Climate Change and Temperature-related Mortality:lmplications for Health-related Climate Policy＊

TON h u,Jor lsen,an atric .Kinney

I idel ecognized tha limat hange i ne o he greatest challengeshumanit acesin the 21st century[1,2]. A he leadin uthority fo he assessment of climat hange, the Intergovernmenta ane n Climat hange(IPCC),i ts specia eport o lobal warmin f 1.5 °C,has conclude ha uman activities have caused approximately 1.0 ° loba armin bov reindustria evels,wit likel ncreas f 1.5°C between 2030 and 2052[3].Climate-relate isks for natural and huma ystem re projecte o increase with global warmin f 1.5°C, and increase furthe ith 2°Cor above. Thes isks depen n the magnitud n at arming,geographic location, level f development an ulnerability,and o h hoice n mplementatio f adaptatio n itigation options.

Levels o reenhouse gase e.g., CO2, CH4, and N2O ontinue o increas n 2020 despit h etbac rom th andemic o oronaviru isease-2019(COVID-19)[4]. The temporary reductio missions in 2020 related t easure aken i espons o COVID-19 is likely to lea nly a slight decreas n the annual growth rate of CO2concentration i he atmosphere,whic ill b ractically indistinguishabl ro h atura nterannual variability driven largel he terrestrial biosphere[4,5].In fact,climate chang ccelerating an asbecom h lobal healt mergency[6].

Climate chang as alread nfluenced an ill increasingl ffec uma ealt hroug irect,indirect, and diffusin athways[7,8]. The direct impacts include increasing mortality, morbidit nd injurie ause y extrem eathe vents(e.g.,heatwaves,storms, bushfires,droughts and flooding); the indirect effect ompris orsenin ir pollution, increase ector-, food-,and waterborne diseases,and illnessesfrom reduce utrition,du hanges in ecosystem hic ecreas ir quality, result i h prea f diseas ectors and reduc rop production; and the diffusing/delayed impacts refe nfluences on socioeconomic infrastructure,natura esource n nvironment,which lea o worsenin onditions in food and wate ecurity, migration,conflict nd war s well a rofoun enta ealth harms from climate disasters(Figure 1). Amon hes ffects,temperature-related health impact ay hav een mos xtensively studie hes ffects ar irect an mmediate[1,7]. However,health risks attributabl emperatur ary wit pace and tim ue t ifferent socio-environmental characteristics, population structure,publi ealth system nd adaptation capabilities.

Severa ssuesneed t e addresse n evaluating the evidenc h irec mpac f exposur o extrem emperature h lobal scale. First,what is the state-of-the-scienc nowledge on temperature-related health impacts?Second, to wha xtent migh he decreases i old-related mortalit ffse he increasesin heat-relate eaths?Third,wha he relative contributio f different temperature exposure i.e.,non-optimal temperature n emperatur ariability o mortality?Fourth, how muc xces ortality is attributabl eatwave n ol pells?Fifth,what are th ompoun ealt ffects of temperatur n i ollution? Finally, is the evidence availabl ate sufficient for developing coordinate n vidence-informed climate policies?

Heat-related Mortality

Heat is amon he deadliest weather-related phenomen lobally, an eat-related mortality may significantl ncreas limat hange proceeds. The 2020 repor heLancetCountdow n health and climat hang eveal ha ulnerabl opulations wer xpose o an additional 475 millio eatwave eventsgloball n 2019, whic as,i urn, reflected in excess mortalit n orbidity[1].Durin h ast 20 years, there ha ee 53.7% increase in heat-related mortalit n peopl lde han 65 years, reachin tota f 296,000 death n 2018. Another international stud xamine he burden o eat-relate ortality attributabl o human-induced climat hang ver recent decades[9]. The oun ha ecent climate chang s already responsibl or a considerable proportion o eat-related excess mortality burden.Thes esult uppor he nee o rgen lobal implementation of stronge itigation strategies towardszero-carbo arget educ urther heating an elate ealt mpacts.

