The inclusion and exclusion criteria were developed not only to ensure methodological rigor but also to align with the review's central theme: understanding the health risks of cassava consumption through the lens of unsustainable food systems in SSA. These criteria ensure that the selected studies reflect the human health consequences of systemic weaknesses in food systems within SSA. Table 2 shows the inclusion and exclusion criteria for the review.

A comprehensive literature search was conducted in the following electronic databases: PubMed, Scopus, Journal Storage (JSTOR), and African Journals Online (AJOL). Prior to the search, an academic librarian was consulted to assist in developing comprehensive search strategies, identifying relevant databases, optimizing keywords and subject heading selection, and ensuring rigorous adherence to the PRISMA guidelines. The search strategy used a combination of relevant keywords and concepts related to cassava and cyanide. The keywords used in the search included cassava, cyanide, hydrogen cyanide, linamarin, lotaustralin, cyanogenic glycosides, konzo, health risk, acute poisoning, toxicity, tropical ataxic neuropathy, and neurological disorder. A full search index and the Boolean search strings used for each database are shown in Table 3. To ensure comprehensive search, no restrictions or filters were placed on publication year, language or country during the search of various databases. The initial search was conducted in January 2025 and updated on 24th June 2025. Two independent reviewers performed the search and screening processes and resolved any disagreements through discussion or consultation with a third reviewer. Supplementary data illustrating cassava production trends were obtained from FAOSTAT (2025).

All database search results were collated, stored, and managed in Mendeley desktop software (version 1.19.8). Multiple folders were created in Mendeley to contain the results of each database search, and then all the results were aggregated into a single folder with duplicates removed. The study selection process was conducted in two phases: an initial screening of titles and abstracts, followed by a full-text review of potentially eligible studies. Data from the included studies were extracted using a standardized data extraction form. The extracted data included: study characteristics (i.e., authors, publication year, study location, study population); cyanide exposure assessment (i.e., thiocyanate concentrations); residual HCN concentrations and cassava processing methods; reported health outcomes or adverse effects; and potential confounding factors.

The findings from the included studies were synthesized based on geographical distribution (country of study population), cyanide exposure levels reported (e.g., in μmol L⁻¹), reported health outcomes (e.g., konzo, tropical neuropathy, acute poisoning), and their underlying factors such as processing methods and HCN concentrations in cassava products. A geographic map and tabular summaries were used as appropriate to the data. The findings have been reported following the PRISMA statement, including a flow diagram of the study selection process, tabular summaries of study characteristics (design, population, and key findings), and narrative syntheses, highlighting patterns, contradictions and gaps, as appropriate.

The initial database search identified a total of 4,964 records: PubMed (n = 631), Scopus (n = 1,446), JSTOR (n = 1,284), and AJOL (n = 1,603). After removing 681 duplicate records, 4,283 unique records were screened for titles and abstracts. From this pool, 472 publications were retrieved for full-text screening. Records were excluded at this stage based on irrelevance to the topic, that is, purely technical focus, use of animal or in vitro models, lack of relevance to cassava as a crop, non-English language, or unavailability of the full text. Following full-text screening, 426 publications were excluded for not addressing health impacts, being review articles or editorials, or being conducted outside SSA. Finally, 45 studies met the inclusion criteria and were included in the final review (Figure 1).

Figure 2 illustrates the frequency of included publications by the country where the studies were conducted. The Democratic Republic of Congo (DRC) had the highest frequency of publications (n = 18), followed by Mozambique (n= 10), Tanzania (n =7), and Nigeria (n = 6). Other countries such as Malawi, Uganda, Zambia, and Cameroon each contributed only one study. This distribution highlights a concentration of research activity in regions with a high burden of cassava-related cyanide exposure, particularly in Central and East Africa.

