Cluster of Nontoxigenic Corynebacterium diphtheriae Infective Endocarditis and Rising Background C. diphtheriae Cases—Seattle, Washington, 2020–2023

Abstract

Background

Nontoxigenic Corynebacterium diphtheriae, often associated with wounds, can rarely cause infective endocarditis (IE). Five patients with C. diphtheriae IE were identified within 12 months at a Seattle-based hospital system. We reviewed prior C. diphtheriae–positive cultures to determine if detections had increased over time and evaluated epidemiologic trends.

Methods

We conducted a formal electronic health record search to identify all patients aged ≥18 years with C. diphtheriae detected in a clinical specimen (ie, wound, blood, sputum) between 1 September 2020 and 1 April 2023. We collected patient demographics, housing status, comorbidities, substance-use history, and level of medical care required at detection. We extracted laboratory data on susceptibilities of C. diphtheriae isolates and on other pathogens detected at the time of C. diphtheriae identification.

Results

Between 1 September 2020 and 1 April 2023, 44 patients (median age, 44 years) had a C. diphtheriae–positive clinical culture, with most detections occurring after March 2022. Patients were predominantly male (75%), White (66%), unstably housed (77%), and had a lifetime history of injecting drugs (75%). Most C. diphtheriae–positive cultures were polymicrobial, including wound cultures from 36 (82%) patients and blood cultures from 6 (14%) patients, not mutually exclusive. Thirty-four patients (77%), including all 5 patients with C. diphtheriae IE, required hospital admission for C. diphtheriae or a related condition. Of the 5 patients with IE, 3 died of IE and 1 from COVID-19.

Conclusions

Findings suggest a high-morbidity outbreak disproportionately affecting patients who use substances and are unstably housed.

Nontoxigenic Corynebacterium diphtheriae detections are increasing among Seattle residents with unstable housing and substance use. Five patients with fulminant C. diphtheriae infective endocarditis presented over a 1-year period, raising concern for a high-morbidity outbreak caused by a frequently overlooked pathogen.

Nontoxigenic Corynebacterium diphtheriae, an aerobic gram-positive bacillus often associated with wound infections, is a rare cause of infective endocarditis (IE) [1, 2]. Unlike toxigenic C. diphtheriae (the toxin-producing form predominantly associated with respiratory disease that can also cause infection at nonrespiratory anatomical sites [eg, skin, conjunctiva, genital mucosa]) [3], nontoxigenic C diphtheriae is neither vaccine-preventable nor nationally reportable despite the risk of invasive disease [4, 5].

Nontoxigenic C. diphtheriae IE, first identified in 1893, has a mortality rate ranging from 35% to 66% based on prior literature [2, 4, 6–10]. Living in poor sanitary conditions and having chronic wounds increases the risk of C. diphtheriae IE, as colonization and skin breakdown can create a portal of entry into the bloodstream [4, 11–15]. Persons with underlying cardiac disease or prosthetic valves, alcohol use, homelessness or housing instability in urban areas, recent immigration or refugee status from high-prevalence areas, and persons who inject drugs are also at increased risk [4, 6, 11, 12]. Corynebacterium diphtheriae IE outbreaks have been reported worldwide, predominantly among patients with the above risk factors but also among patients without cardiac disease or history of injecting drugs [7–10, 16, 17]. Many IE cases are sporadic, but IE outbreaks can be caused by a common clonal lineage [10, 18].

Within Seattle, where our hospital system is based, intermittent cutaneous and respiratory C. diphtheriae outbreaks occurred during the 1970s among persons experiencing homelessness (PEH) who injected drugs or had alcohol dependence, particularly during winter when indoor crowding in shelters is common [13, 19]. The most recent cluster of detections described in the literature began in 2018 among PEH; most patients had cutaneous infection and most isolates were genetically related [13].

Here, we describe 5 patients with C. diphtheriae IE identified over a 12-month period within our hospital system, along with the clinical and laboratory epidemiology of all C. diphtheriae detections over the last 2.5 years.

METHODS

Setting

We identified C. diphtheriae detections in clinical culture specimens during 1 September 2020–1 April 2023 in an academic hospital system in Seattle, Washington. The hospital system includes multiple outpatient clinics and 5 acute care hospitals. One hospital is the only level-1 trauma center in Washington State and serves as a tertiary burn and trauma referral center for Alaska, Montana, and Idaho.

