Past Endemic Izumi Fever or Yersinia pseudotuberculosis Infection Reappears Sporadically

Abstract

Izumi fever (IF), also known as Far East scarlet-like fever (FESLF), is caused by Yersinia pseudotuberculosis and it has clinical features resembling those of Kawasaki disease (KD). As both diseases are rare in adolescents and young adults, it is challenging to recognize them, thus often leading to a delayed diagnosis. We herein present two cases of IF or FESLF (IF/FESLF). The first case was misdiagnosed as KD, which led to a diagnostic delay. The second case was recognized earlier owing to our experience with the first case. Although cultures were negative in both cases, presumably due to the prior use of antimicrobial agents, our clinical suspicion and a paired serological assay for anti-Y. pseudotuberculosis antibodies finally led to a successful diagnosis.

Introduction

Izumi fever (IF), also known as Far East scarlet-like fever (FESLF), is caused by Yersinia pseudotuberculosis and it demonstrates symptoms resembling those of Kawasaki disease (KD). Although it is recognized as a major disease in the differential diagnosis of KD, especially among young children in Japan, the incidence of IF or FESLF (IF/FESLF) has markedly decreased because of improved sanitation. It is challenging to diagnose adolescent and young adult cases of IF/FESLF because both KD and IF/FESLF are rare in this age group and more relevant conditions, such as toxic shock syndrome and drug reactions, should be considered first. We herein present two cases of IF/FESLF in an adolescent and a young adult patient who were both successfully diagnosed using paired serological tests.

Case Reports

Case 1

A 21-year-old Japanese man with a 6-day history of high-grade fever, sore throat, arthralgia, palmar induration, erythema, and diarrhea visited our hospital in mid-October. Just prior to the onset of the symptoms, he ate grilled meat at a restaurant but otherwise denied any exposure to animals or the consumption of raw water. Four days before admission to the hospital, a nearby clinic prescribed levofloxacin, scopolamine, rebamipide, domperidone, and some probiotics for suspected gastroenteritis. However, the patient's symptoms did not improve. He developed KD at the age of six and had bronchial asthma and fair-controlled atopic dermatitis. Physical examination revealed a body temperature of 38.7°C, blood pressure of 117/62 mmHg, and pulse rate of 95 beats/min. Bilateral mild conjunctivitis, a reddish tongue, and oral mucosa (Fig. 1a, b), mild cervical lymphadenopathy, faint truncal erythema, and diffuse mild abdominal tenderness were noted. Furthermore, bilateral palmoplantar induration and erythema were also observed (Fig. 1c). Laboratory data revealed mildly elevated white blood cell (WBC) count (8,040 /μL), elevated C-reactive protein (CRP; 8.36 mg/dL), and erythrocyte sedimentation rate (ESR; 68 mm/h). The patient's hepatic enzyme levels were also elevated: aspartate aminotransferase (AST), 59 IU/L; alanine aminotransferase (ALT), 152 IU/L; lactic dehydrogenase [LD; measured using the Japan Society of Clinical Chemistry (JSCC) method], 301 IU/L; alkaline phosphatase (ALP; measured using the JSCC method), 537 IU/L; and gamma-glutamyl transferase (γ-GT), 209 IU/L. There was no evidence of anemia, thrombocytopenia, or any abnormal kidney function. Serological assays for Epstein-Barr virus (EBV) and cytomegalovirus showed an uninfected pattern, with all negative results. Anti-streptolysin O and anti-streptokinase antibody levels were normal. The soluble interleukin-2 receptor level was elevated (3,721 U/mL). Contrast-enhanced computed tomography (CT) revealed mesenteric lymphadenopathy of multiple small lymph nodes and hepatosplenomegaly (Fig. 2). Toxic shock syndrome was suspected due to staphylococcal or streptococcal skin and soft-tissue infections; therefore, the patient was hospitalized and cefazolin was initiated empirically. Since there was no obvious focus of skin infection, we suspected conditions such as KD, mesenteric lymphadenitis due to bacterial infection, and malignant lymphoma. Cefazolin treatment was discontinued, and the patient was switched to vancomycin because methicillin-resistant Staphylococcus aureus was identified in pharyngeal culture; however, the symptoms did not improve, and fever, diarrhea, and mucocutaneous erythema persisted. On the sixth day of admission, the WBC count and CRP levels were further elevated to 23,150 /μL and 9.91 mg/dL, respectively, and the liver function test results were abnormal. On the 10th day after symptom onset, palmoplantar desquamation appeared (Fig. 3), and the symptoms met all six criteria for KD. While monitoring cardiac involvement using echocardiography, we administered aspirin and intravenous immunoglobulin (IVIG; 1 mg/kg), which is also a common therapy for toxic shock syndrome. Serial echocardiography did not reveal any evidence of cardiac involvement, and treatment did not alleviate the symptoms. Although repeated stool cultures were negative for enteropathogenic bacteria, we suspected a Y. pseudotuberculosis infection because this pathogen is known to cause IF, which was endemic in the past and manifested as serial symptoms resembling scarlet fever and KD. After ceftriaxone initiation, the body temperature decreased, and diarrhea gradually decreased. The patient was discharged on day 25 and was scheduled to be followed-up on an outpatient basis with the continuation of aspirin. A serological investigation of paired serum samples on the day of admission and in the convalescent phase revealed elevated titers of anti-Y. pseudotuberculosis antibody (Table 1). Hence, we diagnosed Y. pseudotuberculosis infection with IF. Aspirin was discontinued and the patient was asymptomatic.

