Journal of Infection (I99x) zz, I53-I6O
S e a s o n a l v a r i a t i o n o f L e g i o n n a i r e s ' d i s e a s e in S c o t l a n d R. S. Bhopal* and R. J. Fallon]" * Division of Epidemiology and Public Health, The Medical School, Framlington Place, Newcastle upon Tyne NEe 4HH and t Department of Laboratory Medicine, Ruchill Hospital, Bilsland Drive, Glasgow G2o 9NB, Scotland, U.K. Accepted for publication Io September I99o Summary A summer and autumn peak in incidence is a characteristic epidemiological feature of Legionnaires' disease, a fact which may help in diagnosis. Previous reports from Scotland have reported a winter peak. This study examined the seasonal variation in Legionnaires' disease in Scotland in each year and by category of diagnosis. A characteristic summer/autumn peak was seen for travel-related infection. For nontravel infection, there was an autumnal peak between I978 and I982 but an early winter peak between I983 and I986. Overall, the pattern could be described as a summer/autumn plateau reaching a peak in early winter. This pattern, seen in men and women, was dearest in the Greater Glasgow area. Nosocomial cases were clustered in the months of October, November and December. In non-travel cases the classical seasonal pattern was not observed and clinicians cannot rely upon this epidemiological feature for diagnosis. However, the relative infrequency of the disease between January and May was a consistent and potentially useful feature. Most reports on seasonal variation are based on aggregated data on travel and nontravel cases. T h e present observations question the hypotheses developed to explain the seasonal pattern and call for further studies.
Introduction L e g i o n n a i r e s ' disease, in b o t h e p i d e m i c and sporadic f o r m , is c o m m o n e r in s u m m e r an d early a u t u m n 1' ~ and, hence, admission of a p n e u m o n i a pat i ent at such a time s h oul d alert the clinician to this diagnosis. 3-~ T h e seasonal peak has b e e n o b s e r v e d for c o m m u n i t y - a c q u i r e d and nosocomial cases 4 and for b o t h travel and n o n - t r a v e l associated cases. 6' 7 E x p l a n a t i o n s given for the seasonal variation in cl ude t hes e : 1egionellae m u l t i p l y faster in the w a r m e r waters o f s u m m e r ; the greater use o f cooling towers in s u m m e r for a i r - c o n d i t i o n i n g m a y p r o v i d e o p p o r t u n i t i e s for d i s s e m i n a t i o n ; th e w a r m e r cooling t o w e r w a t e r o f late s u m m e r allows legionellae to b e c o m e infective and transmissible to h u m a n s ; the life cycle o f species w hi c h gr ow in symbiosis with legioneUae d e t e r m i n e s th e n u m b e r s o f legionellae. F o r example, legionellae are digested by am oebae at low t e m p e r a t u r e s b u t infect a m o e b a e at t e m p e r a t u r e s of 30 degrees cen tig r ad e or m o r e 8 a nd t he n u m b e r o f legionellae rises in association with t he d e g r a d a t i o n o f t he pl a nt Myriophyllum spicatum. 9 oi63-4453/9I/O2OI53 +08 $03.00/0
© I 9 9 I The British Society for the Study of Infection
I54
R. S. BHOPAL AND R. J. FALLON
T w o studies have shown that legionellae are easier to isolate from natural water in summer and autumn than in winter. 1°,11 Climate is important in the web of causation of Legionnaires' disease but in ways which are poorly understood. For example, high winds were associated with the Stafford and Washington outbreaks 12'13 and wind direction appeared to determine the distribution of cases in several outbreaks, e.g. the Dennistoun outbreak in Glasgow 14 and the BBC outbreak in London. 15 Humidity is important for the survival of all organisms pathogenic to man 16 and this has been demonstrated in vitro for legionellae. ~7 Weather, and particularly wind speed, also determines the dissemination and deposition of aerosols. Particles exceeding o.I m m in diameter rapidly settle but droplet nuclei which are less than o'I m m may be suspended in air and carried long distances, particularly by air-currents, but also by diffusion and sedimentation. TM ~8 While weather may explain seasonal variation, its precise role is undetermined. An analysis of cases in Scotland up to I98I showed that, excluding travel cases, there was no clear seasonal variation except that there were fewer cases in the first quarter of the year. '9 Later, Fallon reported that in Scotland the disease incidence peaked in the fourth quarter of the year, and commented that this was paradoxical due to the cooler weather in Scotland and the earlier switching off of cooling towers. ~° These observations, which were confirmed in a study of I984 cases, defied explanation. 21 We now report a more detailed study of the m o n t h - b y - m o n t h variation in Legionnaires' disease in Scotland between I978 and I986. Our emphasis has been on separating travel and nontravel cases, examining the pattern in several areas of Scotland, and assessing the consistency of the pattern over time. Sources of data and methods C a s e l i s t i n g , p a t i e n t d e t a i l s a n d c l a s s i f i c a t i o n o f cases
