Quantification of heavy metals in Chinese jasmine flower cakes and risk assessment

Abstract

Jasmine flower cakes are particularly popular in China; however, they may contain heavy metals, posing health risks. We quantified heavy metals in jasmine flower cakes and assessed associated health risks. Al, Cr, Mn, Ni, Cu, As, Cd, and Pb in 40 samples were quantified via inductively coupled plasma mass spectrometry, revealing concentrations of 9.32, 0.012, 6.64, 0.296, 4.34, 0.002, 0.001, and 0.008 mg/kg, respectively. The detection rates were 100% except for Cd (82.50%) and As (67.50%). Cluster analysis categorised these metals into five groups: As/Cd, Al/Cr/Pb, Mn, Ni, and Cu. Although all samples exhibited mean health indices (HIs) < 1, nine samples (22.5%) exhibited HI > 1. The maximum HI was 2.261. The average estimated carcinogenic risk values were 27.00 × 10⁻⁶ across all samples, with risk ranking as Cd > As > Pb. Overall, the cumulative carcinogenic risk of the cakes is within an acceptable range; however, the risk levels of certain batches warrant further attention. The findings provide a scientific basis for evaluating the safety of jasmine-flavoured cake consumption and establishes a reference analytical framework for risk screening and regulatory strategies of other traditional specialty foods. This supports high-quality development of food industry while safeguarding public health.

Introduction

Jasminum sambac (L.) Aiton is a jasmonium plant in the Luteolaceae family. It is native to tropical and subtropical regions and is widely cultivated in several countries, such as India, Iran, Egypt, Turkey, and China¹. In the food industry, it is used as a natural flavouring agent owing to its unique aroma. It is often used in pastry production, including Jasmine flower cakes. This traditional pastry, primarily made from jasmine flowers and grains, is highly favoured by consumers. Since 1982, the consumption of refined grains in China has increased. From 1985 to 2013, the market share and contribution of refined rice, wheat flour and their products have continued to rise².

Jasmine flower cakes are nutrient-rich. However, owing to the presence of heavy metals, they may also pose health risks. With rapid economic development, China’s industrialisation has resulted in significant energy consumption and severe environmental pollution^3,4. Within the soil–cereal system, diet constitutes the primary route of human exposure to heavy metals, serving as the main source of heavy metal intake^5,6. The contamination of food with heavy metals has raised considerable public concern regarding food safety in China. The overall heavy metal contamination rate of agricultural soils in China is 10.18%, with Cd, Hg, Cu, and Ni accounting for the majority. Approximately 13.86% of grain production is affected by heavy metal contamination in agricultural soils, underscoring the critical role of soil as a major source of heavy metals in crops and the importance of addressing it for food safety^7,8.

Heavy metals, characterised by high atomic weight and mass, are integral components of the natural environment⁹. Even at relatively low concentrations, heavy metal ions interact with biomolecules, leading to irreversible chronic or acute intoxication¹⁰. For example, Hg, Cd, and Mn are known to impair brain development, cognitive function, and motor skills, thereby potentially contributing to neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease¹¹. Pb is an inorganic toxic metal with no biological function in humans, it can interfere with metabolic processes and cause chronic neurotoxicity, posing risks to both ecosystems and human health¹². In particular, Pb exposure is associated with impaired intellectual development in children¹³. Inorganic arsenic exposure can lead to serious health consequences, including skin diseases, cardiovascular diseases, neurological disorders, and an increased risk of cancer¹⁴.

In China, the consumption of Western-style convenience foods, such as bread and biscuits, has increased dramatically due to rapid socioeconomic changes and increased average incomes¹⁵. As a representative product within this category, jasmine flower cake has also experienced rapid market growth. However, to date, no studies have examined the heavy metal content of jasmine flower cakes or the associated health risks. The ICH Q3D Guideline for Elemental Impurities classifies inorganic As, Pb, Cd, and Hg as Class 1 (highly toxic, requiring stringent control), Ni as Class 2A (relatively high toxicity, requiring focused attention), and Cu and Cr as Class 3 (lower relative toxicity, often requiring attention). Elements such as Mn, Al, Fe, and Zn are categorised under “Other”¹⁶. Although these elements exhibit lower toxicity, their established oral reference doses (RfD) are as follows: 0.14 mg/kg-day for Mn; 0.43 mg/kg-day for Al; 0.7 mg/kg-day for Fe; 0.3 mg/kg-day for Zn^14,17. Considering that the production of jasmine flower cakes may involve the addition of Al-based leavening agents, which can contribute to elevated Al levels¹⁸, and that Mn has a comparatively lower RfD, indicating a higher relative risk, both Al and Mn were included in the assessment. However, as jasmine flower cakes are baked goods requiring high-temperature processing, Hg, while a recognised contaminant, is highly volatile upon heating. Most Hg is released at temperatures of approximately 170 °C during volatilisation¹⁹. Therefore, Hg was excluded from the analytical panel.

In this study, we selected eight elements—Al, Cr, Mn, Ni, Cu, As, Cd, and Pb—to prioritise contaminants of highest potential risk in jasmine flower cakes. Currently, for baked goods, Chinese national standards stipulate maximum level requirements only for Al and Pb. Internationally, the Codex Alimentarius Commission General Standard for Food Additives (CXS 192-1995) sets a limit for Al. Neither the Codex Alimentarius Commission (CAC) nor the European Union has established specific maximum levels for the remaining seven elements in baked goods^20–22. To address this gap, we collected 40 samples of commercially available jasmine flower cakes from across China and quantified the levels of Al, Cr, Mn, Ni, Cu, As, Cd, and Pb via inductively coupled plasma mass spectrometry (ICP-MS), followed by a comprehensive health risk assessment (Table 1).

