Fish larvae were collected from the Douro Estuary, a Portuguese urban estuary located in the northwest of Portugal (41°8’36.00”N 8°39’25.00”W). This estuary extends 21.6 km from the estuary mouth until its upstream limit, the Crestuma Lever Dam (Rodrigues et al., 2022 and references therein). It is highly impacted by intense urbanization and touristic activity, with a high density of boat traffic, several freshwater streams, as well as wastewater treatment plant (WWTP) effluents draining into the estuary (Azevedo et al., 2008; Ramos et al., 2015). Monthly sampling surveys were conducted from September 2021 to September 2022 at 11 sampling stations along the estuary ( Supplementary Figure S1 in the Supplementary Material). At each sampling station, planktonic samples were collected using a conical plankton net with a 1 m diameter, 4 m length, and 500 μm mesh size. Subsurface (1–2 m depth) planktonic circular tows were performed during the slack phase of spring tides (i.e., 2 hours before high tide) at a constant velocity of approximately 1ms⁻¹ for 5 min. The volume of water filtered was determined using a Hydro-Bios flowmeter attached to the net. Samples were immediately transferred to flasks and fixed and preserved in 70% ethanol until laboratory analysis. The number of individuals was counted for the entire sample, and using the volume of water filtered, it was standardized to the number of fish larvae per 100 m⁻³.

Fish larvae were sorted and isolated from the collected samples under a stereomicroscope (Nikon SMZ800) and then identified to the highest possible taxonomic separation using specialized literature (Russell, 1976; Ré, 1999; Ré and Meneses, 2008; Rodriguez et al., 2017). The two most abundant and frequent larval fish species observed, P. microps and S. pilchardus ( Figure 1 ), were selected for MPs quantification, due to their contrasting differences in terms of pelagic larval duration and their use of estuaries as habitats. The ontogenetic development stage of each fish larva was also recorded, namely: yolk-sac, preflexion, flexion, and postflexion. Fish larvae in the yolk-sac stage are newly hatched larvae and still have yolk-sac or remnants of yolk; they frequently have an unformed mouth, unpigmented eyes, and undeveloped pectoral fins. Then the subsequent stages are preflexion, flexion and postflexion, which are related to the flexion of the notochord during caudal fin development. The preflexion stage begins once complete absorption of the yolk sac has occurred and ends with the start of notochord flexion. The flexion stage occurs at the beginning of the dorsal bending of the notochord tip and ends when the notochord tip reaches its final position and the principal caudal-fin rays and supporting skeletal elements are in the adult longitudinal position. The postflexion stage begins after the conclusion of notochord flexion and ends at the beginning of metamorphosis (Russell, 1976; Ré and Meneses, 2008).

The presence of MPs in planktonic species has been assessed in different ways, namely by the incidence of ingestion (by analysing organisms individually, e.g., Steer et al., 2017) or by encounter rate (by analysing a pool of organisms, e.g., Desforges et al., 2015; Sun et al., 2017). Here, the presence of MPs in the two selected species was quantified by analysing pools of organisms. All fish larvae belonging to the two species selected were isolated and grouped in pools according to the station and month where they were collected, the species, and the ontogenetic development stage, making a total of 99 samples ( Supplementary Figure S1 in the Supplementary Material). The abundance of MPs per fish in each sample was obtained by dividing the total number of MPs obtained in each sample by the total number of fish larvae digested in that same sample. The average MPs abundance was calculated by dividing the total number of MPs by the total number of fish larvae. Those samples were subjected to a protocol previously optimized to quantify MPs in planktonic organisms, including fish larvae (Rodrigues et al., 2023). In brief, each fish larva was individually selected, thoroughly rinsed with deionized water, and checked for any external MPs using a stereomicroscope to ensure that no particles were attached/adsorbed to their body, and then transferred to the respective vial.