T h e Department of Laboratory Medicine at Ruchill Hospital, Glasgow, has acted as the Scottish reference laboratory for legionella infection since I977. T h e tests done include the indirect fluorescent antibody test ( I F A T ) using polyvalent, heat-killed antigens; culture using buffered-charcoal yeast extract agar, both with and without antibiotic supplements; the direct fluorescent antibody test; and urinary antigen detection. Samples are accepted from other laboratories both for primary testing and confirmation of results. Most, if not all, Scottish laboratories use the acting reference laboratory. Case-lists held at the Ruchill Hospital laboratory were cross-checked with another compiled by the Communicable Diseases (Scotland) Unit, which is responsible for surveillance and where voluntary laboratory returns are collated and a 'master-list' of possible cases prepared. T h e background and clinical details were obtained from laboratory request forms and, for cases occurring after mid-I984, from replies to a questionnaire sent by R. J. F. to the consultant responsible for the care of the patients. Details recorded by R . J . F . during telephone calls and letters to clinicians about patients were also used. Hospital consultants and general practitioners (GPs) were asked to check, and, if missing, to provide details on a computer printout giving the name,
Seasonality of Legionnaires' disease
~t55
address, date of birth, hospital n u m b e r and date-of-onset of illness of patients in their charge. Their view on the validity of the diagnosis was sought. General practitioners were asked about the patient's occupation and hospitalisation or foreign travel before his illness. With permission from GPs some patients were written to, allowing a further check on the address and both the travel and hospitalisation history. T h e date of onset was still not obtained for 20 of 452 potential cases (4"4 %) and was arbitrarily assigned using the date of the first diagnostic test; the onset was said to be 2 weeks before this date. Based on the above information and the case-definition below, each patient was assigned as a probable case (henceforth called case) or as not a case. Further, patients were classified as nosocomial if there was any history of visiting a hospital during the incubation period, as travel-associated if there was travel outside Scotland, and outbreak-related if they were included in the two outbreaks described in Scotland in the period I978-I986.14'22 Data on the n u m b e r of serology tests done was obtained from the register of tests at the Department of Laboratory Medicine, Ruchill Hospital, and supplemented with information obtained from other laboratories. Case definition for Legionnaires' disease Probable diagnosis. A case with a clinical history of an acute pneumonia or acute lower respiratory tract infection and one or more of the following: (I) Culture of the organism. (2) For Legionella pneumophila only: (a) a four-fold rise (or fall) in titre to at least 64 but no other clinical diagnosis ; or (b) a static titre of 256 or more but no other clinical diagnosis; or (c) positive direct fluorescent antibody test on respiratory secretions or tissue using specific reagents; or (d) Legionella-antigen detected in urine. For two cases exceptions were made for the sake of consistency : both had static titres of less than 256 but were categorised as cases during the I984 Dennistoun outbreak. 14 Geographical analysis Addresses were converted (manually) to seven-unit postcodes using postcode directories. T h e postcodes were used to classify patients by health board area of residence (this task was done by the N o r t h e r n Health Boards Operational Research Unit) using a computerised postcode directory. Calculation o f moving averages Moving averages smooth the pattern and help to interpret the trend when there is marked fluctuation in disease frequency. T h e method is best illustrated by an example. T h e 3-monthly moving average for April r978 is the mean of March, April and M a y ; that for May is April, May and June. Values cannot be calculated for January 1978 and December 1986. T h e I2-monthly moving average cannot be calculated for J a n u a r y - M a y r978 and August-December I986.
156
R. S. B H O P A L
A N D R. J. F A L L O N
T a b l e I Monthly Number of cases by category of disease Month Category
J
F
M
A
M
J
J
A S
(a) (b) (c) (d) (e) (f) (g) (h) (1)
15 4 II 9 2
14 14 7 7
8 I 7 6 I
14 6 8 4 4
15 4 ii 9 2
42 IO 32 18 14 25
30 12 i8 17 i 3
47 20 27 18 9 I i
II
2 12
I 6
8
I io
All cases Travel-relatedcases Non-travelcases Men Women Community outbreak Nosocomial outbreak Possible n o s o c o m i a l cases Community-acquired, non-travel, non-outbreak cases
o
7
15
25
38 13 25 20 5 4 o I 20
O
N
D
Total
47 16 31 23 8
55 9 46 30 16
32 3 29 23 6
7 3 21
5 6 35
I 28
357 98 259 184 75 33 I3" 15 198
* T w o o f t h e cases r e p o r t e d b y T i m b u r y et al., 2~ d i d n o t m e e t t h e case definition u s e d in this study.
60
2500
50
2000
40 1500
o* ,~ 30 8
o
1000
20 500
10
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
0
Months
Fig. I. Seasonal variation in cases and serology tests. (IN), Total cases; (11), total sero tests.
Results
T h e r e were 366 patients w h o met our definition of Legionnaires' disease b e t w e e n 1977 and 1986 and the date of onset was k n o w n (or assigned) for 357 (97"5 %). T a b l e I summarises the n u m b e r of cases, b y m o n t h , in each of the categories o f infection. Figure I shows the seasonal variation for all cases in relation to the pattern for the n u m b e r of serology tests. Excepting late I985 and early I986, w h e n the nosocomial outbreak occurred, 22 this pattern of serology tests was consistent over the years.