Infective Endocarditis Case Series

Nontoxigenic C. diphtheriae is not listed as a typical pathogen within the modified Duke criteria [20]. Guided by the definitions for typical pathogens, we defined a confirmed case of C. diphtheriae IE as at least two sets of blood cultures positive for monomicrobial nontoxigenic C. diphtheriae with vegetation(s) on echocardiogram. A probable case of IE was defined as at least two sets of blood cultures positive for monomicrobial nontoxigenic C. diphtheriae with embolic phenomena. As in the modified Duke Criteria, a blood culture set consists of paired, simultaneously drawn aerobic and anaerobic blood culture specimens.

For the case series, we performed chart reviews to describe the clinical course of the 4 of 5 patients with IE from whom we were able to obtain consent. We extracted demographic information, comorbidities, risk factors, presenting symptoms, laboratory and imaging studies, antibiotic treatment, and hospital course. For individual clinical histories, gender is omitted and “they/them” pronouns are used to reduce patient-identifying information.

Broader Epidemiology of Corynebacterium diphtheriae Detections

We conducted a formal electronic health record search to identify all patients aged 18 years and older with C. diphtheriae detected in a clinical specimen (ie, wound, blood, sputum) during 1 September 2020–1 April 2023. Through chart review, we collected patient demographics, comorbidities (including human immunodeficiency virus [HIV], hepatitis C, diabetes, heart disease, and mental health diagnoses [depression, anxiety, bipolar disorder, schizophrenia]), housing status, lifetime history of injection drugs, substances used (methamphetamine, heroin, cocaine, non-prescription opioid, fentanyl) and route of administration (injected or inhaled) within the last 3 months, and history of prior wound infections. We identified the highest level of medical care required at presentation (critical care, inpatient admission, or outpatient or brief emergency department evaluation). Data were compared between 2 authors and managed using REDCap (Research Electronic Data Capture) [21, 22].

To assess for increasing cases over the study period, we calculated a Spearman rank correlation between number of months from September 2020 and the monthly number of patients with C. diphtheriae detected in a clinical specimen. Significance was set at P < .05. Calculations were performed using R version 4.2.3 [23].

Laboratory Evaluation

Corynebacterium diphtheriae detections in cultures were identified by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-ToF Biotyper; Bruker, Billerica MA). We abstracted information on C. diphtheriae detections in blood, wound, and respiratory cultures and data on other bacterial pathogens detected. Antimicrobial susceptibility testing on C. diphtheriae isolates was performed using gradient strips (bioMérieux) according to Clinical and Laboratory Standards Institute (CLSI) M45-A3 guidelines [24, 25]. Minimum inhibitory concentrations (MICs) were interpreted according to CLSI M45-A3 and European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoints [25, 26]. Isolates were forwarded to the Washington State Public Health Laboratory for processing and then to the Centers for Disease Control and Prevention (CDC) Pertussis and Diphtheria Laboratory for polymerase chain reaction testing to screen for the toxin (tox) gene and the Elek test to assess for toxin production.

Patient Consent Statement

For the patients with IE, we contacted those who were still living and next of kin for those who had died to obtain consent for sharing full clinical histories. We attempted contact at least twice. We obtained consent for 4 patients; 1 patient or next of kin could not be reached and their aggregate data were analyzed but their detailed clinical summary is excluded.

Chart abstraction work on C. diphtheriae detections was approved by the University of Washington Human Subjects Division (STUDY00005397).