Figure 1.

Mucocutaneous findings in case 1. The patient presented with mild conjunctivitis (a), a reddish oral mucosa with exudates (b), and palmar induration and erythema (c).

Figure 2.

Abdominal computed tomography findings in case 1. Abdominal computed tomography shows diffuse mesenteric lymphadenopathy without bowel wall thickening.

Figure 3.

Palmar desquamation found in the clinical course of case 1. The palmar erythema subsided and desquamation appeared on the 10th day of symptom onset.

Table 1.

Serological Examination for Anti-Yersinia pseudotuberculosis Antibodies in Case 1.

Case 1		Serotypes
Day of admission		Yp2a		Yp5a
Day 5		1:20		1:20
Day 15		1:160		1:80

According to the reference values of the laboratory where serological assays were conducted, 1:160 or higher was considered positive.

Case 2

Six years after Case 1, we encountered an 18-year-old Japanese man who was referred to our hospital in late December because of fever and rash. He had been taking acetaminophen for two weeks for occasional lower back pain. He noticed mild erythema on the trunk 10 days before his hospital visit, which had spread to the extremities. Three days before the hospital visit, he had developed a high-grade fever with shaking chills, mild sore throat, facial erythema, and swollen hands, for which he had visited a local physician. Rapid antigen detection tests for Streptococcus pyogenes and Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) were negative. Laboratory tests showed a mildly elevated WBC count and CRP level. Levofloxacin and prednisolone were prescribed; however, the patient's symptoms did not improve. Therefore, the patient was referred to our hospital.