Seasonality of Legionnaires' disease
157
50
35-
•
/\.//A-q Jan
Feb
Mar
Apr
May
dun Monfhs
Jul
Aug
/ Sep
Ocf
Nov
Dec
Fig. 2. Seasonal variation by category of infection. ( I ) , Non-travel; ([Z]), Community-
acquired, non-travel, non-outbreak infection; (O), nosocomial; ( ~ ) , travel.
12 I0
8
/'i,,/,,,,,
6
4
_
2
0
te-
A
- - ~
JA 78
JO
JAJO 79
JA 80
JOJ
81
IIIt
A JO
JA 82
OJA 83
JO
I
JAJOJAJOJAJO 84 85
Fig. 3. Seasonal variation for non-travel cases by year: 3 monthly ( ( ...... ) moving averages.
86
) a n d I2 monthly
Travel-related cases (Fig. 2) showed the characteristic s u m m e r / a u t u m n peak, and this pattern was observed throughout the 9 years (data not given). Figure 2 also shows that for non-travel cases the peak occurred in early winter, and this applied to both m e n and w o m e n (Table I). T h i s pattern was accentuated w h e n outbreak and nosocomial cases were excluded (Fig. 2), i.e. for c o m m u n i t y - a c q u i r e d , non-travel, non-outbreak cases ( C A N T N O ) . Seven of the I3 nosocomial outbreak cases occurred in October and five in N o v e m b e r , while of the ~5 possible nosocomial cases three occurred in October, six in N o v e m b e r and one in D e c e m b e r : a distinct late autumn/early
I58
R. S. B H O P A L
AND
R.
J.
FALLON
30
•
Jan
Feb
Mar
Apr
May
dun Jul Months
Aug
Sep
Oct
Nov
Dec
Fig. 4. Seasonal variation by Health Board area. ( l l ) , G G H B = Greater Glasgow Health Board; ([]), L H B = L o t h i a n Health Board; ( • ) , L a n H B = Lanarkshire Health Board; ( ~ ) , other health boards.
winter pattern. However, Fig. 3, which includes all non-travel cases, shows that the pattern was inconsistent over time. Between I978 and I982 there was evidence of a s u m m e r / a u t u m n peak, but between I983 and I985 there was a definite late autumn/early winter peak (the June 1984 peak was the outbreak14). Figure 3 also shows marked variation in time in the n u m b e r of cases and that the disease incidence rose in I983, I984 and ~985. This trend will be discussed in a future paper. Figure 4 shows that the early winter peak was most marked in the Greater Glasgow Health Board area, and was less clear elsewhere. None the less, there was no s u m m e r / a u t u m n peak, rather an a u t u m n / w i n t e r plateau. Discussion
T h e s u m m e r / a u t u m n seasonal peak for Legionnaires' disease has been such a consistent epidemiological feature that its validity has not previously been questioned. Yet, it may be that it is only observed in travel-related infection and is related to the numbers of people travelling, and hence exposed, rather than the degree of exposure. Unfortunately, most published studies do not distinguish between these categories of infection. T h e seasonal variation for non-travel infection in England and Wales, as shown in aggregated data for the period I979-I986, was unremarkable, 23 though evident, when travel-related cases were added. However, in preceding reports showing aggregate data 1977-i984, 24 and I98I-~98225 the seasonal pattern was discernible for both travel and non-travel cases, though more distinctive for the former. It may be that the seasonal pattern has altered in England and Wales; if so this change would be consistent with our observations in Scotland where the seasonal peak
Seasonality of Legionnaires' disease
I59
for non-travel, n o n - o u t b r e a k Legionnaires' disease in the period I 9 8 3 - I 9 8 5 , b u t not b e t w e e n I978 and I983, was in late a u t u m n and early winter. T h e explanation for our findings did not lie in m o r e tests being done in winter, t h o u g h the n u m b e r then was a little higher than at other times, particularly in late I985 and early I986 following the nosocomial outbreak. 22 T h e s e observations defy rational explanation b u t invite speculation. 2°' 21 C o u l d it be that the peak reflects the period w h e n cooling towers are cleaned, a time w h e n legionella concentrations are at their highest? 26 Is there some change in climatology such as wind direction ? F o r example, in s u m m e r the prevailing winds are from the north and the air currents low in humidity, b u t the winter winds are f r o m the west and the air moist (personal communication, M r J. Alladyce, principal meteorological officer), a factor which w o u l d be expected to i m p r o v e the survival of legionellae in air. iv Is it that the conditions which p r o m o t e d legionella g r o w t h in Scotland b e t w e e n I983 and I985 o c c u r r e d late in the year for u n k n o w n reasons? Studies to examine the environmental distribution of legionellae at different times, as previously r e p o r t e d b y Bercovier 1° and T o b i a n s k y , ~1 w o u l d help to interpret these findings. T h i s s t u d y has practical and theoretical applications. Clinicians, at least in Scotland, should not be u n d u l y swayed b y the epidemiological advice that associates p n e u m o n i a cases in s u m m e r with Legionnaires' disease except w h e n there is a history of travel abroad. T h e travel history remains an important pointer to the diagnosis of atypical pneumonia. T h e low n u m b e r of cases in late winter and spring m a y have some clinical value. T h e further examination and future explanation of o u r observations m a y help to u n d e r s t a n d the relationship b e t w e e n sources of legionellae, p r o d u c t i o n of aerosol, virulence of organisms, exposure of patients and infection. W e hope that this reexamination of an established epidemiological observation will spur others to s t u d y the seasonal pattern in different categories of Legionnaires' disease and over several time periods. (We are grateful to the Greater Glasgow Health Board and the Pneumonia Research Trust for financial support with this work. We thank the patients, general practitioners, hospital consultants, medical records officers, laboratory staff, environmental hea!th officers, personnel at premises with cooling towers and other colleagues who helped in our enquiries. Dr D. Parkin gave expert advice on moving averages and the preparation of Fig. 3. Mrs K. Ferguson provided clerical support and Mrs L. Hutchinson prepared the manuscript.) References
i. Band JD, Fraser DW. Legionellosis. In : Reeves DS, Geddes AM, Eds. Recent advances in infection 2, Chapter 5. Edinburgh : Churchill Livingstone, x982 : IoI-I I7. 2. Baxter D. Legionnaires' disease..4 comprehensive description and contemporary bibliography. Occasional Paper No. 9. Manchester (Department of Community Medicine), i985. 3. Bartlett CLR, Macrae AD, Macfarlane JT. Legionella infections. London : Edward Arnold, I986. 4- Helms CM, Viner JP, Weisenburger DD, Chiu LC, Renner ED, Johnson W. Sporadic Legionnaires' Disease: clinical observations on 87 nosocomial and community-acquired cases..4m J Med Sci, I984; 288: 2-I2.