RESULTS

Endocarditis Case Series

Case 1

A person in their 30s with a history of depression presented with a 5-day history of fever, abdominal pain, dyspnea, and left-sided weakness. They had recently injected drugs and were experiencing homelessness. Their temperature was 35.8°C and other vital signs were stable. No wounds were noted on examination. They had leukocytosis (30 000 cells/µL); HIV testing was nonreactive. Urine toxicology was positive for methamphetamine and fentanyl. The patient was empirically treated with intravenous vancomycin and ceftriaxone. Three sets of blood cultures initially grew nontoxigenic C. diphtheriae; subsequent daily blood cultures were negative after 1 day of antibiotics. Imaging demonstrated splenic and renal infarcts and cerebral septic emboli. Transthoracic echocardiogram (TTE) revealed vegetations on the anterior (2.3 × 1.7 cm) and posterior (1.4 × 0.8 cm) mitral valve leaflets. The patient began dual antibiotic therapy with intravenous vancomycin and ampicillin when the isolate was found to have intermediate sensitivity to ceftriaxone (MIC = 2 mg/L). Gentamicin was avoided given the renal infarct. The patient underwent surgical mitral valve replacement given sizeable vegetations with destruction of valve leaflets. They transitioned to vancomycin monotherapy for a planned 6-week course and were discharged in stable condition. However, the patient acquired coronavirus disease 2019 (COVID-19) after their hospitalization and died within 30 days from that infection.

Case 2

A person in their 40s with hepatitis C complicated by cirrhosis presented with a 5-day history of confusion, body pain, and loss of consciousness. They had injected drugs in the prior 3 months and were experiencing homelessness. No wounds were noted on examination. The patient was afebrile, tachycardic (109 beats/min), and tachypneic (24 breaths/min). They had leukocytosis (47 810 cells/µL), thrombocytopenia (11 000 cells/µL), hyperbilirubinemia (16 mg/dL), acute kidney injury (AKI) (creatinine [Cr], 2.21 mg/dL), and severe lactic acidosis (lactic acid, 10.6 mmol/L). A urine toxicology screen was positive for methamphetamines, benzodiazepine, fentanyl, and methadone. Computed tomography (CT) angiography of the chest, abdomen, and pelvis identified hypodense lesions in the aortic leaflet, spleen, and kidney. Head CT showed multiple embolic infarctions. A TTE showed 3-valve endocarditis with aortic (1.4 × 1.5 cm), mitral, and tricuspid valve vegetations. They were empirically treated with intravenous vancomycin, ceftriaxone at central nervous system dosing (2 g every 12 hours), and metronidazole. Three sets of blood cultures grew nontoxigenic C. diphtheriae. Penicillin G was added due to continued critical illness. After 1 day of antibiotics, subsequent daily blood cultures were negative. The patient developed progressive pressor-dependent shock with multisystem organ failure, and died 6 days after admission.

Case 3

A person in their 30s presented with a week-long history of cough, fever, and dyspnea, which progressed to altered mental status. They were unstably housed and had a history of injecting drugs. The patient was obtunded on arrival to the hospital, febrile (38.3°C), tachycardic (144 beats/min [bpm]), hypertensive (193/151 mmHg), and tachypneic (64 breaths/min) with an oxygen saturation of 93%, requiring intubation. No wounds were noted on examination. Urine toxicology screen was positive for methamphetamines, opiates, and cannabinoids. A TTE demonstrated vegetations on the anterior and posterior mitral valve leaflets (2 × 1 cm) and the aortic valve (1.2 × 0.7 cm) with severe aortic insufficiency. Chest CT showed multifocal lung disease. The patient was treated with intravenous vancomycin, cefepime, and metronidazole, which was later expanded to vancomycin, meropenem, micafungin, and azithromycin. Two sets of blood cultures were positive for nontoxigenic C. diphtheriae on the day of admission and the subsequent day. Cultures collected after hospital day 3 were negative. The patient was evaluated by the cardiac surgery team for valvular repair but developed worsening pressor requirements and died within 96 hours of admission.

Case 4

A person in their 70s with a history of atrial fibrillation, chronic systolic heart failure (ejection fraction, 35%), and diabetes mellitus was found unconscious and in respiratory distress. The patient was housed with a remote history (>10 years prior) of injecting drugs. They were afebrile, tachycardic (197 bpm), hypotensive (90/48 mmHg), tachypneic (40 breaths/min), and hypoxic on room air (91%). The patient was intubated, started on 2 pressors, and treated with intravenous vancomycin and cefepime for presumed mixed cardiogenic and septic shock. On examination, they had venous stasis changes and a pretibial ulcer. They had leukocytosis (14 910 cells/µL), AKI (Cr, 4.31 mg/dL), and lactic acidosis (lactic acid, 6.2 mmol/L). A urine toxicology screen was negative. Sputum cultures and two sets of blood cultures were positive for nontoxigenic C. diphtheriae. A TTE demonstrated worsening mitral regurgitation compared with an echocardiogram 1 year prior but no clear vegetations. A TEE could not be performed given his critical illness. Head CT demonstrated multiple infarctions. Abdominal imaging was not done. Lactate increased to 7.7 mmol/L, concerning for acute mesenteric ischemia. The patient developed worsening multisystem organ failure in the setting of persistent bacteremia and died within 48 hours of admission.