Table 1.

Sample information.

Batch	Batch number	Origin	Batch	Batch number	Origin
S1	20250510	Yunnan	S21	20250427	Fujian
S2	20250513	Yunnan	S22	20250425	Fujian
S3	20250508	Yunnan	S23	20250425	Fujian
S4	20250509	Yunnan	S24	20250417	Fujian
S5	20250511	Yunnan	S25	20250516	Fujian
S6	20250507	Yunnan	S26	20250516	Fujian
S7	20250514	Yunnan	S27	20250514	Fujian
S8	20250503	Yunnan	S28	20250425	Zhejiang
S9	20250502	Yunnan	S29	20250510	Zhejiang
S10	20250513	Yunnan	S30	20250426	Zhejiang
S11	20250506	Yunnan	S31	20250506	Jiangsu
S12	20250404	Yunnan	S32	20250121	Guangxi
S13	20250507	Yunnan	S33	20250506	Guangxi
S14	20250424	Yunnan	S34	20250427	Shandong
S15	20250225	Fujian	S35	20250215	Shandong
S16	20250316	Fujian	S36	20250506	Sichuan
S17	20250510	Fujian	S37	20250108	Sichuan
S18	20250505	Fujian	S38	20250507	Hebei
S19	20250508	Fujian	S39	20250513	Anhui
S20	20250428	Fujian	S40	20250501	Beijing

Results and discussion

Heavy metals in Jasmine flower cakes

Validation of the analytical method

The limit of detection (LOD), limit of quantification (LOQ), recovery rates, and relative standard deviations (RSDs) of the analytical method are summarised in Table 2. All elements demonstrated excellent linear relationships with correlation coefficients (r) ranging from 0.9996 to 1.0000. The recovery rates for all elements were between 86.34 and 107.84%. The precision was evaluated by calculating RSD, which ranged from 0.74 to 4.26%. The LOD and LOQ ranges for ICP-MS were 0.0004–0.9 and 0.001–2.9 ng/mL, respectively. The established method satisfies the requirements of this study.

Table 2.

Validation parameters for heavy metal analysis methods.

Element	r	Linear range (ng/mL)	LOD (ng/mL)	LOQ (ng/mL)	Recovery (%)	RSD (%)
Al	1.0000	0–5000	0.9	2.9	107.84	0.74
Mn	1.0000	0–500	0.02	0.06	91.28	2.17
Cu	0.9997	0–500	0.007	0.02	91.11	1.69
Cr	1.0000	0–50	0.004	0.01	98.77	0.76
Ni	1.0000	0–50	0.01	0.03	98.55	2.40
As	1.0000	0–50	0.002	0.008	103.59	1.33
Cd	1.0000	0–50	0.0004	0.001	105.44	4.26
Pb	0.9996	0–50	0.001	0.004	86.34	1.21

LOD: limit of detection, LOQ: limit of quantification, RSD: relative standard deviations.

Heavy metal concentrations

The eight target heavy metals exhibited varied concentrations in the 40 batches of jasmine flower cakes (Tables 3 and 4). The detection rates for Al, Cr, Mn, Ni, Cu, and Pb were 100%, whereas those for Cd and As were 82.50% and 67.50%, respectively. The average concentrations were as follows: Al, 9.32 mg/kg; Mn, 6.64 mg/kg; Cu, 4.34 mg/kg; Ni, 0.296 mg/kg; Cr, 0.012 mg/kg; Pb, 0.008 mg/kg; As, 0.002 mg/kg; Cd, 0.001 mg/kg. The coefficients of variation of the eight heavy metals ranged from 41.7 to 150.0%. The concentrations of Al, Mn, and Cu were significantly higher than those of the other five elements, with concentration ranges of 1.57–38.5, 2.09–18.5, and 2.09–18.6 mg/kg, respectively. At present, China’s national standards set maximum limits only for Al and Pb in baked goods. According to GB 2762–2022, the maximum allowable concentration for Pb is 0.5 mg/kg, and according to GB 2760–2024, that for Al is 100 mg/kg (based on a dry sample). Internationally, the CXS 192–1995 stipulates the same limit for Al as the Chinese national standard. However, neither the CAC nor the European Union has established specific maximum levels for the remaining seven elements in baked goods. The concentrations of Al and Pb in all 40 samples were both below the corresponding Chinese and international standard limits.

Table 3.

Heavy metal concentrations in Jasmine flower cakes (mg/kg).

