Legionella Pneumonia Due to Non-Legionella pneumophila Serogroup 1

Abstract

Legionella pneumophila is one of the important pathogens in community-acquired (CAP) and hospital-acquired pneumonia that can cause severe pneumonia. Early diagnosis and treatment of Legionella pneumonia (LP) are essential because inappropriate therapy for Legionella pneumonia has been reported to worsen the prognosis. The most frequently identified causative pathogen of Legionella pneumonia is Legionella pneumophila serogroup 1. Legionella pneumonia due to non-Legionella pneumophila serogroup 1 is seen in 20% of cases. In diagnosing Legionella pneumonia caused by non-Legionella pneumophila serogroup 1, the urinary antigen test is usually negative; therefore, we need to suspect Legionella pneumonia by clinical information such as symptoms, vital signs, laboratory findings, and radiological findings. Based on our previous report, Legionella pneumonia due to non-Legionella pneumophila serogroup 1 was a mild to severe pneumonia. In addition, in about half of the patients, we could not suspect Legionella pneumonia using a six-point scoring system, which is one of the diagnostic scoring systems. Recently, a new urinary antigen test kit that could theoretically diagnose Legionella pneumonia due to non-Legionella pneumophila serogroup 1 was released in Japan. This can help in early diagnosis of Legionella pneumonia, including the one caused by non-Legionella pneumophila serogroup 1.

1. Introduction

Legionella pneumonia (LP) is caused by Legionella species that are important causative pathogens of community-acquired pneumonia (CAP) and hospital-acquired pneumonia. There are 58 species and three subspecies in the Legionella genus [1]. Legionella species are small to filamentous, Gram-negative rods [2].

The most frequently identified causative microorganism of Legionella pneumonia is Legionella pneumophila serogroup 1, accounting for about 80% of cases [3, 4]. In CAP, the rate of LP is reported to be 0.6–8% [5, 6, 7, 8], although the rate differs in different areas and countries. However, in severe CAP that satisfies the Infectious Diseases Society of America/American Thoracic Society severe pneumonia criteria [9], LP is one of the most important etiologies, because the rate of LP was reported to be 13.5% in 133 patients [10] and 14.1% in 71 patients <60 years old [11]. In addition, inappropriate initial therapy for LP was shown to be one of the independent factors predicting a worse prognosis [12]. Therefore, early and appropriate diagnosis of LP is very important to improve the prognosis of LP patients.

The gold standard in the diagnosis of LP is the identification of Legionella species in respiratory specimens such as sputum and bronchoalveolar lavage fluid. However, some LP patients have no sputum for culture, a dedicated culture medium, such as Wadowsky-Yee-Okuda-α or Buffered Charcoal Yeast Extract-α medium is needed, therefore identification of Legionella species is sometimes difficult, costly and time-consuming.

Currently, a urinary antigen test that detects soluble antigens is widely used for diagnosing LP in daily clinical practice worldwide. This diagnostic method is very useful because the examination procedure is simple and the results are known quickly. In a systematic review and meta-analysis, Shimada et al. reported that the specificity of the Legionella urinary antigen test was 99.1% and sensitivity was 74%; therefore, LP cannot be ruled out if this test is negative. Specifically, the sensitivity of the urinary antigen test for diagnosing LP due to non-L. pneumophila serogroup 1 is low [13]. Therefore, the Legionella urinary antigen test is not useful for diagnosing LP caused by non-L. pneumophila serogroup 1.

The diagnostic key for LP due to non-L. pneumophila serogroup 1 is to suspect Legionella pneumonia based on clinical information such as patients’ symptoms, vital signs, laboratory findings, and radiological findings. Therefore, in this chapter, we describe the clinical characteristics of LP due to non-L. pneumophila serogroup 1 referred to in previous reports.

2. Legionella pneumonia due to non-Legionella pneumophila serogroup 1

2.1 Previous reports

In earlier studies, LP due to non-L. pneumophila serogroup 1 could be a mild to moderate pneumonia [14, 15], not only a severe pneumonia admitted to intensive care unit [16, 17, 18, 19, 20]. Indeed, we reported a case of LP due to L. pneumophila serogroup 9 in which initial treatment with single-dose oral azithromycin appeared useful, although oral levofloxacin was administered subsequently [15].

