The coastal geographic area of Bangladesh is located in a subtropical climate between 21°–23° N and 89°–93° E with a 710-km-long coastline along the Bay of Bengal and is divided into three distinct regions: the steady deltaic western part (Atlantic type), the deltaic eastern section (Pacific type), and the deltaic middle zone. The coastal zone is covered with an extensive river infrastructure (24,000 km in length), a significant number of islands across channels, and an underwater canyon (Swatch of No Ground), in addition to the funnel-like part of the northern Bay of Bengal. Cox’s Bazar, Chattogram, Feni, Noakhali, Laxmipur, Chandpur, Patuakhali, Khulna, Satkhira, and Bagerhat are some coastal districts of Bangladesh. Among them, Cox’s Bazar (21°25′38″N 92°00′18″E), Chattogram (22°20′06″N 91°49′57″E), Feni (23°1.1′N 91°24.6′E), Noakhali (22.70°N 91.10°E), Chandpur (23.2139°N 90.6361°E), and Patuakhali (22.3542°N 90.3181°E) were selected for conducting the present investigation in this study. The study areas are situated in the catchment area of the Meghna River (Noakhali and Chandpur), the Galachipa River and the Tetulia River (Patuakhali), the Feni River (Feni), the Karnaphuli and the Sangu rivers (Chattogram), and the Bara Matamuhuri River, the Naf River, and the Bakkhali River (Cox’s Bazar) (Figure 1). These rivers carried huge amount of wastes that are mainly of different synthetic origins (plastic and other debris) (Hossain et al., 2021). The people of these areas depend on GW for drinking and other household purposes.

A total of 54 GW samples were collected, with three stations from each district. Each sample with a volume of 500 mL was collected in triplicate, resulting in a total volume of 27 L. The samples were collected in 500-mL glass bottles and promptly sealed to prevent airborne contamination. Subsequently, the sealed bottles were preserved in an icebox at −20°C and transferred to the Laboratory for Ecology, Ecosystem, and Biodiversity (LEEB) for further investigation. Upon arrival at the laboratory, the conserved bottles were defrosted at room temperature. The groundwater was then filtered through a 5-micron cellulose nitrate membrane filter. The filter papers containing the filtered material were placed in transparent Petri dishes for microscopic assessment in the subsequent stage.

MP samples were analyzed and identified using established procedures (Barrows et al., 2018; McNeish et al., 2018) at the LEEB, Noakhali Science and Technology University. The protocol is outlined in Supplementary Figure S1. The filter paper was thoroughly examined under a stereomicroscope (Leica EZ4E, Germany) with 8 to 35 x magnification to obtain visual detection and quantification of MPs. We applied laboratory control samples (about two microplastic particles per filter) to adjust the levels of microplastic pollution. An exceptionally high-resolution webcam was used to capture photographs of MPs which was attached to a microscope, and the sizes of MPs were measured by ImageJ software (ver. 2.0.0). We classified particles as fiber, fragment, bead, or film as well as recorded the coloration. Colors are significant because yellow to brown, e.g., is generally associated with fibers from natural sources. Thirty microplastic specimens obtained from water samples were studied using a Fourier transform infrared spectroscope (FTIR, Nicolet 6700, Thermo Fisher Scientific, United States) to detect the chemical structure of the plastic particles.

The following precautions were employed to avoid cross-contamination during the processing and analysis of samples originating via external influences (i.e., atmospheric microplastics). All glass Petri dishes and other necessary equipment were washed with filtered distilled water before and after use. Any types of plastic equipment were skipped over to the greatest extent possible during the measurement of MPs, except during sample collection. When performing laboratory operations, all technicians were obliged to wear button-front cotton lab suits and powder-free nitrile-coated gloves. For control, a blank experiment was conducted following the same procedure, and no synthetic components were found in the blank sample.

The following calculation was used to compute the dietary intake of MPs in water (Almaiman et al., 2021): Exposure = C × M / BW , where C is the recommended consumption of water/person/day, M is the average MP particles/L, and BW is the average body weight for an adult. The World Health Organization’s recommended daily water intake (4.5 L/day) for an adult was used to calculate the per capita exposure estimation (Grandjean and Leparoux, 2004) because reliable information on consuming water in Bangladesh was lacking. The average weight of adult men/women in Bangladesh is 60 kg (Ullah et al., 2017).

