Lead is a factor in about 256,000 US cardiovascular deaths a year. Your cardiologist has never asked about it.

The 2018 mortality study, and what the 412,000 number is.

Bruce Lanphear and colleagues took 14,289 adults aged 20 and older who had a blood lead measurement in the Third National Health and Nutrition Examination Survey (NHANES III, 1988 to 1994) and linked them to the National Death Index through the end of 2011. Median follow-up was 19.3 years. 4,422 of them died, 1,801 from cardiovascular disease and 988 from ischaemic heart disease. The paper was published in Lancet Public Health in April 2018 and is open access.

The geometric mean blood lead in the cohort was 2.71 µg/dL. One in five had a level of 5 or above. Moving from the 10th percentile of blood lead (1.0 µg/dL) to the 90th (6.7 µg/dL) carried a hazard ratio of 1.37 for death from any cause (95% CI 1.17 to 1.60), 1.70 for cardiovascular death (1.30 to 2.22), and 2.08 for ischaemic heart disease death (1.52 to 2.85). Those are the adjusted numbers, after age, sex, race, smoking, diabetes, cholesterol, kidney function, and urinary cadmium.

From those hazards the authors computed a population attributable fraction, the share of deaths that would not have happened if everyone's blood lead sat at the bottom of the range. For all-cause mortality it was 18.0% (10.9 to 26.1), which they translated to about 412,000 deaths a year. For cardiovascular mortality it was 28.7% (15.5 to 39.5), about 256,000 deaths a year. For ischaemic heart disease it was 37.4% (23.4 to 48.6), about 185,000 a year.

This page is about the middle number. The historical curve, how that attributable fraction has moved across every national survey since 1976 and what it adds up to over the decades, is on the NHANES cardiovascular page. The wide confidence interval on the 28.7% is real and the authors print it. The low end of the interval is still 15.5% of American cardiovascular deaths.

The finding did not come out of nowhere. Schober and colleagues at the National Center for Health Statistics had run the same cohort with shorter follow-up in 2006 (Environmental Health Perspectives), 9,757 participants aged 40 and up. Relative to blood lead under 5 µg/dL, all-cause mortality was 1.24 times higher at 5 to 9 (95% CI 1.05 to 1.48) and 1.59 times higher at 10 and above (1.28 to 1.98), with the same trend for cardiovascular deaths. Lanphear extended the follow-up and modeled the dose-response continuously. The answer came out the same direction and larger.

Blood pressure. Forty years of knowing.

Joseph Pirkle and colleagues published the blood pressure finding in the American Journal of Epidemiology in February 1985. The data were NHANES II, collected 1976 to 1980, white men aged 40 to 59. After adjusting for age, body mass index, nutritional factors, and blood biochemistries in a multiple regression, both systolic and diastolic blood pressure rose with blood lead, p less than 0.01. The sentence that matters is short: "There was no evidence of a threshold blood lead level for this relationship."

They then ran the finding forward using the Framingham and Pooling Project risk coefficients. Mean blood lead in adult white men had fallen 37% over the four years of the survey as lead came out of gasoline. From that drop alone they predicted a 4.7% decrease in fatal and nonfatal heart attacks over ten years, a 6.7% decrease in strokes, a 5.5% decrease in all-cause deaths over 11.5 years, and 17.5% fewer men in that age group with hypertension. That was the projection in 1985. The Lanphear cohort, enrolled three years later, is the measurement.

The effect per unit is not large. Nawrot and colleagues pooled 31 studies and 58,518 subjects in 2002 (Journal of Human Hypertension): each doubling of blood lead was associated with a 1.0 mm Hg rise in systolic pressure (95% CI 0.5 to 1.4) and 0.6 mm Hg in diastolic (0.4 to 0.8). Nawrot called it a weak association, and per person that is fair. Navas-Acien's systematic review in 2007 pointed out the estimate is probably biased low, since a single blood lead draw is a noisy measure of the exposure that matters. What turns a small per-person shift into a mortality number is that in the peak years nearly every adult in the country was exposed, and the whole blood pressure distribution moved with them.

Read the chart for shape, not for a personal prediction. A shift of a few mm Hg across a whole population moves a lot of people over the treatment line. Today's adult mean blood lead is under 1 µg/dL. The adults who were children or working-age in the 1970s were not there then, and the lead they took in did not leave.

The bone is the record.

