Can we truly rely on the urinary antigen test for the diagnosis? Legionella case report

Abstract

It is critical to diagnose and treat Legionella pneumonia (LP) immediately after infection because of the associated high mortality. The urine antigen test (UAT) is often used for the diagnosis of LP; however, it cannot detect the serogroups of all Legionella species. A detained medical history and several clinical findings such as liver enzyme elevation and hyponatremia are useful in diagnosis. Some specific types of Legionella are found in compost. Herein, we report a case of LP in which the patient's medical history and several clinical findings were useful for diagnosis.

1. Introduction

The urine antigen test (UAT) is often used in the primary care setting for the diagnosis of Legionella pneumonia (LP); however, the test is not adequately sensitive and the results are frequently uncertain. For the present case, the patient's medical history and several clinical findings proved essential for LP diagnosis. The UAT results could not aid the diagnosis owing to its limited sensitivity.

2. Case Presentation

A 69‐year‐old Japanese man presented to the emergency department for a 3 day history of a fever and chills. He did not have a cough, rhinorrhea, pharyngalgia, dyspnea, nausea, diarrhea, or loss of appetite. He had diabetes mellitus (HbA1c 7.0% 11 days before admission) and atrial fibrillation, and he was receiving glimepiride, sitagliptin, and warfarin medication. In terms of risk factors for LP, he was an ex‐smoker and a gardener, but had not traveled abroad or taken a public bath.

On physical examination, his body temperature was 36.3°C, his heart rate was 91 beats per minute, his blood pressure was 99/68 mm Hg, his respiratory rate was 16 breaths per minute, and his SpO₂ was 98% in room air. Coarse crackles were heard in his right lower lung field. The rest of the physical examination was normal.

His laboratory test results are shown in Table 1. The blood test showed anemia, thrombocytopenia, liver enzyme elevation, hyponatremia, and elevated creatinine phosphokinase. Proteinuria and microhematuria were revealed via urine analysis. The result of a Legionella UAT was negative. We assessed his sputum Gram stain and classified it as Geckler group 6; no bacteria were detected in the smear. The sputum culture test showed the presence of only alpha‐hemolytic Streptococcus and Candida. It was negative for Legionella though we did not use the buffered charcoal yeast extract agar base to detect Legionella species. A chest radiograph and a computerized axial tomography scan showed consolidation in the right lower lobe (Figures 1 and 2).

Table 1.

Laboratory data on admission

Complete blood count		Blood Chemistry Serology		Arterial blood gas (Room air)
White blood cell	7700/μL	T‐protein	6.3 g/dLa	pH	7.433 mm Hg
Red blood cell	356×104/μLa	T‐bil	0.7 mg/dL	PaCO2	36.0 mm Hg
Hemoglobin	11.2 g/dLa	AST	265 U/La	PaO2	72.5 mm Hga
Hematocrit	31.5%a	ALT	61 U/La	HCO3	23.7 mmol/L
Platelet count	11.9×104/μLa	LDH	572 U/La	BE	0.2 mmol/L
		ALP	196 U/L	O2Sat	97.3%
Blood coagulation system		CPK	15510 U/La	AG	12.6 mmol/L
PT	17.7 sa	Amy	47 U/L	Lac	9 mg/dL
%PT	42.0%a	BUN	28 mg/dLa
PT‐INR	1.54a	Cr	1.65 mg/dLa
APTT	53.9 sa	Na	126 mEq/La
		K	3.5 mEq/L
		Cl	92 mEq/La
		Blood sugar	382 mg/dLa
		CRP	19.79 mg/dLa
Urinalysis				Urinary antigen test
Specific gravity	1.020	Ketone body	‐	Pneumococcus	‐
pH	6.0	Blood	3+a	Legionella	‐
Protein	2+a	Leukocyte	‐
Glucose	2+a

^aResults out of the reference range.

Figure 1.

Chest radiography findings. The radiograph shows consolidation in the right lower lobe, which gradually disappeared after day 2

Figure 2.

Computed tomography findings. The image shows the consolidation spanning the entire right lower lobe region

Despite the negative UAT result, we initiated ceftriaxone, azithromycin, and minocycline treatment for LP because of a high clinical suspicion based on the clinical history and physical and laboratory findings. We confirmed the diagnosis using a sputum loop‐mediated isothermal amplification (LAMP) test (Eiken Chemical Co. Ltd, Japan). We administered levofloxacin because his body temperature was elevated (without relative bradycardia) and his bloody sputum persisted. Following treatment, he recovered and he was discharged on day 12 (Figure 3).

Figure 3.

Clinical course of the patient

3. Discussion

LP comprises 2% to 9% cases of community‐acquired pneumonia.1, 2, 3, 4, 5 Generally, legionellosis is caused by infections acquired from a water reservoir or hot spring contamination. The common risk factors of LP are cigarette smoking, chronic lung disease, increasing age, and immunosuppression. This patient had history of smoking and diabetes mellitus, which we considered risk factors of LP.

Generally, in cases of LP, respiratory symptoms such as cough and sputum are not prominent at first. Hyponatremia occurs frequently in legionellosis.1, 6, 7, 8, 9, 10 Renal and hepatic dysfunction, elevated creatinine phosphokinase, thrombocytopenia, and leukocytosis are considered the common laboratory abnormalities. Hematuria and proteinuria are also common abnormalities. In this case, the patient only complained of a fever and chills, and he did not have a cough or sputum. However, he eventually presented all the aforementioned laboratory features. A Gram stain of his sputum resulted in its classification as Geckler group 6, which did not aid the diagnosis.

