Introduction

Lead is a highly toxic substance especially impacting children’s health1. The neurological and behavioral effects of lead are considered to be irreversible1,2. Young children are particularly susceptible to lead poisoning. Children absorb up to 4 times better lead compared to adults, their organ systems, especially their cognitive system is still developing and is negatively affected1. In 2021, the World Health Organization (WHO) published guidelines on management of lead exposure and recommended that a blood lead level of 5 µg/dL in children should trigger an intervention, although it should be noted that there is no known safe level of lead in blood3. In 2020, Pure Earth and UNICEF estimated that one in three children globally suffer from elevated blood lead levels above 5 µg/dL, especially in lower income countries4. The Toxic Truth Report highlighted a number of consumer goods that can contribute to lead poisoning, many of which were included in the Rapid Market Screening (RMS) study4.

There are multiple lead exposure pathways, from consumption of lead-contaminated dust near industrial hotspots, to ingestion of lead-contaminated paint chips to lead-adulterated spices2,5,6,7,8. Prior studies have identified lead contamination in a variety of consumer products such as paints, ceramics, spices, foodware, traditional medicines, and cosmetics1,2,8,9,10,11,12,13. However, the geographic variability and overall distribution of lead in these potential exposure sources have not been adequately characterized, particularly in low- and middle-income countries (LMICs). Given that lead exposure sources can vary considerably by location, it is important to identify local sources of lead exposure, especially for young children and particularly within their homes. Local exposure source assessments are important allowing for more accurate interventions that are designed specifically to target local priority sources, allowing for greater effectiveness and efficiency than interventions that are designed without local data.

Therefore, the aim of this study was to systematically analyze products in a range of LMICs for lead. Ultimately, we wanted to determine which product types are more likely to contain lead and how lead concentrations vary around the world. More specifically, the research questions were as follows:

  • What is the geographic distribution of lead concentrations among product types identified as containing lead in previous studies in the selected countries?

  • How do the lead concentrations in each product type compare to available regulatory standards or health guidelines?

Methods

The research team developed a “Rapid Market Screening” protocolled by some of the authors (EN, AS, GB, JF), former Pure Earth Senior Director of Programs Petr Sharov and other Pure Earth staff with external expert review prior to implementation (see Supplement A). The protocol describes the information to collect about each market, vendor, and item, and includes standardized analytical approaches for each product type.

Country selection

The overall goal was to select 25 geographically diverse LMICs, including at least one country from each of the six World Bank-classified regions that represent the majority of LMICs: Africa, Middle East and North Africa, East Asia and the Pacific, South Asia, Europe and Central Asia, and Latin America and the Caribbean. Within each of these regions, candidate countries were listed and weighted for inclusion based on several factors:

  • Prioritizing countries with evidence of a high prevalence and/or severity of childhood lead poisoning14. For this purpose, an extensive literature review was performed.

  • Within each global region, balancing a mix of countries with and without evidence of lead-tainted products (or countries that are hypothesized to sell these products based on cultural similarities).

  • Priority was given to countries in which Pure Earth employees or its partners have capacities and in which there is sufficient geopolitical stability to be able to work there safely.

The countries were evaluated on the basis of the above-mentioned characteristics and ultimately selected by the project team in a discourse. Consequently, a total of 25 countries, including India (Fig. 1), were chosen for the study. However, recognizing the vast size and diversity of India, the sampling strategy focused on three geographically distinct states within the country. As such, the study covered 27 locations, including Armenia; Azerbaijan; Bangladesh; Bolivia; Colombia; Egypt; Georgia; Ghana; the Indian states of Maharashtra, Tamil Nadu, and Uttar Pradesh; Indonesia; Kazakhstan; Kenya; Kyrgyzstan; Mexico; Nepal; Nigeria; Pakistan; Peru; the Philippines; Tajikistan; Tanzania; Tunisia; Turkey; Uganda; and Vietnam).

Figure 1
figure 1

Countries in which the study was conducted.