Upon admission, he complained of headache, sore throat, dry cough, mild diarrhea, itchy skin rash, and loss of appetite. He had developed suppurative umbilical granuloma 10 months prior to this presentation. His family had a pet dog, but he denied exposure to wildlife, insect bites, or the consumption of raw water. He was administered levofloxacin, prednisolone, rupatadine, tranexamic acid, and carbocysteine for two days. Physical examination revealed the following: body temperature, 40.4°C; blood pressure, 118/57 mmHg; pulse rate, 112 beats/min. Diffuse erythema on the face, macular erythema on the trunk and extremities, and bilateral indurative palmoplantar edematous erythema were also noted (Fig. 4a). He also had mild bilateral non-exudative conjunctivitis and left posterior cervical lymphadenopathy. His abdomen was tender in the right lower quadrant, without either guarding or hepatosplenomegaly. No other superficial lymphadenopathies or oral mucosal changes were observed. Laboratory tests revealed the following: mildly elevated WBC count (9,040 /μL) with neutrophilic predominance (89.3% of WBC); elevated ESR (54 mm/h); CRP level, 10.09 mg/dL, and abnormal liver function test results [AST: 95 IU/L, ALT: 276 IU/L, LD: 242 IU/L (measured with International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) method), ALP: 194 IU/L (measured with the IFCC method), and γ-GT: 945 IU/L]. Anti-streptolysin O and anti-streptokinase antibody levels were normal. An antiviral antibody assay revealed a post-infection pattern of EBV characterized by elevated viral capsid antigen-IgG and EBV nuclear antigen levels, normal viral capsid antigen-IgM levels, and an uninfected pattern of cytomegalovirus. Adverse reactions due to acetaminophen and adult-onset KD were suspected. However, abdominal CT indicated another possibility. It revealed ascending colonic wall thickening with adjacent fat stranding and ileocecal lymphadenopathy (Fig. 5). These CT findings raised the possibility of intestinal infectious diseases, including yersiniosis. IF/FESLF due to Y. pseudotuberculosis was suspected to be the most probable diagnosis. Acetaminophen was discontinued and ceftriaxone (2 g daily) was initiated after stool and blood cultures were obtained. Aspirin and IVIG for KD were pending, while he was monitored for cardiovascular complications using echocardiography. Colonoscopy revealed prominent lymphatic follicles in the terminal ileum, cecum, and the ascending colon. On the fourth day of admission, his body temperature normalized, and the macular rash gradually disappeared, followed by palmoplantar desquamation on the sixth day (Fig. 4b). Cultures from stool, blood, and ileocecal mucosal tissue were negative for Yersinia and other pathogens. The patient was discharged on the eighth day and was followed up on an outpatient basis. Cardiac complications such as enlarged coronary arteries and pericardial effusion were not observed. Paired serum samples for anti-Y. pseudotuberculosis antibodies revealed elevated titers of serotypes 2a, 4a, and 4b (Table 2). Consequently, a diagnosis of IF/FESLF was made, and the follow-up was discontinued.

Figure 4.

Palmoplantar findings in case 2. Indurative, edematous erythema was found at the time of admission (a), which subsided gradually, followed by desquamation on the 6th day of admission (b).

Figure 5.

Abdominal computed tomography findings in case 2. Abdominal computed tomography revealed ascending colonic wall thickening and adjacent fat stranding accompanied by local lymphadenopathy.

Table 2.

Serological Examination for Anti-Yersinia pseudotuberculosis Antibodies in Case 2.

Case 2		Serotypes
Day of admission		Yp2a		Yp4a		Yp4b
Day 2		1:40		-		-
Day 13		1:80		1:80		1:40
Day 20		1:160		1:320		1:160

Discussion

Yersinia pseudotuberculosis causes self-limiting gastroenteritis, mesenteric lymphadenitis, and terminal ileitis. In Japan and eastern Russia, Y. pseudotuberculosis has been reported to cause systemic inflammatory syndrome. Some strains prevalent in the Far East produce a superantigenic toxin called Y. pseudotuberculosis-derived mitogen A (YPMa), which is responsible for a KD-like illness (1).

Izumi et al. reported an unidentified scarlet-like febrile illness that broke out in Japan in 1927, several small outbreaks were sporadically reported until the late 1940s (2,3). The disease was assumed to be infectious and was thus named IF; however, the causative pathogen was not discovered. Further epidemics in Japan in 1977 led to the discovery that the disease was caused by Y. pseudotuberculosis by isolating the organism from the stool specimens obtained from many patients (2). In addition, several epidemics of Y. pseudotuberculosis infection have been reported in Vladivostok and other Russian Far East areas (4) since 1959. Because the initial symptoms resembled scarlet fever, the disease was named FESLF (1).

