Polychlorinated biphenyls in mussels, small pelagic fish, tuna, turtles, and dolphins from the Croatian Adriatic Sea waters: an overview of the last two decades of monitoring

The Adriatic Sea is a semi-enclosed basin of the Mediterranean that extends over 138,000 km². Although the Adriatic is known as an oligotrophic sea, its shallow coastal waters are characterised by higher nutrient input from rivers, lower salinity, and greater zooplankton and pelagic fish spawn area than open waters (26). It is home to a diverse array of marine organisms, ranging from microscopic plankton to large apex predators.

According to the Croatian Bureau of Statistics for 2020, of the 68,232 t of fish caught in the Croatian part of the Adriatic, more than 95 % were pelagic [e.g., sardine (Sardina pilchardus); anchovy (Engraulisen crasicolus), and bluefin tuna (Thunnus thynnus)], of which 77.1 %, were sardines. In 2020, the mariculture production of 18,275 t was dominated by seabream (Sparus aurata, 7,792 t), seabass (Dicentrarchus labrax 6,754 t), and bluefin tuna (2,611 t). Furthermore, Croatia, like other Mediterranean countries, strives for species diversification in aquaculture and has recently introduced meagre (Argyrosomus regius) and common dentex (Dentex dentex) into the Adriatic Sea farms (27).

Although relatively shallow, the Adriatic is known for its biodiversity thanks to a variety of seabed sediments, such as rocky, sandy, and muddy. However, the last few decades have seen changes in its flora and fauna owed to the climate change (especially the rise in temperature and higher salinity), anthropogenic activity, and Lessepsian migration. It faces problems with new allochthonous species that are mostly thermophilic and have expanded their habitats migrating northward. Some of the new species come from the ballast water or aquaculture and are potentially invasive. Overall, the number of Adriatic fish species had grown from 407 in 1996 to 456 in 2017 (28). Since the Croatian part of the Adriatic covers more than 35 % of the total Croatian territory, monitoring changes in marine ecosystems and the conservation of its biodiversity are of great importance (29).

Of the 209 listed PCB congeners, our research has primarily been focused on the six so-called “indicator PCBs” (PCB-28, PCB-52, PCB-101, PCB-138, PCB-153, PCB-180), because they are commonly used as indicators of overall PCB contamination in a given environment. These congeners are often found in commercial PCB mixtures and are prevalent in environmental samples. By monitoring the levels of these indicator PCB, scientists and environmental researchers can gain insight into the extent of PCB pollution and potential risks to ecosystems and human health.

We have also studied 11 toxicologically relevant congeners (PCB-60, PCB-74, PCB-105, PCB-114, PCB-118, PCB-123, PCB-156, PCB-157, PCB-167, PCB-170, and PCB-189). These congeners are distinguished from indicator congeners, insofar as they are known to have greater toxic or bioaccumulation potential and are of greater concern in risk assessment and regulatory contexts. They are often analysed separately to better understand their impact and to develop appropriate management strategies.

The concentrations of either indicator or toxicologically relevant PCBs in various marine organisms are often given either for each individual congener or for their sums (∑Ind PCB and ∑ToxRel PCB, respectively).

In our research (9,10,11,12,13,14,15,16,17,18,19,20) all samples were analysed at the Biochemistry and Organic Analytical Chemistry Unit of the Institute for Medical Research and Occupational Health, Zagreb, Croatia. Statistical analysis and modelling with artificial intelligence (AI) algorithms through machine learning were done in collaboration with the University of Belgrade Institute of Physics, Belgrade, Serbia (13,14,15, 17,18,19,20).

Filter feeding organisms, such as mussels, have been proven effective as indicator organisms in environmental monitoring programmes worldwide, owing to their abundance and capacity to accumulate a wide spectrum of contaminants. In order to establish baseline concentrations of PCBs, we investigated wild and farmed Mediterranean mussel (Mytilus galloprovincialis) sampled in the coastal waters of the eastern Adriatic. Wild mussel populations were collected at 14 locations along the Croatian coast in March 2006 (10), and our findings showed that ΣInd PCB levels were consistently higher than the ΣToxRel PCB levels (Table 1). Among indicator PCBs, PCB-138 and PCB-153 dominated with concentrations at least twice as high as any of the rest, followed by PCB-52, PCB-28, PCB-101, and PCB-180. The most dominant ∑ToxRel PCBs were PCB-118, PCB-60, PCB-170, and PCB-114. Spatial PCB distributions were relatively consistent throughout, with particularly high levels observed near areas of intense industrial activity such as large shipyards, yacht marinas, industrial zones, and urban centres. These areas emerged as hot spots for PCB pollution, indicating local sources of contamination. One sampling site located in a major port and industrial hub (Rijeka Bay) recorded the highest PCB levels on the Croatian coast. An earlier monitoring study of this location conducted by Picer and Picer (30) covering the period from 1972 to 1992 concluded that PCB levels remained unchanged over those 20 years and that their sources were atmospheric deposition, local urban and industrial wastewater, and maritime activities.

