Konuspayeva et al. 2009 — Heavy metals in camel milk, shubat, and fodder from Kazakhstan
Konuspayeva and colleagues from Al-Farabi Kazakh National University and CIRAD conducted the first published assessment of heavy-metal contamination in camel milk from Kazakhstan, a country with documented environmental burdens from Soviet-era nuclear testing, polymetal mining, oil extraction, and leaded gasoline still in widespread use at the time of sampling. Sample (i) collected paired water, fodder, raw camel milk, and shubat (fermented camel milk) from 8 farms in 4 regions near identified pollution sources (polymetal industries, oil forages, phosphate manufacturing, the Baikonur cosmodrome). Sample (ii) collected 63 individual raw camel milk samples from 12 different farms in the same 4 regions for Pb, Cr, and Zn. Analytical platform was ICP-AES (Varian Vista MPX – CCD) after wet-oxidation digestion. The headline finding is a mean Pb of 250 ± 56 ppb in the 63-sample dataset, well above tolerable limits used in Turkey (20 ppb), Germany/Netherlands (50 ppb), and Kazakhstan’s own standard (100 ppb), with a maximum individual value of 532 ppb at Atyrau. Lead consistently decreased from raw milk to shubat, which the authors attribute provisionally to lactic-bacteria binding under study elsewhere in the same lab.
Key numbers
All concentrations as reported in the source. Note that Sample (i) values are in ppm (mg/kg) while Sample (ii) Pb and Cr values are in ppb (µg/kg) and Sample (ii) Zn is in ppm. Detection limit: 0.01 ppm for all metals except As (0.1 ppm). Method: ICP-AES, Varian Vista MPX – CCD, wet-oxidation digestion (HNO3/H2SO4/HClO4 mineralisation; for fodder, dry-way mineralisation with HF silica removal).
Sample (i): 8 farms, herd-mixed milk, paired water/fodder/milk/shubat (Table 1)
Mean ± SD (in ppm wet weight = mg/kg):
| Matrix | Cu | Fe | Mn | Zn | As | Pb |
|---|---|---|---|---|---|---|
| Fodder | 10.40 ± 2.93 | 793.69 ± 630.48 | 62.38 ± 20.67 | 32.95 ± 27.15 | 1.03 ± 0.49 | 4.28 ± 9.60 |
| Camel milk | 0.07 ± 0.04 | 1.48 ± 0.53 | 0.084 ± 0.03 | 5.16 ± 2.17 | <0.1 | 0.025 ± 0.019 |
| Shubat | 0.16 ± 0.16 | 1.57 ± 0.46 | 0.088 ± 0.02 | 7.212 ± 2.55 | <0.1 | 0.007 ± 0.005 |
Expressed in ppb (µg/kg) for the seven HMTc-relevant analytes measured in Sample (i):
- Pb (total) in camel milk: mean 25 ± 19 ppb; maximum 60 ppb at farm 4 (Shymkent region, near Kengtaw polymetal industry).
- Pb (total) in shubat: mean 7 ± 5 ppb; not detectable in two samples (farms 4 and 5).
- tAs in camel milk: <100 ppb (under analytical limit). Not separately speciated.
- tAs in shubat: <100 ppb.
- Cu in camel milk: 70 ± 40 ppb; in shubat 160 ± 160 ppb.
- Fe in camel milk: 1,480 ± 530 ppb; in shubat 1,570 ± 460 ppb.
- Mn in camel milk: 84 ± 30 ppb; in shubat 88 ± 20 ppb.
- Zn in camel milk: 5,160 ± 2,170 ppb; in shubat 7,212 ± 2,550 ppb.
Fodder maxima of note: Pb 34.90 ppm (34,900 ppb) at Kengtaw farm 4 near a polymetal industry; Fe 2,533 ppm at the same farm; Mn 82.1 ppm; Zn 94.3 ppm; As 1.6 ppm at the same farm. Water samples contained no detectable contaminants.