The polymer hazard index (PHI) is a tool designed to evaluate the environmental and health-related hazards posed by different types of polymers, likely based on their chemical constituents (Hossain M. B. et al., 2023; Al Nahian et al., 2023). Using the equation provided below, the polymer risk index can be estimated: P H I = ∑ P n × S n , where Sn is the hazard score of the plastic polymers and Pn is the percentage of each plastic polymer type detected (Al Nahian et al., 2023). Applying the PHI score, referred to as the methodology of Al Nahian et al. (2023), the risk grades and hazard classes (Supplementary Table S1) of plastic components in all samples were estimated.

Each station of studied areas contained three replicates in our experiments; therefore, the concentration of MPs was expressed as mean ± standard deviation (SD). The mean abundance of MPs from different stations of each site was compared using the statistical analyses (ANOVA) through PAST (PAleontological STatistics, Version 4.03) statistical software, where a probability level less than 0.05 was considered as significant. Before conducting ANOVA, the pairwise differences were tested by Tukey’s HSD test. The homogeneity of variances was thoroughly checked by Levene’s test during the analysis. If necessary, square root or logarithm was used to transform the data. On the other hand, data heteroscedasticity was tested, even after data transformations. Levene’s test was allowed to reject the hypotheses at levels lower than the p values. Arc-GIS (version 10.7) software was used to show the sampling point in the study area map.

MPs were extracted and identified from all 18 sampling points from six coastal districts of Bangladesh (Figure 2). The abundance of MPs in the GW ranged from 4 to 75 items/L water with an average of 24.63 ± 16.13 items/L water (Figure 3). One-way ANOVA results demonstrated that there were significant variations (F = 2.96, p = 0.003) among the sampling points. Samples from the districts of Chandpur, Noakhali, and Cox’s Bazar had more MPs than those from other districts. Due to the dense population, coexistence of industrial and agricultural areas, as well as commercial and tourism activities, it is possible that these three major towns use a significant number of plastic items. According to several studies, MP pollution has caused the municipal water system to degenerate into a catastrophe (Hossain et al., 2021; Nur et al., 2022). Tukey’s pairwise comparisons revealed that the number of MPs in the GW of Chandpur and Chittagong differed significantly.

The results of this study were contrasted with those of other studies carried out globally (Table 1) and found that the GW of Bangladesh is severely polluted compared to the public drinking water collected and supplied from a variety of sources in United States, Germany, India, Denmark, Mexico, and Sweden (Strand et al., 2018; Panno et al., 2019b; Mintenig et al., 2019; Kirstein et al., 2021; Selvam et al., 2021; Weber et al., 2021). As previously mentioned, GW is predominantly used for drinking and crop cultivation in rural areas of Bangladesh. If contaminated with microplastic particles, it could pose a hazard to human health compared to ingesting seafood contaminated with pollutants (Re, 2019). The occurrence of MPs in GW demonstrates the susceptible nature of fractured reservoirs to sewer lines, sinkholes, and urban wastewater systems (Ganesan et al., 2019). A microplastic debris phase mechanism can be established in agricultural operations if GW polluted by upward microplastic transportation is used for farming and productive land. Excessive contamination by microplastics in GW can be harmful to GW microbes. GW has a diverse range of organisms, such as troglofauna (White, 1988; Breitenbach, 2008; Momot et al., 2008) and stygofauna (Huang et al., 2021), which could be infected with plastic particles. Nevertheless, the precise mechanism by which GW microplastics harm such fauna species is unexplained and requires further investigation.

The MPs detected in the GW samples were grouped into four categories, e.g., fibers, films, fragments, and microbeads (Figure 4A). Fibers were the most abundant (97.6%), followed by films (1.8%), fragments (0.3%), and microbeads (0.3%). Fibers were abundant at every sampling location (Figure 4A) and were filamentous in shape. The maximum number of fibers was recorded from S11 (average 57.33 items/L) and minimum in S4 (average 8 items/L). Microbeads were observed only in S13 (average 1.33 items/L). Although it is still difficult to specifically identify the source of MPs, the morphological characteristic generally identifies the sources of MPs. Fibers might have generated from the disintegration of regular plastic products, clothing materials, and sewage (Hossain et al., 2021). Additionally, fibers from fishing nets and ropes may have accumulated in the GW as fishing is one of the main industries in the studied area. All films and fragments were irregular in shape, and microbeads were round. Previous research studies have hypothesized that fibers may originate from clothing or rope; films may be related to single-use plastic bags, wrappers, and so forth. Microbeads are likely to be found in personal care products sold to consumers (Wang et al., 2020; Nur et al., 2022). When compared, this result is in line with observations from Selvam et al. (2021) and Strand et al. (2018), where fibers predominated over other types of MPs.

