BACKGROUND: Haemorrhagic stroke (HS) is a major cause of mortality and disability. Previous studies reported inconsistent associations between ambient air pollutants and HS risk. OBJECTIVE: We evaluated the association between air pollutant exposure and the risk of HS in a cosmopolitan city in the tropics. METHODS: We performed a nationwide, population-based, time-stratified case-crossover analysis on all HS cases reported to the Singapore Stroke Registry from 2009 to 2018 (n = 12,636). We estimated the risk of HS across tertiles of air pollutant concentrations in conditional Poisson models, adjusting for meteorological confounders. We stratified our analysis by age, atrial fibrillation and smoking status, and investigated the lagged effects of each pollutant on the risk of HS up to 5 days. RESULTS: All 12,636 episodes of HS were included. The median (1st-to 3rd-quartile) daily pollutant levels from 22 remote stations deployed across the island were as follows: (PM(2.5) = 15.9 (12.7-20.5), PM(10) = 27.3 (22.7-33.4), O(3) = 22.5 (17.3-29.8), NO(2) = 23.3 (18.8-28.4), SO(2) = 10.2 (5.6-14.4), CO = 0.5 (0.5-0.6). The median (1st-to 3rd-quartile) temperature (°C) was 27.9 (27.1-28.7), that of relative humidity (%) was 79.4 (75.6-83.2), and that of total rainfall (mm) was 0.0 (0.0-4.2). Higher levels of CO were significantly associated with an increased risk of HS (3rd tertile vs 1st tertile: Incidence Rate Ratio (IRR) = 1.06, 95% CI = 1.01-1.12). The increased risk of HS due to CO persisted for at least 5 days after exposure. Individuals under 65 years old and non-smokers had a higher risk of HS when exposed to CO. O(3) was associated with increased risk of HS up to 5 days (3rd tertile vs 1st tertile: IRR(day 1) = 1.07, 95% CI = 1.02-1.12; IRR(day 5) = 1.07, 95% CI = 1.02-1.13). CONCLUSION: Short-term exposure to ambient CO levels was associated with an increased risk of HS. A reduction in CO emissions may reduce the burden of HS in the population.
Heat waves are unusually high temperature events over consecutive days and may cause adverse impacts such as morbidity and mortality. The interaction between heat waves and urban heat island (UHI) effects has remained a subject of debate, as some studies prove heat wave-UHI synergy while others do not. Furthermore, heat waves affect tropical cities more severely than mid-latitude cities, but there is a disproportionate lack of heat wave studies focusing on tropical cities. We attempt to narrow this gap by studying the heat wave in Singapore in April 2016 using ground observations and the Weather Research and Forecasting (WRF) model. Compared to non-heat wave days, the ground observations show that daytime temperatures can be 3 degrees C higher during the heat wave. Despite the temperature spike, the UHI intensity is not amplified during the heat wave, maintaining its peak near 2.5 degrees C during both heat wave and non-heat wave periods. WRF simulation results also agree well with measurements and predict UHI peaks near 2.5 degrees C during both periods, showing no heat wave-UHI synergy. The spatially averaged UHI intensity also shows no such synergy. There is no significant change of wind speed, soil moisture availability or heat storage flux during the heat wave. Therefore, the lack of heat wave-UHI synergy in our study is consistent with current understanding of factors contributing to UHI. This study shows that the heat wave-UHI interaction in a tropical city can be different from that in cities in the temperate climate zone and more studies should be conducted in tropical cities, which are projected to suffer larger impacts of increasing heat stress.
Understanding extreme temperature variations is important for countries to manage risks associated with climate change. Yet, the characteristics of temperature extremes and possible climate change impacts have not been adequately investigated in Singapore. In this study, we attempted to do so by defining 14 extreme temperature indices (ETIs) for the period of 1982-2018 in Singapore, and investigating the trends of those ETIs using a pre-whitening Man-Kendall test coupled with the Sen’s slope estimator method. The linear and nonlinear relationships between ETIs and El Nino Southern Oscillation (ENSO) were also examined using correlation, composite and wavelet analysis. Our results indicate that trends of temperature extremes varied according to station locations, ETIs and time scales. In all stations, ETIs such as the monthly mean value of the diurnal range between maximum and minimum temperatures (DTR), cool nights (TN10p) and cool days (TX10p) presented decreasing trends, while the rest of them exhibited increasing trends. The composite values varied for different ETIs-meaning that while eight no-threshold ETIs reflected smaller values, other ETIs reflected relatively larger composite values, indicating that ENSO may have affected those ETIs more. The ETIs were mainly statistically and significantly coherent with ENSO at a 2-8 year cycle. We hope that our findings would be beneficial for climate action planning and temperature-related disaster prevention in Singapore.