Sample (ii): 63 individual raw camel milk samples, 12 farms across 4 regions (Table 2)
Mean ± SD by region:
| Region | Pb (ppb) | Zn (ppm) | Cr (ppb) |
|---|---|---|---|
| Almaty | 245.94 ± 34.69 | 48.62 ± 1.52 | 46.91 ± 37.69 |
| Atyrau | 244.18 ± 83.66 | 59.90 ± 1.69 | 43.12 ± 21.03 |
| Aralsk | 249.20 ± 50.29 | 57.08 ± 2.21 | 38.12 ± 15.96 |
| Shymkent | 260.50 ± 35.00 | 52.13 ± 2.49 | 96.19 ± 56.68 |
Statistical notes from source: Shymkent Pb was significantly higher than the other three regions (P < 0.05). Shymkent Cr was significantly higher than the other three regions (P < 0.01). Zn did not differ significantly between regions. Different superscript letters in the source table indicate significance groupings.
Overall Sample (ii) means stated in the abstract and discussion text:
- Pb (total): 250 ± 56 ppb (mean of 63 samples), max 532 ppb at Atyrau.
- Zn: 5.42 ± 0.2 ppm (5,420 ± 200 ppb) overall.
- Cr (total): 59.5 ± 45.4 ppb overall. No reference values were available for camel milk; samples described as “widely higher than the tolerable values in cow milk.”
Internal inconsistency note (Sample ii Zn): The per-region Zn means in Table 2 (48.62–59.90 ppm) are approximately 10× higher than the overall mean stated in the abstract and discussion (5.42 ppm). The Pb and Cr columns in Table 2 are internally consistent with the abstract’s overall means, but the Zn column is off by an order of magnitude. This appears to be a units transcription error in Table 2 (most plausibly the Zn column should be in ppb, or the values should be divided by 10). Source values are reproduced verbatim above per HMI policy; downstream synthesis should treat the Sample (ii) regional Zn means as suspect and rely on the abstract/discussion value (5.42 ppm ± 0.2) for any aggregated use.
Regulatory and literature context cited by the authors
The authors cite the following milk-Pb tolerable values for context (not their own findings):
- Turkey: 20 ppb tolerable lead in milk.
- Germany and the Netherlands: 50 ppb normal/tolerable value.
- Kazakhstan national standard: 100 ppb.
- Kazakh standard for shubat: arsenic not to exceed 55 ppb.
Cow-milk Pb context cited from literature: 1.32 ppb (rural area); 49–67 ppb (industrial area); up to 844 ppb (near zinc and lead smelter); 320 ppb (close to road). Cow-milk As context cited from literature: 0.2 ppb (rural); 21.8–50 ppb (industrial and traffic); up to 684 ppb (industrial). The authors state that camel milk Pb in their Sample (ii) is “quite important and higher than the tolerable value” relative to the cited references.
Mercury (Hg, tHg, MeHg) was NOT measured. Cadmium (Cd) was NOT measured. Antimony (Sb), uranium (U), nickel (Ni), aluminium (Al), and tin (Sn) were NOT measured. Inorganic As (iAs) and hexavalent Cr (Cr-VI) speciation was NOT performed.
Methods (brief)
Sampling. Sample (i): 8 farms in 4 Kazakhstani regions selected for proximity to identified pollution sources — polymetal industries at Tekeli and Cary-Ozek (Almaty region), oil forages (Atyrau region), the Baikonur cosmodrome (Aral region), polymetal industries and phosphate manufacturing at Kengtaw and Aca (Shymkent region). Water, fodder, herd-mixed raw camel milk, and shubat (fermented camel milk) sampled per farm when available; shubat lacking at farms 1 and 6. Sample (ii): 63 individual raw camel milk samples from 12 different farms across the same 4 regions, still close to contaminating sources. Camel populations included dromedary (Camelus dromedarius), Bactrian (Camelus bactrianus), and hybrids.