The pathogenicity of Y. pseudotuberculosis depends on the presence of pYV (a plasmid associated with Yersinia virulence), an essential virulence factor. Genotypes of Y. pseudotuberculosis are divided into six groups based on other virulence factors: the high-pathogenicity island (HPI) and YPM (consisting of three subtypes: YPMa, YPMb, and YPMc) (1,3). HPI and YPMa are closely correlated with the clinical features of Y. pseudotuberculosis. HPI carries a biosynthetic gene cluster for yersiniabactin, which is a molecule involved in siderophore-mediated iron acquisition. YPMa is a superantigenic toxin capable of activating 5-20% of the entire T-cell population by binding directly to the major histocompatibility complex class II (MHC-II) molecule, independent of antigen specificity. Superantigenic toxins are produced by some strains of Gram-positive cocci, such as Streptococcus pyogenes and Staphylococcus aureus, which are the causative pathogens of toxic shock syndrome. YPMa is the only superantigen identified in Gram-negative bacteria. These genetic groups have distinct geographical distributions that contribute to the regional differences in clinical manifestations. The most common genotype in Europe is group 2 (HPI+, YPM-), which causes gastroenteritis and mesenteric lymphadenitis, whereas the major pathogenic strains in Japan and Far Eastern Russia are group 3 (HPI-, YPMa+), which are responsible for the clinical features of IF/FESLF.

IF/FESLF has clinical features similar to those of scarlet fever, toxic shock syndrome, and KD. Interestingly, an association between Y. pseudotuberculosis infection and KD has been previously reported. KD has been suggested to be related to infectious diseases because of epidemic features and seasonal associations in its occurrence. During a nationwide KD epidemic in Japan, 10 outbreaks of Y. pseudotuberculosis have been reported (5). Epidemic KD peaks between October and May in Japan coincide with outbreaks and sporadic occurrences of Y. pseudotuberculosis infection. The age and sex distribution profiles of the two conditions were also similar. KD predominantly affects children aged <5 years, with a male-to-female ratio of 1.7:1, whereas Y. pseudotuberculosis infection mainly affects younger children (with a peak incidence age of 2 years), with a male-to-female ratio of 1.6:1 (5). In a retrospective cohort study involving 452 patients with KD, Y. pseudotuberculosis was found in 9.3% of stool cultures, and these patients had significantly higher rates of coronary artery lesions (6). Among 164 children infected with Y. pseudotuberculosis, 57 (35%) presented with KD. In another study, 29 of 329 patients with Y. pseudotuberculosis infection had KD (5). Horinouchi et al. demonstrated that 10% of patients with KD were positive for anti-Y. pseudotuberculosis and/or anti-YPM antibodies. Cardiac sequelae were more frequent in the positive group than in the negative group (7).

Although Y. pseudotuberculosis infection may partially play a role in KD development, it does not explain the entire course of the illness. KD may develop in genetically predisposed individuals exposed to specific infectious and environmental triggers. This hypothesis therefore explains the genetic and geographical distribution of KD (8).

In Case 2, antimicrobial therapy was prioritized because IF/FESLF was thought to be more likely due to obvious ileocecal intestinal wall thickening and associated lymphadenopathy. IVIG might be used for cases where signs of cardiac complications associated with KD appear while monitoring with frequent echocardiography. As an immediate and good response to ceftriaxone was observed, we did not administer any treatment for KD and continued cardiac monitoring until we found a positive serological test result. It is uncertain whether this therapeutic policy was appropriate because Y. pseudotuberculosis infection is associated with a higher incidence of cardiac sequelae in KD.

In infants and young children, it is not necessary to differentiate between Y. pseudotuberculosis infection and KD. All patients should be treated for KD because they have worse outcomes if untreated, and an effective therapeutic strategy has been clearly established. Even if these patients are infected with Y. pseudotuberculosis, they are more likely to have cardiac involvement; therefore, treatment with aspirin and IVIG is reasonable. However, there is no established evidence on whether IVIG improves the prognosis in adolescents and young adults, because both diseases are not prevalent in this age group. In a previous case series of patients with adult-onset KD, cardiac involvement was less frequent than in children (5% vs. 20%) (9). The mean diagnostic delay was 12.5-13 days, and IVIG was administered to 70-79% of the patients. Despite delayed administration, IVIG was effective in shortening the disease course but it was not significantly effective in reducing cardiac involvement (10,11). Although the clinical features met the criteria for KD in Case 1 of this report, IVIG failed to alleviate the symptoms. We could not find a good reason to administer IVIG to adult patients with KD-like illness, especially those suspected of having Yersinia pseudotuberculosis infection. Considering the possible poor cost efficiency, we could not ignore the high cost of the relatively large amounts of IVIG needed in this age group. We recognized the possibility of KD on the fourth day of onset in case 2 and monitored for cardiac involvement throughout the course, and we could safely preserve IVIG.