I60
R. S. BHOPAL AND R. J. F A L L O N
5. Macfarlane JT, Ward HJ, Finch RG. Hospital study of adult community-acquired pneumonia. Lancet I982; ii: 255-258. 6. Public Health Laboratory Service. Communicable Disease Report. Commun Med 1985 ; 7: 3o4-3o8. 7. Committee of Inquiry. Second report of the Committee of Inquiry into the outbreak of Legionnaires' disease in Stafford in April 1985. London: Her Majesty's Stationery Office, I987. 8. Rowbotham TJ. Pontiac fever, amoebae, and Legionellae. Lancet r98I; i: 4o-4I. 9. Fliermans CB. Philosophical ecology : Legionella in historical perspective. In: Thomsberry C, Ballows A, Feeley JC, Jacubowski W, Eds. Legionella. Proceedings of the 2nd International Symposium. Washington, D.C. : American Society for Microbiology, I984. Io. Bercovier H, Fattal B, Shival H. Seasonal distribution of legionellae isolated from various types of water in Israel. Isr J Med Sci r986; z2: 644-646. I x. Tobiansky L, Drath A, Dubery B, Koornhof HJ. Seasonality of Legionella isolates from environmental sources. Isr J Med Sci r986; zz: 64o-643 . I2. Thacher SB, Bennet JV, Tsai T F et al. An outbreak in I965 of severe respiratory illness caused by the Legionnaires' Disease bacterium. J Infect Dis 1978; x38: 5IZ-519. r 3. Hoyle F, Wickramasinghe NC, Watkins J. Legionnaires' Disease : seeking a wider source. Lancet I985; i: I216--1217. 14. Ad-hoc Committee. Outbreak of legionellosis in a community. Lancet I986; ii: 38o-383. I5. Report of the Westminster Action Committee. Broadcasting House Legionnaires' disease. London: City of Westminster, I988. I6. Bovallius A, Roffey R, Henningston E. Long-range transmission of bacteria. A n n N Y A c a d Sci I98o; 353: I86-I97. I7. Hambleton P, Broster MG, Dennis PJ, Henstridge R, Fitzgeorge R, Conlan JW. Survival of virulent Legionella pneumophila in aerosols. J Hyg (Camb) I983; 9o: 451-46o. i8. Riley RL. Historical introduction (airborne contagion). Ann N Y Acad Sci I98O; 353 : 3-9. 19. Fallon RJ, Abraham WH. Legionella infections in Scotland. J Hyg (Camb) I98Z; 89: 439-448. 20. Fallon RJ. Legionella infections. Br Med J I986; z93: I I75. 2I. Bhopal RS, Fallon RJ. Geographical epidemiology of Legionnaires' Diseases in Glasgow: a preliminary report. Communicable Diseases (Scotland) Unit Weekly Bulletin r988; 22(I7): 7 - I I , and I988; 22(I8): 2. 22. Timbury MC, Donaldson JR, McCartney AC et al. Outbreak of Legionnaires' Disease in Glasgow Royal Infirmary: microbiological aspects. J Hyg (Camb) r986; 97: 393-4o3. 23. Public Health Laboratory Service. Report from the PHLS Communicable Disease Surveillance Centre. Br Med J I988; 296: 778. 24. Public Health Laboratory Service. Report from the PHLS Communicable Disease Surveillance Centre. Br M e d J 1986; z93: 681-682. 25. Public Health Laboratory Service Communicable Disease Centre. Legionella infection I98I-2. Br M e d J I985; z87: 826. 26. Tyndall RJ. Concentrations, serotypic profiles and infectivity of Legionnaires' Disease bacteria populations in cooling towers. J Cooling Tower Inst I982; 3: 25-33.