In each vial, 2 mL of 30% (v/v) H₂O₂ solution was added and placed in the oven at 65°C for a maximum of 7 hours. This protocol guarantees the complete digestion of the organism’s tissues while preserving the characteristics of MPs (e.g., shape, colour) present in the organism. After the digestion, the samples were filtered through a glass filtration system using a 0.45 μm pore size cellulose membrane and stored in a petri dish, which was left to dry at room temperature (20-25°C). Finally, the filters were visually inspected for the presence of MPs; due to equipment restrictions, only plastic particles between 5 mm and 0.01 mm were considered.

MPs recovered from each sample were observed, counted, and characterized under a stereomicroscope (Nikon SMZ800) and classified according to shape (fiber, fragment, film, sphere, and foam) and colour. Those with more than one colour were classified according to the predominant one. MPs were photographed and their precise size was obtained using the stereomicroscope Leica EZ04 software; in the results, MPs were assigned to size classes (0.01-0.05 mm, 0.06-0.5 mm, 0.6–1 mm, 1.1-1.5 mm, 1.6–2 mm, 2.1-2.5 mm and 2.6-3.0 mm).

A subsample of the most representative particles found was selected for polymer identification using a stereomicroscope for further FTIR (Fourier Transform Infrared Spectroscopy) analysis (45% of the total particles found); this follows the recommendation under the European Union Marine Strategy Framework Directive (MSFD) that recommends the examination of at least 10% of the sample (Galgani et al., 2023). Polymer spectra were registered in a PerkinElmer (Waltham, Massachusetts, USA) FTIR Spectrum 2 instrument, coupled with attenuated total reflectance (ATR). The resulting spectra were compared with reference library spectra, and matches with confidence levels of 65% were accepted. Of the particles analysed using FTIR, 85% were successfully confirmed as plastic polymers. The remaining 15% of the particles were not suitable for placement on the crystal surface of the FTIR, and their polymer type could not be confirmed; hence, they were not included in the results presented here. The fibers entangled in a bundle had different colours and may have had differing chemical compositions; however, their polymer was not confirmed using FTIR due to equipment restrictions.

To prevent any contamination, the following QA/AQC measures were taken: i) coloured cotton laboratory coats were worn during all laboratory procedures; ii) all the materials used were made of glass; only in a few exceptions plastic polymers materials were used (e.g., sieving of water samples to collect fish larvae – the plastic polymers of these material were confirmed by FTIR and MP contamination from these materials were not observed); iii) cleaning steps were applied before use including rinsing steps to all the material used in the field and laboratory; v) all the procedures in the laboratory were made inside a fume hood; iv) a petri dish with deionized water was exposed to the environment near the stereomicroscope as an air contamination control and inspected for MPs at the end of all the sorting processes (including fish larvae and MPs); and vi) blank samples with only 30% (v/v) H₂O₂ solution were included, in triplicate, each time the protocol for MPs quantification was performed (these blank samples were subjected to all the steps of the protocol together with the vials with samples). All blanks and petri dish controls were always inspected, and none of them had any type of MP contamination.

Larval length measurements for all specimens of both species were obtained using a stereomicroscope and Leica EZ04 software. The ontogenetic development stage of each fish larva was also recorded, namely: yolk-sac, preflexion, flexion, and postflexion (Russell, 1976; Ré and Meneses, 2008). The relationship between larval length and ontogenetic stage was assessed using linear regression analysis, with statistical significance set at α = 0.05. Two-way analysis of variance (ANOVA) was used to determine the effect of species and larval ontogenetic development stages on MPs concentration, and shape (i.e., number of different shapes observed in that sample), colour (i.e., number of different colours observed in that sample) and size (i.e., number of different sizes observed in that sample) diversity with species and larval ontogenetic development stages as fixed factors. A two-way analysis of variance (ANOVA) was used to ascertain the effect of month and sampling station on MPs concentration, with months and sampling stations as fixed factors. MPs concentration data were square root transformed to fulfil the ANOVA assumptions of homogeneous variance and normally distributed data (Zar, 1996). The normality and homogeneity of the data were verified using the Shapiro-Wilks and Levene’s tests. A significance level of 0.05 was applied to all analyses. The error term accompanying the mean values refers to the standard deviation; however, in some cases, the standard deviation is not presented since only one sample was analysed for that species/ontogenetic development stage/month/site. All tests were performed using TIBCO Statistica™ 14.0 software.