Full size image

The cities were samples were selected were:

  • Ghana – Cape Coast, Koforidua, Kumasi

  • Maharashtra (India) – Kolhapur, Nagpur, Pune

  • Uttar Pradesh (India) – Gaziabad, Lucknow, Prayagraj

  • Tamil Nadu (India) – Kancheepuram, Namakkal, Tiruvallur

  • Indonesia – Makassar, Medan, Surabaya

  • Bangladesh – Barisal, Khulna, Rajshahi

  • The Philippines – Sebu, Davao, Mabalacat, San Fernando

  • Colombia – Barranquilla, Bogota, Bucaramanga, Cali

  • Tajikistan – Bokhtar, Dushanbe, Khudjand, Kulyab

  • Kyrgyzstan – Bishkek, Cholpon-Ata, Osh

  • Kazakhstan – Saryagash, Shymkent, Turkestan

  • Georgia – Mestia, Tbilisi, Zugdidi

  • Armenia – Gyumri, Vagharshapat, Vanadzor

  • Mexico – Cuautia, Huazulco, Puebla, Toluca

  • Peru – Callao, Cusco, Lima, Pucallpa

  • Tanzania – Arusha, Dar Es Salaam, Mwanza

  • Bolivia – Cochabamba, La Paz, Oruro

  • Egypt – Alexandria, Cairo, Giza

  • Kenya—Kisumu, Mombasa, Nairobi

  • Tunisia—Sfax, Tataouin, Tunis

  • Nigeria—Abuja, Lagos, Port Harcourt

  • Uganda—Kampala, Lira, Mbarara

  • Pakistan – Karachi, Lahore, Rawalpindi

  • Nepal – Biratnagar, Kathmandu, Neplagunj

  • Azerbaijan – Baku, Ganja, Imishli

  • Vietnam – Da Nang, Dong Ha, Nam Dinh

  • Turkey – Ankara, Bursa, Eskisehir

Product selection

The general types of products sampled were selected following a series of global desk assessments that reviewed literature on lead concentrations in a variety of products in LMICs14,15,16,17. An initial list of product types and a sample desk assessment were then provided to the researchers in the countries. The initial list was designed to examine common items across geographies, but also allowed for flexibility based on local contexts. The researchers chose any relevant items for their countries, supporting their point with references (both published and unpublished), and carried out initial product screening. Based on this formative research, eleven product types were selected for analysis: ceramic foodware, metal foodware, plastic foodware, cosmetics, toys, paints intended for large surfaces, paints for art and crafts, spices, sweets, staple dry foods, and traditional and herbal medicines.

Product sampling

Sample collectors in each country selected in at least three or four geographically diverse cities, and selected, when possible, at least one wholesale bazaar and one retail market in each city to purchase items. Within each market, sample collectors interviewed the sellers and purchased items from at least two vendors. There are different probabilities of lead contamination between whole, unprocessed spices and processed or powdered spices as well as between loose and packaged spices. We opted for the loose processed spices to reduce the likelihood of getting an invalid result due to technical detection issues if there was indeed a problem. Loose processed spices were usually more homogeneous in composition compared to whole or unprocessed spices. The homogeneity reduced the risk of variability and led to more consistent samples. Samples were collected between September 2022 and May 2023.

Product analysis

The primary screening tools used for the study were portable X-ray fluorescence analyzers (XRF). The portable XRF machine is a valuable screening device for this type of study because it is easy to use, field portable, and gives immediate results that are comparable to laboratory data. Once an XRF device has been purchased, there are few consumable costs, with the exception of ancillary supplies like gloves, bags and labels. The protocol outlined how to prepare each item for testing and specified the number of XRF readings required for each item. In every country except Bangladesh, samples were analyzed with a Thermo Scientific Niton XL3T XRF, using the “Test All” mode, which is designed for consumer goods. In Bangladesh, an Olympus Vanta Series C was used in “Soil” mode, with a subset of those items later also tested with a Thermo Scientific Niton XL3T for confirmation purposes and uniformity of data acquisition. The testers were instructed in the provided protocols and in on-line training sessions to regularly check the accuracy of the XRF using the provided "standard" samples with known lead concentrations. Investigators were instructed to check calibration standards at the beginning of the day, anytime the instrument has been off for 30 min, and at the end of each session. All data collected by the investigators was entered into a central database using the platform SurveyCTO. Subsets of samples from each country were sent to Pure Earth’s Headquarters in New York for quality control (see Supplement B) which consisted of confirmatory testing of 354 representative samples by a certified laboratory and spot checking of field XRF results using the provided protocols and Thermo Scientific Niton XL3T XRF. Local versus New York XRF measurements generally compared well except in the cases of sample heterogeneity (e.g. toys with different colored paints). New York XRF and laboratory data also indicated issues with the field XRF readings from Tajikistan and Kazakhstan, which limited the data from these countries to samples that were sent to New York. Literature shows a XRF and laboratory devices results correlate well in case of soil and sediment samples, however there is not much research done in consumer items18,19. For certain product types, especially powdered spices, XRF lead levels correlated strongly with laboratory results, with correlation coefficients (R2) ranging from 0.70 to 0.98 for plastic toys, cosmetics, plastic foodware and spices. For other product types, however, the agreement between XRF values and laboratory results is not well documented and may depend on the form and preparation of the sample. Literature shows that XRF and laboratory devices results do well correlate in case of soil and sediment samples, however there is not much research done in consumer items20.