Aumber of studie av xamined temperature-relate ortality,an ave evaluated the potential direct healt mpacts o xposur o hea sing scenario-base isk assessment approaches[10-13]. Fo xample, tud rojected future temperature-relate ortality i e ork City[10]. They modelle eat adaptatio y incorporating a nalysi h bserved population response t ea n New York City ove h ours f eigh ecade nd estimate ha he median number o rojected annua eat-relate eaths varie reatly by climate, adaptation and population chang cenario,rangin rom 167 to 3,331 i he 2080 ompared with 638 heat-relate eaths annuall etween 2000 and 2006.AEuropean study foun ha igh summe mbient temperature ave an important impac opulation health, an his impact wil ignificantly increase i he future,accordingt h rojecte ncrease o ean ambient temperature nd frequency, intensit n uration o eathe vents[11]. A stud ssessed th uture impact limate chang eat-relate ortality in 12 UScities[12].Climat hang a rojected to resul n a ncreas n heat-related fatalities over time throughou he 21s entur ll o h ities include he study.Nearly 200,000 heat-related deaths ar rojecte o occu he 12 citie he end of th entury due t limat arming.Another stud stimate ha o h eriod of 2041–2060,relative to 1970–2000,excess heat-relate ortality in the 51 larges itiesi hin e 37,800 (95%CI:31,300–43,500), 31,700(95%CI:26,200–36,600 nd 25,800(95%CI: 21,300–29,800 eath er year unde igh, mediu nd low emission scenario,respectively[13]. ecen tudy projected th xcess cause-,age-,region-,and education-specific mortalit ttributabl utur ig emperature n 161 Chinese districts/counties[14]. The oun hat heat-related exces ortalit rojected t ncrease from 1.9% (95%CI:0.2%–3.3% he 2010 o 2.4%(0.4%–4.1%)i he 2030 nd 5.5%(0.5%–9.9% n the 2090 s under a hig mission scenario.The projected slope r teeper i outhern, eastern,central an orther hina. People with cardiorespirator iseases, females, th lderl nd those wit ow educational attainmen oul e mor ffected.Population ageing amplifie uture heat-relate xcess deaths 2.3-to 5.8-fold under different share ocioeconomic pathways,particularl o h ortheas egion.

Curren reenhous a itigation ambition is consisten ith 3 ° loba ea armin bove preindustrial levels. There i clear need to strengthen mitigation ambition t tabiliz he climate a h aris Agreement goal of warmin f less than 2 °C. Recently, tud xamine he difference ity-leve eat-relate ortality betwee he 3° rajector n armin f 2 ° nd 1.5 °C,using the dat rom 15 U ities[15]. They foun ha atchetin itigation ambition to achieve the 2°Cthreshold coul voi etween 70 and 1980 annua eat-relate eath e it uring extreme events.Achievin he 1.5° hreshold could avoi etween 110 and 2,720 annua eat-related deaths.These result rovid ompelling evidence for the heat-related healt enefitso imitin lobal warmingto 1.5 ° n the United States.

An international collaborative tea erformed a multi-regio nalysis in 451 location n 23 countries wit ifferent climate zones,an valuate hanges i emperature-relate ortalit nde cenarios consistent wit he Pari greemen arget 1.5 and 2°C nd mor xtreme temperatur ncreases(3 and 4°C)[16]. The result ugges ha imiting warming below 2° ould prevent larg ncrease n temperature-relate ortality in mos egions worldwide. The comparison between 1.5 and 2 ° s mor omple nd characterize y higher uncertainty,wit eographical difference hat indicat otential benefit imite reas locate n warme limates, where direct climat hange impactswil or iscernible.

In general,al hese studies hav eporte significan ncrease i xcessmortality associated with heat. Heat-related mortalit s likel ose a serious threa opulatio ealt ithin th ontex f global warming, bu her ill b ubstantia ealth benefit imitin loba armin o 1.5 °C.

