Journal of Medical Microbiology

0 downloads 7 Views 1MB Size Report
1. Comparison of Legionella longbeachae and Legionella pneumophila cases in. 1 .... antibody response to L. pneumophila serogroup 1 (ECDC, 2012). A case ...

Journal of Medical Microbiology Comparison of Legionella longbeachae and Legionella pneumophila cases in Scotland; implications for diagnosis, treatment and public health response --Manuscript Draft-Manuscript Number:

JMM-D-15-00660R2

Full Title:

Comparison of Legionella longbeachae and Legionella pneumophila cases in Scotland; implications for diagnosis, treatment and public health response

Short Title:

Comparison of Legionella longbeachae and Legionella pneumophila infection in Scotland

Article Type:

Standard

Section/Category:

Microbial Epidemiology

Corresponding Author:

Ross Liam Cameron, MPH Health Protection Scotland Glasgow , UNITED KINGDOM

First Author:

R. L. Cameron, MPH

Order of Authors:

R. L. Cameron, MPH K. G. J Pollock, MPH, PhD D. S. J. Lindsay, PhD E. Anderson, MPH, MBChB

Abstract:

The reported incidence of Legionnaires' disease caused by Legionella longbeachae has increased since 2008 in Scotland. While microbiological and epidemiological studies have identified exposure to growing media as a risk factor for infection, little is known about the differences regarding disease risk factors, clinical features and outcomes of infection with L. longbeachae when compared with L. pneumophila. A nested case-case study was performed comparing 12 L. longbeachae cases with 25 confirmed L. pneumophila cases. Fewer L. longbeachae infected patients reported being smokers (27 % [95 % CI 2 %-52 %] vs. 68 % [95 % CI 50 %-86 %], p=0.034) but more L. longbeachae patients experienced breathlessness (67 % [95 % CI 40 %-94 %] vs. 28 % [95 % CI 10 %-46 %]), p=0.036). Significantly more L. longbeachae infected patients received treatment in intensive care (50 % [95 % CI 22 %-78 %] vs. 12 % [95 % CI 0 %-25 %], p=0.036). However, the differences in diagnostic methods between the two groups may have led to only the most severe cases of L. longbeachae being captured by the surveillance system. No differences were observed in any of the other pre-hospital symptoms assessed. Our results highlight the similarity of Legionnaires' disease caused by L. pneumophila and L. longbeachae and reinforce the importance of diagnostic tools other than the urinary antigen assays for the detection of non-L. pneumophila species. Unfortunately cases of community acquired pneumonia caused by Legionella species will continue to be under-diagnosed unless routine testing criteria changes.

Powered by Editorial Manager® and ProduXion Manager® from Aries Systems Corporation

Manuscript Including References (Word document)

Click here to download Manuscript Including References (Word document) Comparison of Legionella longbeachae and

1

Comparison of Legionella longbeachae and Legionella pneumophila cases in

2

Scotland; implications for diagnosis, treatment and public health response

3

R.L. Cameron1, K.G.J. Pollock1, D.S.J. Lindsay2 and E. Anderson1

4

1

Health Protection Scotland, Glasgow, Scotland, UK

5

2

Scottish Haemophilus, Legionella, Meningococcus and Pneumococcus Reference

6

Laboratory, Glasgow, Scotland, UK

7

Corresponding author:

8

Dr Kevin Pollock

9

Meridian Court

10

8 Cadogan Street

11

Glasgow

12

G2 6QE

13

Tel: 0141 300 1913

14

Email: [email protected]

15 16

Keywords: Legionella longbeachae, Legionella pneumophila, symptoms, risk factors

17

Abstract word count: 249

18

Main text: 1761

19 20 21 22

Running title: Legionella longbeachae infection in Scotland

1

23

Abstract

24

The reported incidence of Legionnaires’ disease caused by Legionella longbeachae has

25

increased since 2008 in Scotland. While microbiological and epidemiological studies have

26

identified exposure to growing media as a risk factor for infection, little is known about the

27

differences regarding disease risk factors, clinical features and outcomes of infection with L.

28

longbeachae when compared with L. pneumophila. A nested case-case study was performed

29

comparing 12 L. longbeachae cases with 25 confirmed L. pneumophila cases. Fewer L.

30

longbeachae infected patients reported being smokers (27 % [95 % CI 2 %-52 %] vs. 68 %

31

[95 % CI 50 %-86 %], p=0.034) but more L. longbeachae patients experienced breathlessness

32

(67 % [95 % CI 40 %-94 %] vs. 28 % [95 % CI 10 %-46 %]), p=0.036). Significantly more

33

L. longbeachae infected patients received treatment in intensive care (50 % [95 % CI 22 %-

34

78 %] vs. 12 % [95 % CI 0 %-25 %], p=0.036). However, the differences in diagnostic

35

methods between the two groups may have led to only the most severe cases of L.