A total of 707 fish larvae were collected: 350 fish larvae of S. pilchardus and 357 fish larvae of P. microps ( Figure 1 ). The two species presented different temporal patterns: S. pilchardus was collected mainly between September 2021 and May 2022, with its abundance peaking in October, January, and April, reaching an average of 8.0 ± 13.7 fish larvae per 100 m^-3 in April. This species was not observed in July 2022 or September 2022. Spatially, S. pilchardus was more abundant between 0.5 and 5.5 km from the estuary mouth, corresponding to the lower and middle areas of the estuary, near the sea ( Supplementary Figures S1 , S2 in the Supplementary Material). P. microps was collected in all sampling months and sampling stations, with the highest abundance observed in June, reaching an average of 5.0 ± 12.1 fish larvae per 100 m^-3. Throughout the estuary, P. microps was more abundant between the middle and upper areas of the estuary, at 5.5 and 10.5 km from the estuary mouth ( Supplementary Figures S1 , S2 in the Supplementary Material).

S. pilchardus larvae were identified in three different ontogenetic development stages: 81% at the yolk-sac stage, 15% at the preflexion and 5% in the flexion stage. No S. pilchardus larvae were found in the postflexion stage. P. microps larvae were identified in four stages: 0.3% at the yolk-sac stage (n = 1 fish larvae), 59% at the preflexion stage, 38% at the flexion stage, and 4% at the postflexion stage. A linear relationship between larva length and the ontogenetic stage was observed in both species. P. microps larvae had an average length of 3.8 ± 0.1 mm at the preflexion stage, 8.2 ± 1.1 mm at the flexion stage, and 11.1 ± 0.7 mm at the postflexion stage (r² = 0.97, p < 0.001). S. pilchardus larvae had an average length of 4.7 ± 0.2 mm at the yolk-sac stage, 7.4 ± 0.7 mm, and 11.0 ± 0.8 mm at the flexion stage (r² = 0.93, p < 0.001) ( Supplementary Figure S3 in the Supplementary Material).

This novel MPs research provides much-needed field evidence on MPs contamination in early life stages of fish, as it makes a direct comparison between MPs contamination in two larval fish species with different ecological guilds. Importantly, it is the first study to investigate whether MPs contamination varies throughout the ontogenetic development stages in wild fish larvae. Consequently, we evaluated when MPs contamination of wild larval stages of fish starts. A previous laboratory study reported that MPs contamination in fish larvae starts with the beginning of exogenous feeding, when the yolk reserves are depleted, and the fish larvae need to start predating other organisms (Cousin et al., 2020). However, our results showed that fish larvae in the yolk-sac stage, which were still endogenously feeding, were already contaminated with MPs. At this stage, fish larvae still have their mouths closed, are not yet feeding on prey, and are still being fed endogenously via the yolk-sac. In the case of S. pilchardus, approximately 80% of the fish larvae analysed were at the yolk-sac stage, and an average of 3.7 ± 4.1 MPs per fish larva was observed. S. pilchardus larvae at the yolk-sac stage had between 0.1 and 12 MPs per fish larva, meaning that not all the fish larvae had MPs. However, some of these young larvae had up to 12 MPs per larva, a concentration similar to that reported for adult stages of S. pilchardus (e.g., Lopes et al., 2020; Bouzekry et al., 2023; Trani et al., 2023). Hence, our results importantly demonstrate that MP contamination occurs much earlier in fish development than previously understood, which can have important implications. Finding MPs in this early stage means a higher lifetime exposure to this contaminant than just accumulating them through feeding as adults. Cousin et al. (2020) concluded that this might potentially have a larger impact on fish than exposure only during adulthood, since fish in their early life stages are at their most sensitive stage and typically have less developed detoxification mechanisms and are more sensitive to contaminants. Moreover, MPs found at these stages indicate greater ecosystem contamination than surface-level ingestion, which could ultimately affect the entire marine food web structure, as contaminated larvae serve as prey for numerous species and carry this contamination load throughout their lives.

