Am. J. Trop. Med. Hyg., 78(2), 2008, pp. 318–320 Copyright © 2008 by The American Society of Tropical Medicine and Hygiene
Short Report: Socio-demographic and Climatic Factors as Correlates of Mediterranean Spotted Fever (MSF) in Northern Sardinia Maria Fenicia Vescio, Maria Adriana Piras, Massimo Ciccozzi, Antonina Carai, Francesca Farchi, Michele Maroli, Maria S. Mura, Giovanni Rezza,* and the MSF Study Group† Department of Infectious Diseases, Istituto Superiore di Sanità, Roma, Italy; Department of Infectious Diseases, University of Sassari, Sassari, Italy; Infectious Diseases Division, San Francesco Hospital, Nuoro, Italy
Abstract. Mediterranean spotted fever (MSF) is endemic in the Mediterranean area. We carried out a retrospective study to investigate the association between socio-demographic and climatic factors and MSF incidence in northern Sardinia. We found that maximum temperature levels during the previous summer were associated with increases in MSF incidence. cases was greater than that assumed under the sampling framework, the standardized residuals were plotted in turn against the observed number of cases, the fitted values, and relevant variables. A negative binomial distribution was used in the presence of overdispersion, and potential confounders were included as additional independent variables. Robust SEs were computed to address correlation in the residuals and overdispersion.
INTRODUCTION Mediterranean spotted fever (MSF), or “boutonneuse” fever, is caused by Rickettsia conorii and transmitted by the brown dog tick Rhipicephalus sanguineus.1 The disease is endemic in the Mediterranean area, with incidence in Italy being higher in the southern regions and in the islands. Most cases occur during the summer, with an apparent increase at the end of the 1970s.2,3 Occupation, participation in outdoor activities, and rural environment have been found associated with increased risk of infection with R. conorii,4,5 although the role of climatic change in MSF emergence has also been hypothesized.1 In this study, we evaluated the association between sociodemographic and climatic factors and MSF incidence.
RESULTS The temporal pattern of MSF shows some peaks between 1990 and 2002, which seem to follow 1 year after those observed for the maximum average temperature levels recorded during the summer months (Figure 1). Negative binomial regression models confirmed these findings, showing that increases in the maximum temperature of 1°C corresponded to an IRR of 1.32 (Table 1). There was also evidence of interaction between maximum temperature levels and altitude, this being caused by the effect of higher temperatures at sea level than at higher altitudes. No association was found between the incidence of MSF and rainfall (either in the same year or in previous years). The proportion of men and women used in agricultural occupations was not associated with an increase in IRR, although an increase of one point in the proportion of low educational attainment increased the IRR by 11%. There was a quadratic relationship between MSF relative incidence ratio and population density, with IRR being higher in rural underpopulated areas than in suburban and urban areas. The incidence of MSF was higher in men and women ⱖ 60 years of age (RR, 1.33) but did not vary by sex. No evidence of association was found between the incidence of MSF and the minimum temperature levels between November and March (in the same year or in previous years) when analyzed as the mean value or as the lowest recorded value.
MATERIALS AND METHODS We carried out a retrospective study to identify cases of MSF occurring in the resident population of the provinces of Sassari and Nuoro (North Sardinia) between 1990 and 2002. Cases were collected through a systematic revision of the discharge registries from the hospitals sited in the study area. The diagnosis of MSF was based on clinical and serologic assessment, with information on age, sex, and municipality of residence of the cases also being collected from clinical records. Data on temperature (minimum and maximum average values) and on rainfall for each month, year, and municipality were obtained from the Sardinia climatic unit. Overall, 819 cases were identified (287 women and 532 men). Population figures, the proportion of men and women working in agricultural occupations, the proportion of unemployed, and the proportion of those with low educational attainment were obtained for each municipality from the 1991 census. Poisson and negative binomial regressions were carried out to study the association between socio-demographic and climatic factors with MSF incidence through the calculation of incidence rate ratios (IRRs) and their 95% confidence intervals (CIs). To study whether the variation in the number of
DISCUSSION Raoult and others6 showed graphically that incidence of MSF is correlated with preceding year mean summer temperatures, although their results were not statistically robust because of the low number of cases. We found that an increase of 1°C in the mean maximum temperature levels during summer was associated with an increase of 32% in MSF incidence a year later. The association of summer temperature with MSF incidence was stronger for men and women
* Address correspondence to Giovanni Rezza, Epidemiology Division, Dipartimento di Malattie Infettive, Istituto Superiore di Sanità viale Regina Elena 299, 00161 Roma, Italy. E-mail:
[email protected] † The MSF Study Group includes Claudia Rais, Federica Cambosu, Catello Panu Napodano, Giovanni Satta, Anna Laura Porcu, Marco Cilliano, Maria Laura Pettinato, and Stefano Boros.
