Air quality and particulate matter

Why particulate matter ranks at the very top

The Global Burden of Disease analysis 2021 in the Lancet compared 88 risk factors. At the very top is particulate matter, with 8 percent of all healthy life years lost worldwide, followed by high blood pressure with 7.8 and smoking with 5.7 percent. The WHO estimates 4.2 million premature deaths in 2019 from outdoor air alone, 89 percent of them in poorer countries.

That is a global figure. Does it also apply at Central European levels? On this there is an analysis in the New England Journal of Medicine from 2017: the entire American Medicare population, just under 61 million people over 65, over 13 years. For every 10 micrograms more particulate matter per cubic meter in the annual mean, all-cause mortality rose by 7.3 percent.

The second part of the result is the decisive one. In the regions below 12 micrograms, that is, with clean air by American standards, the effect did not disappear. It was steeper: 13.6 percent per 10 micrograms. According to these data, there is no threshold below which particulate matter becomes harmless.

Where Germany stands

In 2021 the WHO halved its guideline value for the annual mean to 5 micrograms; the EU is lowering its limit value from 25 to 10 from 2030. The mean exposure in the urban background was most recently 9 micrograms, according to the German Environment Agency (Umweltbundesamt); the EU limit value is met everywhere.

The WHO guideline value of 5 micrograms, on the other hand, was exceeded in 2024 by 97 percent of all monitoring stations, that is, practically all of them. We live precisely in the range in which the Medicare curve is steepest.

What the dust does in the body

PM 2.5 refers to particles smaller than 2.5 micrometers. They are small enough for the alveoli, and the smallest pass from there into the blood. The lung is not the main victim: according to the WHO breakdown, 68 percent of deaths are due to heart attack and stroke and only 4 percent to lung cancer. It is a vascular disease.

The strongest source is in the kitchen

According to a large American activity study, we spend around 87 percent of our time in buildings. Indoor air is the outdoor air that comes in, plus what we produce ourselves. The strongest source in a normal home is cooking. A measurement study from Seattle with standardized searing found peaks of 200 to 1,400 micrograms per cubic meter in the kitchen and living room. Depending on ventilation, it takes 1 to 10 hours for this to subside.

With a gas stove, nitrogen dioxide is added. A Stanford group measured in 32 homes: without a range hood, a small kitchen can exceed the American one-hour limit within minutes. Candles are the third source; a Düsseldorf study with 55 volunteers found an indication of a small decrease in lung function after 2 hours of candle smoke. The street outside the window is mainly an ultrafine particle problem: ultrafine particles halve within the first 150 meters.

What a filter device has to achieve

The key figure is the Clean Air Delivery Rate, or CADR: clean air per hour. The US Environmental Protection Agency calculates around 110 cubic meters per hour for 9 square meters. A bedroom of 20 square meters therefore needs about 240. Devices from about 150 euros often manage this, but only on the high setting, and that is loud. On the night setting, only a fraction of it remains.

A particle filter cannot remove gases, neither nitrogen dioxide from the stove nor carbon dioxide. On this there is a blinded Harvard study with 24 office workers over 6 days in a climate chamber: for every 400 parts per million more carbon dioxide, performance in a decision-making test fell by 21 percent; the worst condition was 1,400, the best around 500. It was funded by the air-conditioning group United Technologies; the effects are so large that they should be read with caution. The remedy costs nothing: open the window.

What the studies show

Medicare cohort study, New England Journal of Medicine 2017

Harvard analysis of the entire American Medicare population, just under 61 million people over 65, over 13 years. For every 10 micrograms more particulate matter per cubic meter in the annual mean, all-cause mortality rose by 7.3 percent. In the regions below 12 micrograms, the association was not weaker but steeper, namely 13.6 percent per 10 micrograms.

Shanghai 2015, randomized and double-blind

Device versus dummy in dormitory rooms, 35 students. 48 hours of filter operation lowered particulate matter by 57 percent, from 96 to 41 micrograms per cubic meter. Several inflammatory markers in the blood fell markedly, systolic blood pressure by just under 3 percent.