Taking into consideration previous laboratory studies on zooplankton and fish, the presence of MPs in the yolk-sac fish larvae, a development stage when larvae are not yet feeding on prey, can occur through transfer to offspring from the mother (Dai et al., 2025). Martins and Guilhermino (2018) reported the transfer of MPs (around 0.005 mm) to the F₁ generation of Daphnia magna and that exposure to MPs caused reduced fertility and increased mortality among the offspring. Moreover, the offspring growth rate decreased, indicating a reduction in population fitness within just one generation. Pitt et al. (2018) observed the transfer of plastic particles to fish offspring, and the presence of plastic particles (around 42 nm) in the yolk-sac of F₁ embryos of the zebrafish Danio rerio. Also, Teng et al. (2022) observed maternal transfer of plastic particles (around 44 nm) when F₀ D. rerio were exposed to these particles, which were subsequently detected in the intestine, liver, and pancreas of F₁ D. rerio. Malafaia et al. (2022) showed that plastic particles (around 26 nm) were transferred to embryos after gestational exposure of the guppy Poecilia reticulata females. Hence, these studies demonstrated not only the transfer of plastic particles to offspring but also their impacts on the offspring. Even more concerning, species may require several generations to recover completely from developmental and reproductive negative effects induced by this contaminant (Martins and Guilhermino, 2018). Nonetheless, these laboratory studies used particles several orders of magnitude smaller than the MPs recovered in this study, and the mechanisms underlying maternal transfer of plastic particles remain unclear. Some of those studies reported that plastic particles were able to penetrate the chorion of the egg (Pitt et al., 2018) while others suggested accumulation in the maternal gametes through binding to lipids and vitellogenin, which facilitates transfer to oocytes and subsequently to the embryonic yolk sac. Given the above, it is essential to include the analysis of eggs in future studies on MPs contamination and to evaluate other contamination pathways in subsequent research. These results demonstrate the complexity and magnitude of the impacts caused by plastic particles on fish larvae, highlighting the need for new investigations, not only to understand the mechanisms of uptake of plastic particles but also the impacts of this transfer on fish growth and development. It is of note that the larval stages of fish are their most sensitive stage, during which they use estuaries as nursery areas to maximize survivorship (Ramos et al., 2010, and references therein). However, as those areas tend to be highly contaminated by MPs (Ramos et al., 2024), the estuarine nursery function, which is associated with enhanced growth and survival during the initial development stages of fish, can be compromised by these contaminants.

The understanding of whether MPs contamination profile changes throughout the ontogenetic development stages was another fundamental question approached in this study. P. microps and S. pilchardus have a pelagic larval duration of 10 and 40 days, respectively. In that period since hatching, the newly hatched larvae already had MPs, with an increase of 47% from preflexion to postflexion in the case of P. microps and reaching an average of 5.6 ± 2.5 MPs per P. microps larva in the stage of postflexion and 3.0 ± 1.8 MPs per S. pilchardus larva in the flexion stage. However, MPs’ concentration did not vary significantly throughout the development stages of each species. Moreover, throughout ontogenetic development, MP diversity remained similar across all stages, in terms of number of shapes, colours and sizes. A similar diversity of MPs was observed in the stage of endogenous feeding and well as in the stages of exogenous feeding. Even in the stages where fish larvae start to prey more actively, there was no change in the diversity of MPs towards a specific MP type of interest.

S. pilchardus has a pelagic larval duration of 40 days, while that of P. microps is of 10 days, which implies important differences in terms of morphological and physiological development (Russell, 1976). For example, S. pilchardus newly hatched larvae are very rudimentary and the mouth is not open, and the feeding is internal and based on yolk reserves. In contrast, P. microps hatched larvae are more developed, with eyes that are already pigmented and have reduced yolk reserves, needing external feeding in an earlier stage (Russell, 1976). Concerning the use of estuaries as habitats, these species belong to two different ecological guilds according to their estuarine use pattern, P. microps being as an estuarine species (ES) and S. pilchardus a marine migrant (MM; a species that spawn at sea and regularly enter estuaries in large numbers, using estuaries as nursery grounds) (Franco et al., 2008). Fish larvae at the yolk-sac stage still have yolk-sac or remnants of yolk, and they frequently an unformed mouth, unpigmented eyes, and undeveloped pectoral fins (Russell, 1976; Ré and Meneses, 2008). Although these two species present different morphological and physiological developments and periods of ontogenetic development, our results indicate that, in that early stage of fish life, MPs contamination appears to occur independently of species-specific traits.

