We collected the data from original research articles published between January 2010 and September 2024 using the following six academic databases: Web of Science, ScienceDirect, Springer, Google Scholar, Wiley Online Library, and the China National Knowledge Infrastructure (CNKI). The search strategy employed the keywords (“cadmium” OR “Cd”) AND (“straw return” OR “straw incorporation”) AND (“rice” OR “paddy”). Additionally, the reference lists of relevant publications were manually screened to identify any studies that had not been captured by the initial search. Publications qualified for inclusion if they fulfilled the subsequent standards: (1) The research focused exclusively on rice, excluding other crops (e.g., wheat, rapeseed), vegetables (e.g., radish, cabbage, lettuce), or hyperaccumulator plants. (2) The experimental design incorporated a control group (no straw addition) and a treatment group with straw return, each replicated at least three times. (3) Agronomic procedures including rotation schedules, nutrient applications, and water management remained uniform between experimental treatments. (4) Rice was cultivated for one complete growing season following straw application, excluding studies based on microcosm soil incubation experiments. (5) The available Cd in soil samples was extracted through calcium chloride (CaCl₂), ethylenediaminetetraacetic acid (EDTA), or diethylenetriaminepentaacetic acid (DTPA). When measurements were reported over multiple years, data for each year were recorded, and for studies with multiple growth stages in a single season, only data at maturity were collected. Date in figures were extracted using the freeware digitizing software WebPlotDigitizer (version 4.6) (Burda et al., 2017). When only standard errors were available, we converted them to standard deviations using SD = SE × √n. Where neither measure was provided, we estimated variability by applying the mean coefficient of variation calculated across all studies to the reported means (Van Groenigen et al., 2017; Qian et al., 2020). The data selection process is illustrated in Supplementary Figure S1. Of the 2,383 articles initially identified, only 35 met the inclusion criteria, with the study locations depicted in Supplementary Figure S2.

To evaluate the potential driving factors influencing the response to straw return, the data were classified into several subcategories: (1) Initial soil pH: Categorized as strongly acidic (<5.5), acidic (5.5–6.5), neutral (6.5–7.5), and alkaline (>7.5). (2) Initial SOM content: Classified into two levels: low (≤30 g kg^-1) and high (>30 g kg^-1). (3) Initial soil total Cd concentration: Based on the Chinese GB15618-2018 standard for soil Cd pollution risk control, categorized as < 0.85 mg kg^-1 (risk control for rice potted plants), 0.85–3.53 mg kg^-1, and >3.53 mg kg^-1 (strict risk control for field rice). Note that only one publication reported a total Cd concentration below 0.3 mg kg^-1, limiting the representativeness of the lowest concentration category. (4) Background Cd concentration in straw: Divided into three levels: <0.3, 0.3–0.8 and >0.8 mg kg^-1. (5) Experimental type: Field and pot experiments. (6) Straw type: Classified as rice, wheat, rapeseed, and Solanum L. (7) Amount of straw returned: Categories as < 3,500, 3,500–7,000, and >7,000 kg hm^-2. For pot experiments, the straw amount was calculated by multiplying the proportion of straw returned by the soil weight per hectare (2.25 × 10⁶ kg) (Nie et al., 2021). (8) Duration of straw returned: Categorized as < 0.5, 0.5–1, and >1 year.

We quantified straw return impacts on soil available and total Cd, grain/straw/root Cd concentrations, soil pH, and SOM using natural logarithm-transformed response ratios (RR) following Equation 1: RR = ln X t ¯ X c ¯ = ln   X t ¯ − ln   X c ¯ (1)

Here, Xt and Xc denote treatment and control means for straw incorporation versus removal treatments. Response ratio variance (v) was determined using Equation 2: v = S D t 2 n t X t 2 + S D c 2 n c X c 2 (2) where SDt ² and SDc ² represented the squared standard deviations, and nt and nc were the number of experiments conducted for the straw return treatment and the control group, respectively. A categorical random-effects model was employed to combine the effect sizes of individual study into a weighted mean response ratio (RR₊₊) as described in Equation 3: RR + + = ∑ R R i × w i / ∑ w i (3)