Statistical methodology

The values were determined using the XRF; they were recorded in Survey CTO, and the statistical analysis was conducted in Microsoft Power BI. Minimum, maximum, quartile and median levels of lead in various products were determined. There was a significant portion of the XRF results below the XRF’s lower limit of detection (LOD). LOD refers to the minimum concentration of an element detectable with reasonable confidence. It signifies whether an element is present or absent, without guaranteeing the accuracy of the obtained value. Typically, the LOD is quantified as three standard deviations (or 1.5 times the reported measurement error) observed in a sample containing either none or only a trace amount of the analyte. For this reason, and since the data was skewed to the left, we do not report mean, but median values. For product types that were expected to be largely homogenous, such as ground spices or paint samples, researchers were requested to take only one reading. For product types with the potential for a high level of heterogeneity, such as toys, ceramics, which could have multiple colors glazes, or metal foodware, which could have components like rivets, researchers were requested to take between 3 and 5 readings. Due to this potential for heterogeneity and the lack of certainty of how each reading would contribute to the exposure profile for that item, the highest of the readings for a given item was used in the analysis. Furthermore, where there are existing regulations for the items included in the study, these regulations apply to all components of an item, rather than an average.

Reference values

To provide context to the concentrations of lead found in the various products, a “reference level” for each product type was selected. These reference levels serve as thresholds indicating where United Nations (UN) agencies including WHO or particularly well-resourced regulatory authorities have established public health guidance, a level of concern, or a regulatory limit for lead in each product class. Although countries often have unique standards for lead concentrations in different products, a uniform reference level for each product type was applied for this study to facilitate comparisons across countries21. The inclusion of uniform reference levels is not a suggestion that any one guidance value or regulatory standard is superior to any other, or that concentrations below the reference levels are safe. Rather, the reference levels used here are simply an attempt to contextualize concentrations and highlight particularly concerning results. We selected existing regulatory standards and guidance values promulgated by UN agencies, the European Union (EU), and the United States (US), prioritized in that order. We could not identify existing reference levels for total lead in foodware (items used to cook, serve, consume, and store food). While standards for leachable lead from foodware exist, field testing of leachable lead in foodware was not possible. We engaged in a substantial effort to test the degree to which lead leaches from metal foodware (mostly aluminum) with a variety of lead concentrations under various cooking scenarios. Results of leachate testing are unpublished as of February 2024, but are being prepared for publication. For this assessment a reference level of 100 ppm of total lead for all types of foodware was applied. This reference level for total lead is not based on an existing regulatory standard, and lead doses per use likely vary between product types of foodware, and indeed between individual products. The lead dose per use is likely also affected by the type of food prepared, the method and duration of cooking, and other contextual factors. In the absence of any available standards for total lead content for these product types of foodware, we used the US Consumer Product Safety Commission total lead standard for “children’s products'' (also used in this study for the product type of toys). The use of 100 ppm total lead as a conservative threshold is supported by leaching tests on over 100 metal foodware items, which indicated that pots with total lead below 100 ppm leached less than the World Health Organization (WHO) drinking water standard of 10 ppb (17 out of 18 pots tested)22.