S e a s o n a l v a r i a t i o n o f L e g i o n n a i r e s ' d i s e a s e in S c o t l a n d R. S. Bhopal* and R. J. Fallon]" * Division of Epidemiology and Public Health, The Medical School, Framlington Place, Newcastle upon Tyne NEe 4HH and t Department of Laboratory Medicine, Ruchill Hospital, Bilsland Drive, Glasgow G2o 9NB, Scotland, U.K. Accepted for publication Io September I99o Summary A summer and autumn peak in incidence is a characteristic epidemiological feature of Legionnaires' disease, a fact which may help in diagnosis. Previous reports from Scotland have reported a winter peak. This study examined the seasonal variation in Legionnaires' disease in Scotland in each year and by category of diagnosis. A characteristic summer/autumn peak was seen for travel-related infection. For nontravel infection, there was an autumnal peak between I978 and I982 but an early winter peak between I983 and I986. Overall, the pattern could be described as a summer/autumn plateau reaching a peak in early winter. This pattern, seen in men and women, was dearest in the Greater Glasgow area. Nosocomial cases were clustered in the months of October, November and December. In non-travel cases the classical seasonal pattern was not observed and clinicians cannot rely upon this epidemiological feature for diagnosis. However, the relative infrequency of the disease between January and May was a consistent and potentially useful feature. Most reports on seasonal variation are based on aggregated data on travel and nontravel cases. T h e present observations question the hypotheses developed to explain the seasonal pattern and call for further studies.
Introduction L e g i o n n a i r e s ' disease, in b o t h e p i d e m i c and sporadic f o r m , is c o m m o n e r in s u m m e r an d early a u t u m n 1' ~ and, hence, admission of a p n e u m o n i a pat i ent at such a time s h oul d alert the clinician to this diagnosis. 3-~ T h e seasonal peak has b e e n o b s e r v e d for c o m m u n i t y - a c q u i r e d and nosocomial cases 4 and for b o t h travel and n o n - t r a v e l associated cases. 6' 7 E x p l a n a t i o n s given for the seasonal variation in cl ude t hes e : 1egionellae m u l t i p l y faster in the w a r m e r waters o f s u m m e r ; the greater use o f cooling towers in s u m m e r for a i r - c o n d i t i o n i n g m a y p r o v i d e o p p o r t u n i t i e s for d i s s e m i n a t i o n ; th e w a r m e r cooling t o w e r w a t e r o f late s u m m e r allows legionellae to b e c o m e infective and transmissible to h u m a n s ; the life cycle o f species w hi c h gr ow in symbiosis with legioneUae d e t e r m i n e s th e n u m b e r s o f legionellae. F o r example, legionellae are digested by am oebae at low t e m p e r a t u r e s b u t infect a m o e b a e at t e m p e r a t u r e s of 30 degrees cen tig r ad e or m o r e 8 a nd t he n u m b e r o f legionellae rises in association with t he d e g r a d a t i o n o f t he pl a nt Myriophyllum spicatum. 9 oi63-4453/9I/O2OI53 +08 $03.00/0
© I 9 9 I The British Society for the Study of Infection
I54
R. S. BHOPAL AND R. J. FALLON
T w o studies have shown that legionellae are easier to isolate from natural water in summer and autumn than in winter. 1°,11 Climate is important in the web of causation of Legionnaires' disease but in ways which are poorly understood. For example, high winds were associated with the Stafford and Washington outbreaks 12'13 and wind direction appeared to determine the distribution of cases in several outbreaks, e.g. the Dennistoun outbreak in Glasgow 14 and the BBC outbreak in London. 15 Humidity is important for the survival of all organisms pathogenic to man 16 and this has been demonstrated in vitro for legionellae. ~7 Weather, and particularly wind speed, also determines the dissemination and deposition of aerosols. Particles exceeding o.I m m in diameter rapidly settle but droplet nuclei which are less than o'I m m may be suspended in air and carried long distances, particularly by air-currents, but also by diffusion and sedimentation. TM ~8 While weather may explain seasonal variation, its precise role is undetermined. An analysis of cases in Scotland up to I98I showed that, excluding travel cases, there was no clear seasonal variation except that there were fewer cases in the first quarter of the year. '9 Later, Fallon reported that in Scotland the disease incidence peaked in the fourth quarter of the year, and commented that this was paradoxical due to the cooler weather in Scotland and the earlier switching off of cooling towers. ~° These observations, which were confirmed in a study of I984 cases, defied explanation. 21 We now report a more detailed study of the m o n t h - b y - m o n t h variation in Legionnaires' disease in Scotland between I978 and I986. Our emphasis has been on separating travel and nontravel cases, examining the pattern in several areas of Scotland, and assessing the consistency of the pattern over time. Sources of data and methods C a s e l i s t i n g , p a t i e n t d e t a i l s a n d c l a s s i f i c a t i o n o f cases
