Global warming and neurological practice: systematic review

Inclusion criteria

a. Studies published in English and listed in MEDLINE (PubMed) and Embase from January 1, 2000.

b. Non-English publications if an English abstract and a reliable translation of the manuscript into English were available

c. Cross-sectional or longitudinal, retrospective or prospective, observational clinical and research studies presenting original data about major neurological disorders (Alzheimer’s disease, stroke, epilepsy, neuroinfectious diseases, Parkinson’s disease and multiple sclerosis)—morbidity, frequency and/or mortality or any subtype—in response to an increase to ambient temperature were included.

d. Cross-sectional or longitudinal, retrospective or prospective, observational clinical and research studies presenting original data about major neurological disorders (headache/migraine, Alzheimer’s disease, stroke, epilepsy, neuroinfectious diseases, Parkinson’s disease or multiple sclerosis) in refugee and/or migrant populations.

Exclusion criteria

a. Studies published prior to January 1, 2000.

b. We excluded commentaries, summaries, reviews, editorials, animal studies, toxicological studies, duplicates, and studies that reported another association not related to our objectives.

c. In cases of missing data (those we cannot extract), the authors were contacted for additional information. If no answer was obtained, the study was excluded.

Inclusion criteria were human studies, published in English (or when a reliable English translation was available) and listed in MEDLINE (PubMed) or Scopus from January 1st, 2000 onwards; we included cross-sectional or longitudinal, retrospective or prospective, observational clinical and research studies presenting original epidemiological and/or clinical data about major neurological disorders as outlined above and their morbidity and/or mortality and/or prevalence/incidence, in relation to an increase in ambient temperature (any), respectively, refugee and/or migrant populations (irrespective of whether comparisons were made between migration groups with populations from the place of origin or place of departure).

Exclusion criteria were publications prior to January 1st, 2000; lack of a reliable English translation; and we excluded commentaries, summaries, reviews, editorials, animal studies, toxicological studies, and duplicate studies; and in cases of missing data, the authors were contacted for additional information. If no answer was obtained, studies were excluded as well.

Alzheimer’s and non-Alzheimer’s dementia

Eleven studies investigated the effects of high ambient temperatures on patients with Alzheimer’s and non-Alzheimer’s dementia (Cornali et al., 2004; Hansen et al., 2008; Page et al., 2012; Zanobetti et al., 2013; Trang et al., 2016; Culqui et al., 2017; Conti et al., 2007; Lee et al., 2018; Linares et al., 2017; Zhang et al., 2016b; Xu et al., 2019). The studies were from eight countries with a total case population of N = 961,582. One study did not specify the number of cases (Hansen et al., 2008).

Maximum temperatures reported were >30 °C in nine studies (Cornali et al., 2004; Hansen et al., 2008; Zanobetti et al., 2013; Trang et al., 2016; Culqui et al., 2017; Conti et al., 2007; Linares et al., 2017; Zhang et al., 2016b; Xu et al., 2019), and respectively 18 °C and >24.5 °C in two of the studies (Page et al., 2012; Lee et al., 2018). In nine of the 11 studies, high ambient temperatures were associated with worsening of symptoms including agitation, as well as an increase in hospitalizations and/or mortality (Cornali et al., 2004; Hansen et al., 2008; Zanobetti et al., 2013; Culqui et al., 2017; Conti et al., 2007; Lee et al., 2018; Linares et al., 2017; Zhang et al., 2016b; Xu et al., 2019). Five studies suggested that high temperature was attributable to worsening of symptoms and increased risk of admission (Cornali et al., 2004; Culqui et al., 2017; Lee et al., 2018; Linares et al., 2017; Zhang et al., 2016b) while Hansen et al. found a significantly increased rate of both hospital admissions and mortality in 94,447 patients with dementia during periods of heat waves (Hansen et al., 2008). Zanobetti et al., (2013), Conti et al. (2007) and Xu et al. (2019) also found significantly increased mortality during hot months. In the remaining two studies, no associations were found between high ambient temperatures and hospital admissions or death in people with dementia (Page et al., 2012; Trang et al., 2016). Overall, a negative impact of increased temperatures on hospitalization and mortality in patients with dementia was reported in nine of the 11 studies.

Epilepsy

Only one epilepsy study by Rakers et al. (2017) met inclusion criteria, a case-crossover study with 604 patients. The results indicated a 46% lower risk of admissions for epileptic seizures one day after exposure to temperatures above 20 °C. Further, an inverse association was found between epileptic seizures and low atmospheric pressure and high humidity (Rakers et al., 2017).