BACKGROUND: Global incidence of dengue has surged rapidly over the past decade. Each year, an estimated 390 million infections occur worldwide, with Asia-Pacific countries bearing about three-quarters of the global dengue disease burden. Global warming may influence the pattern of dengue transmission. While previous studies have shown that extremely high temperatures can impede the development of the Aedes mosquito, the effect of such extreme heat over a sustained period, also known as heatwaves, has not been investigated in a tropical climate setting. AIM: We examined the short-term relationships between maximum ambient temperature and heatwaves and reported dengue infections in Singapore, via ecological time series analysis, using data from 2009 to 2018. METHODS: We studied the effect of two measures of extreme heat – (i) heatwaves and (ii) maximum ambient temperature. We used a negative binomial regression, coupled with a distributed lag nonlinear model, to examine the immediate and lagged associations of extreme temperature on dengue infections, on a weekly timescale. We adjusted for long-term trend, seasonality, rainfall and absolute humidity, public holidays and autocorrelation. RESULTS: We observed an overall inhibitive effect of heatwaves on the risk of dengue infections, and a parabolic relationship between maximum temperature and dengue infections. A 1 °C increase in maximum temperature from 31 °C was associated with a 13.1% (Relative Risk (RR): 0.868, 95% CI: 0.798, 0.946) reduction in the cumulative risk of dengue infections over six weeks. Weeks with 3 heatwave days were associated with a 28.3% (RR: 0.717, 95% CI: 0.608, 0.845) overall reduction compared to weeks with no heatwave days. Adopting different heatwaves specifications did not substantially alter our estimates. CONCLUSION: Extreme heat was associated with decreased dengue incidence. Findings from this study highlight the importance of understanding the temperature dependency of vector-borne diseases in resource planning for an anticipated climate change scenario.
OBJECTIVES: Evidence of the relationship between climate variability, air pollution and human parainfluenza virus (HPIV) infections has been inconsistent. We assessed this in a paediatric population from a highly urbanized tropical city-state. METHODS: We analysed all reports of HPIV infections in children <5 years old obtained from a major specialist women and children's hospital in Singapore. Assuming a negative binomial distribution and using multivariable fractional polynomial modelling, we examined the relations between climate variability, air quality and the risk of HPIV infections, adjusting for time-varying confounders. RESULTS: We identified 6393 laboratory-confirmed HPIV infections from 2009 to 2019. Every 1 °C decline in temperature was associated with a 5.8% increase (RR: 0.943, 95% Confidence Interval [95% CI]: 0.903-0.984) in HPIV infection risk 6 days later. Every 10% decrease in relative humidity was associated with a 15.8% cumulative increase in HPIV risk over the next 6 days (cumulative RR: 0.842, 95% CI: 0.771-0.919). Rainfall was positively associated with HPIV risk 2 days later (RR: 1.021, 95% CI: 1.000-1.043). A within-year seasonal rise of HPIV was driven by HPIV-3 and HPIV-1 and preceded by a seasonal decline in temperature. Gender was an effect modifier of the climate-HPIV relationship. Air quality was not associated with HPIV risk. CONCLUSIONS: This study demonstrates a close association between HPIV infection risk and tropical climate variability. The climate dependence and seasonal predictability of HPIV can inform the timing of community campaigns aimed at reducing infection risk and the development of hospital resources and climate adaption plans.
BACKGROUND: Acute respiratory infections (ARIs) are among the most common human illnesses globally. Previous studies that examined the associations between climate variability and ARIs or ARI pathogens have reported inconsistent findings. Few studies have been conducted in Southeast Asia to date, and the impact of climatic factors are not well-understood. This study aimed to investigate the short-term associations between climate variability and ARIs in Singapore. METHODS: We obtained reports of ARIs from all government primary healthcare services from 2005 to 2019 and analysed their dependence on mean ambient temperature, minimum temperature and maximum temperature using the distributed lag non-linear framework. Separate negative binomial regression models were used to estimate the association between each temperature (mean, minimum, maximum temperature) and ARIs, adjusted for seasonality and long-term trend, rainfall, relative humidity, public holidays and autocorrelations. For temperature variables and relative humidity we reported cumulative relative risks (RRs) at 10th and 90th percentiles compared to the reference value (centered at their medians) with corresponding 95% confidence intervals (CIs). For rainfall we reported RRs at 50th and 90th percentiles compared to 0 mm with corresponding 95% CIs. RESULTS: Statistically significant inverse S-curve shaped associations were observed between all three temperature variables (mean, minimum, maximum) and ARIs. A decrease of 1.1 °C from the median value of 27.8 °C to 26.7 °C (10th percentile) in the mean temperature was associated with a 6% increase (RR: 1.06, 95% CI: 1.03 to 1.09) in ARIs. ARIs also increased at 23.9 °C (10th percentile) compared to 24.9 °C of minimum temperature (RR: 1.11, 95% CI: 1.07 to 1.16). The effect of maximum temperature for the same comparison (30.5 °C vs 31.7 °C) was non-significant (RR: 1.02, 95% CI: 0.99 to 1.05). An increase in ambient temperature to 28.9 °C (90th percentile) was associated with an 18% decrease (RR: 0.82, 95% CI: 0.80 to 0.83) in ARIs. Similarly, ARIs decreased with the same increase to 90th percentile in minimum (RR: 0.84, 95% CI: 0.80 to 0.87) and maximum (RR: 0.89, 95% CI: 0.86 to 0.93) temperatures. Rainfall was inversely associated with ARIs and displayed similar shape in all three temperature models. Relative humidity, on the other hand, exhibited a U-shaped relationship with ARIs. CONCLUSION: Our findings suggest that lower temperatures increase the risk of ARIs. Anticipated extreme weather events that reduce ambient temperature can be used to inform increased healthcare resource allocation for ARIs.