T h e Department of Laboratory Medicine at Ruchill Hospital, Glasgow, has acted as the Scottish reference laboratory for legionella infection since I977. T h e tests done include the indirect fluorescent antibody test ( I F A T ) using polyvalent, heat-killed antigens; culture using buffered-charcoal yeast extract agar, both with and without antibiotic supplements; the direct fluorescent antibody test; and urinary antigen detection. Samples are accepted from other laboratories both for primary testing and confirmation of results. Most, if not all, Scottish laboratories use the acting reference laboratory. Case-lists held at the Ruchill Hospital laboratory were cross-checked with another compiled by the Communicable Diseases (Scotland) Unit, which is responsible for surveillance and where voluntary laboratory returns are collated and a 'master-list' of possible cases prepared. T h e background and clinical details were obtained from laboratory request forms and, for cases occurring after mid-I984, from replies to a questionnaire sent by R. J. F. to the consultant responsible for the care of the patients. Details recorded by R . J . F . during telephone calls and letters to clinicians about patients were also used. Hospital consultants and general practitioners (GPs) were asked to check, and, if missing, to provide details on a computer printout giving the name,
Seasonality of Legionnaires' disease
~t55
address, date of birth, hospital n u m b e r and date-of-onset of illness of patients in their charge. Their view on the validity of the diagnosis was sought. General practitioners were asked about the patient's occupation and hospitalisation or foreign travel before his illness. With permission from GPs some patients were written to, allowing a further check on the address and both the travel and hospitalisation history. T h e date of onset was still not obtained for 20 of 452 potential cases (4"4 %) and was arbitrarily assigned using the date of the first diagnostic test; the onset was said to be 2 weeks before this date. Based on the above information and the case-definition below, each patient was assigned as a probable case (henceforth called case) or as not a case. Further, patients were classified as nosocomial if there was any history of visiting a hospital during the incubation period, as travel-associated if there was travel outside Scotland, and outbreak-related if they were included in the two outbreaks described in Scotland in the period I978-I986.14'22 Data on the n u m b e r of serology tests done was obtained from the register of tests at the Department of Laboratory Medicine, Ruchill Hospital, and supplemented with information obtained from other laboratories. Case definition for Legionnaires' disease Probable diagnosis. A case with a clinical history of an acute pneumonia or acute lower respiratory tract infection and one or more of the following: (I) Culture of the organism. (2) For Legionella pneumophila only: (a) a four-fold rise (or fall) in titre to at least 64 but no other clinical diagnosis ; or (b) a static titre of 256 or more but no other clinical diagnosis; or (c) positive direct fluorescent antibody test on respiratory secretions or tissue using specific reagents; or (d) Legionella-antigen detected in urine. For two cases exceptions were made for the sake of consistency : both had static titres of less than 256 but were categorised as cases during the I984 Dennistoun outbreak. 14 Geographical analysis Addresses were converted (manually) to seven-unit postcodes using postcode directories. T h e postcodes were used to classify patients by health board area of residence (this task was done by the N o r t h e r n Health Boards Operational Research Unit) using a computerised postcode directory. Calculation o f moving averages Moving averages smooth the pattern and help to interpret the trend when there is marked fluctuation in disease frequency. T h e method is best illustrated by an example. T h e 3-monthly moving average for April r978 is the mean of March, April and M a y ; that for May is April, May and June. Values cannot be calculated for January 1978 and December 1986. T h e I2-monthly moving average cannot be calculated for J a n u a r y - M a y r978 and August-December I986.
156
R. S. B H O P A L
A N D R. J. F A L L O N
T a b l e I Monthly Number of cases by category of disease Month Category
J
F
M
A
M
J
J
A S
(a) (b) (c) (d) (e) (f) (g) (h) (1)
15 4 II 9 2
14 14 7 7
8 I 7 6 I
14 6 8 4 4
15 4 ii 9 2
42 IO 32 18 14 25
30 12 i8 17 i 3
47 20 27 18 9 I i
II
2 12
I 6
8
I io
All cases Travel-relatedcases Non-travelcases Men Women Community outbreak Nosocomial outbreak Possible n o s o c o m i a l cases Community-acquired, non-travel, non-outbreak cases
o
7
15
25
38 13 25 20 5 4 o I 20
O
N
D
Total
47 16 31 23 8
55 9 46 30 16
32 3 29 23 6
7 3 21
5 6 35
I 28
357 98 259 184 75 33 I3" 15 198
* T w o o f t h e cases r e p o r t e d b y T i m b u r y et al., 2~ d i d n o t m e e t t h e case definition u s e d in this study.
60
2500
50
2000
40 1500
o* ,~ 30 8
o
1000
20 500
10
0
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
Dec
0
Months
Fig. I. Seasonal variation in cases and serology tests. (IN), Total cases; (11), total sero tests.
Results
T h e r e were 366 patients w h o met our definition of Legionnaires' disease b e t w e e n 1977 and 1986 and the date of onset was k n o w n (or assigned) for 357 (97"5 %). T a b l e I summarises the n u m b e r of cases, b y m o n t h , in each of the categories o f infection. Figure I shows the seasonal variation for all cases in relation to the pattern for the n u m b e r of serology tests. Excepting late I985 and early I986, w h e n the nosocomial outbreak occurred, 22 this pattern of serology tests was consistent over the years.