Urinary thiocyanate levels, a biomarker of cyanide exposure from cassava consumption, were reported in 28 of the 45 included studies (Table 6). The levels varied widely across populations, health conditions, and intervention periods. Several studies, particularly those conducted in the DRC, Mozambique, and Tanzania, reported extremely high thiocyanate levels, far above the World Health Organization's safety threshold of 350 μmol/L for urinary thiocyanate. The highest concentrations were observed in individuals diagnosed with konzo - a neurological disorder linked to cyanide toxicity. For instance, (Tshala-Katumbay et al. 2001) reported concentrations of 1,128 μmol/L among individuals with konzo in the DRC, while (Mlingi et al. 1991) reported a mean value of 1,730 μmol/L at the onset of a konzo outbreak in Tanzania. Similarly, (Tylleskär et al. 1992) found levels of 757 μmol/L in affected children compared to 50 μmol/L in unaffected peers. In addition, neurological conditions (e.g., tropical ataxic neuropathy), cognitive impairment, and acute intoxication events were reported.

Several studies demonstrated the impact of interventions aimed at improving cassava processing. For example, (Banea et al. 2013) reported a reduction from 930 μmol/L pre-intervention to 150 μmol/L post-intervention. Comparable declines were observed in (Nhassico et al. 2015) (from 530 μmol/L to 180 μmol/L) and (Banea et al. 2015a) (from 444 μmol/L to 191 μmol/L). In contrast, populations with safer cassava processing practices or lower cassava consumption exhibited significantly lower thiocyanate levels. For instance, (Chiwona-Karltun et al. 2000) and (Onabolu et al. 2001) reported levels as low as 50 μmol/L and 17 μmol/L, respectively, in non-endemic or low-consumption areas (Table 6).

Table 7 shows substantial variability in residual Hydrogen Cyanide (HCN) concentrations across different cassava processing methods and duration. Raw bitter cassava roots contained the highest HCN levels, with concentrations reaching 327 ppm (Ministry of Health, 1984). In contrast, properly processed cassava flour showed significantly reduced HCN levels, depending on the method and duration of processing. Sun-drying alone reduced HCN levels to between 30 and 95 ppm, while heap fermentation was more effective, lowering concentrations to as low as 17–28 ppm (Cardoso et al., 2004; Ernesto et al., 2002). The combination of heap fermentation and sun-drying further reduced HCN to < 10–20 ppm, particularly under lab-controlled fermentation conditions for 6 d (Chabwine et al., 2011).

Studies involving wetting methods (i.e., 2 h sun and 5-h shade exposure) as a cassava processing intervention consistently achieved the lowest HCN levels. (eg.,) (Banea et al. 2012, 2013, 2015a) and (Nhassico et al. 2015) reported reductions from pre-intervention levels of 22–74 ppm to post-intervention levels of 4–25 ppm. In contrast, studies that documented unspecified or shortened processing methods consistently reported higher cyanide levels. (Bumoko et al. 2015); (Kambale et al. 2017); and (Kabamba et al. 2023) all recorded ranges between 30 and 200 ppm.

This systematic review highlights the persistent public health risks associated with cyanide exposure from cassava consumption in SSA, despite the crop's critical role in regional food security. The findings reveal a complex interplay between agricultural practices, processing methods, and structural vulnerabilities within food systems that collectively shape the health outcomes of cassava consumers. In countries like the DRC and Mozambique, cassava was a critical staple, especially during periods of conflict, drought, or economic hardship. However, in these contexts, cassava plays a dual role: while it is a resilient and accessible food source, it also poses significant health risks due to its high cyanogenic glycoside content, particularly in bitter varieties. These risks were amplified when cassava was consumed without adequate processing, which remains common in many resource limited settings like SSA.

Despite the availability of effective processing methods such as wetting and heap fermentation, their adoption remains limited due to weaknesses and disruptions in SSA's food systems. For instance, wars and conflicts profoundly disrupt agricultural activities, food distribution, and household food security (Weldegiargis et al., 2023; Muriuki et al., 2023), forcing communities to rely heavily on inadequately processed cassava (Cliff et al., 1997). Socio-economic stress often forces households to rely on shortened or informal processing methods, which are less effective at detoxifying cassava. Moreover, the intentional cultivation of bitter varieties due to their pest resistance, storability, and yield further complicates efforts to reduce cyanide exposure (Imakumbili et al., 2019; Chiwona-Karltun et al., 1998; Kapinga et al., 1997). These food system challenges have intensified in recent years across SSA, exacerbated by climate change impacts such as prolonged dry spells, El Niño events, and cyclones (World Food Programme, 2024; United Nations Office for the Coordination of Humanitarian Affairs, 2022), all of which further entrench the over-reliance on cassava as a food security crop.

