> top > projects > LitCovid-sentences > docs > PMC:7786642 > annotations
LitCovid-sentences  

PMC:7786642 JSONTXT 9 Projects

Annnotations TAB TSV DIC JSON TextAE

Id Subject Object Predicate Lexical cue
T1 0-70 Sentence denotes Long-term exposure to air-pollution and COVID-19 mortality in England:
T2 71-102 Sentence denotes A hierarchical spatial analysis
T3 104-114 Sentence denotes Highlights
T4 115-200 Sentence denotes • We downscaled the coarse spatial resolution of deaths to mitigate ecological bias.
T5 201-286 Sentence denotes • We accounted for confounding, spatial autocorrelation and pre-existing conditions.
T6 287-355 Sentence denotes • We found some evidence of an effect of NO2 on COVID-19 mortality.
T7 356-421 Sentence denotes • The effect of long-term PM2.5 exposure remains more uncertain.
T8 422-501 Sentence denotes • Our spatial model captured strong patterns likely reflecting disease spread.
T9 503-511 Sentence denotes Abstract
T10 512-611 Sentence denotes Recent studies suggested a link between long-term exposure to air-pollution and COVID-19 mortality.
T11 612-798 Sentence denotes However, due to their ecological design based on large spatial units, they neglect the strong localised air-pollution patterns, and potentially lead to inadequate confounding adjustment.
T12 799-933 Sentence denotes We investigated the effect of long-term exposure to NO2 and PM2.5 on COVID-19 mortality in England using high geographical resolution.
T13 934-1110 Sentence denotes In this nationwide cross-sectional study in England, we included 38,573 COVID-19 deaths up to June 30, 2020 at the Lower Layer Super Output Area level (n = 32,844 small areas).
T14 1111-1213 Sentence denotes We retrieved averaged NO2 and PM2.5 concentration during 2014–2018 from the Pollution Climate Mapping.
T15 1214-1363 Sentence denotes We used Bayesian hierarchical models to quantify the effect of air-pollution while adjusting for a series of confounding and spatial autocorrelation.
T16 1364-1602 Sentence denotes We find a 0.5% (95% credible interval: −0.2%, 1.2%) and 1.4% (95% CrI: −2.1%, 5.1%) increase in COVID-19 mortality risk for every 1 μg/m3 increase in NO2 and PM2.5 respectively, after adjusting for confounding and spatial autocorrelation.
T17 1603-1704 Sentence denotes This corresponds to a posterior probability of a positive effect equal to 0.93 and 0.78 respectively.
T18 1705-1808 Sentence denotes The spatial relative risk at LSOA level revealed strong patterns, similar for the different pollutants.
T19 1809-1899 Sentence denotes This potentially captures the spread of the disease during the first wave of the epidemic.
T20 1900-2044 Sentence denotes Our study provides some evidence of an effect of long-term NO2 exposure on COVID-19 mortality, while the effect of PM2.5 remains more uncertain.
T21 2046-2061 Sentence denotes 1 Introduction
T22 2062-2196 Sentence denotes As of 30th of June 2020, COVID-19 has caused more than 500,000 deaths globally, with an estimated case fatality of 1–4% (Hauser et al.
T23 2197-2203 Sentence denotes 2020).
T24 2204-2418 Sentence denotes The UK is one of the countries most affected, with an estimated 57,300 more deaths in England and Wales than it would be expected from mid-February to end of May 2020 had the pandemic not taken place (Kontis et al.
T25 2419-2425 Sentence denotes 2020).
T26 2426-2514 Sentence denotes Established risk factors of COVID-19 mortality include age, sex and ethnicity (Wu et al.
T27 2515-2521 Sentence denotes 2020).
T28 2522-2686 Sentence denotes Previous studies have observed a correlation between pre-existing conditions such as stroke, hypertension and diabetes (Williamson et al., 2020, Yang et al., 2020).
T29 2687-2854 Sentence denotes Long-term exposure to air-pollution has been hypothesised to worsen COVID-19 prognosis: either directly, as it can suppress early immune responses to the infection (E.
T30 2855-2871 Sentence denotes Conticini et al.
T31 2872-3026 Sentence denotes 2020), or indirectly, as it can increase the risk of stroke, hypertension and other pre-existing conditions (Giorgini et al., 2016, Scheers et al., 2015).
T32 3027-3280 Sentence denotes Previous studies suggested an effect of long-term exposure to air-pollution on COVID-19 mortality (Cole et al., 2020, Liang et al., 2020, Travaglio et al., 2020, Wu et al., 2020), however several methodological shortcomings limit their interpretability.
T33 3281-3456 Sentence denotes They were based on data aggregated on large spatial units and thus suffer from ecological fallacy (grouped levels association do not reflect individual ones) (Wakefield 2008).
T34 3457-3634 Sentence denotes Air pollution is characterised by high spatial variability, making the availability of mortality data at the same high spatial resolution crucial (Villeneuve and Goldberg 2020).
T35 3635-3811 Sentence denotes In addition, a coarse geographical resolution might lead to inadequate adjustment for confounders, when these are available at higher resolution (Villeneuve and Goldberg 2020).
T36 3812-4043 Sentence denotes Most previous studies assessed cumulative deaths until mid or end of April and thus the generalisability of their results is limited to the early stages of the epidemic (Liang et al., 2020, Travaglio et al., 2020, Wu et al., 2020).
T37 4044-4104 Sentence denotes One study had data available up to June 5, 2020 (Cole et al.