318
319
CLIMATE AND MSF IN NORTH SARDINIA
FIGURE 1. Incidence of MSF in North Sardinia (provinces of Sassari and of Nuoro) between 1990 and 2002 and maximum temperature levels between May and August by year of study.
living at sea level than for those living at higher altitudes. We also found that the incidence of MSF was higher in rural than in urban and suburban areas and in areas with low schooling, which have low standards of hygiene and parasite control.6 Conversely, we did not find evidence of an association between the proportion of men and women working in agricultural occupations and the incidence of MSF in the area, although we did not have individual level data and therefore were not able to ascertain which subjects were exposed to high-risk occupations (i.e., occupations that require contact with animals or outdoor activities in natural areas). Another factor that was associated with MSF was age, with older men and women being at higher risk of contracting MSF, presumably because of the higher proportion of occupational risk in this age group. Similar results were observed elsewhere.6 Finally, consistent with the results of a study conducted in Spain,5 we did not find evidence of an association between incidence of MSF and sex. Diseases caused by pathogens that spend part of their cycle outside warm-blooded hosts are geographically and temporally limited by environmental variables such as climate (rainfall conditions, the local temperature, the level of humidity, and soil moisture).7 Warm weather can influence the duration of the active period of mature and immature ticks (i.e., larvae and nymphs) and the generation time to the extent of increasing the number of possible generation cycles to two per year. Therefore, it is possible that exposure to high temperatures during spring-summer leads to increases in the activity of larvae and nymphs a year later (these are known to be responsible for the spread of infection at our latitude because they are active in summer, whereas adults are active in spring).8,9 On the other hand, warm weather may reduce tick activity if very high temperatures in spring-summer are combined with low relative humidity levels.10,11 Hence, we found that hot seasons had relatively low incidences, but MSF incidence did not vary with precipitation levels and there was no evidence of an interaction between maximum temperature and rainfall. Similarly, MSF incidence and rainfall were not correlated in a study carried out in Spain.12 Because R. sanguineus is the only vector and host of MSF,9,13 the incidence of the disease depends on the state of the active part of the tick population. Studies carried out in
central Europe and in Sweden on the Ixodes ricinus suggest that more warm spring days and fewer cold winter days may be associated with increases in tick density.11,14–17 However, whether the population density of R. sanguineus increases with temperature or whether the prevalence of infected ticks and their infectivity varies in relation to climatic factors is still not clear. Although the mechanisms are not completely clear, climatic changes have been related to the resurgence of MSF in Spain,12 to changes in the occurrence of Rocky Mountain spotted fever in the oak-hickory-pine–zone of the United States,18 and to increases in the incidence of tick-borne encephalitis in Sweden and central Europe.16 Concern exists that ticks may spread to regions at higher latitude and altitude and become more abundant.14,16,19,20 It is also possible that climatic changes may lengthen the transmission season in some areas.21 Lack of information is a major concern in our study. However, we compared the number of cases by year of occurrence with those collected by the Ministry of Health through the compulsory notification system, which has been in place for MSF since 1990, and found that between 38% and 87% of the cases that occur every year in the provinces of Nuoro and Sassari were under-reported to the national surveillance system. Exceptions to this were 1996 and 1997, when we identified ∼6% fewer cases than the Ministry of Health. Although our survey was able to identify overall more cases than the official sources, it is still possible that cases with milder symptoms were not recorded if they did not reach the hospital, as has also been found elsewhere.6,22 In addition, it is possible that for some cases, the area of residence did not coincide TABLE 1 Incidence rate ratios of rickettiosis in North Sardinia between 1990 and 2002 by individual and area level (municipalities) explanatory variables Variables