Headache and migraine

Two studies from two countries including N = 7,156 cases reported on the effects of high temperatures on headache (33% migraines) (Mukamal et al., 2009; Neut et al., 2012). The largest study by Mukamal et al. analyzed data from 7,054 patients seen in an emergency department between 2000–2007 (Mukamal et al., 2009). In this study, higher ambient temperatures in the 24 h preceding hospital presentation increased the risk of acute headache requiring emergency evaluation with 7.5% for each 5 °C increment in temperature. Neut et al. interviewed 102 children and adolescents with migraine and/or their parents about triggering factors precipitating migraine attacks. 70% stated warm climate could trigger their migraine, and 24% reported warm climate was often or very often a trigger factor for migraine attacks (Neut et al., 2012).

Multiple sclerosis

Six studies from five countries including N = 5,305 addressed the effects of high temperatures on patients with MS (Ogawa, Mochizuki & Kanzaki, 2004; Simmons et al., 2004; Tataru et al., 2006; Leavitt et al., 2012; Roberg & Bruce, 2016; Stellmann et al., 2017). Four studies found high temperatures to be associated with worsening of symptoms in these patients (Ogawa, Mochizuki & Kanzaki, 2004; Simmons et al., 2004; Leavitt et al., 2012; Stellmann et al., 2017). Stellmann et al. (2017) included 1,254 MS patients and showed that mobility deficits as measured by the ‘timed 25 foot walking test’ were increased so that every 10 °C increase of ambient temperature prolonged walking time by 0,35 s. In an anonymized self-reported survey by Simmons et al. (2004) 70% of multiple sclerosis participants reported worsening of their symptoms with high temperatures. Poor cognitive performance significantly correlated with higher outdoor temperatures in another group of 40 multiple sclerosis patients (Leavitt et al., 2012). Furthermore, the latter study also found a significant correlation between poor cognitive performance and high temperatures over a period of 6 months, suggesting a negative impact of warm temperatures on cognitive function in MS patients both when analyzed cross-sectionally and longitudinally (Leavitt et al., 2012). Higher frequencies of MS attacks were recorded during the warmest months in a small sample of 34 patients (Ogawa, Mochizuki & Kanzaki, 2004). The remaining two studies did not show any effects of high temperatures on MS symptomatology (Tataru et al., 2006; Roberg & Bruce, 2016). Thus, majority of the studies included (4 of 6 studies) reported progression of symptoms in MS patients during warm periods.

Parkinson’s disease

Three studies from three countries (FR, ES and US) with a total N = 204,656 evaluated the effects of high temperatures on Parkinson’s disease (Zanobetti et al., 2013; Pathak et al., 2005; Linares et al., 2016). Results from one study indicated a correlation between high ambient temperature (>34 °C) and an increase risk of excess morbidity and mortality in 3,287 PD patients (Linares et al., 2016). In a small French study with 36 PD patients, there was a trend towards more frequent autonomic failure during heat waves (Pathak et al., 2005). In contrast, Zanobetti et al. investigated the effects of extreme hot days (maximum temperature of 31,7 °C) on mortality in 201,333 patients with a diagnosis of PD but found no association between mortality and ambient temperatures (Zanobetti et al., 2013).

Stroke, overview

In total, 58 papers dealing with the effects of ambient temperatures on cerebrovascular morbidity and mortality were included (Zanobetti et al., 2013; Dawson et al., 2008; Feigin et al., 2000; Stafoggia et al., 2008; Vaneckova et al., 2008; Basu & Ostro, 2008; Knowlton et al., 2009; Lin et al., 2009; Wang et al., 2009; Green et al., 2010; Huang, Kan & Kovats, 2010; Basagaña et al., 2011; Wichmann et al., 2011b; Wichmann et al., 2011a; Basu et al., 2012; Cowperthwaite & Burnett, 2011; Ha et al., 2014; Lim, Kim & Hong, 2013; Shaposhnikov et al., 2014; Vaneckova & Bambrick, 2013; Wang, Gao & Wang, 2013; Chen et al., 2017a; Chen et al., 2017b; Chen et al., 2013; Chen et al., 2018; Wang & Lin, 2014; Zhang et al., 2014; Méndez-Lázaro et al., 2016; Vodonos et al., 2016; Guo et al., 2017; Han et al., 2017; Ponjoan et al., 2017; Zhang et al., 2016a; Zhou et al., 2017; Bai et al., 2018; Fu et al., 2018; Luo et al., 2018; Sherbakov et al., 2018; Bao et al., 2019; Shimomura et al., 2019; Ostro et al., 2010; Tarnoki et al., 2017; Oudin, Strömberg & Jakobsson, 2010; Salam et al., 2019; Goggins et al., 2012; Mostofsky et al., 2014; Hori, Hashizume & Tsuda, 2012; Matsumoto, Ishikawa & Kajii, 2010; Hong et al., 2003; Myint et al., 2007; Zheng et al., 2016; Çevik et al., 2015; Kyobutungi et al., 2005; Bobb et al., 2014; Yang et al., 2018; Harlan et al., 2014; Qian et al., 2008; Yang et al., 2016). The studies were from 18 different countries with a total case population of N = 5,869,284. Ischemic stroke was reported in 1,812,457, hemorrhagic stroke in 446,407, and subarachnoid hemorrhage (SAH) in 7,160 cases. In the remaining 3,603,260 cases, stroke subtypes were not specified.