In both cases in this report, levofloxacin had been prescribed at other clinics before hospital admission. Although the antimicrobial agent is supposed to be effective against Y. pseudotuberculosis, it was found to be clinically ineffective. One reason for this phenomenon could be antimicrobial susceptibility; however, unfortunately, pathogenic microorganisms were not detected in either case. We did not assume that antimicrobial resistance could explain the ineffectiveness because, to the best of our knowledge, we could not find either fluoroquinolone or third-generation cephalosporin resistance in the literature. While levofloxacin was administered orally, compared with intravenous ceftriaxone, bioavailability did not seem to be the reason for this therapeutic failure because fluoroquinolones are well absorbed in the setting of infectious enterocolitis. Another reason for antimicrobial ineffectiveness could be explained immunologically. Hashimoto et al. described the case of a patient with Y. pseudotuberculosis bacteremia who presented with septic shock on the fifth day despite receiving supposedly effective antimicrobial therapy (12). Lemaitre et al. demonstrated a discrepancy between in vitro susceptibility and in vivo ineffectiveness of antimicrobials in Y. pseudotuberculosis infected mice (13). The clinical features of IF/FESLF are mainly caused by an immunological reaction triggered by the superantigenic YPMa. Antimicrobials can reduce the burden of the pathogenic bacteria and their toxins, however, they are unable to halt the immune response. The delayed administration of ceftriaxone in Case 1 would have led to a delayed clinical improvement.

As in these cases, the clinical diagnosis is particularly important because a large number of cases could be prescribed antibiotics at some clinics in advance without sampling cultures, which would presumably lead to negative culture results afterwards. As KD in adolescents and young adults is rare, physicians should be prepared to provide other differential diagnoses concurrently, including IF/FESLF. The previous recognition of the illness script may have contributed to the dramatically earlier diagnosis that was made in Case 2.

The paired serological assay revealed Y. pseudotuberculosis serotype 2a in Case 1, whereas Case 2 was positive for the three serotypes. Previous outbreaks in Japan were caused by various serotypes including 1b, 3, 4b, 5a, and 5b, with a particular predilection for 4b, 5a, and 5b (14). In contrast, an anti-Y. pseudotuberculosis antibody survey conducted on patients with KD was positive for 1a, 1b, 2b, 3, 5a, 5b, and 6, with no predilection for specific strains (7). Notably, some cases were positive for multiple serotypes. The reasons for the multi-serotype positivity were speculated to be that the serum antibody was cross-reactive with the experimental antigens and/or insufficient precision of the experimental antigen purification. The laboratory commissioned to conduct the tests reported that both phenomena occurred frequently. Combined infection with multiple strains was thought to be less likely; therefore, serotype 4a, which had the highest increase in antibody titer, was assumed to be responsible for Case 2.

There are some limitations associated with our study. Namely, we were unable to isolate Y. pseudotuberculosis from stool cultures and could not perform an anti-YPMa antibody assay. Therefore, our diagnostic reasoning for IF/FESLF was based on the clinical features.

Conclusions

We described the cases of an adolescent and a young adult with IF/FESLF who presented with clinical features resembling those of KD. Although the diagnosis of a KD-like illness among adolescents and young adults is challenging because of the rarity of the disease and the higher prevalence of other conditions, it should be considered in the differential diagnosis. Although rare in this age group, IF/FESLF should be included as the recognition of the illness script could contribute to a precise diagnosis. It is uncertain whether patients with IF/FESLF should be treated with aspirin and IVIG. However, it is clear that they should be managed with appropriate antimicrobial therapy and careful monitoring of cardiac involvement.

The authors state that they have no Conflict of Interest (COI).