Continuing our investigation, we aimed to compare the results obtained from the analysis of both wild and farmed Mediterranean mussels. The farmed mussels were collected at 15 shellfish breeding farms situated along the central and southern Croatian Adriatic coast in 2010. Our findings (11) were similar to those observed in wild mussel populations (10), with ∑Ind PCB exceeding ∑ToxRel PCB at all sampling locations (Table 1). The dominant compounds were PCB-138 and PCB-153, followed by PCB-52, PCB-28, PCB-101, and PCB-180. The most prominent congeners in the ToxRel PCB group were PCB-123 and PCB-170. The prevalence of hexachlorinated congeners PCB-153 and PCB-138 in mussels aligns with reports for other Mediterranean locations (31,32,33).

We also observed a significant seasonal variation in contaminant concentrations in farmed mussels; higher levels were recorded in the summer, and lower levels in the winter. However, the contaminant levels remained within tolerable daily intake limits of 2 pg/kg per day (34). Like us, Milun et al. (22) observed seasonal variations in PCB concentrations in wild and cultivated bivalve molluscs (Mytilus galloprovincialis, Ostrea edulis, Venus verrucosa, Arca noae, and Callista chione) collected at 11 wild and two mariculture sites along the Croatian coast in May and November 2012. They also observed variations between sampling sites and suggested potential pollution sources, such as historical industrial activities, long-term urban and industrial wastewater discharge, and the proximity to a nautical marina and local shipyard. In another study, Milun et al. (23) found a significant increase in PCB concentrations, most notably PCB-138 and PCB-153, in wild Mediterranean mussel from Kaštela Bay, known for its steel and cement plant, brewery, food and beverage production, port, shipyard, and a former PVC chlor-alkali facility.

Mussels were also monitored for PCB contamination within the frame of the MYTIAD project focused on the accumulation of PCB-28, PCB-31, PCB-52, PCB-101, PCB-105, PCB-118, PCB-138, PCB-153, PCB-156, and PCB-180 (25). High PCB levels were reported for Taranto, Italy, with concentrations reaching 114.4 µg/kg dry mass, followed by Baošići and Kotor in Montenegro, where PCB levels slightly exceeded 60 µg/kg dry mass.

The research conducted between 2014 and 2016 on sardines, anchovy, round sardinella (Sardinella aurita), chub mackerel (Scomber japonicus), and horse mackerel (Trachurus trachurus) caught along the eastern Adriatic (19) revealed the prevalence of indicator congeners PCB-153, PCB-138, and PCB-180, whereas toxicologically relevant PCB-118 and PCB-170 levels were consistent with previous reports for the Adriatic (35) and Ionian fish species (36). Vuković et al. (14) observed the following pattern of accumulation: anchovy<chub mackerel<horse mackerel<sardines. To assess the health risks for consumers of small pelagic fish, we adopted the Risk Assessment Information System models to address specific local conditions and ran a comprehensive benefit-risk analysis based on essential fatty acid content in these fish (20). The richest sources of omega-3 fatty acids were anchovy and round sardinella, while chub mackerel and sardine followed closely. Sardines had the highest levels of docosahexaenoic acid and eicosapentaenoic acid, both essential for human health. It was reassuring to find out that PCB levels in these fish species did not pose a significant health risk to consumers, and the benefit-risk ratios were consistently below 1, which means that the consumption of these fish varieties has more benefits than potential risks for human health.

Our research extended further to predatory species to gain insights into the distribution of PCBs across the trophic levels of the Adriatic Sea and to understand how the levels measured in predatory species may impact human health. To this end we analysed PCBs in the muscle (because it is consumed as food), liver (because it can tell us a lot about PCB metabolism), and gills (as the entry point of organic contaminants from water) of bluefin tuna farmed in the central Croatian Adriatic in January 2015 (15). Indicator PCB-153, PCB-138, PCB-153, and PCB-180 dominated in all tissues, whereas PCB-118, PCB-123, and PCB-170 dominated among the toxicologically relevant PCBs in tuna liver and muscle, but not in the gills. The concentrations (ng/g lipid mass) of ∑Ind PCB and ∑ToxRel PCB followed the muscle>liver>gill sequence, correlating with the percentage of tissue lipids. Vizzini et al. (37) reported a similar pattern of higher muscle than liver contamination in tuna farmed in Sicily, Italy, although the reported concentrations of Σ43 PCBs (which includes the indicator and toxicologically relevant ones) were significantly higher than our findings (371.74 ng/g per lipid mass in the liver and 1916.98 ng/g in the muscle). Furthermore, the highest concentrations of ΣInd PCBs we found in farmed tuna muscles (15) and small pelagic fish (19) was lower than the maximum permissible level of 75 ng/g wet mass in fish set by the European Commission (38).

