Nitrogen Dioxide, Traffic Pollution, and Headache Risk
Nitrogen dioxide is a reddish-brown irritant gas produced almost entirely by combustion, and in most cities the dominant source is road traffic. Its defining property is an extraordinarily steep spatial gradient: concentrations beside a busy road can be two or three times what they are 300 metres away, which means your exposure is set by geography measured in street widths rather than by the city-wide number your air quality app shows you. The evidence linking short-term nitrogen dioxide rises to headache and migraine presentations is positive but modest, and it is confounded by everything else that travels in the same exhaust plume.
That last point deserves stating early. Nitrogen dioxide is rarely alone. It arrives mixed with ultrafine particles, carbon monoxide, benzene and partly burnt hydrocarbons, plus the noise and vibration of the traffic producing it. Studies reporting a nitrogen dioxide association are usually reporting an association with the whole traffic package, using one gas as its label.
Which is still useful. If the label tracks the package reliably, and it largely does, the number tells you something actionable even when the causal arrow points at a dozen substances at once.
A pollutant measured in street widths
Almost every other environmental exposure discussed in migraine circles is regional. A heat dome covers several countries, a smoke plume a continent, and barometric pressure is essentially uniform across a whole city on any given afternoon.
Traffic pollution is not like that. It is the most spatially concentrated common exposure there is.
Near-road concentrations fall away fast. Measurements consistently show nitrogen dioxide and ultrafine particle counts dropping steeply with distance from a major road, approaching urban background somewhere between 150 and 500 metres depending on wind, terrain and building layout. A flat on the third floor above a six-lane arterial is in a genuinely different exposure environment from an identical flat two streets back, even though both would report the same number if you looked up the city average.
Street canyons trap it. A narrow road with continuous tall buildings on both sides behaves like a box. Air recirculates in a vortex rather than flushing through, and concentrations on the leeward pavement can be several times those on the windward one.
Concentrations also decline with height above the road, so ground and first floor are worse than the fifth.
The practical implication is uncomfortable but honest: for traffic pollution, the single largest determinant of your personal exposure is not the weather, not the season and not the day. It is where you live, where you work, and which route you take between them.
What nitrogen dioxide is and where it comes from
Combustion at high temperature forces atmospheric nitrogen and oxygen to combine. The immediate product is mostly nitric oxide, which oxidises in the air to nitrogen dioxide within minutes to hours. Together the pair are referred to as nitrogen oxides.
Diesel engines are the disproportionate contributor in fleets where diesel is common, because the combustion conditions that make diesel efficient also make it a strong nitrogen oxide producer. Real-world emissions from older diesel vehicles have repeatedly been found to exceed laboratory test figures, which is one reason urban nitrogen dioxide fell more slowly than regulators projected through the 2010s.
Other sources exist and are not trivial. Gas boilers, gas cooking, industrial furnaces and power generation all produce nitrogen oxides. Indoors, an unvented gas hob can push kitchen nitrogen dioxide above anything you would encounter on the street outside, which is a point worth holding onto.
Nitrogen dioxide also has an atmospheric afterlife. It is the precursor that sunlight splits to manufacture ground-level ozone, and it oxidises further into nitrate that condenses into fine particulate. So the same gas that irritates you at the kerbside on a winter morning contributes to a summer ozone episode 40 kilometres downwind and to the secondary particulate load region-wide.
The daily and weekly rhythm
Unlike ozone, which follows the sun, nitrogen dioxide follows the traffic and the atmosphere's willingness to disperse it.
Two peaks on a weekday. A sharp morning rush peak, typically the higher of the two because the overnight boundary layer is still shallow and the air has not begun mixing. Then a broader evening peak, usually somewhat lower because afternoon mixing depth is greater even though the traffic volume is similar.
A midday trough as the boundary layer deepens and dilutes everything.