Aggregate Data From Chart Abstractions

Summary of Infective Endocarditis Cases

Four patients with IE were male and one patient identified as a transgender woman. Two patients had cardiac risk factors: 1 had an underlying prosthetic valve replacement and 1 had heart failure. Four patients had likely recent skin injury: 2 recently injected drugs, 1 had chronic wounds, and 1 had a traumatic injury. During their clinical IE course, 3 patients had cerebral embolic infarctions, 3 had renal or splenic infarcts, and 2 underwent cardiac surgery (Table 1).

Table 1.

Patient Characteristics, Antibiotic Regimens, Blood Culture Clearance, and Outcomes for 5 Patients With Infective Endocarditis in Our Hospital System—April 2022–April 2023, Seattle, Washington

Patient	Age group, y	Antibiotic Regimen and Transitions	Clearancea	Cerebral Septic Emboli Present	Cardiac Surgery Performed	Died? Cause, Time
1	30s	Initial: vancomycin, ampicillin Subsequent: vancomycin monotherapy	1 d	Yes	Yes	Yes Non-IE in 30 d
2	20s	Vancomycin monotherapy	1 d	No	Yes	No
3	40s	Initial: vancomycin, ceftriaxone, metronidazole Subsequent: vancomycin, ceftriaxone, metronidazole, penicillin	1 d	Yes	No	Yes IE in 6 d
4	30s	Initial: vancomycin, cefepime, metronidazole Subsequent: vancomycin, meropenem, micafungin, azithromycin	3 d	No	No	Yes IE in 96 h
5	70s	Vancomycin, cefepime	No clearance	Yes	No	Yes IE in 48 h

Abbreviations: IE, infective endocarditis.

^aTime after initiation of antimicrobial therapy when subsequent blood cultures had no growth.

Broader Epidemiology of Corynebacterium diphtheriae Detections

During 1 September 2020–1 April 2023, 44 patients with cultures positive for C. diphtheriae were identified, including the 5 patients with IE. Detections have been increasing since March 2022 (Figure 1). The Spearman’s rho for the study period was 0.6 (P = .002), consistent with an increasing trend in cases over time. The patients’ median age was 44 years (range, 23–73 years). Most patients were male (n = 33; 75%), White (n = 29; 66%), unstably housed or experiencing homelessness (n = 34; 77%), and had a mental health diagnosis (n = 38; 86%) (Table 1). Eleven (25%) patients, including the 5 patients with IE, required critical care admission at the time of C. diphtheriae detection, and 23 (52%) were admitted to the general inpatient unit. Most patients (n = 41; 93%) had skin breakdown or a wound on presentation, and 35 (80%) had prior wound infections within the last year. While most (n = 33; 75%) had injected drugs during their lifetime, the most common route of substance administration within the past 3 months was smoking/inhalation (n = 29/39; 74%) rather than injection (n = 16/39; 41%). Demographics, comorbidities, and risk factors are summarized in Table 2.

Figure 1.

Epidemiologic curve of 44 patients with Corynebacterium diphtheriae detections within a university-affiliated hospital system, color-coded by highest risk detection present (endocarditis, bloodstream, sputum, wound): 1 September 2020–1 April 2023.

Table 2.