NO	Al	Cr	Mn	Ni	Cu	As	Cd	Pb
S1	22.8	0.018	7.55	0.306	3.90	0.002	0.001	0.009
S2	38.5	0.017	5.86	0.511	3.66	0.003	0.002	0.016
S3	9.73	0.012	6.88	0.232	3.01	0.002	0.002	0.006
S4	8.82	0.010	6.92	0.209	2.31	0.001	0.003	0.006
S5	7.25	0.017	7.72	0.266	2.94	0.001	0.002	0.009
S6	6.37	0.009	6.21	0.217	2.41	–	0.001	0.014
S7	13.3	0.005	5.17	0.178	2.09	0.001	0.001	0.008
S8	7.45	0.012	7.98	0.244	2.69	0.002	0.002	0.011
S9	8.10	0.008	7.71	0.223	2.70	0.002	0.002	0.008
S10	12.3	0.022	6.85	0.363	3.08	0.003	0.002	0.013
S11	7.43	0.009	7.41	0.211	2.59	0.001	0.001	0.012
S12	11.2	0.018	6.28	0.248	2.92	0.003	0.002	0.013
S13	8.13	0.012	8.05	0.258	2.55	0.002	0.001	0.007
S14	14.9	0.014	4.91	0.215	2.53	0.002	0.002	0.008
S15	14.9	0.022	9.16	0.703	4.54	0.003	0.001	0.017
S16	17.2	0.014	5.78	0.422	3.14	0.001	0.002	0.053
S17	4.07	0.010	7.99	0.320	6.83	0.010	0.005	0.006
S18	9.76	0.021	8.62	0.363	6.24	0.012	0.005	0.017
S19	19.2	0.013	18.5	0.194	6.40	0.001	0.001	0.006
S20	16.5	0.015	12.4	0.284	4.25	0.002	0.002	0.008
S21	2.73	0.006	6.09	0.550	3.84	–	0.001	0.006
S22	5.17	0.010	3.57	0.204	3.28	–	0.001	0.003
S23	5.52	0.014	5.01	0.283	3.09	–	0.001	0.004
S24	22.9	0.015	3.45	0.254	2.22	0.006	–	0.006
S25	8.01	0.009	2.09	0.203	5.31	–	–	0.003
S26	4.09	0.009	2.23	0.334	18.6	–	0.001	0.004
S27	2.13	0.006	3.74	0.163	2.45	0.004	–	0.008
S28	2.16	0.010	5.27	0.875	4.00	–	–	0.003
S29	11.1	0.012	12.7	0.277	2.62	–	0.001	0.004
S30	2.34	0.008	4.10	0.206	9.53	0.005	0.002	0.003
S31	7.39	0.010	4.79	0.199	2.41	0.001	0.002	0.005
S32	5.35	0.005	7.08	0.191	2.70	–	0.002	0.003
S33	1.57	0.013	5.06	0.257	2.37	0.001	–	0.004
S34	4.40	0.010	3.73	0.153	2.63	0.001	0.001	0.002
S35	6.81	0.012	6.02	0.230	3.90	–	0.002	0.009
S36	4.04	0.008	4.64	0.273	2.57	–	0.002	0.004
S37	8.45	0.013	14.5	0.260	3.50	0.001	0.002	0.004
S38	3.31	0.005	3.24	0.461	9.67	–	–	0.003
S39	4.17	0.006	6.57	0.233	15.2	–	–	0.007
S40	3.58	0.021	3.63	0.246	2.71	–	0.001	0.003

“–” indicates that the element was not detected in the sample.

Table 4.

Concentration ranges of heavy metals in Jasmine flower cakes (mg/kg).

Index	Al	Cr	Mn	Ni	Cu	As	Cd	Pb
Maximum	38.5	0.022	18.5	0.875	18.6	0.012	0.005	0.053
Minimum	1.57	0.005	2.09	0.153	2.09	0.000	0.000	0.002
Median	7.44	0.012	6.15	0.251	3.04	0.001	0.001	0.006
Mean	9.32	0.012	6.64	0.296	4.34	0.002	0.001	0.008
SD	7.27	0.005	3.27	0.146	3.45	0.003	0.001	0.008
CV (%)	78.0	41.7	49.2	49.3	79.5	150.0	100.0	100.0

SD and CV represent the standard deviation and coefficient of variation, respectively.

Characteristics of heavy metal associations

To investigate the interrelationships among heavy metals in jasmine flower cakes, a stepwise multivariate statistical analysis was performed. The results from Pearson correlation analysis, hierarchical cluster analysis (HCA), and principal component analysis (PCA) collectively revealed consistent elemental association patterns, providing a data-driven foundation for further exploration of potential pollution sources (Fig. 1).

Fig. 1.

Sampling distribution map. Base map source was sourced from the Resource and Environment Science and Data Center, Chinese Academy of Sciences (https://www.resdc.cn). Sampling sites were plotted using ArcGIS 10.8 (https://www.esri.com).

Pearson correlation analysis (Table 5) indicated significant positive correlations among several heavy metals. The strongest correlation was observed between As and Cd (r = 0.664, P < 0.01), suggesting a highly synchronous variation pattern of these two elements in the samples. Furthermore, significant positive correlations (P < 0.05) were found between Al and Cr, Al and Pb, and Cr and As. These linear associations imply that the corresponding elemental pairs may be influenced by certain common factors.

Table 5.

Correlations among heavy metals in Jasmine flower cakes.

	Al	Cr	Mn	Ni	Cu	As	Cd	Pb
Al	1	0.486**	0.285	0.130	− 0.185	0.186	0.136	0.396*
Cr		1	0.250	0.241	− 0.210	0.342*	0.291	0.319*
Mn			1	− 0.003	− 0.110	0.102	0.298	0.070
Ni				1	0.113	0.035	− 0.080	0.233
Cu					1	0.036	− 0.040	− 0.115
As						1	0.664**	0.187
Cd							1	0.220
Pb								1

*Indicates significant correlation (P < 0.05), **indicates extremely significant correlation (P < 0.01).