Here, RRi and wi denote the individual effect magnitudes and corresponding weights for each included study. Statistical significance of differences between subcategories was determined when the 95% confidence interval (CI) of the pooled RR++ excluded unity. Straw return effects were considered statistically meaningful if the 95% CI did not encompass zero (p < 0.05). The confidence interval calculation followed Equation 4: 95   %   C I = RR + + ± 1.96 1 / ∑ w i (4)

Heterogeneity tests (Qm) evaluated how variables responded differently to available and total soil Cd across treatment groups following straw incorporation. A significant Q _m value (p < 0.05) indicated that the explanatory variable (categorical factor) has a significant effect on the response ratios. For clearer interpretation the response of variables to soil available Cd and soil total Cd, we converted the weighted response ratio (RR₊₊) into a percentage form by the formula in Equation 5: Percentage change = exp RR + + − 1 × 100   % (5)

The statistical analyses were performed using the R package metafor (v. 4.3.1) with the escalc and rma. mv functions (Viechtbauer, 2010). Robustness checks—including the Egger test, Fail-Safe N calculation, and funnel plot analysis—were also conducted in R (see Supplementary Table S1; Supplementary Figure S3).

Linear regression and mixed-effects meta-regression models were employed to investigate relationships between response ratios (RRs) for soil available and total Cd versus soil pH, SOM, plant tissue Cd concentrations (grain, straw, root), and initial soil characteristics (Viechtbauer, 2010). The glmulti package in R (Calcagno and Mazancourt, 2010) was used to evaluate the relative importance of multiple factors controlling Cd concentrations in soil fractions (available and total) and plant tissues (grain, straw, root). Predictor variables included initial soil properties (pH, SOM, total Cd), straw type, and response ratios of soil pH and SOM following straw return. Factors with summed Akaike weights ≥0.8 across all candidate models were considered most influential (Terrer et al., 2016; Van Groenigen et al., 2017).

Straw return significantly increased soil available Cd by 12.7% (95% CI: 8.0%–16.7%) and soil total Cd by 7.7% (95% CI: 3.5%–12.0%) compared with no-straw controls. In addition, notable increases were observed in soil pH (1.8%, 95% CI: 0.9%–2.7%) and soil organic matter (SOM) (12.3%, 95% CI: 8.0%–16.7%). These changes were paralleled by corresponding elevations in Cd concentrations across various rice tissues (Figure 1). Robustness assessments—including the Egger test, Fail-Safe N calculation, and funnel plot analysis—revealed no evidence of publication bias (Supplementary Table S1; Supplementary Figure S3). As anticipated, significant residual heterogeneity was observed in the random-effects meta-analysis for both the soil available Cd (I ² = 94.93%, Qt = 1007.475, p < 0.0001) and soil total Cd datasets (I ² = 88.71%, Qt = 303.587, p < 0.0001), indicating that additional moderators may explain the variability (Supplementary Table S1).

pH significantly modulated straw-return effects on available Cd concentrations (p < 0.01), whereas no significant influence was observed for total Cd levels (p > 0.05) (Figure 2; Supplementary Table S2). Under strongly acidic conditions (pH < 5.5), straw application enhanced available and total Cd concentrations by 13.9% (p < 0.001) and 11.3% (p < 0.01), respectively, compared to control treatments. Similar increases of 15.5% and 9.7% (both p < 0.01) were observed in moderately acidic conditions (pH = 5.5–6.5). And these enhancement effects diminished progressively with increasing pH (Figure 2).

In contrast, initial SOM, initial soil Cd, and initial Cd content in straw exhibited minor impacts on soil Cd availability (p > 0.05) (Figure 2a; Supplementary Table S2). As these the content of factors rise, the effect of straw return on Cd availability of soil all showed a decreasing trend (Figure 2a). Notably, the initial soil Cd concentration strongly influenced the total Cd content following straw return (p < 0.001), whereas the initial Cd content in straw had a minimal impact with higher initial straw Cd content resulted in greater soil total Cd concentrations following straw incorporation. (p > 0.05) (Figure 2b; Supplementary Table S2). And the impact of straw return on soil total Cd increases with higher initial soil total Cd and straw Cd content (Figure 2b).