T38 4105-4232 Sentence denotes 2020) and another up to June 12, 2020 (Statistics 2020), capturing a proportion COVID-19 deaths attributable to the first wave.
T39 4233-4465 Sentence denotes In this nationwide study in England, we investigated the effect of long-term exposure to air pollution on COVID-19 mortality during the entire first wave of the epidemic, after accounting for confounding and spatial autocorrelation.
T40 4466-4579 Sentence denotes We focused on exposure to NO2 and PM2.5 (atmospheric particulate matter that has a diameter of less than 2.5 µm).
T41 4580-4791 Sentence denotes We downscaled the LTLA geographical information to the Lower Layer Super Output Area (LSOA) to alleviate the effect of ecological bias and exploit the variability of the exposure at high geographical resolution.
T42 4793-4803 Sentence denotes 2 Methods
T43 4805-4826 Sentence denotes 2.1 Study population
T44 4827-4919 Sentence denotes We included all COVID-19 deaths as reported to Public Health England (PHE) by June 30, 2020.
T45 4920-5169 Sentence denotes These include deaths that had a laboratory confirmed report of COVID-19 (including at post-mortem) (EpiCell 2020), as well as suspected COVID-19 deaths, defined as deaths without a positive test but with mention of COVID-19 in the death certificate.
T46 5170-5254 Sentence denotes These definitions were consistent during the study period and over the study region.
T47 5255-5318 Sentence denotes The main outcome of this study was laboratory confirmed deaths.
T48 5319-5564 Sentence denotes We selected COVID-19 deaths up to June the 30th to ensure we captured COVID-19 deaths attributable to the first wave of epidemic that in England and Wales was over by the end of May, when all-cause mortality was no longer elevated (Kontis et al.
T49 5565-5571 Sentence denotes 2020).
T50 5572-5770 Sentence denotes Individual data on age, sex, ethnicity, lower tier local authority (LTLA) of the residential address and type of residence type (i.e. nursing homes, prisons, medical facilities etc.) were available.
T51 5771-5860 Sentence denotes Population at risk in England was available through Office for National Statistics (ONS).
T52 5861-6020 Sentence denotes Information at the LSOA level about age and sex was available for 2018, whereas about ethnicity for 2011 (the most recent years available at time of analysis).
T53 6022-6038 Sentence denotes 2.2 Downscaling
T54 6039-6111 Sentence denotes There were 317 LTLAs in England in 2019 (Supplemental Material Fig. S1).
T55 6112-6226 Sentence denotes Such a coarse geographical unit is not expected to capture the strong localised spatial patterns of air-pollution.
T56 6227-6298 Sentence denotes We thus downscaled the LTLA geographical information to the LSOA level.
T57 6299-6413 Sentence denotes LSOAs are high resolution geographical units in England (32,844 units in 2011, see Supplemental Material Fig. S2).
T58 6414-6623 Sentence denotes The median population per LSOA in 2018 was 1617, varying from 591 to 14,696 (min to max) (Supplemental Material Fig. S3), and the median area per LSOA was 0.4km2, varying from 0.0002km2 to 68.4km2(min to max).
T59 6624-6700 Sentence denotes The LTLA boundaries are revised every year, whereas the LSOA ones at census.
T60 6701-6979 Sentence denotes Let l~m denote that the l-th LSOA belongs to the m-th LTLA, nijkm the number of deaths in the m-th LTLA and Pijkl the population in the i-th age group (1<, 1–4, 5–9, …, 85–90, >90), j-the sex (male or female), k-th ethnic group (White, Mixed, Asian, Black, Other) and l-th LSOA.
T61 6980-7148 Sentence denotes We sampled nijkm individual deaths at the l-th LSOA level from a Multinomial distribution with probabilities:πijkl=Pijkl/∑l~mPijkl,and repeated the procedure 100 times.
T62 7150-7163 Sentence denotes 2.3 Exposure
T63 7164-7235 Sentence denotes We considered exposure to NO2 and PM2.5 as indicators of air pollution.
T64 7236-7273 Sentence denotes We selected these pollutants because:
T65 7274-7689 Sentence denotes 1) they reflect different sources of air-pollution (NO2 reflects traffic related air-pollution, whereas PM2.5 is a combination of traffic and non-traffic sources), 2) they were considered in previous studies (Cole et al., 2020, Liang et al., 2020, Travaglio et al., 2020, Wu et al., 2020), and 3) they are responsible for the highest number of years of life lost compared to other pollutants in Europe (Ortiz 2019).
T66 7690-7813 Sentence denotes We retrieved NO2 and PM2.5 concentration in England from the Pollution Climate Mapping (PCM; https://uk-air.defra.gov.uk/).
T67 7814-7928 Sentence denotes The PCM produces annual estimates during 2001–2018 for NO2 and 2002–2018 for PM2.5 at 1x1km resolution for the UK.
T68 7929-8092 Sentence denotes The PCM model is calibrated using monitoring stations across the nation and has high predictive accuracy, R2 = 0.88 for NO2 and R2 = 0.63 for PM2.5 (Brookes 2017).
T69 8093-8243 Sentence denotes We defined long-term exposure to these compounds as the mean of the past 5 years for which data was available at the time of analysis, i.e. 2014–2018.
T70 8244-8413 Sentence denotes An alternative is calculating the median, however the distribution of the air-pollutants using any of these metrics is almost identical, (Supplemental Material Fig. S4).