A weekend signal. Concentrations drop noticeably at weekends, particularly where the weekday peak is dominated by commercial diesel traffic. If your headaches cluster midweek and vanish on Sundays, traffic exposure is worth considering alongside the more obvious explanations of work stress and sleep timing.
A strong winter maximum. Not because there is more traffic, but because the atmosphere disperses less. Cold, still, high-pressure conditions with a surface inversion cap the emissions in a shallow layer near the ground. The worst nitrogen dioxide episodes in temperate cities happen in still winter weather with a flat, high barometer.
How traffic exhaust might reach a headache
There is no clean single mechanism, and anyone who offers one is simplifying.
Direct airway irritation. Nitrogen dioxide is an oxidising gas that irritates the respiratory epithelium and increases airway responsiveness. Controlled exposure studies find measurable inflammatory changes in the airways at concentrations found beside busy roads. Airway irritation involves trigeminal nerve endings, and the trigeminal system is the one that generates migraine pain, so a short and plausible route exists.
The ultrafine particle problem. Fresh exhaust contains enormous numbers of particles in the sub-100-nanometre range. These contribute almost nothing to the mass-based PM2.5 reading, so a near-road environment can look unremarkable on a particulate monitor while delivering a very large particle count. Ultrafine particles are the fraction most able to cross from the alveoli into the circulation, and translocation into the bloodstream is the step that turns a lung exposure into a systemic one.
Systemic inflammation and vascular effects. Short-term traffic exposure has been associated in controlled studies with changes in vascular reactivity, blood pressure and markers of inflammation. Migraine is a disorder of a lowered threshold, and inflammatory signalling is one of the things known to lower it.
The non-chemical confounds. Traffic also delivers continuous low-frequency noise, disrupted sleep for those living beside roads, and vibration. Noise exposure alone has documented associations with headache and poor sleep quality. Disentangling the gas from the noise in an observational study is close to impossible, because the two have identical geography.
Being honest about the evidence
The literature on nitrogen dioxide and headache is thinner and messier than the literature on fine particulate.
Several time-series studies of emergency presentations report a positive association between short-term nitrogen dioxide and headache or migraine visits. Several others find nothing once other pollutants are in the model. Effect estimates, when positive, are small at population level.
Three specific problems recur. First, collinearity: nitrogen dioxide, carbon monoxide and ultrafine particles rise and fall together, so statistical models cannot easily assign credit. Second, exposure misclassification: studies assign people the reading from the nearest monitor, but with a gradient this steep, two people assigned the same value can have exposures differing by a factor of three. That kind of error systematically biases results toward finding nothing. Third, meteorological confounding: high nitrogen dioxide days are stagnant, cold, inversion days, and stagnation brings its own physiological baggage.
The reasonable position is that traffic exhaust is a plausible contributor to headache burden in people who live and travel in high-exposure microenvironments, that the population-level signal is weak, and that the individual-level signal for a susceptible person could be considerably stronger than the population average implies.
The pressure and weather overlay
This is where a barometric log earns its keep, because the meteorology of a bad traffic-pollution day is highly specific.
The worst episodes need a surface temperature inversion: cold air pooled at ground level under warmer air aloft, usually forming on a clear, calm night under a ridge of high pressure and persisting through a short winter day. Vertical mixing effectively stops. Emissions from the morning rush accumulate in a layer perhaps 100 to 300 metres deep instead of dispersing through a kilometre or more. Basin and valley cities suffer most, because terrain adds cold-air drainage to the effect. A winter inversion in Salt Lake City or Denver is the textbook version of this, and the same physics operates on a smaller scale in any river valley.
The barometric signature is a high, flat pressure trace lasting days. Nothing is moving. If your only tracked variable is pressure change, an inversion week reads as uneventful while your exposure quietly triples. That mismatch is exactly the pattern a migraine tracker app is useful for surfacing: Pressure Pal logs pressure and its trend beside your symptoms, and a run of flat high pressure sitting next to a run of bad days is a signal in itself, pointing you toward the air rather than the barometer.