The burden of cassava-related health risks was concentrated in a few countries, most notably the DRC, Mozambique, Tanzania, and Nigeria. However, the absence of epidemiological data from other high cassava-consuming countries, such as Ghana and Malawi (Adebayo, 2023), may not necessarily indicate a lack of risk. In some cases, this may reflect weak surveillance systems to detect sub-lethal cyanide exposure or milder neurological symptoms that are not present as overt clinical cases. Nonetheless, it may also reflect the absence of key contributing factors such as wars and conflicts or the widespread use of safer processing practices, as observed in Malawi (Chiwona-Karltun et al., 2000). Furthermore, many of the studies included in this review are decades old, suggesting either a decline in research interest—possibly due to reduced funding—or a genuine reduction in incidence, potentially linked to increased awareness and improved cassava processing practices. This gap highlights the urgent need for renewed investment in research and the strengthening of surveillance systems to capture both acute and chronic health effects of cyanide exposure in cassava-consuming populations.

The review revealed a consistent association between cassava consumption and neurological conditions such as konzo and tropical ataxic neuropathy (TAN). Konzo is characterized by a sudden onset of non-progressive spastic paraparesis, leading to severe motor impairment and often resulting in irreversible disability (Kashala-Abotnes et al., 2019; World Health Organization, 1996; Tylleskär et al., 1991). In contrast, TAN presents as a progressive condition involving both sensory and motor deficits, while broader neuropathies may arise from prolonged, low-level cyanide exposure (Adamolekun, 2011).

Biomarker data, particularly urinary thiocyanate levels, provided strong evidence of chronic cyanide exposure in affected populations (Cliff et al., 2011). In several studies (Bumoko et al., 2014, 2015; Tshala-Katumbay et al., 2001; Mlingi et al., 1992), levels exceeded WHO recommended safety limit of 350 μmol/L (World Health Organization, 1996). In contrast, lower thiocyanate levels (< 350 μmol/L) were typically observed in populations with access to adequately processed cassava, reduced reliance on cassava as a staple, or the use of low-cyanide cassava varieties. These findings highlight the effectiveness of targeted mitigation strategies in reducing cyanide exposure and preventing associated health outcomes.

Processing methods play a critical role in determining the residual HCN levels in cassava products. Raw bitter cassava roots were found to contain extremely high HCN concentrations, reaching up to 327 ppm (Ministry of Health, 1984). In contrast, more effective processing techniques, particularly heap fermentation and the wetting method consistently reduced HCN levels to below the WHO's recommended safety limit of 10 ppm for cassava flour and dried cassava products (FAO/WHO, 2011). Post-intervention studies, particularly using the wetting method, demonstrated substantial reductions in HCN content, with some achieving levels as low as 4–10 ppm (Banea et al., 2012, 2013). However, the review also found that shortened or improperly applied methods, such as soaking for only 1–2 days, often resulted in dangerously high HCN levels—sometimes exceeding 400 ppm. These inadequate practices were frequently driven by time constraints, limited awareness, or lack of access to water. Such findings underscore the urgent need to promote improved cassava processing techniques, particularly the wetting method, and to strengthen community education on safe detoxification practices to reduce the risk of cyanide exposure.