T71 8414-8723 Sentence denotes We weighted the exposure using a combination of population estimates available from the fourth version of Gridded Population of the World collection at 1x1km grid as of 2020 (Center for International Earth Science Information Network - CIESIN - Columbia University 2018) and from ONS at LSOA level as of 2018.
T72 8724-8828 Sentence denotes Let Xgl be the pollutant and Pgl the population in the intersection of the g-th grid cell and l-th LSOA.
T73 8829-8992 Sentence denotes Assuming the Xg is constant (i.e. Xgl=Xg for all intersections) in the g-th grid cell, we define the population weighted version X¯lof Xgl as:X¯l=∑glP¯glXg∑glP¯gl.
T74 8993-9092 Sentence denotes To calculate P¯gl, we first compute w¯gl=wgl/∑glwgl, where wgl is the area weight per intersection.
T75 9093-9140 Sentence denotes Then calculate the population per intersection:
T76 9141-9153 Sentence denotes Pgl'=Pgw¯gl.
T77 9154-9270 Sentence denotes We then use the Pl (LSOA populations) and obtain P¯gl=vglPl, where vgl is the normalised Pgl', ie vgl= Pgl'/∑glPgl'.
T78 9272-9288 Sentence denotes 2.4 Confounders
T79 9289-9437 Sentence denotes We considered confounders related with meteorology, socio-demographics, disease spread, healthcare provision and health related variables (Table 1).
T80 9438-9651 Sentence denotes As meteorological confounders, we considered temperature and relative humidity and calculated the mean for March-June 2018 as this is the latest year with data available at 1x1km grid retrieved from the MetOffice.
T81 9652-9769 Sentence denotes We weighted temperature and relative humidity using the population weights calculated for the air-pollution exposure.
T82 9770-9899 Sentence denotes As socio-demographical confounders we considered age, sex, ethnicity, deprivation, urbanicity, population density and occupation.
T83 9900-10100 Sentence denotes Information on age (2018), sex (2018), ethnicity (2011), urbanicity (2011) and population density (2018) was available at the LSOA level from ONS (the most recent years available at time of analysis).
T84 10101-10309 Sentence denotes To adjust for deprivation, we used quintiles of the index of multiple deprivation at LSOA level in 2019 (Ministry of Housing, Communities and Local Government), excluding the dimension related to air quality.
T85 10310-10598 Sentence denotes We used estimates of occupational exposures to COVID-19, as calculated by ONS, to adjust for high risk exposure to COVID-19, defined as those with a score higher than 80/100 (corresponding to at least >1 per week exposed to someone infected, Supplemental Material Text S1.1 and Table S1).
T86 10599-10785 Sentence denotes To account for disease progression, we used the number of days since the 1st reported case and the number of positive cases in each LTLA (as of 30th of June 2020, as retrieved from PHE).
T87 10786-10947 Sentence denotes Adjustment for the latter factors is expected to attenuate geographical differences generated due to regional differences about the timing on the pandemic curve.
T88 10948-11091 Sentence denotes For healthcare provision, we used the number of intensive care unit beds per population, in February 2020 per NHS trust, as retrieved from NHS.
T89 11092-11271 Sentence denotes Last, as health-related variables, we considered smoking and obesity prevalence at the GP practice level during 2018–2019, as retrieved from PHE (Supplemental Material Text S1.1).
T90 11272-11314 Sentence denotes Table 1 Data sources used in the analysis.
T91 11315-11377 Sentence denotes Confounders Source Spatial Resolution Temporal Resolution Type
T92 11378-11461 Sentence denotes Temperature MetOfficehttps://www.metoffice.gov.uk/ 1 km2 March-June 2018 continuous
T93 11462-11551 Sentence denotes Relative humidity MetOfficehttps://www.metoffice.gov.uk/ 1 km2 March-June 2018 continuous
T94 11552-11706 Sentence denotes Index of Multiple Deprivation Ministry of Housing, Communities and Local Governmenthttps://www.gov.uk/ Lower layer super output area 2019 rank (quintiles)
T95 11707-11818 Sentence denotes Urbanicity Office for National Statisticshttps://www.ons.gov.uk/ Lower layer super output area 2011 urban/rural
T96 11819-11923 Sentence denotes Days since 1st reported case Public Health England Lower tier local authority Until 30th June continuous
T97 11924-12031 Sentence denotes Number of positive cases Public Health England Lower tier local authority Until 30th June discrete (counts)
T98 12032-12168 Sentence denotes Population density Office for National Statisticshttps://www.ons.gov.uk/ Lower layer super output area 2018 continuous (log transformed)
T99 12169-12326 Sentence denotes Number of intensive care unit beds National Health Servicehttps://www.england.nhs.uk/ National Health Service trust February 2020 continuous (per population)
T100 12327-12456 Sentence denotes Smoking Public Health Englandhttps://fingertips.phe.org.uk/ General practitioner catchment area 2018–2019 continuous (prevalence)
T101 12457-12586 Sentence denotes Obesity Public Health Englandhttps://fingertips.phe.org.uk/ General practitioner catchment area 2018–2019 continuous (prevalence)
T102 12587-12721 Sentence denotes High Risk Occupation Office for National Statisticshttps://www.ons.gov.uk/ Middle layer super output area 2011 continuous (prevalence)
T103 12723-12747 Sentence denotes 2.5 Statistical methods
T104 12748-12891 Sentence denotes We specified Bayesian hierarchical Poisson log-linear models to investigate the association of COVID-19 deaths and NO2 and PM2.5 independently.
T105 12892-13039 Sentence denotes The LSOA specific standardised mortality ratio is known to be an unstable estimator with high variance when the number of expected deaths is small.
