Published since April · 2026
Bali Air Dispatch
Anonymous, non-commercial

Bali Air Dispatch

Dispatch IV — The long-term health effects of high PM2.5.
Edition XVII
Bulletin · Dispatch IV
The long-term health effects — read the standalone dispatch
Dispatch · IV
The long-term health effects of high PM2.5.
Filed · 26 April 2026
Approx. 7-minute read

When the World Health Organization halved its safe-exposure ceiling for fine particulate matter in 2021, it did so because the underlying epidemiology had stopped finding a floor. The harm scales with the dose; the dose is what we breathe.

For every additional ten micrograms a cubic metre of long-term PM2.5, pooled meta-analyses commissioned for the WHO 2021 guideline review report a 9.5% increase in all-cause mortality, a 12.7% rise in circulatory mortality, a 14.3% rise in ischaemic heart disease mortality, a 14.6% rise in cerebrovascular mortality, a 13.8% rise in mortality from chronic obstructive pulmonary disease, and a 9.3% rise in lung-cancer mortality. The all-cause estimate alone implies an attributable fraction of about 8.7% per ten-microgram increment, if the association is causal.

These are not edge-case findings. The Global Burden of Disease project attributes more than four million deaths in 2019 to long-term ambient PM2.5. A separate global modelling study estimates that the open burning of domestic waste alone — the activity overhanging Bali in April of 2026 — produces roughly 270,000 premature deaths a year, although that source-specific figure assumes the toxicity per microgram is the same as that of average urban PM2.5. The toxicology suggests it is probably worse.

Plastic and household-waste smoke is chemically distinct from typical urban PM. It is dominated by carbon — more than seventy per cent in one source-profile study — and enriched in polycyclic aromatic hydrocarbons, polychlorinated and polybrominated dioxins and furans, and transition metals including lead, cadmium, chromium, nickel, copper, zinc, and antimony. Co-emitted gases include benzene, styrene, formaldehyde, vinyl chloride, and 1,3-butadiene. The mixture has high oxidative potential and contains environmentally persistent free radicals; in cell-free assays it generates hydrogen peroxide, hydroxyl radicals, and reactive chlorine species.

The mechanism is not mysterious. Inhaled PM2.5 deposits in the deep lung. From there it triggers oxidative stress and chronic inflammation, with vascular endothelial dysfunction and autonomic-nervous-system imbalance following on. The downstream pathology spans cardiovascular disease, respiratory disease, cancer, and — increasingly — the nervous system. A 2022 meta-analysis reported a 40 per cent increase in dementia risk per ten micrograms of long-term PM2.5; a more conservative 2025 burden-of-proof analysis reports at least a fourteen per cent average increase across the range of exposures observed in the published cohorts. For autism-spectrum disorder, the prenatal-exposure literature suggests a thirty-one per cent increase per ten micrograms during pregnancy and sixty-four per cent during early childhood.

Pregnancy outcomes show a coherent dose response. A 2021 global meta-regression estimated that each ten-microgram rise in ambient PM2.5 during pregnancy reduces birth weight by twenty-two grams, raises low-birth-weight risk by eleven per cent, and raises preterm-birth risk by twelve per cent. These outcomes matter especially for the Bali context, because waste burning happens close to homes, repeats episodically, and falls hardest on neighbourhoods without formal collection.

The Kerobokan sensor recorded a 469-day mean of 21.9 µg/m³ — sixteen-point-nine micrograms above the WHO five-microgram annual guideline. Applied naively to the all-cause coefficient, that delta corresponds to roughly a fifteen per cent excess long-term mortality risk, if the relationship holds at this concentration band. The honest caveat is that the published coefficients are for total ambient PM2.5; the source-specific coefficient for plastic-burning PM has not been estimated. Multiple lines of toxicology suggest the true number is probably higher.