The Mediterranean Sea pollution with PCBs raises particular concern for endangered species, such as wild tunas, sea turtles, and dolphins. PCB levels determined in the muscle tissue of wild bluefin tuna sampled in the Adriatic in 1996 were among the highest reported in the literature (Table 1), raising concern about bioaccumulation through the food web (16). Similarly, Chiesa et al. (39) suggest that the Mediterranean Sea could be more polluted than the Pacific, Indian, and Atlantic Ocean, judging by the highest PCB levels found in the Mediterranean tuna.

Speaking of endangered species, one Adriatic study (9) addressed the anthropogenic impact on loggerhead turtles (Caretta caretta) by measuring PCBs in their fat tissue and provided an additional insight into PCB bioaccumulation along the food web. These measurements were made in 26 turtles found dead between June 2001 and November 2002. PCB-153 median concentrations per lipid mass (244.6 ng/g) were the highest, followed by PCB-138 (169.1 ng/g), PCB-180 (50.4 ng/g), and PCB-118 (33.4 ng/g). These levels are higher than reported for loggerheads from the Italian part of the Adriatic (40) but lower than reported for Greece and Cyprus (41).

We also investigated PCB contamination of the common bottlenose dolphin (Tursiops truncatus) in 13 specimens found stranded at different locations along the northeast coast of the Adriatic Sea between 2000 and 2005 (12). Their PCB profile in all tissues (Table 1) was also dominated by three indicator congeners, PCB-153, PCB-138, and PCB-52, in that order. An analysis of PCBs in the blubber samples of delphinids from the south-eastern Brazilian coast (42) suggests that low-chlorinated PCBs, such as PCB-52 found in our study, travel over long distances as they bind to suspended particulate matter in the atmosphere. In our study (12), PCB congeners were distributed similarly across the analysed dolphin tissues: hexachloro>pentachloro>heptachloro>tetrachloro> trichlorobiphenyls, and their levels decreased in the following order: blubber>muscle>kidney>liver>heart>lung. Furthermore, their lipid-normalised concentrations were significantly higher than in the wild or farmed tuna and loggerhead sea turtles (Table 1).

Our findings are similar to those reported by Genov et al. (43) in common bottlenose dolphins found in the Gulf of Trieste and the adjacent waters in the northern Adriatic between 2011 and 2017, but both are much higher than those reported for dolphins found along the Adriatic coast in Italy (7.25–56.96 µg/g lipid mass for Σ17 PCBs) (44) or those found along the western Mediterranean coast in 2000–2003 (45, 46). According to Storelli and Marcotrigiano (44), sum PCB levels in blubber exceeding 50 µg/g wet mass might present a health risk to cetaceans, yet in our study, only two blubber samples exceeded this threshold.

As for PCB-138 and PCB-153, they resist metabolism up the food web, are regularly found at the highest levels in marine organisms (46,47,48,49), and are suitable for comparison within and between species (50). In this context, our PCB-153 results suggest that the endangered species of wild tuna, loggerhead sea turtles, and dolphins from the eastern Adriatic Sea are among the most affected by PCB pollution worldwide.

Our research of PCB levels in marine organisms from the Croatian Adriatic conducted over the last two decades has shown the dominance of PCB-138, PCB-153, PCB-180, and PCB-170. An interesting finding is that the last congener is more dominant in mussels (both wild and farmed) and farmed tuna than in other species. Small pelagic fish, in turn, revealed seasonal variations and site-specific differences in PCB levels.

Furthermore, our research has shown higher PCB accumulation in wild mussels and tuna than their farmed counterparts, in which the sum of indicator PCB did not exceed the permissible EU limit. This may be owed to controlled feeding and growth conditions.

Even though PCB levels in the Adriatic reflect global pollution, high levels in dolphins and wild tuna are particularly worrying, as they confirm their biomagnification along the food chain and higher pollution in the Mediterranean and Adriatic Sea than the world’s oceans. This is why our future research should focus on PCB exposure and toxicity in the Adriatic predatory species. We recommend establishing routine monitoring programmes specifically targeting predatory species known to bioaccumulate PCBs. These programmes should involve regular sampling of tissues such as muscle, liver, and adipose tissue to assess PCB levels and track trends in PCB contamination over time. This can provide valuable data on the effectiveness of regulatory measures and identify emerging contamination hotspots.

We also recommend informing consumers about healthy seafood choices and providing guidance on safe consumption practices.