Sample preparation. Wet-oxidation digestion: 10 mL milk sample combined with 65% nitric acid (Carlo Erba Reagents SA) in a ceramic capsule on a heating plate at 500°C for 4 hours, with additional nitric acid added to ashes and evaporated to complete destruction of organic matter. Final solution diluted to 10 mL with distilled water containing 1% nitric acid. Authors note the wet-oxidation method was preferred over dry-way to avoid loss of volatile elements. Fodder preparation used dry-way mineralisation with hydrofluoric acid silica removal.
Analysis. Inductively Coupled argon Plasma – Atomic Emission Spectrometer (ICP-AES), Varian Vista MPX – CCD, at CIRAD-UR Qualisud, Montpellier, France. Quantification by standard addition method using 11-point standard curves. Reference materials: AccuTrace Quality Control Standard #1 (Accu Standard) and Laboratory Performance Check Standard (AccuStandard). Detection limits: 0.01 ppm for all metals except As (0.1 ppm).
Statistics. Descriptive statistics (mean, SD) per metal per matrix. One-way ANOVA for region effect. An “index” I = x_ij / mean(x_i) was computed to compare matrix-to-matrix changes irrespective of absolute concentration. Pearson correlation for fodder–milk and milk–shubat relationships. Authors report no overall fodder–milk correlation but flag Cu (r = 0.724) and Mn (r = -0.725) in milk correlated with Zn in fodder, and Pb in milk correlated with Fe (r = 0.897) and Mn (r = 0.815) in fodder.
Limitations stated by authors. Budget constraint limited the analyte set in Sample (ii) to Pb, Cr, Zn. No cadmium or mercury measured anywhere. Detection limit for As (0.1 ppm = 100 ppb) was too coarse to characterize sub-100 ppb As in milk; the authors note that more sensitive colorimetric methods reported in the literature give cow-milk As in the 20–50 ppb range. Sample sizes per region (15–16 per region in Sample ii) limit power for inter-region comparisons.
Implications
Certification (HMTc): One of the first published occurrence datasets for raw camel milk, with Pb, tAs, Cr, Cu, Fe, Mn, Zn coverage in a defined Central Asian supply-chain context. The Sample (ii) Pb mean of 250 ± 56 ppb (max 532 ppb) exceeds every national tolerable value cited by the authors — Turkey 20 ppb, Germany/Netherlands 50 ppb, Kazakhstan 100 ppb — by 2.5- to 12.5-fold at the mean. This is relevant when assessing risk priors for any camel-milk-derived ingredient entering the HMTc program, particularly for products that aggregate raw milk from regions with co-located polymetal or oil-extraction industries. The shubat lead values being consistently lower than paired raw-milk values is a methodologically interesting observation for any future fermentation-as-mitigation discussion. The absence of Cd and Hg measurements leaves a literature gap for those analytes in camel milk that this paper does not fill.
Courses: Useful for teaching environmental-source-to-food-matrix sampling design (paired water/fodder/milk/fermented-product across pollution gradients), and for teaching the limitations of total-element ICP-AES for arsenic and chromium risk assessment (no speciation, As LOD too coarse for sub-100 ppb characterization). The Sample (ii) Zn units discrepancy (Table 2 values 10× higher than the abstract/discussion overall mean) is a worked example of why source values are preserved verbatim and flagged in verification notes rather than silently harmonized.
App: Provides occurrence priors for raw camel milk in a Kazakhstani supply-chain context for Pb (high, mean ~250 ppb with right tail to 532 ppb), tAs (<100 ppb but uncharacterized below that), Cr (mean ~60 ppb), and Zn (likely ~5 ppm overall per abstract; per-region Table 2 values should not be used). The Cu, Fe, Mn values from the smaller Sample (i) herd-mixed milk dataset (n=8) provide lower-confidence priors for the nutritional/trace elements. Camel milk from regions with co-located polymetal or oil-extraction activity should be treated as a meaningful Pb-exposure source in the app’s contamination-likelihood model.