36

longbeachae being captured by the surveillance system. No differences were observed in any

37

of the other pre-hospital symptoms assessed. Our results highlight the similarity of

38

Legionnaires’ disease caused by L. pneumophila and L. longbeachae and reinforce the

39

importance of diagnostic tools other than the urinary antigen assays for the detection of non-

40

L. pneumophila species. Unfortunately cases of community acquired pneumonia caused by

41

Legionella species will continue to be under-diagnosed unless routine testing criteria changes.

42 43 44 45 46 47

2

48

Introduction

49

Legionella pneumophila is the most common cause of Legionnaires’ disease in countries

50

worldwide apart from Australia and New Zealand where Legionella longbeachae infection is

51

more common (O’Connor et al., 2007; Graham et al., 2012). Since 2008 in Scotland,

52

Legionnaires disease cases caused by Legionella longbeachae have been reported every year,

53

with peaks in incidence in 2013 and 2014 (Fig). Previous studies have found cases are

54

commonly keen gardeners with exposure to plant growing media (Pravinkumar et al., 2010;

55

Potts et al., 2013). Indeed, in a survey of 24 bagged potting composts in the UK, 16 % of the

56

plant growing media contained L. longbeachae (Currie et al., 2014). The rest of the UK has

57

not observed an increase in the incidence of L. longbeachae infection, most likely reflecting

58

under ascertainment; and the increase in Scotland is probably due to improved detection

59

(Fig).

60 61

Anecdotal reports from local health protection teams have suggested augmented virulence of

62

infection with L. longbeachae. Accordingly, the aim of this study was to determine the

63

differences, if any, with those infected with L. pneumophila in relation to patient

64

characteristics, outcomes, signs and symptoms. It is important to appreciate these differences

65

to aid in timely, appropriate diagnosis and the detection of potential outbreaks with resultant

66

source elucidation (Pravinkumar et al., 2010; Potts et al., 2013).

67 68

Methods

69

Health Protection Scotland (HPS) undertakes passive, enhanced national Legionella

70

surveillance and collects demographic and clinical information on each Legionnaires’ disease

71

case through collection of surveillance forms completed by public health clinicians in the

3

72

NHS Health Boards. Utilising these data, which are held in a national database, we performed

73

a nested case-case study.

74 75

Ten confirmed and two probable cases of L. longbeachae reported to HPS between 2008 and

76

2014 were identified according to the European Centre for Disease Prevention and Control

77

(ECDC) case definitions (ECDC, 2012). According to ECDC definitions all Legionnaires’

78

disease cases must have a clinical diagnosis of pneumonia (ECDC, 2012). Laboratory

79

diagnosis is required to define whether a case is confirmed or probable with cases being

80

confirmed if i) Legionella is cultured from respiratory secretions or a normally sterile site; ii)

81

if L. pneumophila serogroup 1 antigen is detected in urine; iii) or a case has a specific

82

antibody response to L. pneumophila serogroup 1 (ECDC, 2012). A case is defined as

83

probable if i) L. pneumophila antigen is detected in respiratory secretions or lung tissue; ii)

84

Legionella nucleic acid is detected in a specimen, iii) the patient has a specific antibody

85

response to non-L. pneumophila serogroup 1 iv) or a single high titre to a Legionella species

86

(ECDC, 2012). Additionally, a case can be defined as probable if they meet the clinical

87

criteria and the epidemiological criteria of exposure to the same common source or

88

environmental exposure. Epidemiological criteria are commonly utilised in outbreak

89

situations. Cases included in the study were diagnosed through a combination of culture,

90

serology, PCR and, in the situation of most L. pneumophila serogroup 1 cases, urinary

91

antigen testing. In Scotland, the majority of clinicians utilise the urinary antigen assay

92

routinely for diagnostic testing of suspected Legionella infection, however, clinical protocols

93

in some NHS Health Boards alert clinicians to perform PCR and culture on samples from

94

patients with severe pneumonia (HPS, 2014). Two L. longbeachae cases in this study were

95

culture negative but through a combination of Legionella species nucleic acid detection in

96

respiratory secretions and evidence of a immune response (either a four-fold rise in titre or a

4

97

single high titre of antibodies to L. longbeachae), these cases were identified as probable

98

cases by the Scottish Haemophilus, Legionella, Meningococcus and Pneumococcus

99

Reference Laboratory (SHLMPRL).

100 101

The twelve cases of L. longbeachae were compared to 25 confirmed cases of L. pneumophila

102

reported to HPS between 2008 and 2014. These cases were also defined according to the

103

ECDC classifications (ECDC, 2012). Cases of both L. pneumophila and L. longbeachae

104

infection were excluded if pneumonia status and clinical date of onset was unknown or not

105

reported, if symptoms were not recorded in the database or if the cases had travelled to

106

another country two to ten days before onset of symptoms. As the majority of Legionnaires’

107

disease cases are travel related in Scotland, only a small number of cases were eligible to be

108

included in the study. Further, L. pneumophila cases were excluded if they were part of an

109

outbreak, as case definitions may have been altered during the incident. Data were entered

110

into SPSS 21 (SPSS Inc., Chicago IL) for analysis. The Fisher’s exact test and the

111

independent-samples t-test was undertaken to assess differences between cases, and p