T106 13040-13369 Sentence denotes To overcome this problem, we used a well-established hierarchical framework, specifying spatially structured and unstructured random effects, so that the model borrows strength from the other areas across the entire study region, as well as from the neighbouring ones (Best et al., 2005, Wakefield et al., 2000, Wakefield, 2006).
T107 13370-13541 Sentence denotes We model these random effects using a re-parametrisation of the Besag-York-Molliè conditional autoregressive prior distribution (Besag et al., 1991, Simpson et al., 2017).
T108 13542-13574 Sentence denotes We fitted four models including:
T109 13575-13804 Sentence denotes 1) each pollutant (model 1), 2) each pollutant and the spatial autocorrelation term (model 2), 3) each pollutant and all confounders (model 3) and 4) each pollutant, the spatial autocorrelation term and all confounders (model 4).
T110 13805-13970 Sentence denotes All models were adjusted for age, sex and ethnicity using indirect standardisation; we used the English population as the standard population to calculate the rates.
T111 13971-14113 Sentence denotes We do not report results from the joint analysis including both pollutants since they are highly correlated (Supplemental Material Figure S5).
T112 14114-14345 Sentence denotes In order to propagate the uncertainty resulted from the sampling we used for the downscaling, we fitted the models over 100 downscaled samples and then performed Bayesian model averaging to combine the estimates (Gómez-Rubio et al.
T113 14346-14352 Sentence denotes 2020).
T114 14353-14476 Sentence denotes We performed a complete case analysis since for only 1.1% of the cases information about age, sex and ethnicity is missing.
T115 14477-14690 Sentence denotes We report results as posterior median of % increase in mortality risk for every 1 μg/m3 increase in the air-pollutants, 95% credible intervals (CrI) and posterior probability that the estimated effect is positive.
T116 14691-14887 Sentence denotes We also report posterior median of spatial mortality relative risks (exponential of the spatial autocorrelation term) and posterior probabilities that the spatial relative risks are larger than 1.
T117 14888-15005 Sentence denotes The mathematical formulation of the models and prior specifications are given in the Supplemental Material Text S1.2.
T118 15006-15048 Sentence denotes All models were fitted in INLA (Rue et al.
T119 15049-15055 Sentence denotes 2009).
T120 15056-15180 Sentence denotes Covariate data and code for running the analysis are available at https://github.com/gkonstantinoudis/COVID19AirpollutionEn.
T121 15182-15207 Sentence denotes 2.6 Sensitivity analyses
T122 15208-15254 Sentence denotes We performed a series of sensitivity analyses.
T123 15255-15379 Sentence denotes First, we repeated the main analyses using data at the LTLA level with all exposures and confounding weighted by population.
T124 15380-15570 Sentence denotes Second, we examined if there is a differential effect of long-term exposure to air-pollution at the early stages of the epidemic, considering the lockdown (23rd of March 2020) as a landmark.
T125 15571-15822 Sentence denotes Third, we assessed the correlation between the latent field of the full model (model 4) with that of the model excluding or including only covariates indicating disease spread (i.e. number of tested positive cases and days since first reported cases).
T126 15823-15900 Sentence denotes Fourth, we categorised pollutants into quintiles to allow more flexible fits.
T127 15901-15978 Sentence denotes Fifth, we repeated the analysis including the suspected cases to the outcome.
T128 15979-16179 Sentence denotes Sixth, we repeated the analysis changing the definition of long-term exposure to the mean of the past 3 and 10 years for which data was available at the time of analysis, i.e. 2016–2018 and 2009–2018.
T129 16180-16341 Sentence denotes Seventh, we fitted a zero-inflated Poisson model to account for the proportion of zeros in the data (36% in the 100 samples – see Supplemental Material Fig. S6).
T130 16343-16353 Sentence denotes 3 Results
T131 16355-16376 Sentence denotes 3.1 Study population
T132 16377-16499 Sentence denotes We identified 38,573 COVID-19 deaths with a laboratory confirmed test in England between 2nd March and 30th June (Fig. 1).
T133 16500-16627 Sentence denotes The age, sex and ethnicity distribution of the deaths follows patterns reported previously (Supplemental Material Tables S2-3).
T134 16628-16668 Sentence denotes Fig. 1 Flowchart of the COVID-19 deaths.
T135 16670-16683 Sentence denotes 3.2 Exposure
T136 16684-16761 Sentence denotes Fig. 2 shows the population weighted air-pollutants at LSOA level in England.
T137 16762-16906 Sentence denotes We observe that the localised variation of NO2, for instance due to the highways, is adequately captured at the spatial resolution of the LSOAs.
T138 16907-17039 Sentence denotes The mean of NO2 is 16.17 μg/m3 and it varies from 2.99 μg/m3 in highly rural areas to 50.69 μg/m3 in the big urban centres (Fig. 2).
T139 17040-17124 Sentence denotes The mean of PM2.5 is 9.84 μg/m3 with a smaller variation, 5.14–14.22 μg/m3 (Fig. 2).
T140 17125-17170 Sentence denotes Fig. 2 Population weighted exposure per LSOA.
T141 17172-17188 Sentence denotes 3.3 Confounders
T142 17189-17273 Sentence denotes Plots and maps of the confounders can be found in Supplemental Material, Fig. S7-17.