8.7%
All-cause mortality attributable fraction per ten-microgram-per-cubic-metre increment in long-term PM2.5. Pooled estimate from the WHO 2021-aligned meta-analysis.
14.3%
Excess ischaemic-heart-disease mortality per ten-microgram-per-cubic-metre rise. One of the largest single-cause coefficients in the pooled record.
270k
Estimated annual premature deaths globally attributable to PM2.5 from open burning of domestic waste, per CCAC-aligned modelling. Source-equal-toxicity assumption is conservative.
~15%
Approximate excess all-cause mortality risk implied by Kerobokan's 21.9 µg/m³ mean against the WHO 5 µg/m³ annual guideline, applying the pooled coefficient. Plastic-burning PM is plausibly more harmful per microgram than this estimate.

Who pays the toll first.

The vulnerable populations are the predictable ones: children, pregnant people and their fetuses, older adults, anyone with pre-existing cardiopulmonary disease, informal waste workers, and the residents of neighbourhoods without reliable collection. Lower socioeconomic status amplifies the risk through cumulative exposure and reduced access to filtration, medical care, and the option to leave. Across the cohort literature, no clear threshold has been established below which chronic cardiovascular harm disappears.

The 2021 guideline of five micrograms a cubic metre — and the twenty-four-hour ceiling of fifteen — were not picked at random. They were the level at which the meta-analyses stopped showing improvement with further reductions in exposure. Indonesia's national ambient standard, codified in Peraturan Pemerintah 22 of 2021, remains at fifty-five micrograms. Under domestic law, most of Bali's bad-air days are still formally normal.

What the published evidence does not tell you.

The strongest limitation in the literature is that almost every long-term effect estimate is for total ambient PM2.5, not for the fraction specifically derived from plastic and household waste burning. Source-resolved cohort studies near open burns do not yet exist at scale; toxicology suggests the per-microgram harm is worse than average; the actual coefficient is unspecified.

For the purposes of this dispatch, the prudent reading is that chronic exposure to PM2.5 from burning plastics and household waste is unlikely to be safer than typical urban PM2.5, and may be substantially more harmful per unit mass. The exact excess risk in Bali — over a Bali-specific population, over multi-year follow-up, with adjustment for tobacco and cooking exposure — has not been measured. The first step toward measuring it is the same as the first step toward fixing it: a denser network of monitors than this island currently has.

Sources for this dispatch

  1. WHO Global Air Quality Guidelines 2021 — annual mean 5 µg/m³, 24-hour mean 15 µg/m³; harm reported below previously assumed thresholds.
  2. Chen & Hoek 2020 (meta-analysis commissioned for WHO 2021 review) — pooled relative-risk estimates per 10 µg/m³ for all-cause, circulatory, IHD, cerebrovascular, COPD and lung-cancer mortality.
  3. Global Burden of Disease 2019 — ambient PM2.5 attributable mortality.
  4. Kodros et al. 2016 (and follow-ups) — global modelled premature deaths from open burning of domestic waste.
  5. Velis & Cook 2021; Kumar et al. 2023 — chemical profile of plastic and waste-combustion PM (PAHs, dioxins/furans, transition metals, persistent free radicals, oxidative potential).
  6. Wei et al. 2024 (California cohort) — incident AMI, IHD mortality, CV mortality below older regulatory standards.
  7. Wilker et al. 2023; Power et al. 2025 — long-term PM2.5 and dementia risk.
  8. Volk et al. 2013; meta-analyses 2017–2023 — prenatal and early-life PM2.5 and ASD.
  9. Ghosh et al. 2021 — global meta-regression on gestational PM2.5 and birth weight, low-birth-weight, preterm birth.
  10. IARC 2013 — outdoor air pollution and PM as Group 1 carcinogen.
  11. Climate & Clean Air Coalition (CCAC) CLOCC programme — open burning as a dominant PM2.5 source in Indonesian communities with incomplete waste collection.
  12. Peraturan Pemerintah 22 / 2021 (Indonesia) — domestic ambient PM2.5 standard at 55 µg/m³.
Index of dispatches