R-mode clustering (squared Euclidean distance) grouped the eight metals into five clusters (Fig. 2): (Ⅰ) As–Cd, (Ⅱ) Al–Cr–Pb, (Ⅲ) Mn, (Ⅳ) Ni, and (Ⅴ) Cu. Notably, the clustering results were highly consistent with the correlation analysis results: the most strongly correlated pair, As and Cd, formed an independent cluster (Cluster Ⅰ), while Al, Cr, and Pb grouped together into Cluster Ⅱ. This stable elemental association pattern indicates that heavy metal contamination in jasmine flower cakes is non-random and exhibits distinct “fingerprint” characteristics.

Fig. 2.

Dendrogram of R-type hierarchical clustering for heavy metals.

PCA was employed to reduce data dimensionality and quantify the contribution of each association pattern to the overall variance. The dataset satisfied the standard criteria for PCA suitability (Kaiser–Meyer–Olkin measure (KMO) > 0.5, Bartlett’s test P < 0.05). The first three principal components (PCs) cumulatively explained 63.28% of the total variance (Table 6). The loading matrix of each principal component closely corresponded to the cluster grouping identified above. PC1 (31.471% variance contribution), with high loadings (> 0.64) for Al, Cr, Ni, and Pb, primarily represents Cluster Ⅱ (Al–Cr–Pb) and Ni. PC2 (17.071% variance contribution) was dominated by As and Cd (loadings 0.874 and 0.900, respectively), corresponding precisely to Cluster Ⅰ. PC3 (14.746% variance contribution) was mainly associated with Mn and Cu, corresponding to Clusters Ⅲ and Ⅴ.

Table 6.

First three principal components of heavy metal analysis in Jasmine flower cakes.

Heavy metals	Component 1	Component 2	Component 3
Al	0.642	0.095	0.451
Cr	0.645	0.305	0.327
Mn	0.119	0.284	0.524
Ni	0.715	− 0.139	− 0.416
Cu	− 0.055	0.149	− 0.775
As	0.170	0.874	− 0.061
Cd	0.050	0.900	0.137
Pb	0.683	0.140	0.091
Eigenvalue	2.518	1.366	1.179
Variance contribution (%)	31.471	17.071	14.736
Cumulative variance contribution rate (%)	31.471	48.542	63.278

Significant values are in bold.

Risk assessment

Non-carcinogenic risk assessment

The hazard quotient (HQi) and hazard index (HI) results for heavy metals are summarised in Table 7 and Fig. 3. The average HQi values for all batches of jasmine flower cakes were below 1, indicating that the potential health risks posed by the individual heavy metals were relatively low. Notably, significant variations were observed in the Cu concentrations among the different batches. In four batches (S26, S30, S38, and S39), the HQi values for Cu exceeded the safety threshold of 1, with values of 2.127, 1.089, 1.106, and 1.741, respectively. In these same batches, the cumulative risk from all metals, expressed as the HI, also exceeded unity, with values ranging from 1.090 to 2.261. This indicates that the non-carcinogenic risk in these samples was driven not only by elevated Cu levels but also by the additive contribution of other metals. In particular, batch S26 exhibited both the highest HQi for Cu (2.127) and the highest HI (2.261), highlighting a pronounced combined risk scenario. It should be noted that this assessment is based on the dose-additivity assumption and does not account for potential toxicological interactions among the metals.

Table 7.

Exposure assessment and non-carcinogenic risk evaluation.

No	HQi								HI
	Al	Cr	Mn	Ni	Cu	As	Cd	Pb
S1	0.365	0.014	0.123	0.017	0.446	0.017	0.003	0.005	0.991
S2	0.615	0.013	0.096	0.029	0.418	0.022	0.005	0.009	1.208
S3	0.156	0.009	0.112	0.013	0.344	0.014	0.005	0.003	0.656
S4	0.141	0.008	0.113	0.012	0.264	0.006	0.006	0.003	0.553
S5	0.116	0.013	0.126	0.015	0.336	0.011	0.005	0.005	0.627
S6	0.102	0.007	0.101	0.012	0.276	0.000	0.003	0.008	0.509
S7	0.213	0.004	0.084	0.010	0.239	0.005	0.003	0.004	0.563
S8	0.119	0.009	0.130	0.014	0.308	0.013	0.004	0.006	0.604
S9	0.129	0.006	0.126	0.013	0.309	0.016	0.005	0.004	0.609
S10	0.196	0.017	0.112	0.021	0.352	0.022	0.004	0.008	0.731
S11	0.119	0.007	0.121	0.012	0.297	0.011	0.003	0.007	0.577
S12	0.179	0.014	0.103	0.014	0.334	0.023	0.004	0.008	0.678
S13	0.130	0.009	0.132	0.015	0.292	0.015	0.003	0.004	0.600
S14	0.237	0.010	0.080	0.012	0.289	0.012	0.004	0.005	0.649
S15	0.237	0.016	0.150	0.040	0.520	0.021	0.002	0.010	0.997
S16	0.274	0.010	0.094	0.024	0.359	0.010	0.004	0.030	0.806
S17	0.065	0.007	0.131	0.018	0.781	0.074	0.011	0.004	1.090
S18	0.156	0.016	0.141	0.021	0.713	0.088	0.013	0.010	1.157
S19	0.308	0.010	0.302	0.011	0.732	0.006	0.003	0.004	1.376
S20	0.263	0.011	0.202	0.016	0.486	0.017	0.005	0.004	1.006
S21	0.044	0.004	0.099	0.031	0.439	0.000	0.002	0.003	0.624
S22	0.083	0.008	0.058	0.012	0.375	0.000	0.003	0.002	0.540
S23	0.088	0.010	0.082	0.016	0.353	0.000	0.002	0.002	0.553
S24	0.367	0.012	0.056	0.015	0.254	0.044	0.000	0.004	0.751
S25	0.128	0.007	0.034	0.012	0.607	0.000	0.000	0.002	0.789
S26	0.065	0.007	0.036	0.019	2.127	0.000	0.003	0.002	2.261
S27	0.034	0.005	0.061	0.009	0.281	0.030	0.000	0.004	0.424
S28	0.035	0.008	0.086	0.050	0.457	0.000	0.000	0.002	0.638
S29	0.177	0.009	0.207	0.016	0.300	0.000	0.002	0.002	0.714
S30	0.037	0.006	0.067	0.012	1.089	0.037	0.004	0.002	1.254
S31	0.118	0.007	0.078	0.011	0.275	0.010	0.005	0.003	0.508
S32	0.086	0.004	0.116	0.011	0.309	0.000	0.004	0.002	0.531
S33	0.025	0.010	0.083	0.015	0.271	0.006	0.000	0.002	0.410
S34	0.070	0.007	0.061	0.009	0.300	0.007	0.002	0.001	0.458
S35	0.109	0.009	0.098	0.013	0.446	0.000	0.004	0.005	0.685
S36	0.065	0.006	0.076	0.016	0.294	0.000	0.004	0.002	0.463
S37	0.135	0.010	0.237	0.015	0.400	0.009	0.004	0.002	0.812
S38	0.053	0.004	0.053	0.026	1.106	0.000	0.000	0.001	1.243
S39	0.067	0.005	0.107	0.013	1.741	0.000	0.000	0.004	1.937
S40	0.057	0.016	0.059	0.014	0.310	0.000	0.002	0.002	0.460
Mean	0.149 ± 0.116	0.009 ± 0.004	0.108 ± 0.053	0.017 ± 0.008	0.496 ± 0.394	0.014 ± 0.019	0.003 ± 0.003	0.005 ± 0.005	0.801 ± 0.396