T143 17275-17283 Sentence denotes 3.4 No2
T144 17284-17310 Sentence denotes We observe a 2.6% (95%CrI:
T145 17311-17483 Sentence denotes 2.4%, 2.7%) increase in the COVID-19 mortality risk for every 1 μg/m3 increase in the long-term exposure to NO2, based on model 1 (Fig. 3 & Supplemental Material Table S4).
T146 17484-17677 Sentence denotes There is still evidence of an effect, albeit smaller, once we adjust for spatial autocorrelation or confounders, with increases in the long-term exposure to NO2 of, respectively, 1.3% (95% CrI:
T147 17678-17705 Sentence denotes 0.8%, 1.8%), 1.8% (95% CrI:
T148 17706-17736 Sentence denotes 1.5%, 2.1%) for every 1 μg/m3.
T149 17737-17989 Sentence denotes When we adjust for both autocorrelation and confounders the evidence is less strong, with estimates of 0.5% (95% CrI: −0.2%, 1.2%) for every 1 μg/m3 (Fig. 3 & Supplemental Material Table S4) and posterior probability of a positive effect reaching 0.93.
T150 17990-18053 Sentence denotes The spatial relative risk in England varies from 0.24 (95% CrI:
T151 18054-18083 Sentence denotes 0.08, 0.69) to 2.09 (95% CrI:
T152 18084-18130 Sentence denotes 1.30, 3.11) in model 2 and from 0.30 (95% CrI:
T153 18131-18160 Sentence denotes 0.10, 0.84) to 1.87 (95% CrI:
T154 18161-18270 Sentence denotes 1.18, 2.93) in model 4, implying that the confounders explain very little of the observed variation (Fig. 3).
T155 18271-18388 Sentence denotes The variation is more pronounced in the cities and suburban areas (with posterior probability higher than 1; Fig. 3).
T156 18389-18559 Sentence denotes Fig. 3 Density strips for the posterior of COVID-19 mortality relative risk with 1 μg/m3 increase in NO2 (top panel) and PM2.5 (bottom panel) averaged long-term exposure.
T157 18561-18571 Sentence denotes 3.5 Pm2.5
T158 18572-18599 Sentence denotes We observe a 4.4% (95% CrI:
T159 18600-18765 Sentence denotes 3.7%, 5.1%) increase in the mortality risk for every 1 μg/m3 increase in the long-term exposure to PM2.5, based on model 1 (Fig. 3 & Supplemental Material Table S5).
T160 18766-18891 Sentence denotes When we adjust for spatial autocorrelation the effect increases slightly but the credible intervals are wider, 5.4% (95% CrI:
T161 18892-18972 Sentence denotes 2.5%, 8.4%), whereas it is similar when we adjust for confounding 4.9% (95% CrI:
T162 18973-19027 Sentence denotes 3.7%, 6.2%) (Fig. 3 & Supplemental Material Table S5).
T163 19028-19177 Sentence denotes The effect is weak when we account for confounders and spatial autocorrelation 1.4% (95% CrI: −2.1%, 5.1%) (Fig. 3 & Supplemental Material Table S5).
T164 19178-19275 Sentence denotes The posterior probability of a positive effect is lower than observed for NO2, and equal to 0.78.
T165 19276-19437 Sentence denotes The spatial relative risk follows similar patterns as the one reported in the models for NO2, with the posterior median relative risk varying from 0.24 (95% CrI:
T166 19438-19467 Sentence denotes 0.12, 0.46) to 2.26 (95% CrI:
T167 19468-19514 Sentence denotes 1.32, 3.85) in model 2 and from 0.30 (95% CrI:
T168 19515-19544 Sentence denotes 0.15, 0.57) to 1.90 (95% CrI:
T169 19545-19602 Sentence denotes 1.14, 3.17) in model 4 (Supplemental Material, Fig. S18).
T170 19604-19629 Sentence denotes 3.6 Sensitivity analyses
T171 19630-19909 Sentence denotes When LTLAs are the main geographical unit for analysis, the results are consistent, but higher in magnitude, potentially due to inadequate covariate and spatial autocorrelation adjustment due to the coarse geographical resolution (Supplemental Material Tables S6-7, Fig. S19-20).
T172 19910-20105 Sentence denotes Restricting the study period to March 23, 2020 (N = 698) also results in similar estimates for both pollutants, however the uncertainty is higher (Supplemental Material Tables S8-9, Fig. S21-22).
T173 20106-20354 Sentence denotes The latent field of model 4, with NO2 as the pollutant, is similar to the latent fields of the models with and without the disease progression variables, with a correlation coefficient of 0.94 and 0.93 respectively (Supplemental Material Fig. S23).
T174 20355-20462 Sentence denotes The use of quintiles of the pollutants justifies the linearity assumption (Supplemental Material Fig. S24).
T175 20463-20611 Sentence denotes The results are consistent, but the evidence weaker, when suspected COVID-19 deaths are included (Supplemental Material Tables S10-11, Fig. S25-26).
T176 20612-20745 Sentence denotes The results are also similar when we used a 3 or a 10-year mean of the air-pollutants concentration (Supplemental Material Fig. S27).
T177 20746-20866 Sentence denotes The results are consistent when we fitted a zero-inflated Poisson (Supplemental Material Tables S12-13 and Fig. S28-29).
T178 20868-20890 Sentence denotes 3.7 Post-hoc analysis
T179 20891-21033 Sentence denotes In a post-hoc analysis we investigated if the evidence of an effect of NO2 on COVID-19 mortality can be attributed to pre-existing conditions.
