Cause-specific mortality attributable to summer heat in Italy, 2004-2023: a nationwide Bayesian spatiotemporal study

@misc{ceccarelli2026causespecific,
  title = {{Cause-specific mortality attributable to summer heat in Italy, 2004-2023: a nationwide Bayesian spatiotemporal study}},
  author = {E. Ceccarelli and G. Stoppa and M. Blangiardo and G. Konstantinoudis and N. Vanacore and G. Jona Lasinio and G. Minelli},
  howpublished = {medRxiv},
  year = {2026},
  month = oct,
  doi = {10.64898/2026.10.02.26364563},
  url = {https://www.medrxiv.org/content/10.64898/2026.10.02.26364563v1},
  note = {CC BY},
}

TY  - UNPB
TI  - Cause-specific mortality attributable to summer heat in Italy, 2004-2023: a nationwide Bayesian spatiotemporal study
AU  - E. Ceccarelli
AU  - G. Stoppa
AU  - M. Blangiardo
AU  - G. Konstantinoudis
AU  - N. Vanacore
AU  - G. Jona Lasinio
AU  - G. Minelli
PY  - 2026
DA  - 2026/10/02
PB  - medRxiv
DO  - 10.64898/2026.10.02.26364563
UR  - https://www.medrxiv.org/content/10.64898/2026.10.02.26364563v1
AB  - BackgroundItaly experiences a substantial heat-related mortality burden, but evidence on cause-specific mortality across its heterogeneous climatic settings remains limited. We aimed to quantify cause-specific mortality attributable to summer heat in Italy at national and ecological region levels and to map its distribution across municipalities.  MethodsWe analysed daily deaths classified by underlying cause in eight ICD-10 groups among individuals aged 18-64 years and 65 years or older across 7,864 municipalities during June-August, 2004-2023. Mortality data were linked to area-weighted ERA5-Land daily maximum temperature. Nonlinear temperature-mortality associations were estimated with Bayesian hierarchical Poisson models, fitted separately for each cause, age group, and ecological region, with municipality-level attributable fractions.  FindingsHigh temperatures were associated with mortality from a broad range of causes, particularly among individuals aged 65 years or older. Among the study population, heat-attributable fractions were highest for nervous system diseases (14{middle dot}93%, 95% credible inter-val 14{middle dot}22-15{middle dot}69) and mental and behavioural disorders (14{middle dot}47%, 13{middle dot}55-15{middle dot}39). Circulatory diseases accounted for the largest absolute burden, with 85,524 heat-attributable deaths (82,762-88,271) across the summers of the 20 years considered.  InterpretationThe mortality burden associated with summer heat in Italy extended well beyond circulatory and respiratory diseases, highlighting the relevance of a broader range of health outcomes. These findings support heat-health strategies that account for cause-specific and geographical differences in heat-related risk, with greater consideration of neurological and mental health outcomes alongside established cardiorespiratory risks.  FundingNextGenerationEU (CUP: B53C23002660006); Imperial College Research Fellowship; NIHR Imperial Biomedical Research Center (BRC: NIHR203323).  Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe searched PubMed from database inception to 6 August 2026, combining exposure terms ("heat", "high temperature", or "ambient temperature") with mortality terms ("mortality", "cause-specific mortality", or "cause of death"). Additional searches included terms related to attributable burden ("attributable fraction" or "attributable burden") and geographical setting ("Italy" or "Europe"). We considered population-based observational studies examining mortality associated with high ambient temperatures, with particular attention to studies reporting cause-specific associations or attributable burden. Previous studies have consistently associated high ambient temperatures with increased mortality. Cause-specific analyses have shown that the effects of heat extend beyond cardiovascular and respiratory diseases to mental and behavioural, neurological, endocrine and metabolic, genitourinary, digestive, and external causes. However, much of the European evidence has been based on selected cities or relatively large administrative areas, and geographical variation has generally been characterised using administrative regions or broad climatic zones. Evidence on the fine-scale spatial distribution of cause-specific heat-attributable mortality across Italy remains limited, particularly when accounting for geographical differences in temperature-mortality associations.  Added value of this studyTo our knowledge, this is the first nationwide Italian study to estimate cause-specific heat-mortality associations separately across ecological regions and to map the corresponding heatattributable burden at the municipality level. Using daily mortality records from all Italian municipalities over the 20-year period 2004-2023, we estimated heat-mortality associations separately across ecological regions and age groups within a Bayesian modelling framework, propagating uncertainty from the estimated exposure-response associations to the attributable burden. We found consistent heat-related mortality risks across all major causes of death among older adults, with the strongest associations observed for mental and behavioural disorders and diseases of the nervous system.  Implications of all the available evidenceThe available evidence supports considering a broader range of causes than cardiovascular and respiratory diseases alone in heat-health surveillance and prevention. Evidence of increased mortality for selected causes among adults aged 18-64 years also supports extending prevention beyond older populations. Accounting for heterogeneity across ecoregions reflects climatic variability across the country, provides a more accurate assessment of at-risk populations, and offers a framework applicable to other countries and regions characterised by similar climatic variability. These findings can support public health authorities in prioritising prevention measures and allocating resources where they are most needed under a warming climate.
ER  -