Reference levels and measured lead levels in this study are expressed in parts per million (ppm), which is equivalent to milligrams per kilogram (mg/kg). The following reference values used in this study are listed in Table 1.

Table 1 Reference level per product type in ppm.
Full size table

Results

This study analyzed a wide variety of consumer products and materials. Samples included both small batch, informally produced items, as well as large-scale, commercially produced items. There are 382 shopping venues where sampling took place (markets, shopping areas, malls, stores). For metal foodware, total lead levels in 51% of the 520 samples collected were above the reference level of 100 ppm. In 17 locations (countries or, in the case of India, states), even median levels exceeded the reference level and in 9 locations the maximum level exceeded 10,000 ppm (see Supplement D). Of the items in the metal foodware product type that were found to be above the reference level, 69% were pots and pans, 17% were vessels for food or water not intended to be exposed to direct heat, and 14% were cooking utensils. Across all item types, 57% of the items found to be above the reference level were reported to be made of or labeled as aluminum or aluminum alloys. For 35% of the items, we were not able to determine the metal composition based on the item description or label. Other metal types—including brass, copper, and iron alloys—made up the remaining 8% of samples found to be above the reference level. Leachate testing was conducted on more than 100 pots from 25 countries and 5 regions in the US. Lead levels above 100 ppm were common across all regions, indicating the potential to leach lead above the 10 ppb WHO drinking water standard. Notably the 100 ppb threshold could result in a blood lead level of around 0.5 ug/dL based on the US EPA’s IEUBK model, assuming daily intake of 250 g of food at this lead concentration. For ceramic foodware, high lead levels were common across all regions (see Supplement D), with 45% of 308 samples above reference level of 100 ppm. In 11 locations, the median sample exceeded the reference level, suggesting that contaminated items are common. In 25 of the study locations (all but Pakistan and Uttar Pradesh State, India), the maximum lead level was more than 10 times the reference level (see Supplement D). Out of 364 plastic foodware samples, 12% showed lead levels exceeding the reference level of 100 ppm. Unlike ceramic and metal foodware, for which many countries had samples with maximum concentrations above 10,000 ppm, all samples of plastic foodware were below 3300 ppm.

For paints, high lead concentrations were prevalent in new paints even among countries that have already adopted a 90 ppm limit (see Supplement D). The paint samples were divided into two product types: paints intended for use on large surfaces, such as interior and exterior walls, and paints intended for crafts, art, and other specialty uses. This division was based on the recognition that exposure pathways may be different between wall paints, where exposure likely results from deteriorating paint that becomes dust, and specialty paints, where exposure may be more directly related to the application of the paint or use of the painted product (e.g. a toddler getting art paints in the mouth or mouthing a painted toy). Note that we were not able to determine the primary purpose of all paint samples collected, and therefore we have included an additional “unclassified” product type category. We also note that the protocol for testing paint was amended during the RMS study to specify testing only dried paint samples as opposed to allowing analysis of wet samples. In total, 41% of the 437 samples of paint for large surfaces were above reference level of 90 ppm. Out of 70 samples of paint for crafts, art, and other specialty uses, 11% showed lead levels exceeding 90 ppm. Among 102 paint samples not classified, 47% showed lead concentrations exceeding 90 ppm. More than half of the study locations (14 of 27) had maximum lead concentrations exceeding 10,000 ppm, while 10 locations had samples exceeding 20,000 ppm (see Supplement D).

For toys, high lead levels above the reference level of 100 ppm were found in 13% of 781 samples tested. This product type encompasses a variety of hard toys, composed primarily of plastic items, but also including metal, wood or other materials. Some of these toys were also found to have paint or coatings on them, with lead concentrations either in the paint of the underlying substrate. In addition to the variety at the product type level, many toys were heterogeneous, made from a combination of materials. We found many toys to contain internal electronic or metal parts, which were responsible for some of the highest lead readings observed by XRF. Such readings may not necessarily best reflect the potential exposure risk for that item, as the reference level relates to “accessible parts” to children.

For cosmetics, 12% of the 812 samples were above reference level of 2 ppm, across many subcategories. The highest concentrations were reported in traditional eyeliners (see Supplement C). Cosmetics with elevated lead levels were found in 21 of the 25 countries (see Supplement D).