Individual level Age (years) 0–29 30–59 60+ Sex Women Men Area level Maximum temperature level in summer previous year (°C) Minimum temperature level in winter previous year (°C) Altitude (m) 0–224 225–426 427+ Precipitation (mm3/yr) Prevalence of men and women employed in agricultural occupations (%) Prevalence of men and women with a low educational attainment (%) Population density (inhabitants/km2) Population density squared Temperature × altitude 0–224 225–426 427+
IRR
(95% CI)
1 1.12 1.33
(0.92; 1.36) (1.02; 1.73)
1 1.08
(0.92; 1.28)
1.32
(1.11; 1.59)
1.02
(0.97; 1.06)
1 2.68 1.25 1.00
(2.20; 3.27) (0.93; 1.68) (0.99; 1.01)
1.00
(0.95; 1.06)
1.11 0.93 1.00
(1.08; 1.15) (0.90; 0.96) (1.00; 1.00)
1 0.65 0.76
(0.53; 0.79) (0.60; 0.96)
Results from the negative binomial regression models with robust SE.
320
VESCIO AND OTHERS
with the area of infection, resulting in a misclassification of the exposure. Finally, another potential limit concerns the use of terrestrial climate data, which may only be representative of the area in the near proximity of the meteorological stations.23 However, we relied on 42 climatic stations (about 1 every 19 km), ensuring a high coverage of the study area. In conclusion, we found that factors such as exposure to high temperature during summer months in the previous year may increase MSF incidence a year later. Because the ability to predict high and low transmission seasons may help target interventions,24 our results are of possible public health importance. Received December 22, 2006. Accepted for publication April 20, 2007. Authors’ addresses: Giovanni Rezza, Maria Fenicia Vescio, Massimo Ciccozzi, and Francesca Farchi, Epidemiology Division, Department of Infectious Diseases, Istituto Superiore di Sanità, Viale Regina Elena 299, 00161 Roma, Italy. E-mail:
[email protected]. Maroli Michele, Vector-Borne Diseases and International Health, Department of Infectious Diseases, Istituto Superiore di Sanita`, Viale Regina Elena 299, 00161 Roma, Italy. Antonina Carai, San Francesco Hospital, via Mannironi, 08100 Nuoro, Italy. Maria Adriana Piras and Maria S. Mura, Department of Infectious Diseases, University of Sassari, viale San Pietro 43/b, 07100 Sassari, Italy.
8.
9.
10. 11.
12. 13. 14. 15. 16.
REFERENCES 1. Parola P, 2004. Tick-borne rickettsial diseases: emerging risks in Europe. Comp Immmunol Microbiol Infect Dis 27: 297–304. 2. Scaffidi V, 1981. Attuale espansione endemo-epidemica della febbre bottonosa in Italia. Minerva Med 1: 2063–2070. 3. Tringali G, Intonazzo V, Perna AM, Mansueto S, Vitale G, Walker DH, 1986. Epidemiology of Boutonneuse fever in Western Sicily. Distribution and prevalence of spotted fever group Rickettsial Infection in dog ticks (Rhipicephalus sanguineus). Am J Epidemiol 123: 721–727. 4. Mansueto S, Vitale G, Bentivenga M, Tringali G, Di Leo R, 1985. Prevalence of antibodies to Rickettsia conorii after acute attack of boutonneuse fever. J Infect Dis 151: 377. 5. Bernabeu-Wittel M, del Toro MD, Nogueras MM, Muniain MA, Cardenosa N, Maequez FJ, Segura F, Pachon J, 2006. Seroepidemiological study of Rickettsia felis, Rickettsia typi, and Rickettsia conorii infection among the population of southern Spain. Eur J Microbiol Infect Dis 25: 375–381. 6. Raoult D, Tissot Dupont H, Caraco P, Brouqui P, Drancourt M, Charrel C, 1992. Mediterranean spotted fever in Marseille: descriptive epidemiology and the influence of climatic factors. Eur J Epidemiol 8: 192–197. 7. Jacobs PA, Fourie LJ, Horak IG, 2004. A laboratory comparison
17. 18. 19.
