Burden of fine air pollution on mortality in the desert climate of Kuwait

The study addresses how short-term exposure to fine particulate matter (PM2.5) affects mortality in Kuwait, a desert country with frequent, severe dust storms and substantial petrochemical activity. Kuwait experiences extremely high dust deposition and frequent exceedances of WHO PM2.5 guidelines. While adverse health effects of air pollution are well documented globally, the Middle East lacks robust local epidemiology due to sparse monitoring networks and limited historical exposure data, with potential inaccuracies during extreme dust events. Advances in satellite data and spatiotemporal modeling now enable exposure estimation in under-monitored regions. Prior Kuwaiti studies relied on PM10, visibility, seasonal dust indicators, or examined upstream outcomes (e.g., respiratory admissions). This study leverages a historical, high-resolution PM2.5 model to quantify acute mortality risks overall and by cause, sex, age, and nationality, to inform locally relevant air quality standards and health policy.

The paper situates its research within global evidence linking PM2.5 to mortality (e.g., multi-country studies) and regional gaps in the Middle East due to limited monitoring and historical data. Prior Kuwaiti and regional work used proxies (PM10, low visibility, seasonal dust trends) or focused on hospital admissions rather than mortality. The authors note uncertainties regarding differential toxicity of PM2.5 components (e.g., fossil-fuel carbonaceous vs. crustal dust), with inconclusive evidence on a toxicity hierarchy. Comparisons with studies from Europe, North America, Australia, Japan, and the Eastern Mediterranean suggest Kuwait’s effect estimates are within the range reported elsewhere, and meta-analytic evidence from Iran shows similar magnitudes. The literature also indicates heightened vulnerability among the elderly and potential disparities affecting migrant workers in Kuwait, motivating subgroup analyses by age and nationality.

Design and setting: Time-series analysis of daily mortality in Kuwait from January 1, 2001 to December 31, 2016, assessing acute (short-term) effects of PM2.5. Exposure assessment: Daily PM2.5 estimated using a hybrid model (machine learning plus generalized additive mixed models) at 1×1 km spatial and daily temporal resolution across Iraq and Kuwait (R²=0.71), utilizing visibility, satellite aerosol optical depth, land use, and ground observations. Daily PM2.5 averaged over all urban pixels in Kuwait excluding barren desert lands. Meteorology: 24-hour average temperature (°C) and relative humidity (%) obtained from the national meteorological authority. Outcomes: Daily non-accidental all-cause mortality (ICD-10 A00–R99), cardiovascular mortality (I00–I99), and respiratory mortality (J00–J99). Subgroups: sex (male/female), age (children <15, adults 15–65, elderly 65+), nationality (Kuwaiti/non-Kuwaiti). Statistical analysis: Quasi-Poisson time-series regression with model choices guided by quasi-AIC. Linearity of PM2.5 assessed using penalized splines in GAMs. Lag structures examined via distributed lag models (3- and 5-day lags) and moving averages (2-, 3-, 5-, 7-day windows); the best-fit model used a 3-day moving average (lag0–3). Time-varying confounding adjusted using natural splines with 5 df per year for seasonality/long-term trends. Temperature controlled using a distributed lag nonlinear model (DLNM) with a cross-basis: temperature-response modeled via natural splines (3 df), lag-response over 7 days (also assessed 14 days) with 2–3 df equally spaced on the log lag scale. Additional covariates: day of week (categorical) and relative humidity (penalized spline). Effect estimates expressed as percent change in mortality per 10 µg/m³ increase in PM2.5: (exp(β×10)−1)×100%. Health burden quantified as preventable deaths per year for a 10 µg/m³ reduction: (exp(β×10)−1)×(total deaths/years). Differences across subgroups tested via two-sample z-tests using point estimates and standard errors (e.g., Kuwaitis vs non-Kuwaitis; females vs males; adults vs elderly). Analyses conducted in R 4.0.3.