HI: Hazard Index, HQi: Hazard quotient.

Fig. 3.

Hazard quotient (HQi) and hazard index (HI) results of heavy metals in jasmine flower cakes.

Regarding the overall health risks arising from multi-metal exposure, the average HI value across all samples was below 1. However, nine samples (S2, S17, S18, S19, S20, S26, S30, S38, and S39) exhibited HI values greater than 1, accounting for 22.5% of the total batches. This finding demonstrates that the cumulative risk from multi-metal exposure can reach levels at which adverse health effects cannot be excluded, even when individual metal risks are low, underscoring the importance of assessing aggregate exposure.

Carcinogenic risk assessment

The results of the carcinogenic risk (CR) and cumulative CR assessments are summarised in Table 8 and Fig. 4. Among the 40 samples analysed, the average calculated carcinogenic risk (CCR) was 27.00 × 10⁻⁶. Four samples exhibited a CCR < 1 × 10⁻⁶, accounting for 10% of the total; 35 samples exhibited a CCR between 1 × 10⁻⁶ and 1 × 10⁻⁴, accounting for 87.5% of the total; and one sample exhibited a CCR > 1 × 10⁻⁴, accounting for 2.5% of the total. Further analysis revealed that the average CR values for Cd, As, and Pb were 20.69 × 10⁻⁶, 6.15 × 10⁻⁶, and 0.16 × 10⁻⁶, respectively, with CR ranking as Cd > As > Pb. Thirty-six samples had a CCR greater than 1 × 10⁻⁶, with Cd and As identified as the dominant contributors to CR. These high-risk samples were obtained from various regions of China, including Yunnan Province, Zhejiang Province, Guangxi Zhuang Autonomous Region, Hebei Province, Anhui Province, and Beijing Municipality.

Table 8.

Carcinogenic risk assessment and cumulative carcinogenic risk evaluation.

No	CR (× 10−6)			CCR (× 10−6)
	As	Cd	Pb
S1	7.44	17.43	0.18	25.05
S2	9.80	32.02	0.31	42.13
S3	6.45	29.24	0.12	35.81
S4	2.70	36.83	0.11	39.64
S5	4.75	33.16	0.18	38.09
S6	0.00	15.34	0.27	15.62
S7	2.03	17.66	0.15	19.84
S8	5.94	23.93	0.21	30.08
S9	7.12	32.46	0.15	39.74
S10	9.86	23.68	0.26	33.80
S11	5.03	18.66	0.23	23.92
S12	10.35	25.07	0.26	35.68
S13	6.97	18.52	0.13	25.62
S14	5.26	21.39	0.16	26.81
S15	9.52	14.80	0.34	24.65
S16	4.47	23.57	1.04	29.08
S17	33.15	66.60	0.12	99.87
S18	39.53	76.45	0.33	116.31
S19	2.89	19.17	0.12	22.18
S20	7.83	30.68	0.15	38.67
S21	0.00	11.92	0.12	12.03
S22	0.00	15.95	0.06	16.01
S23	0.00	11.12	0.07	11.19
S24	19.99	0.00	0.12	20.11
S25	0.00	0.00	0.06	0.06
S26	0.00	18.40	0.08	18.48
S27	13.47	0.00	0.15	13.61
S28	0.00	0.00	0.06	0.06
S29	0.00	11.41	0.08	11.50
S30	16.66	24.36	0.07	41.08
S31	4.62	29.32	0.10	34.04
S32	0.00	24.04	0.06	24.11
S33	2.51	0.00	0.07	2.58
S34	3.27	13.52	0.04	16.82
S35	0.00	23.55	0.17	23.72
S36	0.00	26.64	0.08	26.72
S37	4.18	25.74	0.07	29.99
S38	0.00	0.00	0.05	0.05
S39	0.00	0.00	0.13	0.13
S40	0.00	14.92	0.05	14.97
Mean	6.15 ± 8.57	20.69 ± 15.81	0.16 ± 0.16	27.00 ± 22.47

CR: carcinogenic risk, CCR: calculated carcinogenic risk.