T180 21034-21419 Sentence denotes We selected hypertension, chronic obstructive pulmonary disease (COPD) and diabetes, because of 1) indications of previous literature that they increase the COVID-19 mortality risk (Williamson et al., 2020, Yang et al., 2020), 2) previous literature that suggest an effect with long-term exposure NO2 (Balti et al., 2014, Cai et al., 2016, Zhang et al., 2018) and 3) data availability.
T181 21420-21612 Sentence denotes We retrieved prevalence data for these pre-existing conditions from PHE available at the GP practice level during 2018–2019 (https://fingertips.phe.org.uk/), Supplemental Material Fig. S30-32.
T182 21613-21827 Sentence denotes The effect of NO2 remains similar, 0.6% (95% CrI: −0.1%, 1.3%) with the posterior probability being 0.94 whereas the spatial relative risk highlights the same geographical locations, Supplemental Material Fig. S33.
T183 21829-21842 Sentence denotes 4 Discussion
T184 21844-21862 Sentence denotes 4.1 Main findings
T185 21863-22026 Sentence denotes This is the first nationwide study in England investigating the effect of long-term exposure to NO2 and PM2.5 during 2014–2018 on COVID-19 mortality at LSOA level.
T186 22027-22171 Sentence denotes The unadjusted models indicate that for every 1 μg/m3 increase in the long-term exposure to NO2 and PM2.5 the COVID-19 mortality risk increases.
T187 22172-22363 Sentence denotes After considering the effect of confounding and spatial autocorrelation there is still some evidence of an effect, albeit is less strong, for NO2, while for PM2.5 there is larger uncertainty.
T188 22364-22513 Sentence denotes The spatial relative risk has strong spatial patterns, identical for the different pollutants, potentially highlighting the effect of disease spread.
T189 22515-22568 Sentence denotes 4.2 Comparison with previous studies focusing on NO2
T190 22569-22714 Sentence denotes Our study is comparable with previous studies in the US, England and the Netherlands assessing the long-term effect of NO2 in COVID-19 mortality.
T191 22715-22801 Sentence denotes The study in the US focused on deaths reported by April 29, 2020, using 3122 counties.
T192 22802-22987 Sentence denotes For the exposure, they calculated the mean of daily concentrations during 2010–2016 as modelled by a previously described ensemble machine learning model (R2 = 0.79) (Di et al., 2019a).
T193 22988-23034 Sentence denotes They reported a 7.1% (95% Confidence Interval:
T194 23035-23191 Sentence denotes 1.2%, 13.4%) increase in mortality per 4.5 ppb (1 ppb = 1.25 μg/m3) increase in NO2 after adjusting for confounders and spatial autocorrelation(Liang et al.
T195 23192-23247 Sentence denotes 2020)(that is approximately 1.3% increase per 1 μg/m3).
T196 23248-23403 Sentence denotes A study in England, with partly overlapping data as in our analysis, also reported a significant association between NO2 and COVID-19 mortality (p < 0.05).
T197 23404-23576 Sentence denotes For the analysis they focused on COVID-19 deaths reported in England up to April 10, 2020, used 317 LTLAs, and did not account for spatial autocorrelation (Travaglio et al.
T198 23577-23583 Sentence denotes 2020).
T199 23584-23731 Sentence denotes The study in the Netherlands using 335 municipalities, mean exposure during 2015–2019 and COVID-19 deaths up to June 5, 2020 reported 0.35 (95% CI:
T200 23732-23934 Sentence denotes 0.04, 0.66) additional COVID-19 deaths for every 1 μg/m3 increase in NO2 after adjusting for confounders and certain spatial controls, such as transmission beyond the Dutch national borders (Cole et al.
T201 23935-23941 Sentence denotes 2020).
T202 23942-24109 Sentence denotes Since the mean number of deaths in their sample is 16.86, the above estimate translates to a 2.0% increase in the COVID-19 mortality for every 1 μg/m3 increase in NO2.
T203 24110-24377 Sentence denotes An ONS report in England using 175 sampling units, 10-year averaged NO2 exposure (PCM) and COVID-19 deaths up to June 12, 2020 found a 0.6% (95% CI: −0.1%, 2.2%) increase in the COVID-19 mortality for every 1 μg/m3 increase in averaged NO2 exposure (Statistics 2020).
T204 24379-24434 Sentence denotes 4.3 Comparison with previous studies focusing on PM2.5
T205 24435-24543 Sentence denotes Our study is comparable with previous studies assessing the long-term effect of PM2.5 on COVID-19 mortality.
T206 24544-24647 Sentence denotes The aforementioned study in the US also assessed the effect of PM2.5 on COVID-19 mortality(Liang et al.
T207 24648-24654 Sentence denotes 2020).
T208 24655-24756 Sentence denotes Their exposure model was previously validated having an R2 = 0.89 for the annual estimates (Di et al.
T209 24757-24764 Sentence denotes 2019b).
T210 24765-24989 Sentence denotes The evidence for PM2.5 was weak, namely 10.8% (95% CI:-1.1%, 24.1%) per 3.4 μg/m3 increase in PM2.5 concentration (that is approximately 3.2% increase per 1 μg/m3) after adjusting for confounding and spatial autocorrelation.
T211 24990-25164 Sentence denotes The ONS report in England found a 1% (95% CI: −3%, 6%) increase in the COVID-19 mortality for every 1 μg/m3 increase in the 10-year averaged PM2.5 exposure (Statistics 2020).
T212 25165-25319 Sentence denotes Our study comes in contrast with another study in the US that used deaths reported until April 22nd, 2020 and counties as the geographical unit (Wu et al.