Both pot and field experiments consistently demonstrated that straw return significantly increased both available Cd and total Cd concentrations in paddy soils, with no significant differences between the two experimental setups (Figure 3; Supplementary Table S2). Notably, straw type exerted a pronounced effect on sil available Cd following straw return (p < 0.001), whereas its impact on total Cd was limited and not statistically significant (p > 0.05) (Figure 3; Supplementary Table S2). Compared with no straw return, rice straw incorporation increased soil available Cd by 20.0% (p < 0.001), while wheat and rapeseed straw amendments resulted in only marginal increases of 5.7% (p > 0.05) and 2.75% (p > 0.05), respectively. Interestingly, Solanum L. straw application conversely decreased available Cd by 11.2%, though this reduction lacked statistical significance (p > 0.05) (Figure 3a). Similarly, the addition of rice straw resulted in a substantial elevation of soil total Cd by 10.1% (p < 0.001), whereas wheat and rapeseed straw applications did not significantly alter soil total Cd content (Figure 3b).

Neither the quantities nor the duration of straw incorporation significantly influenced available Cd and total Cd concentrations in paddy soils (p > 0.05) (Figure 3; Supplementary Table S2). Specifically, available Cd initially decreased with increasing straw amounts before subsequently rising, and exhibited a transient increase followed by a decrease with prolonged application duration (Figure 3a). Conversely, total Cd displayed a progressive increase with both greater straw amounts and longer implementation periods (Figure 3b).

When pooling all data, our meta-regression analysis showed that the initial soil pH was significantly negatively correlated with the RR of soil available Cd (p < 0.01), soil total Cd (p < 0.05), and root Cd (p < 0.001) (Figures 4a–c). In contrast, the RR of grain Cd was significantly positively correlated with the change in soil pH (p < 0.05) (Figure 4d). Root Cd also exhibited a negative correlation with initial soil total Cd (p < 0.001) (Figure 4e). In addition, the RR of soil available Cd was significantly positively correlated with log10-transformed soil total Cd (p < 0.05) (Figure 4f). Furthermore, the RR of soil available Cd was positively associated with the RRs of grain Cd (p < 0.001), straw Cd (p < 0.01), and root Cd (p < 0.001) (Figures 5a–c). By contrast, no significant correlations were detected between the RR of soil total Cd and Cd accumulation in grain, straw, or root (Figures 5d–f).

Furthermore, we conducted path analysis to elucidate the relationships among soil available Cd, soil properties, and grain Cd accumulation. The path analysis revealed that the path coefficient between total soil Cd and available soil Cd was the highest (0.84), followed by initial soil pH (0.45) (Figure 6a). A significant positive correlation was observed between available soil Cd and total soil Cd, whereas a significant negative correlation existed between initial soil pH and available soil Cd. The RR of soil pH exerted a significant indirect effect on available soil Cd through initial soil pH (Figure 6a). Similarly, the direct effect of soil available Cd on grain Cd accumulation was particularly strong, with the highest path coefficient of 0.84, whereas the corresponding coefficient between soil total Cd and grain Cd was much weaker (0.17) (Figure 6b). Both soil available Cd and soil total Cd exhibited significant positive correlations with grain Cd content. Notably, soil total Cd had a significant indirect effect on grain Cd content through soil available Cd, while the RR of soil pH and RR of soil OM exerted significant indirect effects on grain Cd content through initial soil pH (Figure 6b).

Model selection approach indicated that the initial soil pH was the most primary predictor governing Cd bioavailability following straw incorporation (Figure 7a). However, none of the predictors could explain the changes in soil total Cd levels in responses to straw return with the straw type ranked first (Figure 7b). Furthermore, we conducted a subgroup analysis on how initial soil pH affects the changes in soil pH under straw return. As anticipated, under strongly acidic conditions (initial soil pH < 5.5), straw return led to a significant 4.5% increase in soil pH relative to the no-straw control (p < 0.001) (Figure 7c). No statistically significant differences in pH changes were observed for soils with higher initial pH values (Figure 7c).