Fig. 4.

Carcinogenic risk (CR) and calculated carcinogenic risk (CCR) results of heavy metals in jasmine flower cakes.

Overall, the cumulative CR associated with jasmine flower cakes was within acceptable public health limits (< 1 × 10⁻⁴). However, CCR varied substantially among batches, with one sample exceeding 100 × 10⁻⁶. This heterogeneity indicates that, while the population-level risk under typical consumption scenarios is low, the risk associated with specific high-concentration batches may be elevated, underscoring the importance of quality control to mitigate batch-to-batch variability. Specifically, the average CR value for Cd was approximately 3.4-fold higher than that for As and 129-fold higher than that for Pb, making Cd the primary contributor to CR. The contribution of Pb to CR is minimal and may be considered negligible. Therefore, management of CRs associated with jasmine flower cake should primarily focus on Cd and As. Compared with previous studies, our study reveals a distinct heavy metal risk contribution pattern that differs from those reported for common food products. For instance, Taher and Abojassim (2024) reported that Zn was the primary CR contributor in commercially available biscuits, followed by Cd and As²³. Similarly, Li et al. (2022) conducted a heavy metal risk assessment in rice from five regions in Guizhou Province of China and found that the average CR values for As and Pb exceeded the upper limit of the acceptable range (1 × 10⁻⁴); in their study, approximately 88.90% of CR was attributed to As²⁴.

In contrast, our study identifies Cd as the predominant contributor to CR. These findings provide insights into the heavy metal risk profiles of traditional specialty foods. Notably, Cd exhibited substantial variability (SD = 15.81), indicating marked heterogeneity in its concentration across the different samples. This high dispersion suggests potential differences in the soil contamination levels at the sites of raw material cultivation. For example, natural Cd enrichment in soils from regions such as Yunnan and Guangxi may have contributed to the elevated Cd levels^25,26.

Limitations and conclusions

While this study provides a systematic assessment of heavy metals in commercially available jasmine flower cakes, several limitations should be acknowledged. First, the samples were collected from the market at a single time point (May 2025). As processed food products, the composition of jasmine flower cakes may vary with production batches, raw material sourcing cycles, and inventory turnover. Thus, our findings primarily reflect products available at that specific time and may not fully capture variations across different production periods. Second, although samples were sourced from ten producing provinces in China, their geographical distribution was uneven (e.g. a larger proportion from Yunnan Province). This imbalance may affect the generalisability of the average concentrations and risk estimates to all regional products. Finally, while the sample size (n = 40) was sufficient to identify major contamination patterns and exceedance batches, it remains insufficient for reliably characterising the distribution of low-frequency, high-concentration contamination events.

Furthermore, several methodological considerations should be considered when interpreting the results. First, the investigation was confined to the analysis of the final food product. While multivariate statistical analysis revealed distinct elemental association patterns (e.g. As–Cd or Al–Cr–Pb), no environmental samples (e.g. cultivation soil or irrigation water) or processing-chain samples were analysed. Consequently, the identified patterns, while valuable as contamination fingerprints, cannot be definitively attributed to specific sources such as soil geochemistry or industrial processing without further dedicated source-apportionment studies. Second, the dietary exposure assessment relied on pasta consumption data as a proxy, which may overestimate the long-term average intake for this specialty food and thus introduces uncertainty into the absolute risk values; however, the relative risk conclusions remain robust. Third, the health risk assessment was conducted based on the total concentration of arsenic and chromium, without including an analysis of the elemental forms. Because the toxicity of these elements depends strongly on their chemical forms, the actual risk may be lower than the reported value. Future studies should therefore include arsenic and chromium speciation analysis to obtain more accurate data^27,28. Finally, the risk assessment conclusions were derived using a standard adult body weight and daily intake rate. As both the model and its parameters are designed for the standard adult population, the findings are therefore applicable only to this specific demographic group.

Within the scope of this market survey and based on the applied risk assessment model, Jasmine flower cakes generally comply with national standards for individual heavy metals. Al and Mn exhibited the highest average concentrations in jasmine flower cakes; however, due to their relatively high health-based guidelines, their estimated risk level after intake dose normalisation was relatively low, and they did not constitute key risk drivers in this study. In contrast, although the average concentrations of Cd and As were extremely low, they played a dominant role in health risk assessment, primarily due to carcinogenic slope factors. Even in trace amounts, these elements still contribute to the main carcinogenic risk index under a conservative risk assessment framework.

In summary, this study is the first to combine ICP-MS technology, chemometric analysis, and health risk assessment models to systematically investigate the contamination characteristics and health risks of heavy metals in jasmine flower cakes. The results indicate that the average carcinogenic risk associated with heavy metals in jasmine flower cakes is within acceptable limits, although certain batches exhibit elevated non-carcinogenic risks. These findings suggest a shift in regulatory focus from overall compliance toward the targeted identification and dynamic control of high-risk batches. This study provides robust data support and a scientific basis for the quality and safety evaluation of jasmine flower cakes, and offers an analytical framework applicable to other traditional specialty foods. This contributes to safeguarding public health while promoting the high-quality and sustainable development of the food industry.