There have been many case reports of LP caused by non-L. pneumophila serogroup 1, but there have been few case series. Therefore, we investigated the clinical characteristics of LP due to non-L. pneumophila serogroup 1 and compared with LP due to L. pneumophila serogroup 1 [21]. There were 11 patients with LP due to non-L. pneumophila serogroup 1 between March 2001 and June 2016 in our hospital. Their age range was 58–82 years, and eight patients (72.7%) were male. The most common comorbidities were diabetes mellitus, chronic liver disease, and malignant disease in each of the two patients. The most common symptom was fever (72.7%), followed by cough (54.5%), and sputum (54.5%). The distribution of bacterial strains was L. pneumophila serogroup 3 in six patients, L. pneumophila serogroup 9 in three patients, L. pneumophila serogroup 6 in one patient, and L. longbeachae in one patient. As to the severity of pneumonia, about half of the cases (5/11) were mild to moderate according to the pneumonia severity index (PSI) [22], whereas most cases (10/11) were mild to moderate based on CURB-65 [23]. Five patients whose PSI class was less than IV are all improved with oral azithromycin or oral levofloxacin. In contrast, four patients were admitted to the intensive care unit, and all four patients were administered appropriate empiric antimicrobial agents, but three patients died.

2.2 Diagnostic scoring system

We cannot rule out LP by a negative result of the urinary antigen test because the sensitivity of this test is not very good. To diagnose LP with a negative urinary antigen test, we need to suspect it based on the symptoms, vital signs, laboratory examinations, and radiological findings.

In 1998, Cunha advocated a diagnostic scoring system for LP called the “Winthrop-University Hospital (WUH) criteria” [24]. The WUH criteria comprised 15 clinical findings and seven laboratory findings, and it was therefore thought to be too complicated to use in the daily clinical practice.

In 2009, Fiumefreddo proposed a six-point scoring system for predicting LP [25], and this scoring system was validated by Haubitz [26]. This scoring system comprised one symptom, one vital sign, and four laboratory findings. The criteria for the six items are listed in Table 1. A score ≥5 had very high specificity (99.0%) and a high positive predictive value (17.4%), whereas a score <2 had high sensitivity (94.4%) and a high negative predictive value (99.6%). In our previous reports [21], using a cutoff value of ≥2 points, the sensitivity of this scoring system was 54.5% for non-L. pneumophila serogroup 1 patients and 95.7% for L. pneumophila serogroup 1 patients. Therefore, we could not rule out LP due to non-L. pneumophila serogroup 1 using this six-point scoring system. In Figure 1, the patient number and total scores of the six-point scoring system in LP due to non-L. pneumophila serogroup 1 and L. pneumophila serogroup 1 are shown [21].

Temperature	>39.4°C
C-reactive protein	>187 mg/L
Lactate dehydrogenase	>225 mmol/L
Platelets	<171 × 10^9/L
Serum sodium	<133 mmol/L
Unproductive cough

Table 1.

Criteria for each item in the six-point scoring system.

Figure 1.

Total scores of the six-point scoring system in Legionella pneumonia due to non-L. pneumophila serogroup 1 and L. pneumophila serogroup 1.

3. Future perspective

Patients with LP have a worse prognosis if they are not treated with appropriate antibiotic therapy as soon as possible. Some patients with LP due to L. pneumophila serogroup 1 have a negative urinary antigen test, and patients with LP due to non-L. pneumophila serogroup 1 are usually negative on this test. Therefore, it is important to suspect LP based on the clinical findings. However, as shown in our previous report, there are some LP patients in whom we cannot suspect LP based on the clinical findings specific to LP due to non-L. pneumophila serogroup 1. Thus, a simple method and a rapid test kit for diagnosing LP due to non-L. pneumophila serogroup 1 are needed.

In February 2019, Asahi Kasei Pharma Corporation released a urinary antigen test kit that could diagnose LP due to non-L. pneumophila serogroup 1, not only due to L. pneumophila serogroup 1. This kit uses an immunochromatographic method that has a monoclonal antibody recognizing a ribosomal protein L7/L12 unique region of L. pneumophila serogroups 1–15.

In the future, we expect that early diagnosis of LP including non-L. pneumophila serogroup 1 will be possible using this test kit.

4. Conclusion

LP due to non-L. pneumophila serogroup 1 can be a mild to severe pneumonia. To diagnose LP, there are some patients with LP caused by non-L. pneumophila serogroup 1 that could not be suspected to have LP based on their clinical findings, although diagnostic scoring systems have been reported to be useful for predicting LP. We need to investigate the usefulness of the new urinary antigen test kit that could theoretically diagnose these patients.

Acknowledgments

The authors would like to thank all of our colleagues who recruited and treated the patients. They would also like to thank Hiroyuki Fujii from the Department of Clinical Laboratory, Ohara Healthcare Foundation, Kurashiki Central Hospital, for performing sputum culture for Legionella species identification; Dr. Hiroshi Nakajima from the Department of Bacteriology, Okayama Prefectural Institute for Environmental Science and Public Health; and Dr. Junko Amemura-Maekawa from the Department of Bacteriology I, National Institute of Infectious Diseases, for performing Legionella species identification.

Conflict of interest

The authors declare no conflict of interest.