Straw return is widely implemented to enhance soil nutrient dynamics, improve crop productivity, and mitigate environmental pollution from agricultural residue burning (Ye et al., 2015; Berhane et al., 2020; Jin et al., 2020). Our Limited data also suggest that straw return may increase the yield of rice grain (data not shown, indicative of potential publication bias), this finding is align with previous studies (Cheng et al., 2025; Jin et al., 2020; Shan et al., 2021; Zhou H. et al., 2024). However, straw return also resulted in significant increases in both soil available Cd and total Cd, alongside notable elevations in SOM and soil pH, thereby posing a potential risk of enhanced Cd accumulation in various rice tissues (Figure 1).

Straw serves as a major source of SOM, and its incorporation into fields effectively increases SOM content (Bu et al., 2020; Jing et al., 2021). The increased SOM influences Cd behavior through multiple mechanisms. On one hand, small molecular organic acids organic acids released during straw breakdown create chelating associations with Cd, enhancing its activation and bioavailability (Borggaard et al., 2019; Yuan et al., 2021). On the other hand, the humification of straw generates humic substances rich in reactive sites (e.g., carboxyl and phenolic hydroxyl groups) that establish stable organometallic complexes with Cd, thereby increasing Cd adsorption and reducing its bioavailability (Zeng et al., 2011; Gao et al., 2022) (Figure 8). Our model selection analysis indicated that variations in root and straw Cd were best predicted by the initial SOM, whereas changes in grain Cd were primarily associated with alterations in SOM (Supplementary Figure S4). Collectively, these findings demonstrate that changes in soil organic matter indirectly influence Cd uptake and accumulation in plants by modifying the soil chemical environment and associated microbial activity, which in turn alters soil pH (Figure 6b) (Jin et al., 2023; Li Q. et al., 2024).

The increase in soil pH following straw return to paddy fields can be explained by an integrated mechanism involving several processes. The decomposition of rice straw releases alkaline cations (e.g., Ca²⁺, Mg²⁺, K⁺) that neutralize soil acidity (Liang et al., 2023). In addition, microbial ammonification converts organic nitrogen in straw into ammonium (NH₄ ⁺), a process that consumes H⁺ ions and further elevates soil pH (Li and Zhong, 2021). Under anaerobic conditions, microbial denitrification transforms soil nitrates (NO₃ ⁻) into nitrogen gases (N₂ or N₂O), thereby further reducing H⁺ concentrations (Li and Zhong, 2021). Moreover, organic acids released during straw decomposition can solubilize soil iron hydroxides, with the released iron ions reacting with acidic components to further ameliorate soil acidity (Zeng et al., 2011; Yuan et al., 2021). Subgroup analysis of the change in soil pH revealed that significant pH increases occurred only in strongly acidic soils (pH < 5.5) after straw return (Figure 7c). Acidic soils with initially low pH values exhibit more pronounced alkalinization following straw incorporation.

Typically, elevated soil pH is expected to reduce Cd solubility by promoting the transformation of Cd from readily exchangeable forms into less mobile fractions, thereby enhancing its retention in the soil solid phase (Zeng et al., 2011; Wei et al., 2023). However, our meta-analysis revealed that straw return significantly increased soil available Cd, despite the concurrent rise in soil pH. This seemingly contradictory response is consistent with the findings of Wang et al. (2015), who showed that incorporation of Cd-contaminated rice straw simultaneously increased soil pH and EDTA-extractable Cd in previously uncontaminated soils. One plausible explanation is that straw addition stimulates microbial activity, particularly under slightly acidic to neutral conditions, thereby accelerating organic matter decomposition and altering Cd speciation (Ren et al., 2018; Zhao et al., 2019; Su et al., 2021). These microbial-mediated processes, together with enhanced dissolution of organic substrates, may promote the formation of soluble Cd–DOM complexes and increase the proportion of unbound Cd in soil solution, ultimately elevating extractable soil available Cd (Wei et al., 2023; Xu et al., 2023). Importantly, our moderator analyses, meta-regression, and model selection consistently identified initial soil pH as the dominant driver of straw-induced changes in soil available Cd, with the strongest enhancement occurring in acidic soils and progressively weakening as initial pH increased (Figure 2a; Figures 4a, 7a). Under low-pH conditions, straw decomposition is typically accelerated and accompanied by greater production of dissolved organic matter and low-molecular-weight ligands, which can complex Cd and increase its mobility even when bulk soil pH rises.