Demographics of Patients With Corynebacterium diphtheriae–Positive Cultures in Wound, Blood, or Pulmonary Specimens (n = 44) Within a University-Affiliated Hospital System, September 2020-April 2023

Demographic Characteristics	Values
Age, median (range), y	44 (23–73)
Gender, n (%)
Male	33 (75%)
Race, n (%)
American Indian/Alaska Native, Asian or Asian American, multi-race, other	9 (20%)
Black or African American	6 (14%)
White	29 (66%)
Ethnicity, n (%)
Non-Hispanic	32 (73%)
Hispanic	3 (7%)
Unknown	9 (20%)
Unstable housing,a n (%)	34 (77%)
Comorbidities, n (%)
Mental health conditionb	38 (86%)
Hepatitis C	23 (52%)
Untreated hepatitis C	13 of 23 (57%)
HIV, virally suppressed	4 (9%)
Heart disease	7 (16%)
Chronic kidney disease or end-stage renal disease	6 (14%)
Diabetes mellitus	6 (14%)
History of prior wound infection within the last 1 year	35 (80%)
History of prior injection-site infection within the last 1 year	12 (27%)
Substance use, n (%)
Smoking (current or former tobacco use)	44 (100%)
Alcohol abuse	19 (43%)
Any lifetime history of injecting drugs	33 (75%)
Documented illicit substance use (within in the last 3 months)	39 (89%)
Documented route of administration (not mutually exclusive), (n = 39)
Inhaled	29 (74%)
Injected	16 (41%)
Documented substances (not mutually exclusive), (n = 39)
Methamphetamine	34 (87%)
Fentanyl	22 (56%)
Heroin	9 (23%)
Cocaine	8 (21%)
Highest level of medical care required at time of C. diphtheriae detection, n (%)
Critical care admission or emergency critical care	11 (25%)
General inpatient admission	23 (52%)
Outpatient or noncritical emergency room visit	10 (23%)
Culture results, n (%)
Polymicrobial wounds with C. diphtheriae	36 (77%)
Concomitant organisms (not mutually exclusive; most prevalent shown), (n = 36)
Staphylococcus aureus	29 (81%)
Group A Streptococcus	27 (75%)
Proteus species	9 (25%)
Arcanobacterium haemolyticum	5 (11%)
Polymicrobial blood cultures with C. diphtheriae	6 (14%)
Concomitant pathogens (not mutually exclusive), (n = 6)
Staphylococcus aureus	5 (83%)
Group A Streptococcus	3 (50%)
Coagulase-negative Staphylococcus	1 (17%)
Monomicrobial blood cultures	6 (14%)
Endocarditis	5 of 6 (83%)
C. diphtheriae in 1 of 2 bottles	1 of 6 (17%)
Sputum culture	1 (2%)
Clinical presentation, n (%)
Wound/skin breakdown	41 (93%)
Etiologies (not mutually exclusive), (n = 41)
Chronic wounds	28 (68%)
Acute traumatic injury	10 (24%)
Surgical wound	4 (11%)
Burn injury	3 (7%)
Endocarditis	5 (11%)
Deaths	9 (20%)

Abbreviation: HIV, human immunodeficiency virus.

^aUnstable housing includes persons experiencing homelessness (living in shelter or street) as well as persons with informal temporary housing.

^bMental health condition includes depression, anxiety, bipolar disorder, schizophrenia, and borderline personality disorder.

Although we do not have access to vital statistics and cannot confirm the cause of death, 9 (20%) of 44 patients with C. diphtheriae detections had death documented in the medical record at the time of chart review. Four were patients with IE, of whom 3 died from IE-related complications within 1 week of presentation.

Laboratory Evaluation

Most patients with a C. diphtheriae detection had polymicrobial wound cultures with C. diphtheriae on presentation (n = 36; 82%); 29 of 36 (81%) had concomitant Staphylococcus aureus and 27 (75%) had Group A Streptococcus (not mutually exclusive).

Six (14%) patients had polymicrobial bacteremia with C. diphtheriae. Three of 6 patients (50%) had concomitant S. aureus and Group A Streptococcus bacteremia; 2 (33%) had S. aureus bacteremia; and 1 patient had coagulase-negative staphylococcal bacteremia. Four of 6 patients (66%) with polymicrobial bacteremia also had polymicrobial wound cultures. Blood cultures from all 6 patients were negative for C. diphtheriae after 1 day of antibiotics.

Six (14%) patients had monomicrobial C. diphtheriae blood cultures, of whom 5 (83%) are the patients with IE described above. The patient without IE was an outpatient with C. diphtheriae in 1 of 2 blood culture specimens and a concomitant polymicrobial wound infection from which C. diphtheriae was not recovered.