Culex mosquitoes are important vectors of West Nile Virus (WNV), St. Louis Encephalitis Virus (SLEV) and Japanese Encephalitis Virus (JEV). Climate change is expected to alter their ability to spread diseases in human populations. Studies examining the influence of climate variability on Culex mosquitoes in South East Asia are scarce. We examined the influence of climate variability on reported Culex mosquito larval habitats from 2009 to 2018 in Singapore. We analysed the non-linear immediate and lagged weather dependence of Culex habitats over 5 weeks in negative binomial regression models using nationally representative data. We adjusted for the effects of long-term trend, seasonality, public holidays and autocorrelation. There were 41,170 reported Culex larval habitats over the study period. Non-residential premises were associated with more reports of habitats compared to residential premises [Rate Ratio (RR): 113.9, 95% CI: 110.9, 116.9]. Larvae in more than 90% of these habitats were entomologically identified as Culex quinquefasciatus. In residences, every 10 mm increase in rainfall above a 90 mm threshold was associated with a 10.1% [Incidence Rate Ratio (IRR): 0.899, 95% CI: 0.836, 0.968] cumulative decline in larval habitats. Public holidays were not significantly included in the model analysing larval habitats in residences. In non-residences, a 1 °C increase in the ambient air temperature with respect to the mean was associated with a 36.0% (IRR: 1.360, 95% CI: 1.057, 1.749) cumulative increase in Culex larval habitats. Public holidays were associated with a decline in Culex larval habitats in the same week. Our study provides evidence of how ambient air temperature and rainfall variability influences the abundance of Culex mosquito larval habitats. Our findings support the utility of using weather data in predictive models to inform the timing of vector control measures aimed at reducing the risk of WNV and other Culex-borne flavivirus transmission in urban areas.
The need for healthcare workers (HCWs) to wear personal protective equipment (PPE) during the coronavirus disease 2019 (COVID-19) pandemic heightens their risk of thermal stress. We assessed the knowledge, attitudes, and practices of HCWs from India and Singapore regarding PPE usage and heat stress when performing treatment and care activities. One hundred sixty-five HCWs from India (n = 110) and Singapore (n = 55) participated in a survey. Thirty-seven HCWs from Singapore provided thermal comfort ratings before and after ice slurry ingestion. Differences in responses between India and Singapore HCWs were compared. A p-value cut-off of 0.05 depicted statistical significance. Median wet-bulb globe temperature was higher in India (30.2 °C (interquartile range [IQR] 29.1-31.8 °C)) than in Singapore (22.0 °C (IQR 18.8-24.8 °C)) (p < 0.001). Respondents from both countries reported thirst (n = 144, 87%), excessive sweating (n = 145, 88%), exhaustion (n = 128, 78%), and desire to go to comfort zones (n = 136, 84%). In Singapore, reports of air-conditioning at worksites (n = 34, 62%), dedicated rest area availability (n = 55, 100%), and PPE removal during breaks (n = 54, 98.2%) were higher than in India (n = 27, 25%; n = 46, 42%; and n = 66, 60%, respectively) (p < 0.001). Median thermal comfort rating improved from 2 (IQR 1-2) to 0 (IQR 0-1) after ice slurry ingestion in Singapore (p < 0.001). HCWs are cognizant of the effects of heat stress but might not adopt best practices due to various constraints. Thermal stress management is better in Singapore than in India. Ice slurry ingestion is shown to be practical and effective in promoting thermal comfort. Adverse effects of heat stress on productivity and judgment of HCWs warrant further investigation.