20. 21. 22. 23. 24.
of the life cycles of the dog ticks Haemaphysalis leachi and Rhipicephalus sanguineus. Onderstepoort J Vet Res 71: 15–28. Gilot B, LaForge ML, Pichot J, Raoult D, 1990. Relationship between the Rhypicephalus Sanguineus complex ecology and Mediterranean Spotted Fever epidemiology in France. Eur J Epidemiol 6: 357–362. Maroli M, Khoury C, Frusteri L, Manilla G, 1996. Diffusione della zecca del cane (Rhipicephalus sanguineus Lateille, 1806) in Italia: un problema di salute pubblica. Annali dell’ Istituto Superiore di Sanita` 32: 387–397. Danielova V, Benes C, 1997. Possible role of rainfall in the epidemiology of tick-borne encephalitis. Cent Eur J Public Health 5: 151–154. Sumilo D, Bormane A, Asokliene L, Lucenko I, Vasilenko V, Randolph S, 2006. Tick-borne encephalitis in the Baltic States: identifying risk factors in space and time. Int J Med Microbiol 296 (Suppl 40): 76–79. Espejo Arenas E, Font Creus F, Bella Cueto F, Segura Ports F, 1986. Climatic factors in resurgence of mediteranean spotted fever. Lancet 1: 1333. Parola P, Paddock CD, Raoult D, 2005. Tick-borne Rickettsioses around the World: emerging diseases challenging old concepts. Clin Microbiol Rev 18: 719–756. Daniel M, 1983. Influence of the microclimate on the vertical distribution of the tick Ixodes ricinus (L.) in central Europe. Acarologia 34: 105–113. Danielova V, Benes C, 1997. Possible role of rainfall in the epidemiology of tick-borne encephalitis. Cent Eur J Public Health 5: 151–154. Lindgren E, Talleklint L, Polfeldt T, 2000. Impact of climatic change on the northern latitude limit and population density of the disease -transmitting European tick Ixodes ricinus. Environ Health Perspect 108: 119–123. Lindgren E, Gustafson R, 2001. Tick-borne encephalitis in Sweden and climate change. Lancet 358: 16–18. Kaplan JE, Newhouse VF, 1984. Occurrence of Rocky Mountain Spotted Fever in relation to climatic, geophysical and ecological variables. Am J Trop Med Hyg 33: 1281–1282. Daniel M, Danielova V, Kriz B, Jirsa A, Nozicka J, 2003. Shift of the tick Ixodes ricinus and tick-borne encephalities to higher altitudes in central Europe. Eur J Clin Microbiol Infect Dis 22: 327–328. Skarphedinsson S, Jensen PM, Kristiansen K, 2005. Survey of tickborne infections in Denmark. Emerg Infect Dis 11: 1055– 1061. Haines A, Campbell-Lendrum D, Corvalan C, 2006. Climate change and human health: impacts, vunerability, and mitigation. Lancet 367: 2101–2109. Raoult D, Nicolas D, De Micco P, Gallais H, 1985. Aspects epidemiologiques de la Fievre boutonneuse Mediterraneenne en corse du sud. Bull Soc Pathol Exotique 78: 451. Randolph SE, Green RM, Peacey MF, Rogers DJ, 2000. Seasonal synchrony: the key to tick-borne encephalitis foci identified by satellite data. Parasitology 121: 15–23. Kovats RS, Bouma M, Hajat S, Worral E, Haines A, 2003. El Nino and health. Lancet 362: 1481–1489.