Stroke, hospitalizations

Associations between stroke-related hospitalization and ambient temperatures were reported in 38 studies. Thirteen studies reported increasing risks for hospitalizations for stroke with higher ambient temperatures (Dawson et al., 2008; Wang et al., 2009; Ha et al., 2014; Shaposhnikov et al., 2014; Chen et al., 2017b; Vodonos et al., 2016; Sherbakov et al., 2018; Bao et al., 2019; Shimomura et al., 2019; Ostro et al., 2010; Tarnoki et al., 2017; Oudin, Strömberg & Jakobsson, 2010; Salam et al., 2019). Maximum temperatures were 12.9 °C in one study (Dawson et al., 2008), 20–27 °C in five studies (Wang et al., 2009; Ha et al., 2014; Sherbakov et al., 2018; Tarnoki et al., 2017; Oudin, Strömberg & Jakobsson, 2010), and >30 °C in five studies (Vaneckova & Bambrick, 2013; Chen et al., 2017b; Vodonos et al., 2016; Bao et al., 2019; Salam et al., 2019). Temperatures were unspecified in two studies (Shimomura et al., 2019; Ostro et al., 2010). Chen et al. (2017a) found excess hospitalizations due to ischemic stroke with high temperatures, while the risk for admissions because of hemorrhagic stroke was increased during both high and low temperatures. Green et al. (2010) and Basu et al. (2012) also found increased risks of hospitalization due to ischemic stroke with high temperatures (25.3 and 30.1 °C, respectively), but a decreased risk for hemorrhagic stroke. Bai et al. (2018) reported an increase in risk of stroke admissions both with high and low ambient temperatures. In 10 studies, lower ambient temperatures (−16.94 to 17 °C) were associated with increased risk of stroke-related hospitalizations (Wang & Lin, 2014; Guo et al., 2017; Luo et al., 2018; Goggins et al., 2012; Mostofsky et al., 2014; Hori, Hashizume & Tsuda, 2012; Matsumoto, Ishikawa & Kajii, 2010; Hong et al., 2003; Myint et al., 2007; Zheng et al., 2016). In the study by Wang, Gao & Wang (2013) there was an inverse correlation between temperatures and ischemic stroke admissions, leading to an increase of admissions during cold spells and a decrease during heat waves. Lin et al. (2009) reported lower risks of stroke admissions associated with high ambient temperatures (31.7 °C), even though the results were not statistically significant. Çevik et al. (2015) only found association between lower temperature and increased risk of SAH, and as in the remaining eight studies no associations between ambient temperatures and stroke-related hospitalizations was found (Feigin et al., 2000; Knowlton et al., 2009; Wichmann et al., 2011a; Cowperthwaite & Burnett, 2011; Vaneckova & Bambrick, 2013; Ponjoan et al., 2017; Kyobutungi et al., 2005; Bobb et al., 2014). Overall, most studies (17 of 24) reported higher risk of stroke-related hospitalizations during periods of high ambient temperatures.

Stroke, mortality

Twenty-one studies reported on the association between ambient temperatures and cerebrovascular mortality (Zanobetti et al., 2013; Stafoggia et al., 2008; Vaneckova et al., 2008; Basu & Ostro, 2008; Huang, Kan & Kovats, 2010; Basagaña et al., 2011; Wichmann et al., 2011b; Lim, Kim & Hong, 2013; Chen et al., 2013; Chen et al., 2018; Zhang et al., 2014; Méndez-Lázaro et al., 2016; Han et al., 2017; Zhang et al., 2016a; Zhou et al., 2017; Fu et al., 2018; Myint et al., 2007; Yang et al., 2018; Harlan et al., 2014; Qian et al., 2008; Yang et al., 2016). In 11 of the 21 studies (Zanobetti et al., 2013; Stafoggia et al., 2008; Vaneckova et al., 2008; Huang, Kan & Kovats, 2010; Basagaña et al., 2011; Lim, Kim & Hong, 2013; Méndez-Lázaro et al., 2016; Zhou et al., 2017; Yang et al., 2018; Harlan et al., 2014; Qian et al., 2008), increased risk of cerebrovascular death was associated with high ambient temperatures, including very high temperatures (>30 °C). Five studies reported an increase in cerebrovascular mortality with both low and high temperatures (Chen et al., 2013; Chen et al., 2018; Han et al., 2017; Zhang et al., 2016a; Fu et al., 2018), while 2 studies found no effect (Basu & Ostro, 2008; Wichmann et al., 2011b). Zhang et al. (2014), Myint et al. (2007) and Yang et al. (2016) were the only three to report significantly higher risks of cerebrovascular death in periods with cold, but not warm temperatures. Overall, most studies (16 of the 21 studies) found a higher risk of cerebrovascular mortality with increasing ambient temperatures.