Blood lead has a half-life of about a month. It tells you what came in recently. Bone lead accumulates over decades and turns over slowly, so a bone measurement is closer to a lifetime exposure total. The two can point in different directions in the same person: a man with a normal blood lead today and a heavy skeletal burden from 1970.

Howard Hu's group tested which one predicted hypertension. The paper ran in JAMA in April 1996, a case-control study nested in the Veterans Administration Normative Aging Study, 590 men. Tibia (cortical bone) and patella (trabecular bone) lead were measured in vivo with K-shell X-ray fluorescence; blood lead by atomic absorption. Blood lead was low, mean 6.3 µg/dL, and both blood and bone lead were higher in the hypertensive men. Then they built a logistic model that started with age, race, body mass index, family history, alcohol, smoking, dietary sodium and calcium, blood lead, tibia lead, and patella lead, and let backward elimination decide what stayed.

The same logic applies to women, with one added step. Bone demineralization after menopause releases stored lead back into circulation. Nash and colleagues examined 2,165 women aged 40 to 59 in NHANES III (JAMA, March 2003). Women in the top quartile of blood lead (4.0 to 31.1 µg/dL) had 3.4 times the odds of diastolic hypertension (95% CI 1.3 to 8.7) compared with the bottom quartile (0.5 to 1.6). Among postmenopausal women the odds ratio for diastolic hypertension climbed by quartile: 4.6, 5.9, then 8.1 (2.6 to 24.7) in the top quartile. The authors' framing was that in these women the skeleton itself had become the exposure source. The women in that survey were born between 1929 and 1954. The women born 1955 to 1975, the highest childhood blood leads of any American cohort, are in that age window now.

How bone lead gets measured, and what the same skeletal store does to the kidney, is on the bone lead and kidney page. Where the burden came from is on the historical exposure page.

How lead raises blood pressure.

On the question of clinical outcomes beyond blood pressure, Navas-Acien was careful and so is this page. The 2007 review found the evidence for hypertension sufficient to infer causation. For heart attack, stroke, and cardiovascular death it found the evidence "suggestive but not sufficient," because the number of outcome studies was small. Lanphear 2018 was published eleven years later and is the largest of those outcome studies. It does not by itself close the question. It does make the current guidance harder to justify.

Why your cardiologist did not ask.

The 2017 American College of Cardiology and American Heart Association hypertension guideline lists the risk factors to ask about and the secondary causes to screen for. Lead is not on either list. The prior guideline (JNC 8, 2014) did not include it either. The hypertension intake form asks about family history, salt, exercise, alcohol, sleep, and stress. It does not ask about year of birth, childhood address, or occupational lead exposure. No clinical society has issued guidance on screening adults born between 1946 and 1980 for skeletal lead.

The cardiologist who diagnoses hypertension and prescribes lisinopril is not wrong. The clinical system is doing what it was designed to do. The design does not include a path from "patient born in 1968 in a city with leaded gasoline and old housing" to "consider cumulative lead burden before attributing this to genetics and diet." No institution has that job.

Part of the reason is practical. Bone lead is measured by K-shell X-ray fluorescence, and the instruments exist in a handful of research labs. A blood lead draw is cheap and any lab runs it, but as Hu showed, it measures the wrong thing for this question. The other part is that lead sits in environmental health and hypertension sits in cardiology, and the two literatures rarely cite each other. The Lanphear paper's own interpretation line calls low-level lead "an important, but largely overlooked, risk factor for cardiovascular disease mortality."

What the coverage of heart disease leaves out.

Heart disease gets more health coverage than any other condition, and the risk factor list in that coverage is stable: diet, exercise, cholesterol, weight, genetics, stress, sleep, smoking. The hypertension pages at WebMD, Healthline, and the Mayo Clinic list risk factors at length; lead is absent from all three. Scientific American ran a piece on the Lanphear paper when it came out in 2018 and did not follow up.

For scale: the CDC counted 919,032 US deaths from cardiovascular disease in 2023. The attributable fraction has been falling as blood lead falls, so the current number is likely below Lanphear's 256,000, and how far below is what nobody is measuring. The one thing that would settle it, bone lead in a representative sample of Americans over 50, has never been collected in a national survey.

Related on this site.

The same birth cohorts carry the IQ loss and the depression and anxiety findings. Ongoing exposure through lead service lines is still adding to the blood side of the ledger.

Citations