The two items with the highest lead concentration were both eyeliners, known as kajal or kohl, from Pakistan. These samples had concentrations of 637,600 ppm (64%) and one million ppm (100%) lead as assessed by XRF, with lower but still extremely high concentrations (29% and 32%) reported by confirmatory laboratory testing. In some cultures, kajal/kohl is applied to infants and children. The item with the third highest lead concentration of lead (128,400 ppm) was face paint intended specifically for children. Among the samples with elevated lead levels, the most common item was nail polish (29 items, maximum lead concentration of 6,751 ppm), followed by lipstick (15 items, maximum lead concentration of 42,350 ppm), and eyeshadow (13 items, maximum lead concentration of 974 ppm). As noted above and in the Quality Control section (see Supplement B), some deviations were observed between the XRF and lab-based measurements of lead concentration at the highest concentrations among the cosmetics. Nevertheless, at such extreme concentrations, the risk is still significant even with a wide margin of error.

For spices, 2% of 1,084 samples were above reference level of 2 ppm. The highest concentrations were found in turmeric and blends (see Supplement C).

Less than 5% of herbal/traditional medicine (4%), sweets (3%) and staple dry food (1%) were above reference levels (for details see Supplement C).

In summary, out of a total of 5,007 product samples from 25 countries, 913 samples had concentrations of lead exceeding the relevant reference level based on XRF readings, representing 18% of all samples. Metal foodware, ceramic foodware, and paints most frequently exceeded the relevant reference levels (see Table 2).

Table 2 Distribution of lead concentrations and the distribution of highly lead-tainted samples above the reference level across the 11 product types, aggregated for all 25 countries.
Full size table

In all 25 countries where this assessment was performed consumer products were identified that exceeded at least some reference levels (see Fig. 2).

Figure 2
figure 2

Percentage of samples above reference level by country.

Full size image

The Table 3 is organized by country and shows percentages of samples of each product and food product type that exceeded the relevant reference level.

Table 3 Percent of samples exceeding reference level per product type and by country.
Full size table

In Table 4 the regional variation of the results is shown. There are regional trends detectable.

Table 4 Distribution of lead concentrations and the distribution of highly lead-tainted samples above the reference level analyzed by region.
Full size table

Discussion

Previous studies have highlighted elevated levels of total and leachable lead in metal foodware made in LMICs, particularly in inexpensive aluminum foodware11,12,31,32. These pots are generally light, inexpensive, and have good conductivity, which helps conserve fuel usage. Such foodware has previously been found to be made from mixed scrap metal from engine parts, radiators, electronic appliances, and aluminum cans31. In all countries, total lead levels were above the reference level of 100 ppm; in South East Asia 70% of the samples were above this level (see Table 4).

The leachable lead from these pots represents an exposure source through ingested food cooked in these pots. Pure Earth has conducted leachability testing of more than 100 aluminum foodware samples to improve our understanding of the allocation of leachable lead in LMICs and potential doses of lead per use. The results of the leachate investigations in aluminum foodware are currently being prepared for publication. Of the 102 pots tested, 45% had lead concentrations in the leachate exceeding 10 µg/L which could result in a blood lead level in children above 0.47 μg/dL based on modeling conducted using IEUBK. The highest leachate concentration found was 2900 μg/L in a pot from Indonesia, a concentration that could result in a BLL of about 45 μg/dL. The pots with the highest leachate concentrations were from South and Southeast Asia. We found a high frequency (45%) and relatively wide geographic distribution of lead in ceramic foodware (see Table 4). Challenges regarding the use of lead-based glazes have been well-documented in Mexico9,33,34, and have been identified elsewhere35, but the RMS study shows a fairly uniform geographic distribution of contaminated items. This does not necessarily mean that these pieces all contribute to exposure equally. As with all forms of foodware, a high lead concentration on exterior surfaces does not tell us how much lead is leaching into food. The type of glaze, temperature in which it is fired, types of food prepared or served, and ways in which the item is used can all affect leachability and thus exposure36. Ceramics with high concentrations of lead were not limited to handmade, artisanal, or traditional pieces, but included mass-produced pieces that may have been imported to the country where they were purchased. The leachability of lead from various ceramic glazes produced and used under different conditions is an area that requires further research that was beyond the scope of this study.