Tick-borne encephalitis

The effects of high ambient temperatures on tick-borne encephalitis were assessed in four studies from three countries (SE, CZ, SI) (Danielová et al., 2008; Lindgren & Gustafson, 2001; Lukan, Bullova & Petko, 2010; Zeman & Bene, 2004). Lindgren & Gustafson (2001) linked an increase in the incidence of tick-borne encephalitis in Sweden, starting in the mid-80s, to climate change with increasingly milder winters and earlier springs. Lukan, Bullova & Petko (2010) and Zeman & Bene (2004) analyzed respectively 1,786 and 8,700 cases of tick-borne encephalitis in the period of 1961–70 to 2004. During 1980 to 2004 they found an increase in tick-borne encephalitis foci in areas of increasing altitudes which corresponded to gradual rises in annual temperatures, indicating an effect of climate change and higher temperatures on local tick-borne encephalitis incidence. Overall, the studies indicated a rise in the total incidence of tick-borne encephalitis with milder temperatures in Sweden and increasing temperatures in higher altitude areas in the Czech Republic and Slovenia.

Alzheimer’s and non-Alzheimer’s dementia

One observational study reported a prevalence of 12.1% of all type dementia in 157 elderly Japanese individuals (age >70 years) who had migrated from Okinawa, Japan, to Brazil (Yamada et al., 2002). In comparison, risk of all type dementia in Japanese individuals living in Okinawa was 7.3%. Prevalence of dementia in five other regions in Japan was between 5.2 to 11.8%. Overall, prevalence of all type dementia was higher in Japanese migrants in Brazil than in Japanese who had stayed in Japan.

Multiple Sclerosis

Five observational studies investigated the effects of migration on the frequency (Guimond et al., 2014; Hammond, English & McLeod, 2000; Nielsen et al., 2019) and morbidity (Merle et al., 2005; Comini-Frota et al., 2013) of MS. In the study by Guimond et al., MS prevalence was higher in migrants who came from Iran to British Columbia (Yamada et al., 2002), whereas in another study the incidence was higher in the population from the home country (UK and Ireland) compared to the incidence in a migrant population settling in Australia (Hammond, English & McLeod, 2000). In one large retrospective study from Denmark (considered a high risk country for MS with a prevalence of >60 pr. 100,000) including more than 9 million individuals between 1968 and 2015 from the Danish Civil Registration System, higher risk of MS in 1st generation immigrants was found compared to the official risk rates of MS in their birth countries, especially if migration was from low- and intermediate risk countries (e.g., MS prevalence of <5 pr. 100,000 and 5–60 pr. 100,000 respectively) (Nielsen et al., 2019). The risk of MS was higher the younger the age at time of migration, while risk of MS was unchanged in people immigrating at age 35 years or above. MS risk among second-generation immigrants whose parents were born in low risk areas was doubled (Nielsen et al., 2019). Merle et al. investigated visual impairment in MS patients from Martinique who had or had not migrated to metropolitan France for at least 1 year before the age of 15 years and found that MS symptoms were more frequent and severe in the non-migrant group (Merle et al., 2005). Yet Comini-Frota et al. (2013) found no difference of MS morbidity between an Italian migrant population in Brazil and non-migrant Italians.

Stroke

Three observational studies compared data on cerebrovascular disease in migrant populations with data from non-migrant populations (Hayfron-Benjamin et al., 2019; Chiu et al., 2010; Wolfe et al., 2006). Two studies reported a higher prevalence of stroke in non-migrant groups living in China (Chiu et al., 2010) and Ghana (Hayfron-Benjamin et al., 2019), as opposed to migrant populations living in Hong Kong, Singapore, Taiwan, Western Europe, and North America. After adjusting for conventional cardiovascular risk factors, results were no longer significant in one study (Hayfron-Benjamin et al., 2019). In contrast, Wolfe et al. observed a higher incidence of stroke in migrants from Barbados moving to South London, including an increased incidence for specific stroke subtypes such as total anterior cerebral infarctions, posterior cerebral infarctions and SAH (Wolfe et al., 2006).