Seasonality of Legionnaires' disease
157
50
35-
•
/\.//A-q Jan
Feb
Mar
Apr
May
dun Monfhs
Jul
Aug
/ Sep
Ocf
Nov
Dec
Fig. 2. Seasonal variation by category of infection. ( I ) , Non-travel; ([Z]), Community-
acquired, non-travel, non-outbreak infection; (O), nosocomial; ( ~ ) , travel.
12 I0
8
/'i,,/,,,,,
6
4
_
2
0
te-
A
- - ~
JA 78
JO
JAJO 79
JA 80
JOJ
81
IIIt
A JO
JA 82
OJA 83
JO
I
JAJOJAJOJAJO 84 85
Fig. 3. Seasonal variation for non-travel cases by year: 3 monthly ( ( ...... ) moving averages.
86
) a n d I2 monthly
Travel-related cases (Fig. 2) showed the characteristic s u m m e r / a u t u m n peak, and this pattern was observed throughout the 9 years (data not given). Figure 2 also shows that for non-travel cases the peak occurred in early winter, and this applied to both m e n and w o m e n (Table I). T h i s pattern was accentuated w h e n outbreak and nosocomial cases were excluded (Fig. 2), i.e. for c o m m u n i t y - a c q u i r e d , non-travel, non-outbreak cases ( C A N T N O ) . Seven of the I3 nosocomial outbreak cases occurred in October and five in N o v e m b e r , while of the ~5 possible nosocomial cases three occurred in October, six in N o v e m b e r and one in D e c e m b e r : a distinct late autumn/early
I58
R. S. B H O P A L
AND
R.
J.
FALLON
30
•
Jan
Feb
Mar
Apr
May
dun Jul Months
Aug
Sep
Oct
Nov
Dec
Fig. 4. Seasonal variation by Health Board area. ( l l ) , G G H B = Greater Glasgow Health Board; ([]), L H B = L o t h i a n Health Board; ( • ) , L a n H B = Lanarkshire Health Board; ( ~ ) , other health boards.
winter pattern. However, Fig. 3, which includes all non-travel cases, shows that the pattern was inconsistent over time. Between I978 and I982 there was evidence of a s u m m e r / a u t u m n peak, but between I983 and I985 there was a definite late autumn/early winter peak (the June 1984 peak was the outbreak14). Figure 3 also shows marked variation in time in the n u m b e r of cases and that the disease incidence rose in I983, I984 and ~985. This trend will be discussed in a future paper. Figure 4 shows that the early winter peak was most marked in the Greater Glasgow Health Board area, and was less clear elsewhere. None the less, there was no s u m m e r / a u t u m n peak, rather an a u t u m n / w i n t e r plateau. Discussion
T h e s u m m e r / a u t u m n seasonal peak for Legionnaires' disease has been such a consistent epidemiological feature that its validity has not previously been questioned. Yet, it may be that it is only observed in travel-related infection and is related to the numbers of people travelling, and hence exposed, rather than the degree of exposure. Unfortunately, most published studies do not distinguish between these categories of infection. T h e seasonal variation for non-travel infection in England and Wales, as shown in aggregated data for the period I979-I986, was unremarkable, 23 though evident, when travel-related cases were added. However, in preceding reports showing aggregate data 1977-i984, 24 and I98I-~98225 the seasonal pattern was discernible for both travel and non-travel cases, though more distinctive for the former. It may be that the seasonal pattern has altered in England and Wales; if so this change would be consistent with our observations in Scotland where the seasonal peak
Seasonality of Legionnaires' disease
I59
for non-travel, n o n - o u t b r e a k Legionnaires' disease in the period I 9 8 3 - I 9 8 5 , b u t not b e t w e e n I978 and I983, was in late a u t u m n and early winter. T h e explanation for our findings did not lie in m o r e tests being done in winter, t h o u g h the n u m b e r then was a little higher than at other times, particularly in late I985 and early I986 following the nosocomial outbreak. 22 T h e s e observations defy rational explanation b u t invite speculation. 2°' 21 C o u l d it be that the peak reflects the period w h e n cooling towers are cleaned, a time w h e n legionella concentrations are at their highest? 26 Is there some change in climatology such as wind direction ? F o r example, in s u m m e r the prevailing winds are from the north and the air currents low in humidity, b u t the winter winds are f r o m the west and the air moist (personal communication, M r J. Alladyce, principal meteorological officer), a factor which w o u l d be expected to i m p r o v e the survival of legionellae in air. iv Is it that the conditions which p r o m o t e d legionella g r o w t h in Scotland b e t w e e n I983 and I985 o c c u r r e d late in the year for u n k n o w n reasons? Studies to examine the environmental distribution of legionellae at different times, as previously r e p o r t e d b y Bercovier 1° and T o b i a n s k y , ~1 w o u l d help to interpret these findings. T h i s s t u d y has practical and theoretical applications. Clinicians, at least in Scotland, should not be u n d u l y swayed b y the epidemiological advice that associates p n e u m o n i a cases in s u m m e r with Legionnaires' disease except w h e n there is a history of travel abroad. T h e travel history remains an important pointer to the diagnosis of atypical pneumonia. T h e low n u m b e r of cases in late winter and spring m a y have some clinical value. T h e further examination and future explanation of o u r observations m a y help to u n d e r s t a n d the relationship b e t w e e n sources of legionellae, p r o d u c t i o n of aerosol, virulence of organisms, exposure of patients and infection. W e hope that this reexamination of an established epidemiological observation will spur others to s t u d y the seasonal pattern in different categories of Legionnaires' disease and over several time periods. (We are grateful to the Greater Glasgow Health Board and the Pneumonia Research Trust for financial support with this work. We thank the patients, general practitioners, hospital consultants, medical records officers, laboratory staff, environmental hea!th officers, personnel at premises with cooling towers and other colleagues who helped in our enquiries. Dr D. Parkin gave expert advice on moving averages and the preparation of Fig. 3. Mrs K. Ferguson provided clerical support and Mrs L. Hutchinson prepared the manuscript.) References
i. Band JD, Fraser DW. Legionellosis. In : Reeves DS, Geddes AM, Eds. Recent advances in infection 2, Chapter 5. Edinburgh : Churchill Livingstone, x982 : IoI-I I7. 2. Baxter D. Legionnaires' disease..4 comprehensive description and contemporary bibliography. Occasional Paper No. 9. Manchester (Department of Community Medicine), i985. 3. Bartlett CLR, Macrae AD, Macfarlane JT. Legionella infections. London : Edward Arnold, I986. 4- Helms CM, Viner JP, Weisenburger DD, Chiu LC, Renner ED, Johnson W. Sporadic Legionnaires' Disease: clinical observations on 87 nosocomial and community-acquired cases..4m J Med Sci, I984; 288: 2-I2.