In the instance of MPs in GW from coastal regions of Bangladesh, colored MPs were predominantly recorded (57.3%), whereas the rest (42.7%) were transparent (Figure 4B). The colors recorded were violet (28.6%), pink (15.3%), blue (5.7%), green (3.9%), and red (3.8%). Transparent and violet-colored MPs were observed in all the sampling points, whereas the others differed from station to station. Since the color of MPs depend on the local climate, they can differ from one location to another due to their quick disintegration after being exposed to heat and sunlight for a long time (Shruti et al., 2020). The sizes of observed MPs were categorized into three groups, and it was interestingly found that smaller-sized MPs (<0.5 mm in size) were predominant (91.7% of the total MPs found), followed by 0.5–1 mm (6.9%) and 1–5 mm (1.4%) (Figure 4C). MPs less than 0.5 mm in size were abundant in all the sampling points. In contrast, MPs with size between 1 and 5 mm were observed occasionally. Our results correspond with those of Połeć et al. (2018) and Strand et al. (2018). The local climate, including temperature degradation, the longevity of the plastics, wind action, UV light intensity, transportation, residency time, and the distance from the plastics’ point of origin in the environment, largely determine the size of MPs (Nur et al., 2022; Hossain M. B. et al., 2023).

The chemical composition of MPs may also partially indicate their origin. In this study, representative plastic particle samples taken from filter paper were analyzed by FTIR, and the resulting spectrum was contrasted with polymer types and additional referent collections. The characteristic of FTIR spectra confirmed that the studied particles were composed of four types of polymers, e.g., low-density polyethylene (LDPE), high-density polyethylene (HDPE), ethylene vinyl acetate (EVA), and polyethylene terephthalate (PETE) in the GW samples of Bangladesh (Supplementary Figure S2). The highest frequencies in the LDPE and HDPE spectra were indistinguishable at 2,916 cm⁻¹, 2,846 cm⁻¹, 2,914 cm⁻¹, and 2,848 cm⁻¹, respectively, indicating C-H stretching. Peaks at 1,466 cm⁻¹ and 1,462 cm⁻¹ in LDPE and 1,470 cm⁻¹ and 1,462 cm⁻¹ in HDPE, on the other hand, showed CH₂ bending. The highest point in LDPE at 1,382 cm⁻¹ wavelength symbolizes CH₃ bending. Additionally, the spectral lines of 730 cm⁻¹, 719 cm⁻¹, and 717 cm⁻¹ in LPDE and HDPE specify CH₂ rock. Conversely, the electromagnetic spectrum of EVA revealed a strong peak at 2,920 cm⁻¹ and 2,850 cm⁻¹ showing C-H bond bending, accordingly, whereas 1,747 cm⁻¹ and 1,769 cm⁻¹ reflect C=O and CH₂/CH₃ stretch, respectively. PETE also revealed strong peaks at wavelengths of 713 cm⁻¹, which represent aromatic C-H bond stretching at 1,241 cm⁻¹ and C-O stretching at 1,035 cm⁻¹. The -CH₂ fingerprint in lipid structures can resemble that of polyethylene (PE), posing a challenge for differentiation. However, while both lipids and PE exhibit -CH₂ stretching vibrations, their exact absorption bands may differ slightly due to variations in the molecular structure and environment. To distinguish between them, we used spectral databases to reference distinctive absorption peaks associated with each compound. Differentiating and quantification of polyesters like PETE from lipids can be difficult due to the presence of ester groups in both. Through careful examination of variations in peak locations, intensities, and shapes within the FTIR spectra of lipids and PETE, we have successfully differentiated between the two substances and evaluated the respective amounts in a given sample. For instance, lipids frequently exhibit wide, distinct peaks as a result of the existence of numerous functional groups and intricate molecular structures. On the other hand, PETE may display more distinct and concentrated peaks that correspond to certain vibrational modes of the ester groups in its polyester backbone. Lipids generally display absorption peaks at specific wavenumbers that correspond to functional groups, such as hydrocarbons (approximately 2,850–2,950 cm⁻¹ for the stretching vibrations of CH bonds), and alkenes (approximately 3,010–3,070 cm⁻¹ for the stretching vibrations of C-H bonds). On the other hand, PETE has absorption peaks that correspond to ester groups in its backbone structure. In addition, lipids generally display very high absorption bands for functional groups such as esters and carbonyls, but PETE may demonstrate somewhat less intense absorption in these locations. The quantification of these polymers was accomplished by establishing a correlation between the absorbance of certain bands and the content of lipids or PETE in the sample, utilizing calibration curves.