T213 25320-25326 Sentence denotes 2020).
T214 25327-25438 Sentence denotes For the exposure, they used previously validated monthly PM2.5 concentrations (R2 = 0.70) (Van Donkelaar et al.
T215 25439-25484 Sentence denotes 2019) and averaged them during 2000 and 2016.
T216 25485-25580 Sentence denotes After adjusting for confounding but not for spatial autocorrelation, they found an 11% (95% CI:
T217 25581-25686 Sentence denotes 6%, 17%) increase in the COVID-19 death rate for an increase of 1 μg/m3 in PM2.5 concentration (Wu et al.
T218 25687-25693 Sentence denotes 2020).
T219 25694-25787 Sentence denotes Our study comes also in contrast with the study in the Netherlands that reported 2.3 (95% CI:
T220 25788-25910 Sentence denotes 1.3, 3.0) additional COVID-19 deaths for an increase of 1 μg/m3 in the averaged long-term PM2.5 concentration (Cole et al.
T221 25911-25917 Sentence denotes 2020).
T222 25918-26096 Sentence denotes Having a mean number of deaths equal to 16.86, the above estimate translates to a 13.6% increase in the COVID-19 mortality rate for an increase of 1 μg/m3 in PM2.5 concentration.
T223 26098-26128 Sentence denotes 4.4 Strengths and limitations
T224 26129-26288 Sentence denotes Our study is the first study to examine the association between long-term exposure to NO2 and PM2.5 and COVID-19 mortality at very high geographical precision.
T225 26289-26587 Sentence denotes The spatial unit of our analysis is LSOAs, for which there are 32,844 in England (~130 000 km2), whereas previous studies have used 317 LTLAs or 175 sampling units in England, counties in the US (3 122 in an area ~9.8 million km2) and municipalities in the Netherlands (334 in an area ~41 500 km2).
T226 26588-26757 Sentence denotes Such high-resolution allows capturing the localised geographical patterns of the pollutants but also ensures adequate confounding and spatial autocorrelation adjustment.
T227 26758-27006 Sentence denotes Our study also covers, so far, the largest temporal window of the epidemic (capturing COVID-19 deaths attributable to the first wave, Supplemental Material Fig. S34), while most previous studies focused on the early to mid-stages of the first wave.
T228 27007-27059 Sentence denotes This ensures better generalisability of the results.
T229 27060-27311 Sentence denotes In addition, physical distancing and other public health interventions were introduced nationwide in England during the first epidemic, mitigating any distortion between air-pollution and COVID-19 mortality due to potential regional level differences.
T230 27312-27453 Sentence denotes Our results are also consistent in a sensitivity analysis focusing on the pre-lockdown period, in the absence of public health interventions.
T231 27454-27624 Sentence denotes Based on the scientific literature, we adjusted for several variables which would act as the confounders of the relationship between air pollution and COVID-19 mortality.
T232 27625-27815 Sentence denotes Nevertheless, since the aetiology and the factors contributing to COVID-19 mortality are not fully understood yet, we included a spatial random effect to capture unknown spatial confounding.
T233 27816-27891 Sentence denotes The spatial random effect was found to be a crucial component in the model.
T234 27892-28082 Sentence denotes Not accounting for spatial autocorrelation, when spatial autocorrelation is present, is expected to give rise to narrower credible intervals and false positive effects (Lee and Sarran 2015).
T235 28083-28119 Sentence denotes Our study has also some limitations.
T236 28120-28198 Sentence denotes The downscaling procedure will likely inflate the reported credible intervals.
T237 28199-28313 Sentence denotes However, this naturally reflects the uncertainty of the place of residence resulted from the downscaling approach.
T238 28314-28471 Sentence denotes Although we consider small areas, the study is still an ecological one and thus the reported effects do not reflect individual associations (Wakefield 2008).
T239 28472-28603 Sentence denotes Case fatality might have been a more appropriate metric for the analysis, since disease spread is accounted for in the denominator.
T240 28604-28818 Sentence denotes Nevertheless, given the asymptomatic infections and the fact that number of reported infections is not a random sample of the general population, the number of COVID-19 cases per LTLA is not reliable at this stage.
T241 28819-29028 Sentence denotes For the same reason, using the number of reported cases to adjust for disease progression and clustering of cases and deaths might not adequately capture disease progression and clustering of cases and deaths.
T242 29029-29111 Sentence denotes However, part of this clustering was captured in the spatial autocorrelation term.
T243 29112-29248 Sentence denotes We did not account for population mobility during 2014–2018 and assumed constant residence and thus levels of exposure to air-pollution.
T244 29249-29510 Sentence denotes While this is a limitation, we believe that it would have a minimal impact on the results given that 1) the exposure period is relatively short and 2) almost 93% of the deaths in our dataset occurred in people 60 years or older (Supplemental Material Table S2).
T245 29511-29615 Sentence denotes This comprises a population less likely to have moved during the past 5 years (Burgess and Quinio 2020).
T246 29616-29685 Sentence denotes We also could not account for non-residential air-pollution exposure.
T247 29686-29950 Sentence denotes Spatiotemporal variation in the strains of COVID-19 can introduce bias (Villeneuve and Goldberg 2020), however at the time of publication there was no evidence supporting that strain types can confound the relationship between COVID-19 mortality and air-pollution.
T248 29952-29971 Sentence denotes 4.5 Interpretation
T249 29972-30176 Sentence denotes Compared to the previous studies, our results are the smallest in magnitude, likely because of the high geographical precision that allows more accurate confounding and spatial autocorrelation adjustment.