Net Balance between Heat- and Cold-related Mortality

Som tudies have examined an ompared variation oth cold- an eat-relate eaths, and have consistently projecte decreas oldrelated health risks, wit concomitan ncreas n death irectly associate it xposur o heat[17-23].For example, a stud oun hat temperatur ncrease o ustrali r xpecte o resul n ignifican ncreas eat-related mortality an decreas old-related mortality as climat hange continues,bu h alance o hese difference aried by city. In particular,th easonal pattern n temperature-related death il hift[17].A stud xamine he total mortalit urden attributable to non-optimu mbient temperature,an h elativ ontributions from heat and col n 384 location cross 13 countrie nd regions[18].They foun ha os h emperature-related mortality burden wa ttributabl he contribution o old.Th ffec ays o xtreme temperature was substantially less tha hat attributabl ilde ut non-optimum weather.Anothe tud nvestigate h ortalit espons o cold an ea n 18 French cities between 2000 and 2010,and observed that 3.9% (95%CI:3.2%–4.6%)o he tota ortalit a stimate e attributed to cold, and 1.2%(1.1%–1.3% eat[19]. However,ther as apid increase in mortality ris ery high temperature above percentile 99).

A projectio tudy estimated curren nd future seasona atterns in temperature-relate ortality in Manhattan, New York,using 16 downscaled global climate model n wo emissio cenarios[20]. All 32 projections yielde ar eason increase n old seaso ecreases in temperature-related mortality,wit tead ncreases in net annual temperaturerelate eath l ases. Thes esult uggest that, ove rang f model nd scenario f future greenhous as emissions,increases i eat-related mortalit av een shown to outweig eductions i old-relate ortality, wit hifting seasonal patterns. A European stud ocuse ong-term projection n cclimatizatio cenarios of temperature-relate ortality, and observe change in the seasonalit f mortality, with maximu onthly incidenc rogressivel hifting from winter to summer[21].The results sho hat the ris eat-relate ortality wil tar o overtake th eductio eath ro old durin h econd hal f th entury, amountin o a stimated average dro n human lifespan o p 3–4 month n 2070–2100. There i lso som iteratur uestioning whether cold-seaso eaths wil ctuall o dow n armin orld[22,23]. Evidenc uppor f this contrarian vie nclude h resenc ubstantial cold-relate ortalit ot arm an old citie n th urren limate[24].

Nevertheless, huma ifespa igh ndeed increas f ubstantia egree o daptatio arm temperature akesplace.Othe esearchersprojected futur emperature-attributabl rematur eaths in 209 U.S. cities, an ound increasin utur remature deaths bu esult arie ocation, wit ome location howin e eductions i remature temperature-attributable deaths with climate change[25]. An internationa tudy quantifie h otal mortality burde ttributabl o non-optimum ambien emperature, an rojecte xces ortality for cold an eat an hei e hange in 1990–2099 under climat hang cenario n 23 countrie cross nine region h orld[26]. Results indicate, on average, et increase i emperature-relate xcess mortalit nder high-emissio cenarios, with disproportionatel arge impacts i he warme nd poore egion h orld.

A oted, most literature suggest ha her ill be an increase in excessdeath uringsumme n reductio n winter.I eneral,studie roject a net increase of excess mortalit ith climat hange,particularl nde igh-emissio cenarios.However,estimate he net impac xces ortality depen n locatio nd scenario.

Excess Mortality Associated with Temperature Variability and Extreme Temperature Events

Climat hang o nl ncrease verage global surfac emperature, but als nhancesit ariability and lead ncrease requenc nd intensit f temperatur xtremes[3]. Therefore, it i ssentia o evaluate exces ortality associate ith temperature variabilit nd extrem eathe vents. umber of studie ave examine he mortality impact emperatur ariability. Fo xample, a stud ssesse he associatio emperature chang etwee eighborin ays(TCN), an indicato o reflec udde emperature variation,with exces ortality[27]. They used daily mortality an eteorological dat rom 106 communitie f United State nd found tha he relativ is o ag 0–21 day as 0.63(95%CI:0.59–0.68 or extremel egative TC 1st percentile nd 1.46(1.39–1.54)for extremely positiv CN(99th percentile on-accidenta ortality. The lso found prominen ffect xtrem CN n mortality fo ardiovascular,respiratory,pneumonia,an OPDdiseases.People ≥75 year nd thos ith respirator isease, especiall neumonia-deaths,were identifie particularly vulnerable population t CN. Th CN-mortalit ssociation wasmodifie easo n egion.