Methods

Sample collection

In May 2025, a total of 40 samples of jasmine flower cakes were collected from different production enterprises and batches across the Chinese market. All the samples were sourced from 32 distinct production enterprises, including 10 provinces such as Yunnan, Fujian, Zhejiang, Jiangsu, Guangxi, Shandong, Sichuan, Hebei, Anhui, and Beijing. The sampled enterprises were diverse in scale, encompassing both large-scale, well-established food manufacturers with high brand recognition and numerous small- and medium-sized enterprises. Detailed information on the samples is presented in Table 1 and Fig. 1.

Reagents and instrumentation

Multi-element standard solutions for As, Cd, Cr, and Pb (GNM-M041940-2013; batch number: 24D80143; concentration: 1000 µg/mL) were sourced from the National Analysis and Testing Center for Nonferrous Metals and Electronic Materials. The Al standard solution (GBW30019-1000-NC-20; batch number: B1906197; concentration: 1000 µg/mL) was sourced from Tanmo Quality Inspection Standard Material Center. The Mn standard solution (CHAA-B-025120; batch number: 2427227; concentration: 1000 µg/mL) was obtained from Shanghai Anpon Cuisi Standard Technology Service Co., Ltd. The Ni standard solution (GBW(E)080128; batch number: 22011; concentration: 100 µg/mL) was sourced from the National Institute of Metrology, China. The Cu standard solution (GBW(E)082780; batch number: B22070053; concentration: 1000 µg/mL) was obtained from Tanmo Quality Inspection—Standard Material Center. An internal standard solution containing Li, Sc, Ge, Rh, In, Tb, Lu, and Bi (mixed internal standard solution; batch number: 54-124CRY2; concentration: 100 mg/L) was purchased from Agilent Technologies. Analytical grade nitric acid was obtained from MERCK (Merck KGaA, Darmstadt, HE, Germany). Deionised water was employed in all experiments, prepared using an Aike AK-RO-UP-200 Pure Water System.

Analyses were performed on an Agilent 7700X inductively coupled plasma mass spectrometer (Agilent Technologies Inc.) following sample digestion with an Anton Paar Multiwave 3000 microwave digestion system (Anton Paar GmbH). An MS-204 (Mettler Toledo) analytical balance with a precision of 0.00001 g was employed for sample weighing.

Heavy metal analysis

Sample digestion and preparation

The samples were thoroughly crushed and homogenised. Precisely 0.5 ± 0.0001 g of each sample was placed in a digestion vessel, followed by the addition of a 65% nitric acid solution (7.5 mL). The vessel was sealed and left undisturbed for 24 h, after which the following digestion procedure was performed:

• Stage 1: Ramp from room temperature to 130 °C over 5 min and hold for 5 min.

• Stage 2: Ramp from 130 to 150 °C over 5 min and hold for 5 min.

• Stage 3: Ramp from 150 to 190 °C over 5 min and hold for 25 min.

After cooling, the vessels were removed, and the lids were carefully opened to release any residual gases. The inner lids were rinsed with a small amount of water, and the digested samples were evaporated to remove any excess acid. The resulting digested solutions were transferred to 25-mL volumetric flasks. The digestion vessels were washed thrice with a small amount of water, and the washings were combined in the volumetric flasks and brought to volume. The solutions were then thoroughly mixed. A blank control was prepared using the same procedure without the addition of any sample.

Preparation of mixed standard solutions

Mixed standard solutions were prepared to cover a range of concentrations suitable for constructing calibration curves for the analysis of heavy metals in the jasmine flower cake samples. The standard solutions were therefore diluted using a 2% nitric acid solution to prepare mixed calibration solutions with the following concentrations: 100, 500, 1000, 3000, and 5000 ng/mL for Al; 10, 50, 100, 300, and 500 ng/mL for Mn and Cu; 1, 5, 10, 30, and 50 ng/mL for Cr, Ni, As, Cd, and Pb, respectively.

Analysis and method validation

By including parameters such as radio frequency (RF) power, extraction voltage, ion lens settings, sampling depth, gas flow rates, and integration time, the methodology is ensured to be transparent and reproducible. The validation process described here helps to establish the sensitivity, precision, and robustness of the method, which are critical for credible scientific research.

ICP-MS detection conditions

The detection conditions for the target metal elements using ICP-MS were as follows: RF power: 1400 W; extraction voltage: − 142 V; ion lenses 1, 2, and 3: 2.2, − 23, and − 173.5 V, respectively; sampling depth: 62 mm; nebuliser chamber temperature: 2 °C; plasma gas flow rate: 15 L/min; auxiliary gas flow rate: 1 L/min; carrier gas flow rate: 1 L/min; dilution gas flow rate: 1 L/min; number of replicates: 3; integration time: 0.3 s; tuning mode: He mode.

Matrix effects were corrected by the online addition of an internal standard solution. Spectral interferences were corrected by employing a He collision mode and the following interference correction equations:

1

2

3

Method validation

The method was validated in terms of linearity, precision, LOD, LOQ, and recovery rates. To determine the LOD and LOQ, blank samples were analysed 10 times, and the SD value was calculated. The LOD and LOQ were three and ten times the SD, respectively.

Risk assessment

The health risk assessment in this study adopted a screening-level, conservative approach. The objective was to provide an initial, health-protective estimate of potential risk by employing a series of precautionary assumptions regarding exposure and toxicity. Under these conservative conditions, if risks are deemed acceptable, they would be expected to be lower under more realistic exposure scenarios.