T250 30177-30298 Sentence denotes In addition, we report weak evidence of an effect, which could also be due to lack of power and individual exposure data.
T251 30299-30483 Sentence denotes Nevertheless, as for NO2 we find a high posterior probability of an effect on mortality, we argue that a potential explanation might be the mediation effect of pre-existing conditions.
T252 30484-30744 Sentence denotes While in our analysis the inclusion of area-level prevalence of hypertension, diabetes and COPD did not change the results, the ecological nature of the pre-existing conditions data does not allow us to account for the mediation effect at the individual level.
T253 30745-30976 Sentence denotes Our study focuses on the mortality after contracting SARS-CoV-2, however we cannot rule out individual susceptibility to becoming infected as an explanation to the uncertainty in the effect estimates (Villeneuve and Goldberg 2020).
T254 30977-31096 Sentence denotes Such susceptibility can reflect immunosuppression, leading to later increases in inflammation (Edoardo Conticini et al.
T255 31097-31251 Sentence denotes 2020) and thus worse prognosis, or even disease spread, as recent studies have suggested that PM2.5 can proliferate COVID-19 transmission (Bianconi et al.
T256 31252-31258 Sentence denotes 2020).
T257 31259-31312 Sentence denotes Our analysis captured strong spatial autocorrelation.
T258 31313-31556 Sentence denotes The observed pattern could reflect residual variation from a potential inadequate covariate adjustment (including disease spread), spatial variation of pre-existing conditions, other unknown spatial confounders or a combination from all above.
T259 31557-31805 Sentence denotes In a sensitivity analysis, we observed that the factors associated with disease transmission left the latent field unchanged (Supplemental Material Fig. S21), as did the inclusion of hypertension, diabetes and COPD (Supplemental Material Fig. S33).
T260 31806-31976 Sentence denotes When we restricted the analysis to the pre-lockdown period, the latent field for both pollutants captured London and Birmingham, i.e. the cities with the first outbreaks.
T261 31977-32199 Sentence denotes Considering the above, and the fact that COVID-19 is an infectious disease, we believe that large variation of Fig. 4 is likely due to disease spread, which is not adequately captured in the disease progression covariates.
T262 32200-32462 Sentence denotes Fig. 4 Median posterior spatial relative risk (exponential of the spatial autocorrelation term) and posterior probability that the spatial relative risk is larger than 1 for the models with NO2 and a spatial autocorrelation term and the fully adjusted NO2 model.
T263 32464-32477 Sentence denotes 5 Conclusion
T264 32478-32660 Sentence denotes Overall, this study provides some evidence of an association between averaged exposure during 2014–2018 to NO2 and COVID-19 mortality, while the role of PM2.5 remains more uncertain.
T265 32662-32702 Sentence denotes CRediT authorship contribution statement
T266 32703-32730 Sentence denotes Garyfallos Konstantinoudis:
T267 32731-32864 Sentence denotes Conceptualization, Methodology, Software, Formal analysis, Writing - original draft, Writing - review & editing, Funding acquisition.
T268 32865-32882 Sentence denotes Tullia Padellini:
T269 32883-32950 Sentence denotes Methodology, Software, Formal analysis, Writing - review & editing.
T270 32951-32965 Sentence denotes James Bennett:
T271 32966-33006 Sentence denotes Writing - review & editing, Methodology.
T272 33007-33021 Sentence denotes Bethan Davies:
T273 33022-33049 Sentence denotes Writing - review & editing.
T274 33050-33063 Sentence denotes Majid Ezzati:
T275 33064-33155 Sentence denotes Conceptualization, Project administration, Writing - review & editing, Funding acquisition.
T276 33156-33173 Sentence denotes Marta Blangiardo:
T277 33174-33254 Sentence denotes Conceptualization, Methodology, Writing - review & editing, Funding acquisition.
T278 33256-33289 Sentence denotes Declaration of Competing Interest
T279 33290-33460 Sentence denotes The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
T280 33462-33496 Sentence denotes Appendix A Supplementary material
T281 33497-33574 Sentence denotes The following are the Supplementary data to this article:Supplementary data 1
T282 33576-33592 Sentence denotes Acknowledgements
T283 33593-33664 Sentence denotes GK is supported by an MRC Skills Development Fellowship [MR/T025352/1].
T284 33665-33757 Sentence denotes MB and TP are supported by a National Institutes of Health, grant number [R01HD092580-01A1].
T285 33758-34016 Sentence denotes JB and ME are supported by the Pathways to Equitable Healthy Cities grant from the Wellcome Trust [209376/Z/17/Z] and by a grant from the US Environmental Protection Agency (EPA), as part of the Centre for Clean Air Climate Solution (assistance agreement no.
T286 34017-34026 Sentence denotes R835873).
T287 34027-34082 Sentence denotes This article has not been formally reviewed by the EPA.
T288 34083-34200 Sentence denotes The views expressed in this document are solely those of the authors and do not necessarily reflect those of the EPA.
T289 34201-34292 Sentence denotes The EPA does not endorse any products or commercial services mentioned in this publication.
T290 34293-34435 Sentence denotes Infrastructure support for this research was provided by the National Institute for Health Research Imperial Biomedical Research Centre (BRC).
T291 34436-34595 Sentence denotes This work was part supported by the MRC Centre for Environment and Health, which is currently funded by the Medical Research Council (MR/S019669/1, 2019-2024).
T292 34596-34711 Sentence denotes Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.envint.2020.106316.