Blair 2017 - U.S. peanut heavy metals

Blair and Lamb analyzed mercury, cadmium, lead, and arsenic in U.S. peanut crop samples collected at farmer-stock buying points during the 2013-2015 crop years. The sampling design covered three major U.S. growing regions and used censored-data methods for analytes with results below detection limits. The arsenic testing did not differentiate organic from inorganic forms, so arsenic is treated here as total/unspecified arsenic.

Key numbers

Table 1 reports 290 heavy-metal samples across 16 buying points: Southeast runner peanuts from 10 buying points (71 samples in each of 2013, 2014, and 2015), Southwest runner/Virginia/Spanish peanuts from 3 buying points (15, 12, and 16 samples by year), and Virginia-Carolina runner/Virginia peanuts from 3 buying points (12, 12, and 10 samples by year).

Table 8 summarizes population mean estimates and 95% upper confidence limits in ppm:

Cadmium had no censored observations. Table 2 reports the all-year/all-region Cd minimum 0.020 ppm, maximum 0.328 ppm, mean 0.0768 ppm, 95% confidence interval ( 0.0714, 0.0821 ), and UCL95 0.0814.

Arsenic was censored below 0.02 ppm. Table 3 reports all-year/all-region arsenic with 109 censored observations out of 290, maximum 0.41 ppm, ROS mean 0.0282 ppm, 95% confidence interval ( 0.0258, 0.0312 ), and UCL95 0.0308. The article states that the 0.41 ppm sample in 2015 was a Southwest-region outlier.

Lead was censored below 0.02 ppm. Table 5 reports all-year/all-region lead with 94 censored observations out of 290, maximum 0.64 ppm, ROS mean 0.0322 ppm, 95% confidence interval ( 0.0282, 0.0377 ), and UCL95 0.0367. The 0.64 ppm value was a 2013 Virginia-Carolina outlier.

Mercury was too heavily censored for a mean estimate. Table 7 reports probability intervals for testing below the 0.01 ppm detection limit: 2013 all-region samples had 98 censored results out of 98, with interval ( 0.970, 1.00 ); 2014 had 88 out of 95, with interval ( 0.854, 0.970 ); 2015 had 95 out of 97, with interval ( 0.927, 0.997 ).

Methods (brief)

Samples were collected during farmer-stock grading at 16 buying points across three crop years. Each 2 kg sample was composited and homogenized by Eurofins. Heavy metals were measured by ICP-MS on a Perkin Elmer NexION 300D after microwave-assisted acid digestion with nitric acid and hydrogen peroxide. NIST SRM 1568B and 1515 were analyzed as laboratory control samples, and recoveries for analytes of interest were within 10% of 100. For arsenic, the testing did not differentiate between organic and inorganic forms. The analysis used Kaplan-Meier, regression on order statistics (ROS), and maximum likelihood estimation for censored observations, with discussion focused on ROS results.

Implications

This source provides U.S.-market peanut occurrence evidence for Cd, total/unspecified As, Pb, and heavily censored total/unspecified Hg. The routeable peanut values are national crop-level mean estimates and UCL95 summaries, not single-state or product-brand measurements. Mercury should remain as censored total-Hg context because the source does not provide a concentration distribution suitable for mean estimation and does not measure methylmercury.

Verification notes

• PDF text was extracted with pdftotext -layout; Tables 1, 2, 3, 5, 7, 8, and 9 were checked in extracted text.
• No DOI for the Blair/Lamb article was found in the source text; the DOI-looking string in the extracted text belongs to a cited USGS reference, not this paper.
• Raw handle MFK_blair2017, raw SHA, and cite-key checks found no existing source page before creation.
• Units are preserved as source-reported ppm; no conversion to mg/kg was performed.
• Speciation: arsenic testing did not differentiate organic and inorganic forms, so arsenic is routed as tAs. Mercury is total/unspecified Hg; methylmercury was not measured.
• Brand firewall: no product brands are tied to contamination values. Acknowledged industry entities are sample-logistics acknowledgements only, not contamination rankings.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.