The results were supported by those of Mintenig et al. (2019) and Selvam et al. (2021), who also claimed that PE was more common. Since PE dominated among the various polymer categories, it has been hypothesized that the main sources of MPs in the area include agricultural wastes, synthetic fabrics, landfill leaching, and municipal effluents (Shruti et al., 2020; Selvam et al., 2021). According to Ganesan et al. (2019), polyethylene (PE), polystyrene (PS), and polypropylene (PP), which are frequently used for packaging, were found to be the most prevalent polymers in the environment. Fishing ropes and nets are also constructed from low-density polymers, which might be a reasonable explanation for the high abundance of the polymer LDPE in this study (Hossain M. B. et al., 2023). In addition, stretchy PE (LDPE and LLDPE) films with a short life cycle may be released into the environment as a result of the breakdown of agricultural films, plastic bags, wrapping materials, or industrial usage (Hossain M. B. et al., 2023).

We assessed the daily water consumption of MP per person in order to assess human health exposure to MP pollutants and generate dietary guidelines for public health strategies. Using the mean recommended daily consumption of water (4.5 L per day as suggested by the WHO) for an adult (60 kg), the mean value of MPs (24.63 particles/L) in water and the per capita daily dietary intake of MPs have been determined to be 1.85 particles per person per day. These figures represent an average yearly MP intake of 675 particles per individual. According to Cox et al. (2019), the estimated average yearly intake of MPs in the US was 127,000 for male adults and 93,000 for female adults who only drank bottled water. The variation in the detection rate, and therefore consumption may be related to the size of the sample used.

The hazard scoring model using PHI based on their chemical constituents has been applied in the present study in order to identify and categorize the corresponding risk level (Figure 5). The results indicated varying levels of risks to human health with the polymer levels. PETE contributed around 13% of the identified microplastics, commonly originated from fabrics and packaging wastes showed a medium risk (Type-II) to human health (Figure 5). Additionally, the other three polymers, HDPE (39%), LDPE (31%), and EVA (17%), showed minor risk (Type-I) (Figure 5), which would cause less harm due to drinking water from studied sources.

The variation in risk levels among these polymers highlights the significance of not just considering the prevalence of microplastics but also recognizing their potential risks to human health. While PETE poses a moderate risk, the other polymers present a more favorable risk profile. This nuanced comprehension can steer future endeavors in handling and alleviating the impact of microplastics on water quality, placing specific attention on the types of polymers predominantly identified in the studied sources. The findings stress the necessity for targeted interventions to tackle the unique challenges posed by various microplastics, especially those originating from packaging waste. Furthermore, it emphasizes the ongoing importance of research to improve our understanding of the potential health hazards linked to microplastics and to guide effective strategies for water resource management and pollution prevention.

To obtain a thorough understanding of the health hazards associated with MPs, it is imperative to include the mass of MPs and increasing sample sizes. The precision of risk assessments is improved by these measurements, which allow for a more accurate evaluation of toxicity levels and bioaccumulation potential. Nevertheless, a number of reasons, including the goals of the study, available resources, technological difficulties, and methodological limitations, have made it difficult for this study to include mass measurements of MPs. It is possible to underestimate the actual health effects of microplastic exposure if the mass of MPs is ignored. Because of their higher surface area-to-volume ratios, smaller particles could be more dangerous per unit mass. Therefore, the inclusion of mass of MPs into MP research enhances the accuracy of risk evaluations and streamlines the creation of focused mitigation solutions.