A ustralia tud xamine ortalit isk and associated mortality burde rom bot on-optimal temperatures an emperature variability[28]. They collecte ail ime-serie at ll-cause deaths an eathe ariable o h ive mos opulous Australian citie Sydney, Melbourne,Brisbane,Adelaide,and Perth), from 2000 to 2009. Temperature variability wa alculate ro h tandar eviation of hourl emperaturesbetwee wo adjacent days.It wa bserve hat heat,cold an emperature variabilit ogethe esulte n 42,414 death uring th tud eriod,accountin or about 6.0% of all deaths.61.4%, 10.6%, and 28.0% of deaths were attributabl o cold, heat an emperatur ariability,respectively. Thes inding ugges hat,in additio o hea nd cold, temperatur ariabilit eed e considered i ssessing an rojectin he health impact limatechange.

Recently, an international stud rojected the effect o iurnal temperature range(DTR)on future mortalit cross 445 communitie n 20 countries and regions[29]. The ound that,under the unmitigated climat hang cenario(RCP 8.5),the futur verag T s projecte o increase in most countries an egions(by ?0.4 to 1.6°C), particularly i he USA, south-central Europe, Mexico, and South Africa. Th xcess death urrentl ttributable to DT er stimated t e 0.2%–7.4%. Furthermore,th TR-related mortalit isk increase h ongter verage temperature increased; in the linear mixed mode ith th ssumption o n interactive effect wit ong-term averag emperature, they estimated 0.05% additiona T ortalit isk per 1 ° ncrease i verag emperature. Based o he interactio ith long-term averag emperature, the DTR-relate xcess death re projected t ncrease i l ountries o egions by 1.4%–10.3%in 2090–2099. This study suggest ha lobally, DTRrelated exces ortalit igh ncrease under climat hange,and this increasin atter s likely t ar etwee ountrie n egions.

A limat hang roceeds,man ountries experience more an or ntensifie eatwaves,and increasing research attention ha ee aid to the healt mpact f heatwaves. ystematic revie ndicate ha eatwaves significantly increase mortality, bu h agnitud f th ffects varies under differen eatwav efinitions[30].Heatwav ntensity seems t lay elativel ore important rol ha uration i etermining heatwave-related deaths. ew studie ave projecte h mpact f heatwave he future[31,32]. For example,a stud rojecte he heatwave trend n 82 U ommunitie n 2061–2080 unde w C Representative Concentration Pathway cenarios of climat hange(RCP4.5,RCP8.5),tw S Share ocioeconomic Pathway)scenario opulation change(SSP3,SSP5), and thre cenarios o ommunity adaptation t ea none, lagged,on-pace)[31]. They defined a high-mortalit eatwav ncrease ortality ris y ≥ 20%.More high-mortalit eatwave re expected unde h CP8.5 versu CP4.5 scenario,and on-pac daptatio a ignificantl educ he impact o igh-mortalit eatwaves. A nternational stud rojecte xces ortalit n relation to heatwave he future unde ifferent scenario f greenhouse ga missions, wit wo assumption or adaptatio n daptation an ypothetical adaptation nd thre cenarios o opulation chang high,median,an o ariant n 412 communities within 20 countries/regions[32].The mortality change n 2031–2080 versus 1971–2020 rang ro pproximately 2,000% in Colombia to 150% in Moldov nder th ighest emissio nd high-varian opulation scenario,withou ny adaptation. However,th ncreas oul uch smalle he scenario o ower emission nd hypothetical adaptatio dopted.

Compare it eatwaves, les esearch has focuse h mpact of multi-da old spells.A time-series stud xamine h elationship betwee ol pell nd mortality in 31 capita ities across China[33]. They foun hat th elativ isks(RR)o on-accidental mortality for col pell ere 1.03 (95%CI: 1.01–1.05), 1.27(1.19–1.35)and 1.55(1.40–1.70)a ag 0, lag 0–14 and lag 0–27 days,respectively. The greatest effec stimate f cold spell ere found amon otal respirator iseases and COPD, wit R of 1.88(1.65–2.11 nd 1.88(1.58–2.19),respectively. The elderly,less-educated individual n esident outher hin ere mor ulnerabl ol pells. Anothe tudy estimate nd projected mortalit ttributabl o col aves during 1960–2050 in 209 US cities[34].The eported that th ol aves were associated with an increase u mal isk o ortality. The association aried substantiall cros limate regions. Th isk increase it he duratio nd intensit ol aves. Th rojecte ortality relate o cold waves woul ecrease from 1960 to 2050. Suc decrease, however,i mal nd may no ble t ffse h otentia ncrease in heatrelated deaths if the adaptatio o heat i ot adequate. A systematic revie ndicate hat cold spells were associate it ncreased mortalit ates bu her assubstantial heterogeneit n estimated effectsamon he studies[35].