- Study period deaths: 70,321 over 16 years (2001–2016); cardiovascular accounted for 47.9%; respiratory 7.8%; 53.2% Kuwaitis, 46.8% non-Kuwaitis. - Exposure and meteorology: Mean urban PM2.5 = 46.2 µg/m³ (SD 19.8). Mean temperature 27.1 °C (SD 9.8); mean relative humidity 34% (SD 21.4%). - Linearity: Penalized spline fits suggested limited deviation from linear exposure-response (effective df <3). - Primary association (lag0–3, 3-day moving average): • All-cause mortality: 1.19% increase per 10 µg/m³ (95% CI: 0.59, 1.80%). • Cardiovascular mortality: 0.95% per 10 µg/m³ (95% CI: 0.12, 1.78%). • Respiratory mortality: not statistically significant. - Subgroups (all-cause): • Males: 1.17% (95% CI: 0.42, 1.93%); Females: 1.18% (95% CI: 0.28, 2.09%). • Children (<15): 1.87% (95% CI: 0.23, 3.54%); Adults (15–65): 1.87% (95% CI: 0.23, 3.54%); Elderly (65+): 1.09% (95% CI: 0.23, 1.97%). • No statistically significant differences between Kuwaitis vs non-Kuwaitis (p=0.92), males vs females (p=0.94), or adults vs elderly (p=0.94). - Preventable deaths for a 10 µg/m³ annual reduction in PM2.5: • Overall: 52.3 deaths/year (95% CI: 25.7, 79.1). • Kuwaitis: 28.6 (95% CI: 10.3, 47.0); Non-Kuwaitis: 23.9 (95% CI: 6.4, 41.5). • Children: 9.4 (95% CI: 1.2, 17.8); Elderly: 20.9 (95% CI: 4.3, 37.6). • From figure approximations: Cardiovascular ≈20/year; Respiratory ≈4/year; Males ≈31/year; Females ≈21/year; Adults ≈24/year; Elderly ≈21/year; Kuwaitis ≈29/year; Non-Kuwaitis ≈24/year.

The study demonstrates that short-term increases in PM2.5 are associated with higher all-cause and cardiovascular mortality in Kuwait’s dusty, petrochemical-influenced environment, addressing a major evidence gap in the Middle East where monitoring has been sparse. Despite the predominance of crustal dust, the magnitude of the all-cause effect (1.19% per 10 µg/m³) lies within the range reported internationally. Comparisons with multi-country and regional studies indicate heterogeneity across countries, likely reflecting differences in demographics, infrastructure, healthcare access, and possibly PM2.5 composition. The lack of significant differences in risk across sex, nationality, and between adults and elderly was notable. While literature often shows higher vulnerability among the elderly, similar or higher risks among adults here could reflect differing exposure patterns (e.g., more outdoor work), though this remains speculative. Contrary to prior evidence of greater vulnerability among migrant (non-Kuwaiti) populations, comparable risks were observed by nationality overall; suggestive evidence pointed to higher vulnerability among non-Kuwaiti children, warranting further investigation. Overall, the findings provide locally relevant effect estimates and health burden metrics that can inform air quality standards and risk management in desert climates with limited ground monitoring.

High-resolution, satellite- and model-based PM2.5 estimates enabled quantification of acute mortality risks and preventable deaths attributable to PM2.5 in Kuwait. A 10 µg/m³ reduction in PM2.5 could prevent roughly 52 deaths annually, including among children and the elderly. These locally derived estimates can support evidence-based air quality policies in dusty, petrochemically influenced regions. Future research should examine pollutant source contributions and PM2.5 composition, refine exposure assessment, evaluate occupational and outdoor exposure patterns, and further assess health equity and susceptibility, particularly among migrant and pediatric subpopulations.

- Exposure assessment relied on modeled PM2.5 (hybrid machine learning/GAM approach, R²=0.71), implying measurement error that could bias estimates. - The analysis focuses on short-term (acute) associations with daily mortality and does not address long-term exposure effects. - Respiratory mortality associations were not statistically significant, which may reflect limited power or differential sensitivity by cause. - Subgroup comparisons found no significant effect modification overall; the authors note some interpretations (e.g., adult exposures) are speculative and call for further research to clarify equity and susceptibility patterns. - Findings are from Kuwait and may have limited generalizability to settings with different pollutant mixtures, demographics, or healthcare systems, and uncertainties remain regarding differential toxicity of PM2.5 components (crustal dust vs. carbonaceous sources).