Wiki pages this source may touch
Verification notes
The cite-key uses konuspayeva2010-camel-milk-kazakhstan to match the publication-year-on-cover (2009 issue, Volume 1, pages 112–118; received March 2009, accepted November 2009) while preserving the raw_handle filename verbatim (MFD_konuspayeva2010-camel-milk-heavy-metals). The “2010” in the handle reflects the discover skill’s filename heuristic; the journal record gives 2009. The year frontmatter is set to 2009 to match the journal record. Future referencers should be aware both years appear in different bibliographic systems for this paper.
Sample (ii) Zn units discrepancy. Table 2 per-region Zn means (48.62, 59.90, 57.08, 52.13 ppm) average to ~54.4 ppm, while the abstract and discussion text state an overall Zn mean of 5.42 ± 0.2 ppm. The Pb and Cr columns in Table 2 align with the abstract’s overall means within rounding, so this is not a global units-on-table issue but a probable single-column transcription error in the Zn column (most plausibly the values should be ppm/10, or the column is mislabeled and should be in some other unit). Source values are preserved verbatim above; the discrepancy is documented here so that no downstream synthesis silently averages 54 ppm Zn into a camel-milk prior. The abstract/discussion value of 5.42 ppm aligns with the Sample (i) herd-mixed milk Zn mean (5.16 ppm), so 5.42 ppm is the value to carry forward.
Brand firewall (Part 12). No brand names appear in the source. Sampling is aggregated to farm (F1–F8) and region (Almaty, Aral, Atyrau, Shymkent) level. Brand-firewall compliant. Method-vendor names reproduced under the methods exception (Varian Vista MPX – CCD, Carlo Erba Reagents SA, AccuTrace, AccuStandard).
Speciation absence. tAs and Cr values are total-element only. Any future use of these values in iAs or Cr-VI synthesis would require an explicit assumption about speciation fractions and should be flagged as inferred, not measured. The As LOD of 0.1 ppm (100 ppb) is too coarse to characterize sub-100 ppb As concentrations in milk; the authors themselves note that more sensitive methods would be needed to confirm the actual As level.
Cd and Hg absence. Cadmium and mercury (any form) were not measured in this study, despite being two of the four core HMTc dairy analytes. This is a corpus gap for Kazakh camel milk that remains open after this ingest.
Matrices vocabulary. shubat (fermented camel milk) and fodder are used as matrix slugs. shubat follows the precedent in pankiewicz2012 (kefir, buttermilk, cottage-cheese as matrix slugs without dedicated ingredient pages); fodder is included because the fodder data is part of the published table, although fodder is a feed matrix rather than a food matrix and would not normally route to product pages.
Sampling year. Manuscript received 31 March 2009, accepted 10 November 2009. Actual sample collection year is not stated in the paper; sampling_year_range left null.
Open Access licensing. The paper is explicitly Creative Commons Attribution Non-Commercial (CC BY-NC 3.0) per the footer of the published PDF. Bentham Open journal.
Audit subagent (general-purpose, fresh context, 2026-05-30) findings applied. Verdict PROMOTE with two ⚠️ concerns: (1) the Implications section originally cited “Codex Stan 193-1995 raw-milk Pb limit of 20 ppb” alongside the national tolerable values — the audit correctly noted that Konuspayeva et al. do not cite Codex, only Turkey/Germany/Netherlands/Kazakhstan national values. Verified against the source (pp.116–117): no Codex reference appears in the paper. Editorial Codex reference removed; statement now ties the comparison only to the national tolerable values the authors themselves cite. (2) Abstract-vs-Table 1 minor precision differences for Sample (i) shubat Cu (abstract 0.163 ± 0.164 vs Table 1 0.16 ± 0.16) and Zn (abstract 7.217 ± 2.55 vs Table 1 7.212 ± 2.55). Verified against the PDF: both differences are real paper-internal rounding/transcription variations. The wiki page uses Table 1 values per HMI policy of preferring source tables over abstract restatements; the differences are too small (≤0.005 absolute) to be load-bearing for any downstream synthesis.
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.