Emergin videnc uggest ha h ncreas n th requency,intensity,an uratio xtreme temperatur vents du limate change wil e more pronounce ha xpected[36]. Thi ill affect huma ell-bein n ortality mor ha hat estimate oday"s modellin cenarios. Thus, it is importan roperl stimat he healt mpact f extrem emperature event n evelo mproved weather/health earl arnin ystem WHEWSs). The applicatio f WHEWSs i art o rimar reventive measures. If WHEWS re designed an mplemented properly, they will likel educ h ublic health risk a ocal,national, and internationallevels.

O he othe and,huma opulation ave bee dapting t eat, cold and the temperature variability. Severa tudies examined change n temperatur elate ortalit ver the las entury,an ost o hem reported eclinin rend[37-39].For instance, Carso l. analysed weekl ime serie at n London, an oun decreasing vulnerability t emperatur elated mortality during the 20thcentury[37].Another stud nvestigated whethe ensitivit o episode f extrem eat and cold ha hange tockholm, Sweden, for the period 1901–2009,and showed tha h elativ isk o ortality durin xtrem emperatur vents appears t ave fallen[38]. Th ajo easons for such ren nclud mproved sanitation an ealthcare,epidemiologica ransitio e.g., hea elated deaths bein ause y infectiou isease suc s diarrheal disease or septicaemia)in th irst part of th entur ecome les mportant), an h se of ai onditioning. However,ho eopl dap o climat hange varie ith spac nd time.For example, ecent study analyse he historical relationship between weathe luctuations and mortality evolutio n Chin o he 1964–2008 period[40].Th esults sho hat an additional 1 °C ris verag emperatur given yea ncreased th nnua ortality rat y an averag f 3.2%. In contrast,col ay i ot have thes ffects. In addition, by introducin nteractions between temperature n otential modifier uc ccess t octors, hospita eds, ai onditioning and refrigerators, the etermine ha nly residential ai onditionin laye n importan ol n mitigating the temperature-mortalit elationship.However,the penetration o i onditionin n China is still relatively low. Thi a xplain why there i ittl videnc veral daptation t igh temperature n China.

Interactive Effects of Temperature and Air Pollution on Mortality

A number o tudies hav ssesse nteractive effects of temperatur n ir pollution on mortality[41-44]. Most o hese studie av ocused o singl ity bu few studies have use ulticity data. Fo xample, tudy examined whether the mortalit ffects o ir pollution ar odified by temperature an ic ers igh uropean cities[45]. Th esult how tha ig emperature could modif h ffect ir pollutio n daily mortality and high air pollutio igh nhance the ai emperatur ffects. Anothe tud ssessed whethe emperatur nd community-leve ariables modify th ssociation betwee ong-ter M2.5exposures and mortalit cross 207 U ities[46]. They found ighe ssociatio etwee ong-ter M2.5exposure and mortalit n warmer cities. tratified time-serie nalysis wa onducted t xamine whether the effect f particulat atte ess than 10μm i erodynamic diamete PM10ortality wa odifie y temperatur cross eigh hinese cities[47].Th esults indicat ha xtrem igh temperature increase h ssociations o M10with dail ortality.Al hes inding ay have implication o he health impact ssociate ith bot i ollutio n loba limat hange.

Implications for Health-related Climate Policy

Table 1 show he agend o ormulating future health-related climate policy. The pattern f temperature-related mortalit ar ith space and tim n t isessentia ddres hi ssu ithi he loca ontext, whil lso considerin elevan rend t the regional,national an loba evels. Additionally,it implie hat socio-environmental condition t different scales shoul e take nt ccount i he developmen daptation strategie ope wit he increasin mpact limatechange.