Detroit 2018, double-blind at low exposure

40 older women and men in a social housing complex, 3 days per condition. Baseline exposure was already low at 15 micrograms per cubic meter and fell to 7 with a HEPA filter. Systolic blood pressure fell by 3.2 millimeters of mercury.

Two meta-analyses on blood pressure

A systematic review from 2020 with 10 studies and 604 participants found an average of minus 4 millimeters of mercury systolic. A second meta-analysis of 14 crossover studies arrived at minus 2.3 millimeters and explicitly rates the certainty of the evidence as very low.

Ulaanbaatar, largest study on a hard endpoint

540 pregnant women, HEPA device versus no intervention, target outcome birth weight. Particulate matter fell by 29 percent, considerably less than the 57 percent in the dormitory trial. Birth weight did not change.

Where the data stop

For hard endpoints there is nothing. No randomized study shows that an air purifier lowers heart attacks, strokes or mortality. What is available is a surrogate: blood pressure and inflammatory markers over days instead of events over years, in a few hundred people. The second meta-analysis itself rates this evidence as very uncertain and warns against advertising claims. The largest study with a hard endpoint, 540 pregnant women in Ulaanbaatar, showed no effect on birth weight.

How to do it

Start at the source: switch on the range hood and open the window when frying. In the Seattle measurement, standardized searing reached peaks of 200 to 1,400 micrograms per cubic meter, and depending on ventilation it took 1 to 10 hours for this to subside. If you have a gas stove, nitrogen dioxide is added, which pushes a small kitchen without a hood above the American one-hour limit within minutes.

You size a HEPA device by its clean air delivery rate, CADR: the US Environmental Protection Agency calculates around 110 cubic meters per hour per 9 square meters, so about 240 for a bedroom of 20 square meters, which devices from around 150 euros often only manage on the loudest setting. In the studies, the devices ran continuously in the living room and bedroom. What is established is the reduction in concentration, in Shanghai by 57 percent, in Detroit from 15 to 7 micrograms, and a systolic blood pressure lower by about 2 to 4 millimeters of mercury. It is not established that this prevents heart attacks or deaths: the largest study, 540 pregnant women in Ulaanbaatar, lowered particulate matter by 29 percent without changing birth weight. Devices that generate ozone, that is, ionizers, plasma and some UV devices, are excluded, and against carbon dioxide no filter helps, only the window.

Safety

Pure filter devices bring hardly any risks of their own. The exception is devices that generate ozone, that is, ionizers, plasma devices and some UV devices: ozone is a lung irritant, and the US Environmental Protection Agency explicitly advises against them. The second risk is false security, for example a device in the living room while frying takes place in the kitchen without a hood. In the studies, the devices ran continuously in the bedroom and living room, for between 2 days and an entire pregnancy, without reported harm.

Sources

• Global Burden of Disease 2021 — 88 risk factors, The Lancet 2024
• WHO, fact sheet on outdoor air — 4.2 million premature deaths in 2019
• WHO, Global Air Quality Guidelines 2021 — guideline value of 5 micrograms per cubic meter annual mean for PM 2.5
• Medicare cohort study — Di et al., New England Journal of Medicine 2017, 60.9 million persons, 2000 to 2012
• Umweltbundesamt (German Environment Agency), Air quality in Germany 2024 — final evaluation (particulate matter exposure and monitoring stations)
• Dormitory crossover study Shanghai — Chen et al., Journal of the American College of Cardiology 2015, 35 students, 48 hours
• Air purifier study Detroit — Morishita et al., JAMA Internal Medicine 2018, 40 older women and men
• Systematic review 2020 on air purifiers and blood pressure — Walzer et al., Hypertension, 10 studies, 604 participants
• Climate chamber study on carbon dioxide — Allen et al., Environmental Health Perspectives 2016, Harvard, 24 office workers