The influence of Fe, Mn, and Cu on Cd accumulation in rice grains can be attributed to their competitive interactions during uptake and translocation processes, as well as their roles in altering soil redox conditions and Cd bioavailability (Suda and Makino, 2016; Khaliq et al., 2019; Han et al., 2021). Liu et al. (2017) observed that Cd accumulation in rice grains was not directly linked to total or available soil Cd, but was significantly correlated with Fe, Mn, and Cu levels during rice growth under straw return conditions. In contrast, our data demonstrate a significant positive relationship between changes in soil available Cd and those in grain, straw, and root Cd, while no such relationship was evident for changes in soil total Cd (Figure 5). These results, consistent with previous studies (Luo et al., 2023; Zhang et al., 2018; Zhou H. et al., 2024), underscore the critical role of soil available Cd as an indicator of Cd contamination risk in rice.

Initial pH serves as a key factor governing Cd bioavailability following straw incorporation, controlling metal mobilization and plant absorption processes (Siddique et al., 2022). Acidic soil conditions accelerate straw decomposition, leading to increased dissolved organic carbon (DOC) concentrations, which can complex with free Cd²⁺ to form highly mobile Cd-organic complexes, thereby enhancing Cd availability (Yuan et al., 2021; Wei et al., 2023). Our results indicate that under acidic conditions (pH < 6.5), straw return significantly increases both available and total Cd concentrations in paddy soil compared to the no-straw control (Figure 2). Additionally, we observed a negative correlation between changes in soil available Cd, total Cd and root Cd accumulation with initial soil pH (Figures 4a–c), suggesting that under lower pH conditions, crop residue decomposition releases humic acids and other SOM components, thereby enhancing the availability of soil Cd and leading to greater accumulation of Cd in plant roots (Zeng et al., 2011; Gao et al., 2022). Furthermore, the Cd content in grains is also coupled with changes in soil pH (Figure 4e), indicating that Cd translocation and accumulation within rice plants are also modulated by initial soil pH (He et al., 2021; Liao et al., 2021). Compared to grain and straw Cd, root Cd uptake was more sensitive to micro-environmental changes induced by straw return (Supplementary Figures S5, S6). Straw return preferentially stimulated Cd absorption in roots, which was subsequently translocated to aboveground tissues through biochemical processes (Figure 8) (Bu et al., 2020; Zhao and Wang, 2020). This highlights the critical role of rhizosphere interactions in regulating Cd mobility within the plant-soil continuum.

In addition to soil pH, initial soil Cd concentrations significantly influenced total Cd dynamics following straw return (Figure 2b; Supplementary Table S2), with changes in total soil Cd being strongly coupled to variations in available Cd (Figure 4f). Numerous studies have established that Cd accumulation in rice is positively correlated with initial soil Cd concentrations (Cao et al., 2014; Mu et al., 2019). Thus, initial soil pH and Cd content jointly regulate straw decomposition, influencing Cd accumulation across different rice plant components (Figure 8; Supplementary Table S3).

Interestingly, when both native soil Cd and Cd concentrations in returned straw exceeded certain thresholds, Cd activation following straw return decreased (Figure 2a), leading to reduced Cd uptake in rice plants (Supplementary Figure S6). This could be attributed to the toxic effects of high Cd concentrations on soil microbial communities, which may impair microbial-mediated straw decomposition (Yuan et al., 2021; Xu et al., 2023). Under such conditions, returned straw likely behaves as semi-decomposed humus, promoting Cd adsorption by soil solids and thereby reducing Cd solubility (Wang et al., 2021; Ye et al., 2015; Zhou H. et al., 2024). Additionally, we found a negative correlation between initial soil Cd concentration and root Cd accumulation (Figure 4e), indicating that soils with excessively high Cd levels exhibited lower bioavailable Cd concentrations, ultimately limiting root Cd uptake (Figure 2a; Supplementary Figure S6). Notably, when straw Cd concentrations exceeded 0.3 mg kg^-1, straw return significantly increased soil available Cd concentrations (Figure 2a), further demonstrating the role of straw composition in Cd mobilization.