Even thoug or eath re attributable t old than hea an artso he worl he moment,thi attern will change a limate change proceeds[1,26,32].Accumulating evidence indicates that, i igh-emissio cenarios, most region re projecte xperience a stee is n heat-related mortalit ha il ot b alanced b reduction in cold-relate eaths. A result, i ill result i substantia e ncreas n mortality globally.However, the potentia mpact varies acros reas,an opulation ivin armer and poorer regions ar xpected t ustai ncrease urden. Urban dwellers ar ikel o experience more heat exposure ha ura ounterpart u o hea sland effect nd densel opulation.Furthermore,the increas n temperature-relate ortality would be substantiall educed i cenarios involving mitigatio trategie o limi reenhous as emissions. This evidenc s crucial for the development of coordinated and evidence-based climat n ublic health policies,and fo nforming th ngoing internationa iscussio h ealth impact f climat hange.Meanwhile,unintended consequence f mitigation policies shoul ls e evaluated.

Th egativ ealt mpacts o limat hange woul isproportionately affect poorer region f th orld[1,7,8]. Small island nation n ountrie n tropi n ub-tropic areas wil urt most. Social an nvironmenta nequalit houl e fairly addressed. Compariso it ifferen mission scenarios highlights the importanc nd urgenc f implementin itigation policie o imitin lobal warming and reducing the associate ealt isk or both hig n o ncom ountries.

Thus, it become itally importan nd exceedingl rgen o cur arbo missions.However, even if carbo mission ere completely stoppe oday,globa armin nevitabl ve he nex e ecade ue to th on tmospheric residence time o arbo n h nertia o he earth system. Therefore, it is necessar o call fo ctive adaptatio contex limat hange,an he strategie or mitigatio n daptatio eed t e develope arallel. I undamentally important t esign,develop and implement the best adaptatio trategies.

The implications o he IPC pecia eport on global warmin f 1.5 °Care indee ire[3].Rising seas,heatwaves, wildfires, floods,drought nd extreme storm il appenin t a highe requenc n greater intensit ha hey ar ow. Thus, th ext fe ears ar robabl h ost importan n our history.Evidenc hat climat hang ose erious health risk s overwhelmin u eing considered insufficiently i nalyzin n evelopin limate policie ate.Greenhous as emission ave reached a ll-time hig n limat hange is acceleratin n a globa cale[48]. Whether human health wil ontinu o flourish o il nsteaddeteriorate depend n action ake now. Medical professionals have eading rol rotecting populatio ealt ro h mpact limate change.I ctio ake o mitigat arbon emissions, we ma is h ast opportunit ackl limate change. It is encouraging to se ha anada"s municipalities ar eadin h ffort te he harmful effect h limat mergenc n human health an us o ealth-base limate action[49]. It i nticipated tha or unicipalitie il ollo he suit, and mor oncrete an ecisive action(e.g.,net zer arbo ociety i wo-to-three decade arly)wil ak lace.

Table 1. Agenda fo ormulatin ealth-relate limat olicy

Conclusion

Global pattern emperature-relate ortality ar apidl ovin oward or eat-related and les old-relate eaths. Such rend wil ecome more apparen he futur hig mission and no adaptatio cenario i dopted. Through this scopin eview, we believe tha t i ndispensably importan educe carbon footprint rom the individual to th lob Table 2). It is also imperative to note that the “net zero carbon” climat olic ill reduce ai ollutio n rin ubstantia ealt obenefits. Meanwhile, government t al evel eed to develop an mplemen daptation plans t ope wit he increasin mpact f climat hange.Urgen ction r equired fro h l alk he society, particularly medical professionals as carers o atients,to preven rreversible an atastrophic climate-induce ealth risks.

Table 2. Highlightsbox

Conflicts of Interest

The author eclare that there ar o conflict f interest.

Authors’ Contributions

S onceived the stud nd wrot h riginal draft. J n eviewed and edite he manuscript.

#Correspondenc hould b ddressed t ON hi Lu,Prof,MD,PhD, Tel:86-21-38625635,E-mail: tongshilu@scmc.com.cn

Biographical note o h irs uthor: TON h u,male,bor n 1955,MD,PhD,majoring i pidemiology and environmental health.

Received: January 20, 2021;

Accepted:April6,2021