One patient (2%) had a polymicrobial sputum specimen (including Staphylococcus and Pseudomonas) from a ventilator-associated pneumonia.

Corynebacterium diphtheriae isolates were intermediate (95%) or resistant (5%) to ceftriaxone, with 100% sensitivity to vancomycin (Table 3).

Table 3.

Antimicrobial Susceptibility Testing of Corynebacterium diphtheriae Isolates From September 2020 to April 2023—Seattle, Washington

Antimicrobial Agent (No. Tested)	Interpretive Criteria (S)	MIC Range, mg/L	MIC50, mg/L	MIC90, mg/L	%S	%I	%R
Amoxicillin/clavulanic acid (n = 39)a	≤1 mg/L	0.25–1	0.5	0.5	…	…	…
Ceftriaxone (n = 39)b	≤1 mg/L	2–4	2	2	…	95	5
Clindamycin (n = 39)b	≤0.5 mg/L	0.125–1	0.5	0.5	97	3	…
Erythromycin (n = 30)b	≤0.5 mg/L	0.03–0.32	0.032	0.064	100	…	…
Levofloxacin (n = 39)a	…	0.125–0.25	0.12	0.25	…	…	…
Linezolid (n = 32)b	≤2 mg/L	0.25–1	0.5	0.5	100	…	…
Tetracycline (n = 37)b	≤4 mg/L	0.06–16	0.25	0.25	100	…	…
Trimethoprim/sulfamethoxazole (n = 13)b	≤2 mg/L	0.125–1	0.12	1	100	…	…
Vancomycin (n = 37)b	≤2 mg/L	0.5–2	1	1	100	…	…

Abbreviations: EUCAST, European Committee on Antimicrobial Susceptibility Testing; CLSI, Clinical and Laboratory Standards Institute; MIC, minimum inhibitory concentration; MIC₅₀, lowest antibiotic concentration at which 50% of isolates were inhibited; MIC₉₀, lowest antibiotic concentration at which 90% of isolates were inhibited; S, sensitive; I, intermediate; R, resistant.

^aAccording to EUCAST breakpoints.

^bAccording to CLSI M45 third-edition breakpoints.

DISCUSSION

We report 4 confirmed and 1 probable case of nontoxigenic C. diphtheriae IE over a 1-year period, in the context of 39 other patients with C. diphtheriae detections over a 2.5-year period. Thirty-four patients (77%) with C. diphtheriae required either acute or critical care for a related or concomitant condition, and 9 died. The patients with IE had an 80% all-cause, 30-day mortality rate. The substantial increase in cases over time, and the clustering in time and space of IE cases, raises concern for an ongoing high-morbidity C. diphtheriae outbreak within Seattle and King County.

Although nontoxigenic C. diphtheriae infections have been identified in Seattle for over 40 years [19], C. diphtheriae detections increased during the past 3 years in western Washington State among PEH and populations with a history of substance use [27]. Nontoxigenic C. diphtheriae detections are also rising globally [12, 28, 29], which may merit enhanced surveillance for this public health threat. While highly sensitive culture systems and MALDI-ToF have enhanced the ability to detect C. diphtheriae [30], the increasing reports of invasive disease suggests that these detections are not solely attributable to changes in diagnostic methods.

Many non-IE C. diphtheriae detections were from polymicrobial infections, with S. aureus and Group A Streptococcus frequently identified as co-isolates, consistent with prior literature [31–33]. Corynebacterium diphtheriae, often found in chronic nonhealing wounds, may impair wound healing and enhance secondary colonization and infection by other invasive bacteria [7, 15, 34]. Consequently, our hospital laboratory identifies any “diphtheroids” to the species level in wound cultures if the organism is predominant, in pure culture, or present in cultures with S. aureus and Group A Streptococcus, rather than considering them to be contaminants [33].

The virulence of nontoxigenic C diphtheriae is poorly understood. The organism has been shown to produce biofilms and fibrin deposition that may facilitate cytokine activation and tissue invasion [35, 36]. Nontoxigenic C. diphtheriae bloodstream isolates exhibit superior adherence and microtubule-dependent internalization by endothelial cells in in vitro studies, which may be related to pilin, adhesin, iron-uptake genes, and collagen-binding proteins [30, 37]. While our IE cases may represent increasing C. diphtheriae transmission locally, examining virulence factors may help identify if molecular changes are also facilitating increases in invasive disease.