It is of note that the most abundant MPs present in fish larvae in this study, namely in shape and colour, tend to be the predominant ones reported in water in different types of the aquatic environments worldwide, including estuaries (e.g. Ramos et al., 2024; Botterell et al., 2022; Tang et al., 2023), suggesting that MPs contamination in fish larvae may happen incidentally. For example, fibers, which are one of the most abundant types of MPs reported in estuarine waters (Ramos et al., 2024), represented more than 50% of the MPs recovered, except in the flexion stage of S. pilchardus. Moreover, fibers presented a higher diversity of sizes and the bigger MPs recovered from fish larvae in this study were fibers. Previous research has shown that fibers can be more easily mistaken for natural prey and have a shape that facilitates ingestion, even accidentally (Tang et al., 2023). Additionally, fibers tend to become entangled and often cluster together in bundles within the aquatic environment. These factors, combined with fibers being the most abundant microplastic shape in water, likely explain their high concentration across all developmental stages of fish larvae, including the yolk-sac stage, and their diverse size ranges (Tang et al., 2023). Other studies have reported colour-selective ingestion by certain organisms due to resemblance to natural prey (e.g., Wright et al., 2013). Hence, the colour of MPs seems to be an important characteristic since colours as green, transparent and blue could potentially resemble fish larvae prey, increasing the active uptake of MPs. Nonetheless, throughout the ontogenetic development stages of the two species studied here, a similar diversity of colours was observed and the MPs’ colours in higher abundance were the same as the ones in water, again reinforcing that contamination occurs randomly and is environment-dependent rather than driven by species-specific characteristics as prey preferences. Regarding the size, Ding et al. (2023) observed an exponential relationship between MP concentration and the size ingested and the adult wild fish age. MPs size diversity and the average size were not different among the different ontogenetic development stages in our study, indicating that MPs size is similar in all stages and there was no relationship between the size and the ontogenetic development stages, possibly suggesting that environmental availability rather than selective feeding drives contamination patterns. These results increase the much-needed understanding of the contamination mechanisms in wild organisms, suggesting that fish larvae contamination happens randomly and is environment-dependent rather than species-specific traits. Hence, environmental management needs to be redirected from focusing on species-specific mitigation strategies towards prioritizing the reduction of overall MP pollution in aquatic environments.

Another key objective of this study was to determine whether the species-specific MPs contamination profiles documented in adult fish extend to the early larval stages. Our results revealed that both species exhibited similar MPs concentrations and MPs types, despite their distinctly different ecological characteristics. Although P. microps had a higher diversity of MP shapes and sizes compared to S. pilchardus, this difference possibly reflects their contrasting habitat exposure rather than selective ingestion. P. microps is an estuarine species that spends its entire life in estuaries; they are always exposed to the highly contaminated estuarine waters of the Douro estuary (Rodrigues et al., 2019), possibly explaining the higher diversity of MPs shapes and sizes in their larvae. In fact, the predominance of transparent and blue fibers in both species,which mirrors the most abundant types of MPs reported in estuarine environments (Ramos et al., 2024), suggests a more passive contamination mechanism. Both Steer et al. (2017) and Tang et al. (2023) also reported a similarity between MPs in water and those ingested by fish larvae and zooplankton, respectively. The similarity between MPs in water and fish larvae may indicate that MP presence in fish larvae represents primarily accidental ingestion driven by environmental availability rather than active selection based on ecological guild or any physical characteristics. Hence, while Siddique et al. (2024) previously demonstrated that adult fish feeding behaviour and physical characteristics such as mouth size significantly influence MP ingestion, our study reveals a different scenario in the early life stages, since larval fish contamination appears to be independent of species-specific traits or ecological guilds.