As of January 2024, 48% of countries in the world had legally binding controls on lead concentrations in new paints37. Many of these have adopted regulations based on a model law establishing a maximum lead concentration of 90 ppm26. However, many of the paint samples analyzed through the RMS study that exceeded 90 ppm were collected from countries that have a 90 ppm regulatory limit. In eight of the countries (Colombia, Georgia, Kenya, Kyrgyzstan, Mexico, Pakistan, Philippines, Vietnam) and Indian States (Maharashtra, Tamil Nadu Uttar Pradesh) were such regulations apply, more than 10% of paint samples had lead concentrations above 90 ppm (details see Supplement C). This suggests a considerable enforcement gap in these locations.

In cosmetics, in addition to high lead levels in kajal/kohl, elevated lead levels were found in other traditional products, including henna and kumkum (a red powder made of turmeric and other ingredients and used for social and religious purposes in India)38. Lead levels above the reference level were also found in a variety of conventional cosmetics, such as nail polish, lipstick, and eyeshadow as previously described, as well as face powder, mascara, eyeliner, liquid foundation/concealer, and hair products23.

Previous studies have identified elevated lead levels in certain spices from countries around the Mediterranean, the Caucasus, and South Asia, among other39,40,41,42,43. In several countries, prior programs have confirmed that elevated lead concentrations were the result of producers adding lead-based pigments to spices to make their colors brighter8. The RMS study was not designed to focus specifically on countries known to have contaminated spices, nor to focus solely on the types of spices that have been identified as more often contaminated. Rather, the RMS study includes a broad range of spice types from countries that were selected based on product-agnostic criteria. As a result, the findings generally show low levels of lead in spices. The minimum detection level for the XRF is often between 2 to 4 ppm for spices, and thus it is possible that some samples had a reading of “non-detect” but exceeded the reference level of 2 ppm.

To our knowledge, this screening effort encompasses sampling from the widest range of consumer product types and geographies to date. Still, there are limitations in sample size for any given product type due to budgetary and logistical constraints. In some instances, there are only a few samples of a certain product type for a specific country. This study includes samples from 25 countries, which were selected at convenience. From each country, mainly three or four geographically diverse major cities were selected. The authors do not purport that these findings can be considered definitive for a country or region or globally, but they provide indications of potential sources of concern and hotspots geographically where supply chains do not effectively regulate for lead contamination.

Conclusion

This screening effort revealed elevated lead levels among multiple product types and across diverse geographies. Total lead in metal foodware exceeded the 100 ppm reference value in more than half of the metal foodware tested, with samples above the reference level in all 25 countries in this study. The situation is similar with total lead content in ceramic foodware. In both cases, it is the amount of lead leaching out of these items into food that represents a potential exposure source. XRF analysis of foodware for total lead is relatively cheap and simple, especially compared to heated leachate testing and associated laboratory costs for analysis with limited resources in LMICs. Further studies on leachability—especially of metal and ceramic foodware—are required to establish usable total lead reference values. A transfer factor for lead from foodware to food would allow more LMICs to better protect consumers from lead exposure from foodware.

The problem of lead paint is very well known, but as our results show existing national and international measures to eliminate lead in paint are not sufficient. In this study, we bought new paints on the markets 41% of which contained lead above the threshold value of 90 ppm. This means that lead-based paints are still being sold and used today to paint large surfaces in homes, exposing more generations to lead in the air and dust that is harmful to the development of the nervous system. Better lead paint laws need to be enacted and/or regulations enforced. There is also a need to analyze the supply chains for cosmetics to determine where these high lead cosmetics come from and how they are traded to end up in local markets.

With so many consumer products in 25 countries containing significant amounts of lead, there is a clear global public health impact as lead from these products exposes pregnant women, infants, children, adolescents and adults. Few LMICs conduct large surveys or ongoing monitoring of children’s blood lead levels. The result is that there is little visibility into the prevalence, severity, and geographic distribution of lead poisoning for many countries. This study highlights the allocation of potential lead exposure sources and the importance of blood lead level surveillance and related source apportionment to prioritize product- and non-product lead exposure sources and their elimination in low- and middle income-countries.