I60
R. S. BHOPAL AND R. J. F A L L O N
5. Macfarlane JT, Ward HJ, Finch RG. Hospital study of adult community-acquired pneumonia. Lancet I982; ii: 255-258. 6. Public Health Laboratory Service. Communicable Disease Report. Commun Med 1985 ; 7: 3o4-3o8. 7. Committee of Inquiry. Second report of the Committee of Inquiry into the outbreak of Legionnaires' disease in Stafford in April 1985. London: Her Majesty's Stationery Office, I987. 8. Rowbotham TJ. Pontiac fever, amoebae, and Legionellae. Lancet r98I; i: 4o-4I. 9. Fliermans CB. Philosophical ecology : Legionella in historical perspective. In: Thomsberry C, Ballows A, Feeley JC, Jacubowski W, Eds. Legionella. Proceedings of the 2nd International Symposium. Washington, D.C. : American Society for Microbiology, I984. Io. Bercovier H, Fattal B, Shival H. Seasonal distribution of legionellae isolated from various types of water in Israel. Isr J Med Sci r986; z2: 644-646. I x. Tobiansky L, Drath A, Dubery B, Koornhof HJ. Seasonality of Legionella isolates from environmental sources. Isr J Med Sci r986; zz: 64o-643 . I2. Thacher SB, Bennet JV, Tsai T F et al. An outbreak in I965 of severe respiratory illness caused by the Legionnaires' Disease bacterium. J Infect Dis 1978; x38: 5IZ-519. r 3. Hoyle F, Wickramasinghe NC, Watkins J. Legionnaires' Disease : seeking a wider source. Lancet I985; i: I216--1217. 14. Ad-hoc Committee. Outbreak of legionellosis in a community. Lancet I986; ii: 38o-383. I5. Report of the Westminster Action Committee. Broadcasting House Legionnaires' disease. London: City of Westminster, I988. I6. Bovallius A, Roffey R, Henningston E. Long-range transmission of bacteria. A n n N Y A c a d Sci I98o; 353: I86-I97. I7. Hambleton P, Broster MG, Dennis PJ, Henstridge R, Fitzgeorge R, Conlan JW. Survival of virulent Legionella pneumophila in aerosols. J Hyg (Camb) I983; 9o: 451-46o. i8. Riley RL. Historical introduction (airborne contagion). Ann N Y Acad Sci I98O; 353 : 3-9. 19. Fallon RJ, Abraham WH. Legionella infections in Scotland. J Hyg (Camb) I98Z; 89: 439-448. 20. Fallon RJ. Legionella infections. Br Med J I986; z93: I I75. 2I. Bhopal RS, Fallon RJ. Geographical epidemiology of Legionnaires' Diseases in Glasgow: a preliminary report. Communicable Diseases (Scotland) Unit Weekly Bulletin r988; 22(I7): 7 - I I , and I988; 22(I8): 2. 22. Timbury MC, Donaldson JR, McCartney AC et al. Outbreak of Legionnaires' Disease in Glasgow Royal Infirmary: microbiological aspects. J Hyg (Camb) r986; 97: 393-4o3. 23. Public Health Laboratory Service. Report from the PHLS Communicable Disease Surveillance Centre. Br Med J I988; 296: 778. 24. Public Health Laboratory Service. Report from the PHLS Communicable Disease Surveillance Centre. Br M e d J 1986; z93: 681-682. 25. Public Health Laboratory Service Communicable Disease Centre. Legionella infection I98I-2. Br M e d J I985; z87: 826. 26. Tyndall RJ. Concentrations, serotypic profiles and infectivity of Legionnaires' Disease bacteria populations in cooling towers. J Cooling Tower Inst I982; 3: 25-33.