In addition to traditional or conventional processing methods, recent innovations in cassava breeding and processing technologies offer promising solutions. Notably, bio fortified cassava varieties such as UMUCASS-38 (TMS 01/1371) and NR-8082, developed through collaborative efforts involving the [International Institute of Tropical Agriculture (IITA), 2012] and Harvest Plus, combine high yield, drought tolerance, and enhanced pro vitamin A content with low cyanogenic potential (Odoemelam et al., 2020). These varieties have been shown to contain cyanide levels of approximately 17.02 ± 0.02 ppm (UMUCASS-38) and 18.01 ± 0.01 ppm (NR-8082) in flour form, which are well below the WHO's safety threshold of 50 ppm for fresh cassava roots (FAO/WHO, 2011). Technologies such as hydraulic pressing, mechanical peeling, and grated fermentation have also been developed to enhance detoxification and reduce processing time [International Institute of Tropical Agriculture (IITA), 2012]. However, adoption of these innovations remains limited in many rural areas due to affordability constraints, weak extension services, and low awareness (Nweke et al., 2002). Cultural food preferences and the time-intensive nature of some improved methods may also hinder uptake. Expanding access to improved varieties, strengthening extension systems, and promoting behavior change communication are essential for scaling up safe cassava utilization.

Protein-deficient diets are a major underlying contributor to cyanide-related health risks. Many populations consuming bitter cassava lacked access to protein-rich foods that provide sulfur-containing amino acids, such as methionine and cysteine, critical for detoxifying cyanide into thiocyanate (Richie et al., 2023). Without sufficient intake of quality protein, the risk of conditions like konzo increased significantly (Tylleskär et al., 1995; Cliff et al., 1985; Ministry of Health, 1984). For instance, the [Integrated Food Security Phase Classification (IPC), 2022a] estimated that approximately 26.4 million people in the DRC experienced high levels of acute food insecurity (IPC Phase 3 or above) between July and December 2022, driven by widespread poverty, conflict, displacement, low agricultural production, high food prices, and poor infrastructure [Integrated Food Security Phase Classification (IPC), 2022b]. For instance, in Malawi – a country where cassava is consumed – protein poverty is evident, as most dietary protein comes from cereals that are low in essential amino acids like lysine and tryptophan [International Food Policy Research Institute (IFPRI), 2020a,b; Muleya et al., 2022]. In addition, poor dietary diversity further contributes to deficiencies in micronutrients like selenium, calcium, and zinc, which are essential for neurological health and resilience to cyanide toxicity (Kurmi et al., 2023).

This review also underscores how cassava toxicity reveals broader systemic vulnerabilities within SSA's food systems. In the context of worsening climate change, persistent conflict, economic fragility, and weak infrastructure, cassava remains one of the few accessible, drought-resilient crops, even when its consumption carries serious health risks. During the 2022/23 period, over 55 million people in 12 countries within the [Southern African Development Community (SADC), 2022] were food insecure. Additionally, 35 non-international armed conflicts were active across Africa in 2024 (Geneva Academy of International Humanitarian Law, and Human Rights, 2025). These persistent challenges reduce the viability of other staples, disrupt markets and agricultural production, and limit access to safe, nutritious diets. These pressures are unlikely to subside in the near future, reinforcing the chronic nature of cyanide toxicity. Strengthening food systems through continuous surveillance, support for diversified and climate-resilient agriculture, and the promotion of food sovereignty where communities control their food production and processing is essential for long term mitigation.

This review has several strengths that enhance the credibility and comprehensiveness of its findings. First, it employed a broad search strategy across multiple databases—PubMed, Scopus, JSTOR, and AJOL, ensuring the inclusion of country and region-specific studies. Second, the methodological quality of all included studies was rigorously assessed using validated tools: JBI Critical Appraisal Checklist for Case Series and NIH Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies. Third, the review adhered to the PRISMA guidelines, promoting transparency, reproducibility, and methodological rigor. Finally, the involvement of multiple reviewers in both the screening and risk of bias assessment processes helped enhance consistency and reduce the risk of selection bias. Despite these strengths, the review has a limitation. The exclusion of non-English language publications may have introduced language bias. However, to minimize this effect, studies with English titles and abstracts were screened for relevance, even if the full text was not available in English. While this approach helped reduce the risk of overlooking important studies, it may still have led to the omission of valuable data published in other languages.