubMed] [Google Scholar]
- 7. Tamma PD, Aitken SL, Bonomo RA et al. Infectious Diseases Society of America guidance on the treatment of AmpC β-lactamase-producing Enterobacterales, carbapenem-resistant Acinetobacter baumannii, and Stenotrophomonas maltophilia infections. Clin Infect Dis 2022; 74: 2089–114. 10.1093/cid/ciab1013 [DOI] [PubMed] [Google Scholar]
- 8. Nicolau DP, Freeman CD, Belliveau PP et al. Experience with a once-daily aminoglycoside program administered to 2,184 adult patients. Antimicrob Agents Chemother 1995; 39: 650–5. 10.1128/AAC.39.3.650 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Kohlmann R, Bähr T, Gatermann SG. Species-specific mutation rates for ampC derepression in Enterobacterales with chromosomally encoded inducible AmpC β-lactamase. J Antimicrob Chemother 2018; 73: 1530–6. 10.1093/jac/dky084 [DOI] [PubMed] [Google Scholar]
- 10. Derrick C, Bookstaver PB, Lu ZK et al. Multicenter, observational cohort study evaluating third-generation cephalosporin therapy for bloodstream infections secondary to Enterobacter, Serratia, and Citrobacter species. Antibiotics (Basel) 2020; 9: 254. 10.3390/antibiotics9050254 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Slain D, Howard C, Cooper CG. An antimicrobial treatment assessment of Serratia marcescens bacteremia and endocarditis. Front Antibiot 2022; 1: 942721. 10.3389/frabi.2022.942721 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.