Our meta-analysis demonstrated that, compared to no straw return, the incorporation of rice straw significantly increased both soil available Cd and total Cd concentrations in paddy fields, whereas the return of wheat and rapeseed straw had no significant effect (Figure 3a). This distinct response is likely stems from the unique biochemical composition and decomposition characteristics of rice straw (Mussoline et al., 2013; Xia et al., 2018; Yan et al., 2019).

Rice is recognized as the cereal crop with the highest Cd accumulation capacity, leading to inherently elevated Cd concentrations in rice straw (Dong et al., 2024; Wang S. et al., 2023; Xue et al., 2022). When rice straw is returned to the field, its decomposition releases the accumulated Cd back into the soil, thereby enhancing its bioavailability (Su et al., 2021), and our model selection indicated that the straw type was the first predictor influencing soil total Cd responses to straw return (Figure 6b); Additionally, rice straw possesses a high content of lignin (10%–20%), cellulose (35%–45%), and silicon (9%–13%) (Bhattacharyya et al., 2020; Sharma et al., 2023; Nakyp et al., 2024). Lignin, a structurally complex aromatic polymer, is resistant to microbial degradation, while silicon primarily exists in rice straw as silicon dioxide (SiO₂), which accumulates in plant cell walls, increasing straw rigidity (Bhattacharyya et al., 2020; Nakyp et al., 2024). The interaction between silicon, lignin, and cellulose forms a dense matrix that further impedes microbial decomposition (Ma et al., 2015; Asiri et al., 2023). Furthermore, the rice straw exhibits a high carbon-to-nitrogen ratio (C/N) (Wang W. et al., 2023). A high C/N ratio indicates nitrogen limitation, which restricts microbial proliferation and slows mineralization, resulting in the gradual release of available Cd into the soil (Luo et al., 2023). The slow decomposition process is further exacerbated under the anaerobic conditions prevalent in flooded paddy fields, beneficing to a prolonged release of Cd and an increase in its bioavailability (Yuan et al., 2021).

In contrast, wheat and rapeseed straw contain lower levels of lignin, cellulose, and silicon, decompose more rapidly, and do not contribute significantly to Cd mobilization in the soil. Consequently, their return does not lead to the same increase in soil available Cd as observed with rice straw. These findings highlight the crop-specific effects of straw return on Cd dynamics and underscore the necessity of considering straw composition when implementing straw management practices in paddy fields.

In our meta-analysis, all data were collected from studies examining the application of straw under rice cultivation conditions and its impact on Cd-related indicators in paddy fields, excluding experimental incubations with microorganisms. This methodological approach enhances the authenticity and reliability of our findings. However, the limited sample size included in the analysis presents a potential constraint, as our conclusions may be influenced by the results of a single study. In addition, most of the included observations were derived from paddy systems in China, whereas data from other major rice-producing regions (e.g., Southeast Asia and South Asia) remain scarce. In addition, key mechanistic variables—such as soil redox potential (Eh), detailed Cd speciation, mineral interactions (Fe, Mn, Cu), and microbial community dynamics—were not consistently available across studies, limiting a more comprehensive mechanistic interpretation of Cd mobilization processes.

Despite these limitations, our synthesis reveals a consistent trend that straw return may elevate Cd bioavailability and promote Cd accumulation in rice tissues, highlighting potential food-safety concerns in contaminated areas. Given that straw return is widely practiced across major rice-growing regions in Asia, the risk patterns and mechanistic insights identified here are likely relevant beyond China and may inform sustainable straw management in Cd-contaminated paddies worldwide. Accordingly, straw incorporation should be implemented cautiously in Cd-affected paddies, and straw removal (including root residues) may be advisable under high-risk conditions (Li C. et al., 2024; Tian et al., 2025). From an engineering and management perspective, combining straw return with feasible amendments (e.g., lime, zeolite, pumice) or converting straw into biochar via pyrolysis represents practical strategies to stabilize Cd while maintaining soil organic carbon benefits (Chen et al., 2016; Huang et al., 2024a).

Furthermore, alternative straw utilization pathways, such as biomass energy production or industrial raw-material applications, may enhance economic viability while minimizing environmental risks (Weiser et al., 2014; Bhattacharyya et al., 2020). Collectively, these approaches provide a sustainability-oriented framework for optimizing straw management in Cd-contaminated rice systems, balancing agronomic productivity, environmental protection, and food safety.