Despite this, the average MPs concentration in this study (1.0 MPs in the case of S. pilchardus and 2.0 MPs in the case of P. microps) was much higher than in other studies focused on fish larvae, namely 0.005 MPs (Zavala-Alarcón et al., 2023) and 0.57 MPs per fish larva (Rashid et al., 2021) ( Table 2 ). In contrast, two other studies reported higher MPs concentrations, namely 5.2 MPs (Steer et al., 2017) and 7.5 MPs per fish larva (Carrillo-Barragan et al., 2024) ( Table 2 ). Despite this, the concentration reported in this study is of concern since it is higher or in the same range as the contamination values reported in other Portuguese estuaries for adult fish (1.67 MPs per adult fish in species as Dicentrarchus labrax, Diplodus vulgaris, and Platichthys flesus collected in the Mondego estuary – Bessa et al., 2018, and between 2 and 8 MPs per adult fish in species as Cyprinus carpio, Mugil cephalus and P. flesus collected in the Minho Estuary – Guilhermino et al., 2021). However, it is important to emphasize that these comparisons should be treated with caution due to possible differences in the MPs methodologies used in each study.

S. pilchardus is a marine migrant, using estuaries occasionally and migrating into the estuary if the environmental conditions, such as salinity, temperature, and river flow, are favourable (Morais et al., 2009; Ramos et al., 2009). The present results support this behaviour since S. pilchardus was only observed in part of the year and areas close to the sea. This species showed a higher MPs ingestion in December, March and August. It is important to highlight that in some S. pilchardus larvae, contamination reached 12 MPs per larva, a concentration higher than reported for adult S. pilchardus in other studies on Portugal (0.16 MPs per fish – Lopes et al., 2020), Moroccan Mediterranean coast (9.6 MPs per fish - Bouzekry et al., 2023), Adriatic Sea (4.6 MPs per fish - Trani et al., 2023) and Italy (6.2 MPs per fish - Trani et al., 2023). S. pilchardus has a very high socio-economic importance as a major food-fish species and also plays a vital role in marine ecosystems via predation by larger fishes and seabirds (Whitfield et al., 2022). There is a risk to survival and healthy growth since early stages can impact the adult stocks and ultimately their role as one of the main consumed fish by humans (Bouzekry et al., 2023). P. microps, an estuarine species, was collected all year round and in every area of the estuary and the peak of MPs ingestion was observed for a longer period (between December and May). It is important to note that the pelagic larval duration for P. microps is only 10 days; however, in that small period, fish larvae of this species already had an average of 2.0 MPs per fish larva and reached until 12 MPs per fish larva. This species is a central species in estuarine food webs as it is preyed upon by crustaceans, larger fishes and shorebirds (Whitfield et al., 2022) and, hence, any contamination is passed through that food chain.

Laboratory studies suggest that MPs uptake and effects will depend on MPs characteristics (e.g., Pannetier et al., 2020; Kim et al., 2022; Yap and Yasid, 2023). A great variety of shapes and sizes were reported in this study, which can potentially cause different effects with different levels of harm to different species, namely as deposition of particles in the liver, gill and gut in the case of fibers or, damage to the intestine and morphological deformations in the case of fragments (Chouchene et al., 2023). Nonetheless, other studies reported no significant effects or minimal impact on fish larvae (e.g., Karami et al., 2017; Sleight et al., 2017). It is of note that all the fish larvae collected in this study were alive when collected, implying that they were not carrying a lethal load of MPs, which may be higher than that encountered here. However, this study did not assess the health status of the fish larvae or the toxicity of MPs ingested. It will be important in future research to investigate the health status of fish larvae contaminated with MPs, together with the potential sublethal effects of these contaminants. Our findings reveal a widespread occurrence of MPs in field fish larvae, emphasizing bioavailable transfer through food chains by MPs pollution.