The reverse case is equally instructive. A vigorous frontal passage tightens the pressure gradient, brings wind and deepens the mixing layer. Nitrogen dioxide falls sharply. So the day the barometer moves is usually the day the traffic pollution clears, and if you get a headache on that day, the exhaust is not a plausible culprit.
Reducing your dose without moving house
Change the route before you change the hour. Unlike ozone, where timing dominates, route choice does most of the work here. A parallel residential street one block from an arterial can cut your walking or cycling exposure substantially for a minute or two of added journey time. Cyclists benefit most, because exertion multiplies the inhaled volume.
Do not sit in the queue with the vents open. Concentrations inside a vehicle in slow traffic are frequently higher than on the pavement beside it, because you are directly in the exhaust plume of the car ahead. Switching ventilation to recirculate in congestion, and back to fresh air once moving, meaningfully reduces in-cabin exposure. Tunnels and multi-storey car parks are the extreme case.
Put the bedroom on the quiet side. If your home fronts a busy road and you have a choice of rooms, sleeping at the back addresses gas concentration, particle count and traffic noise simultaneously. This is a larger intervention than any filter.
Ventilate against the rush, not with it. Opening windows during the morning peak on a still winter day imports the worst air of the day. Midday and late evening are better.
Deal with your own combustion. Use the extractor hood every time you cook with gas. A hob in an unventilated kitchen is a nitrogen dioxide source you fully control, and for many people it is the largest single contributor to their personal exposure. Filtration will not help you here: HEPA removes particles, but nitrogen dioxide is a gas and passes straight through.
Frequently asked questions
How far from a road do I need to be for it to matter?
Most of the excess concentration is gone by roughly 150 to 300 metres downwind, though the exact distance depends on traffic volume, wind direction and whether buildings trap the air. Being on the upwind side of a road is better than the downwind side at the same distance.
Is nitrogen dioxide worse than fine particulate for headaches?
The evidence base for fine particulate is stronger and more consistent. Nitrogen dioxide is best understood as a marker for the near-road mixture rather than as the sole active agent, which is precisely why it is hard to study on its own.
Does an electric vehicle fleet solve this?
It removes the exhaust component, which is most of the nitrogen dioxide. It does not remove brake, tyre and road surface particulate, which is a substantial share of near-road particle mass, nor the noise. Electrification improves the picture considerably without eliminating the near-road effect.
Why is my indoor reading high when I live somewhere quiet?
Almost always gas cooking, and occasionally an unflued heater or an attached garage. A gas hob running for twenty minutes without extraction can produce kitchen concentrations well above roadside levels.
Does a cloth or surgical mask help while cycling?
Not for nitrogen dioxide, which is a gas that passes through fabric untouched. A well-fitted respirator reduces particle intake but does nothing meaningful for gases without a carbon layer. Route choice is more effective than any mask.
Should I stop cycling in the city?
The published risk-benefit analyses consistently favour continuing to cycle, because the cardiovascular and general health gains outweigh the pollution penalty in almost all urban settings. Choosing quieter routes captures most of the benefit of stopping without any of the cost.
Where this leaves you
Traffic pollution is the most local of the common air exposures. Nitrogen dioxide is a reliable signpost for it rather than the whole story, and the honest evidence for a headache link is suggestive rather than settled. What is not in doubt is that exposure varies enormously across a few hundred metres, that it peaks on still winter mornings under a flat high barometer, and that most people can move their personal dose substantially with route, ventilation and cooking decisions that cost almost nothing.
If your bad days cluster on cold, calm, unremarkable-looking weekdays with nothing at all happening on the pressure trace, the traffic corridor you walk through twice a day is a reasonable place to start looking. A city-wide air quality figure will not tell you that. Your own street will.
Pressure Pal tracks barometric pressure and its trend for your location beside your logged symptoms, which is how a week of flat high pressure and bad days stops looking like a mystery and starts looking like a stagnation episode.