Currently, there are no standard treatment recommendations for patients with nontoxigenic C. diphtheriae IE, and C. diphtheriae is not listed as an IE-associated microorganism by the 2023 modified Duke criteria [1, 4, 6, 20]. Prior C. diphtheriae IE treatment regimens reported in the literature have included vancomycin monotherapy, penicillin and gentamicin dual therapy, and other combinations involving vancomycin, ceftriaxone, and aminoglycosides [4]. However, due to the rarity of this infection, there is no strong evidence for the superiority of 1 regimen over another. Some reports note rising rates of resistance to gentamicin, β-lactams, quinolones, and daptomycin [1, 14]. Most concerning are the reports of reduced susceptibility to β-lactams, and none of our isolates were susceptible to ceftriaxone. Meropenem may be a treatment alternative in such cases [38]. While our patients with IE had isolates susceptible to vancomycin, 3 died despite treatment with vancomycin and escalation to combination therapy. Comprehensive management guidelines for invasive C. diphtheriae are urgently needed, including recommendations on echocardiography in the setting of C. diphtheriae bacteremia to rule out endocarditis, empiric antimicrobial therapy to optimize bactericidal activity given reports of increasing antimicrobial resistance, considerations for dual therapy with progressive disease, and the potential inclusion of C. diphtheriae as an IE pathogen in the Duke criteria.

Our study has several limitations. Medical charts may be incomplete and do not consistently describe environmental risk factors, living conditions, or housing instability. Record quality is dependent on the patients’ self-reported symptoms and risk factors, engagement with healthcare systems over time, and provider documentation. Patients at highest risk for C. diphtheriae infection (PEH with chronic wounds and substance use) are a vulnerable and stigmatized population with significant barriers to care, which may limit testing and identification of C. diphtheriae. With regard to laboratory detection, cases are not identified when wounds are not cultured or diphtheroids are not identified to the species level. The small number of patients with C. diphtheriae IE precludes any robust statistical analysis on risk factors or clinical outcomes for this local outbreak. Finally, genomic analysis is essential to determine the relatedness of clinical isolates in this outbreak: this is in process and will be reported elsewhere as part of a broader public health investigation.

Despite these limitations, our study highlights the considerable morbidity and mortality associated with rising nontoxigenic C. diphtheriae detections and invasive disease within our local setting. Greater attention to wound care, linkage to care, and improved sanitation may aid in prevention. As an emerging invasive pathogen, nontoxigenic C. diphtheriae merits increased surveillance, guidance on patient management, and potential designation as a notifiable disease [26]. As next steps, the Public Health Seattle & King County and Washington State Department of Health, supported by the CDC, are evaluating risk factors and opportunities for the prevention of C. diphtheriae statewide.

Notes

Author Contributions. E. N. K., T. F., S. T. H., C. F., and P. S. P. conceptualized the study. K. F. L. and H. N. K. coordinated systematic electronic medical record case identification and database management. E. N. K., T. F., S. T. H., and E. J. C. designed the chart abstraction tool; E. N. K. and T. F. performed chart abstractions; and E. N. K., T. F., S. T. H., and H. Z. S. provided case series data. J. M., L. B., T. T., and F. C. F. provided laboratory data, analysis, and interpretation. C. D. and N. G. provided subject matter expertise. E. N. K. wrote the initial draft and performed epidemiologic analysis. P. S. P. provided supervision. All authors provided critical revision.

Acknowledgments. The authors acknowledge Jennifer Lenahan, Seattle & King County Communicable Disease Epidemiology and Immunizations Section; Sofia Husain and Scott Lindquist, Washington State Department of Health; Commander Michelle L. Holshue, Centers for Disease Control and Prevention; and the patients and families affected by C. diphtheriae within Seattle & King County.

Financial support. This work was supported by the National Institutes of Health (grant numbers 5T32AI007140-45 to E. N. K. and T. F.; 5T32AI007044-47 to S. T. H.; 5T32AI7044-48 to H. Z. S.).