Dietary exposure

To evaluate dietary exposure and potential health risks, dietary exposure was estimated using the following equation:

4

where C is the concentration of heavy metals in the sample (mg/kg), dIR is the average daily intake of food (kg/day), and BW is the body weight (kg).

According to the World Health Organisation (WHO) guidelines for risk assessment, when data on consumption of specific foods are unavailable, it is reasonable to use consumption data from similar foods (e.g. cereals and their products). In the absence of specific survey data on Jasmine flower cakes consumption and the high similarity between Jasmine flower cakes and traditional pasta products in terms of primary ingredient composition (mainly wheat flour) and dietary usage, the dIR value of pasta products was adopted as the exposure assessment parameter for Jasmine flower cakes. Based on the data from the China National Center for Nutrition and Health Monitoring, the dIR value for this food category is 0.1372 kg/day, and the BW value is 60 kg²⁹.

Notably, specific, survey-based consumption data for jasmine flower cakes are currently unavailable. Given that the primary ingredient of jasmine flower cakes is wheat flour, similar to many pasta and bakery products, the daily intake rate for pasta products (dIR = 0.1372 kg/day) was adopted as a reasonable proxy for this assessment. This approach is consistent with common practice in dietary exposure studies of specialty grain-based foods. We acknowledge that this assumption may represent a conservative (health-protective) overestimate of long-term average daily intake for this occasionally consumed food item, and this uncertainty was considered in the interpretation of the risk estimates.

Non-carcinogenic risk assessment

The non-carcinogenic risk was assessed using standard procedures established by the United States Environmental Protection Agency (US-EPA) to evaluate the HQi and HI values. HQi quantifies the exposure risk of individual heavy metals, indicating the relative contribution of each metal to non-carcinogenic risk. It is calculated as follows:

5

where RfD is the daily oral reference dose (mg/kg/day), based on technical reports from US-EPA. The RfD values for Al, Cr, Mn, Ni, Cu, As, Cd, and Pb are 1.43 × 10⁻¹, 3 × 10⁻³, 1.4 × 10⁻¹, 4 × 10⁻², 2 × 10⁻², 3 × 10⁻⁴, 1 × 10⁻³, and 4 × 10⁻³ mg/kg/day, respectively. However, because multiple heavy metals are typically present in food simultaneously, evaluating a single metal alone may underestimate the actual risk. The HI is the sum of all HQi values of the heavy metals assessed in the non-carcinogenic risk evaluation, affording a combined exposure risk from multiple heavy metals, and is calculated as follows:

6

An HI value ≤ 1 indicates no significant adverse effects. HI > 1 indicates potential non-carcinogenic adverse effects, whereas HI ≥ 10 indicates a higher likelihood of chronic toxic effects¹⁷. The combined approach of evaluating both HQi and HI enables evaluation of both individual and cumulative heavy metal exposure risks, providing a more comprehensive scientific basis for the health risk assessment of jasmine flower cakes.

Carcinogenic risk assessment

The CR value assesses the probability of cancer development in populations exposed to heavy metals for long periods. For heavy metals that are more likely to pose potential CRs (e.g. Pb, As, and Cd), the CCR value, used for CR characterisation, is determined using the following equation:

7

According to the technical reports from US-EPA, a CCR value < 1 × 10⁻⁶ indicates a low CR, while 1 × 10⁻⁶ ≤ CCR ≤ 1 × 10⁻⁴ indicates a moderate CR, which is within an acceptable range for carcinogenic levels. CCR > 1 × 10⁻⁴ indicates a high CR, warranting attention and further investigation³⁰.

Statistical analysis

All statistical analyses were performed using SPSS (version 25.0; IBM Corp., Armonk, NY, USA) and Origin 2021 (OriginLab Corporation, Northampton, Massachusetts, USA). Descriptive statistics (mean, SD, median, and range) were calculated for heavy metal concentrations.

Pearson correlation analysis was conducted to examine the linear relationships among the eight heavy metals. The significance of r was tested using a two-tailed test. Correlations were considered significant at P < 0.05 and highly significant at P < 0.01.

Multivariate statistical analyses, including HCA and PCA, were employed to explore the associations and potential sources of heavy metals.

For HCA, the square Euclidean distance was used as the similarity measure, and the average linkage method was applied to perform R-mode clustering on the metal variables, thereby generating a dendrogram.

For PCA, the suitability of the data for dimensionality reduction was first assessed using the KMO measure and Bartlett’s test of sphericity. A KMO value > 0.5 and a significant Bartlett’s test (P < 0.05) were considered acceptable. Components with eigenvalues > 1 were retained (Kaiser’s criterion). Varimax rotation was applied to facilitate interpretation of the principal components. Heavy metals with absolute loading values > 0.5 on a given component were considered significant contributors to that component.

Author contributions

Conceptualisation, Gang Deng and Liqun Shen; methodology, Jianjun Huang and Shuang Liang; sample analysis, Jianjun Huang and Guoyi Wu; validation, Jianjun Huang; data curation, Yanyun Liu and Shaodong Liang; project administration, Gang Deng and Liqun Shen; writing—original draft, Jianjun Huang; writing—review and editing, Shuang Liang, Guoyi Wu, Shaodong Liang, Yanyun Liu, Gang Deng, and Liqun Shen. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the General Project of Guangxi University of Traditional Chinese Medicine (2022MS026).

Data availability

Data is contained within the article; further inquiries can be directed to the first author.

Declarations

Competing interests

The authors declare no competing interests.