While the mortality rate among patients with community‐acquired legionellosis ranges from 16% to 30%, with or without treatment using inactive antibiotics such as beta‐lactam agents, it reduces to less than 10% in patients with community‐acquired LP treated with potent therapies.11, 12, 13, 14, 15 In many countries, such as Japan, USA, and UK, cases of legionellosis should be reported immediately to the public health department. This is because earlier detection and appropriate diagnosis and treatment are very important for reducing the associated high mortality risks.

The commercial UAT (Alere BinaxNOW^® Legionella Urinary Antigen Card) is an immunochromatographic membrane assay that is used to detect the Legionella pneumophila serogroup 1 soluble antigen in human urine.16 A vast majority of community‐acquired LP cases are caused by this species and serogroup.4, 17, 18 Although it is very useful to define a diagnosis of LP, it is of limited use to detect non‐serogroup 1 L. pneumophila and other species. The sensitivity and specificity of UAT for detecting other serogroups and species are currently unknown.

LAMP (Loopamp Legionella Detection Kit E) is a nucleic acid amplification method19 to specifically amplify the target gene using four primers specific to six distinct regions. The whole amplification reaction continuously takes place under isothermal conditions. We used this method to support our legionellosis diagnosis. The sensitivity and specificity of this method are 91.3% (21/23) and 100% (112/112), respectively.20 The correlation between UAT, LAMP, and culture tests is shown in Table 2; almost all patients (21/22) with positive cultures tested positive for LP via LAMP, while only 59.1% of patients (13/22) with positive cultures tested positive for LP via a UAT.19 We hypothesized that the cause of this case of LP was non‐serogroup 1 L. pneumophila infection or an infection from another species, because the positive results were obtained using the LAMP method, and negative results were observed using the UAT.

Table 2.

Correlation between UAT, LAMP, and culture tests19

		UAT			LAMP
		+	−	Total	+	−	Total
Culture	+	13	9	22	21	1	22
	−	0	113	113	0	113	113
		13	122	135	21	114	135

Inhalation or ingestion of potting soil is associated with LP. We identified several cases of legionellosis with causes related to gardening (see Table 3). While a few patients were infected by L. pneumophila, most patients were infected by L. longbeachae. This species was first isolated in 1980 from a patient with pneumonia in Long Beach, California.21 Interestingly, the UAT was negative for all patients infected with L. Longbeachae. We also assembled data regarding the distribution of Legionella species in Japanese potting soils (see Table 4). Similar reports have been published for distribution in other areas such as America, Australia, and Europe.22, 23 These reports consistently indicate that that L. pneumophila is not commonly found in the soil. Therefore, we should consider other Legionella species as the causative pathogen particularly when the soil is the possible pathway of infection, as in this case. Although we were not able to identify which Legionella species was the pathogen in this case, we speculated it was L. longbeachae or other related species and not L. pneumophila, because this patient was a gardener. This speculation explains both the negative result with the UAT and the positive result of the LAMP test.

Table 3.

Legionella cases caused by gardening

Time	Nation	Character	Outcome	UAT	Species	References
1996/7	Japan	52, m	Death		L. longbeachae	24
1999/7	Japan	83, m	Survive	‐	L. longbeachae	25
2000/5	USA	77, f			L. longbeachae	22
2000/5	USA	45, m	Death		L. longbeachae
2000/6	USA	46, f			L. longbeachae
2000/8	Netherlands	81, m	Death		L. longbeachae	26
2000/8	Netherlands	69, m	Death		L. longbeachae
2004/12	Netherlands	67, m	Death		L. longbeachae
2001/5	Australia	40s, m	Survive	+	L. pneumophila	27
2004/8	Japan	72, m	Survive	‐	L. longbeachae	28
2009/10	Japan	74, m	Survive	‐	L. longbeachae	29
Not noted	New Zealand	79, f	Death		L. longbeachae	30
Not noted	Switzerland	60, m		+	L. pneumophila	31

Table 4.

Legionella species occurring in Japanese compost

	References:32	References:24
L. bozemanii	9	13
L. longbeachae	8	9
L. micdadei	5	7
L. pneumophila	2	1
L. cincinnatiensis	1	2
L. birminghamensis	1	1
L. gormanii	1	1
L. oakridgensis	0	2
L. spp.(cannot identify)	4	8
Legionella positive	16	22
Sample size	17	30

L. pneumophila is not the main species found in compost.

In conclusion, the medical history and clinical findings are key to diagnosing LP, because the UAT is not always helpful owing to its limited sensitivity. In particular, when gardening is the route of infection, because the occurrence of L. pneumophila in the soil is rare, other species such as L. longbeachae should be considered as the potential causative pathogen for LP. In addition, in such cases, the clinicians should be careful in interpreting the result of the UAT, because the results can be negative even when these species are present. The LAMP test is useful with higher sensitivity and specificity than UAT when the clinical suspicion of legionellosis is high although UAT is negative.

Conflict of Interest

The authors have stated explicitly that there are no conflicts of interest in connection with this article.

Miyata J, Huh JY, Ito Y, Kobuchi T, Kusukawa K, Hayashi H. Can we truly rely on the urinary antigen test for the diagnosis? Legionella case report. J Gen Fam Med. 2017;18:139–143. https://doi.org/10.1002/jgf2.16