Paul et al. 2025 — Arsenic in small indigenous freshwater fish, Lower Meghna River, Bangladesh
Ten small indigenous fish species (SIS) from Bangladesh’s Lower Meghna River were systematically sampled across three seasons (pre-monsoon, monsoon, post-monsoon) and ten stations to quantify total arsenic in muscle tissue and assess associated human health risks. Five of the ten species — Mystus vittatus, Glossogobius giuris, Lepidocephalichthys guntea, Neotropius atherinoides, and Apocryptes bato — exceeded the WHO safe consumption threshold of 1 mg/kg tAs in muscle tissue, at least during the pre-monsoon season. Arsenic levels in water and sediment were within non-polluting limits by standard indices, yet fish-muscle concentrations were substantially higher, demonstrating significant bioaccumulation from environmental pathways. Pre-monsoon concentrations were consistently highest across all species, followed by post-monsoon, then monsoon, with statistically significant seasonal differences (p < 0.05) but no spatial gradient across the three river zones (p > 0.05). Benthic carnivores and scaleless/fusiform species accumulated more arsenic than pelagic-benthic omnivores, indicating that morphology, habitat, and feeding behavior modulate bioaccumulation. Children were identified as the population most at risk from fish consumption in this region.
Key numbers
n = 300 fish muscle samples: 10 species × 10 stations × 3 seasons.
WHO safe consumption threshold for arsenic in fish: 1 mg/kg.
Species exceeding 1 mg/kg (pre-monsoon means):
- Apocryptes bato (Ab): mean 1.35 ± 0.20 mg/kg (pre-monsoon); range 0.99–1.67 mg/kg
- Neotropius atherinoides (Na): mean 0.97 ± 0.15 mg/kg (pre-monsoon); range 0.80–1.17 mg/kg
- Lepidocephalichthys guntea (Lg): mean 0.95 ± 0.13 mg/kg (pre-monsoon); range 0.79–1.17 mg/kg
- Mystus vittatus (Mv): mean 0.82 ± 0.16 mg/kg (pre-monsoon); range 0.53–1.03 mg/kg
- Glossogobius giuris (Gg): mean 0.92 ± 0.13 mg/kg (pre-monsoon); range 0.76–1.15 mg/kg
Note: all five exceed 1 mg/kg within their pre-monsoon range. Ab mean is above the threshold; others approach or cross it at upper range.
Species below 1 mg/kg across all seasons (selected):
- Puntius sophore (Ps): pre-monsoon mean 0.40, monsoon 0.20, post-monsoon 0.29 mg/kg
- Amblypharyngodon mola (Am): pre-monsoon mean 0.59, monsoon 0.28, post-monsoon 0.44 mg/kg
- Devario devario (Dd): pre-monsoon mean 0.70, monsoon 0.44, post-monsoon 0.60 mg/kg
- Chela cachius (Cc): pre-monsoon mean 0.61, monsoon 0.24, post-monsoon 0.42 mg/kg
- Gudusia chapra (Gc): pre-monsoon mean 0.66, monsoon 0.31, post-monsoon 0.44 mg/kg
All species showed significant seasonal variation (one-way ANOVA p = 0.0 for all), with pre-monsoon > post-monsoon > monsoon in most cases. No significant spatial variation across upper, mid, and lower river zones (p > 0.05 for all species).
Water arsenic: pre-monsoon mean 0.012 ± 0.003 mg/L (highest), monsoon mean 0.002 ± 0.0007 mg/L, post-monsoon mean 0.006 ± 0.0018 mg/L; all seasons within Bangladesh Environmental Conservation Rules limit (0.05 mg/L) and WHO drinking water guideline (0.01 mg/L is lower — pre-monsoon exceeds WHO drinking water guideline but not the national standard). Seasonal variation significant (p = 0.0); spatial variation not significant (p = 0.89).
Sediment arsenic: pre-monsoon mean 6.88 ± 0.98 mg/kg, monsoon 4.94 mg/kg, post-monsoon 5.55 mg/kg; seasonal variation significant (p = 0.0); spatial variation not significant (p = 0.5). Below the shale standard of 13 mg/kg.
LOD/LOQ: fish muscle LOD = 0.02 µg/g (20 ppb), LOQ = 0.06 µg/g (60 ppb). All fish samples above LOQ. Water LOD = 0.1 µg/L, LOQ = 0.3 µg/L. Sediment LOD = 0.5 µg/kg, LOQ = 1.5 µg/kg.
CRM recoveries: 93–106% for fish tissue (NIST SRM 2976), sediment (NIST SRM 2702), and water.
Methods
Total arsenic only; no speciation between iAs and organic As species reported. Fish muscle dissected, freeze-dried (Zirbus VaCo 2), 0.5 g aliquots digested in microwave digestion system (WX-6000) with ultrapure HNO3, analyzed by ICP-MS (PerkinElmer ELAN9000). Water: 0.45 µm membrane filtration, HNO3 acidification, ICP-MS (ELAN9000). Sediment: dried at 50°C, muffle-furnace ignition to 450°C, digestion with 50% HNO3, AAS (SHIMADZU AA-7000). QA/QC: CRM 320 (Merck), acid-cleaned labware, 10% duplicate analyses; RSD ≤ 11%. Calibration linearity R² > 0.998.
Limitation: total arsenic only — no iAs/tAs split. The WHO 1 mg/kg threshold referenced is for total arsenic in fish; the more health-relevant metric would be inorganic As fraction, which is typically 10–30% of total As in freshwater fish. Reported values therefore likely overstate inorganic-As risk. No data on cooking method effects (boiling may reduce As by ~20–35%). The 10 fish specimens per species per station is a small sample for statistical power at within-station level. Results reflect the Lower Meghna specifically, which sits in one of Bangladesh’s most arsenic-contaminated groundwater zones; these findings are not representative of freshwater fish globally.
Implications
Certification: this paper is relevant context for product categories using freshwater fish from arsenic-contaminated river basins (Bangladesh, similar groundwater-impacted regions in South/Southeast Asia). Total As in muscle tissue of certain SIS species exceeds 1 mg/kg during pre-monsoon; the absence of iAs speciation limits direct application to regulatory limits but flags these species and regions as high-risk.
Courses: the paper is a strong teaching case for the interaction between groundwater arsenic contamination, seasonal hydrology (pre-monsoon dilution vs. concentration effects), fish morphology (scaleless, benthic, fusiform body shapes concentrate more As), and food safety. The seasonal dynamic — highest risk during pre-monsoon when groundwater inputs dominate river flow — is a critical concept for sourcing-season risk management.
App: freshwater fish ingredient page should note that species from Bangladesh and similar highly arsenic-contaminated South Asian groundwater basins can carry total As well above 1 mg/kg in muscle tissue, particularly during dry seasons; speciation data are needed to compute iAs-specific risk.
Microbiome: no direct microbiome content in this paper.